Characterization of In-Use Emissions of Hybrid …docs.trb.org/prp/16-5789.pdf · 3...
Transcript of Characterization of In-Use Emissions of Hybrid …docs.trb.org/prp/16-5789.pdf · 3...
1
2
Characterization of In-Use Emissions of Hybrid Electric and 3
Plug-in Hybrid Passenger Vehicles 4
5
6
7
Reza Farzaneh, Ph.D., P.E. * 8
Texas A&M Transportation Institute 9
505 E. Huntland Dr., Suite 455, Austin, TX 78752 10
Tel 512.407.1118 | Fax 512.467.8971 | Email: [email protected] 11
12
Jeremy Johnson 13
Texas A&M Transportation Institute 14
15
Doh-Won Lee, Ph.D. 16
Texas A&M Transportation Institute 17
18
Chaoyi Gu 19
Texas A&M Transportation Institute 20
21
and 22
23
Tara Ramani 24
Texas A&M Transportation Institute 25
26
27
28
29
30
31
32
33
34
35
36
37
38
Total words: 5,501 + (4 figures and 4 tables) × 250 = 7,501 39 40 41
* Corresponding author
Farzaneh, Johnson, Lee, Gu, and Ramani
2
ABSTRACT 1
The research presented in this paper is part of a broad study that investigated the 2
implications of Hybrid Electric Vehicles (HEVs) and Plug-in Electric Hybrid Vehicles 3
(PHEVs) in terms of mobile source emissions which were used to develop an approach 4
for incorporating EVs into regional emissions inventory procedures. The main objective 5
of the work presented in this paper was to characterize the tailpipe emissions of HEVs 6
and PHEVs. The research team focused on the Texas context, specifically with regards to 7
practices of conducting regional emissions inventories using the United States 8
Environmental Protection Agency’s (EPA) MOVES model. Researchers conducted an 9
extensive vehicle activity data collection exercise from a sample of EVs in major Texas 10
metropolitan areas. These data was used to develop more than 60 representative Texas-11
specific HEV and PHEV drive schedules. In-use emissions testing of four HEVs and four 12
PHEVs was then conducted using portable an emissions measurements system and 13
engine control unit loggers to obtain operating-mode-based emissions and electricity 14
consumption rates, which were combined with the drive schedules to obtain 15
distance-based emissions rates for each HEV and PHEV by speed bin and road class. 16
3
INTRODUCTION 1
The term electric vehicles (EVs), also sometimes termed electrified vehicles, refers broadly to 2
vehicles that obtain at least a part of the energy required for their operation from electricity. In 3
this research, EVs were defined as including hybrid electric vehicles (HEVs), plug-in hybrid 4
electric vehicles (PHEVs), and battery electric vehicles (BEVs). EVs were estimated to account 5
for 7.5 percent of annual new passenger car sales and 0.3 percent of annual new passenger truck 6
sales in 2013 (1,2). With new models coming into the market and increasing availability of 7
supporting infrastructure, it is expected that the number of EVs in the U.S. vehicle fleet will 8
continue to grow in the future. 9
By obtaining part or all of the energy needed for propulsion from electricity, EVs can potentially 10
achieve higher energy efficiencies and result in less exhaust emissions when compared with 11
conventional vehicles powered solely by internal combustion engines. However, there is limited 12
research on the emissions implications of increased EVs in the vehicle fleet. Depending on the 13
type of vehicle, the emissions that need to be considered include vehicle exhaust (tailpipe) 14
emissions as well as the emissions associated with the electricity generation used for the 15
charging of vehicle batteries in the case of plug-in hybrid and battery electric vehicles. Only 16
tailpipe emissions were considered in the work presented in this paper. 17
From the perspective of transportation agencies, the mobile source emissions component of EVs 18
exhaust emissions is particularly relevant, especially in nonattainment and attainment 19
maintenance areas needing to meet transportation conformity requirements. EPA’s MOVES 20
model forms the basis for emissions estimations used in the conformity analyses, state 21
implementation plan (SIP) development, and other mobile source emissions estimations 22
conducted in Texas and much of the United States. The current state of the practice in the use of 23
MOVES does not account for electric vehicles with regard to location-specific driving 24
characteristics, emission rates, and market penetration. However, MOVES provides a platform 25
and has the flexibility to accurately incorporate these aspects into emissions estimations. 26
The main objective of the work presented here was to characterize the tailpipe emissions of 27
HEVs and PHEVs in the context of regulatory-mandated regional emissions inventories in 28
Texas. A series of field studies were designed and executed to collect and analyze in-use data on 29
HEVs and PHEVs activity patterns and emission characteristics. Using GPS units, a series of 30
vehicle activity data collection was conducted from a sample of HEVs and PHEVs in major 31
Texas metropolitan areas. Researchers used the data to develop representative Texas-specific 32
electric vehicles drive schedules. The research team also conducted in-use emissions and power 33
consumption testing of the test vehicles using portable emissions measurements systems and 34
engine control unit (ECU) loggers to obtain operating-mode-based emissions and electricity 35
consumption rates. These data were analyzed according to the operating mode concept of the 36
MOVES model and then combined with the drive schedules to obtain distance-based emissions 37
rates for HEVs and PHEVs by speed bin and road class. 38
STATE-OF-THE-PRACTICE 39
First generation EVs were available in the U.S. in the late 1800s and early 1900s and they 40
actually outsold other types of cars between 1899 and 1900 (3). However, the rapid proliferation 41
Farzaneh, Johnson, Lee, Gu, and Ramani
4
of internal combustion engine-powered vehicles as a result of their affordability and extended 1
travel range forced the EVs out of the market by 1935. EVs started to reappear in the U.S. 2
market around the year 2000 mainly in response to high oil prices and air quality concerns and 3
the number of EVs in the U.S. have increased steadily since then. 4
Form the different types of EVs, HEVs are now the fastest-growing segment of the light-duty 5
vehicle market. The number of registered HEVs in the U.S. grew to nearly 2 million in 2011. 6
Virtually all major vehicle manufacturers are offering or are planning to offer plug-in EVs 7
(PHEVs and/or BEVs) in the U.S. market in addition to their HEV offerings. The major factors 8
influencing the current EV market share have been identified as energy cost (4,5), battery cost 9
and capacity (6,7), charging infrastructure (8,9), and government incentive programs (10,11). 10
The main distinction between HEVs and plug-in EVs is that HEVs do not use electricity from a 11
power grid, i.e., their batteries cannot be charged by plugging them into a power outlet (12). 12
HEVs use regenerative brakes and/or their internal combustion engine to charge their batteries 13
which have usually less than 8 kWh of energy capacity. HEVs use electricity when the vehicle is 14
traveling at a low speed and has the potential to reduce tailpipe emissions. Compared with 15
HEVs, a PHEV’s battery capacity is much larger, usually more than 20 kWh, and can obtain 16
electricity from a power grid in addition to regenerative braking and internal combustion engine. 17
Therefore, PHEVs’ average tailpipe emissions are lower than HEVs and conventional gasoline-18
powered vehicles because electricity is their major energy source. BEVs solely run on electricity 19
and therefore do not produce any tailpipe emissions. 20
The literature identifies two electric vehicle operating modes to describe the electric portion of 21
PHEV and BEV operations. Charge depletion (CD) mode is when the battery is charged above a 22
threshold and the vehicle is powered solely by the battery. Charge sustaining (CS) mode is when 23
the battery is discharged below a threshold and the vehicle is powered intermittently by a 24
gasoline-powered engine and the battery (Error! Bookmark not defined.). In the CS mode, the 25
internal combustion engine is working and emitting tailpipe emissions while in the CD mode 26
there are no tailpipe emissions. BEVs operate solely in the CD mode. PHEVs operate in the CD 27
mode when the battery is charged above a threshold, and after that, they operate in the CS mode. 28
Therefore, when only mobile source emissions are considered—which is the system boundary of 29
transportation air quality conformity analysis—EVs could potentially provide benefits in 30
reducing regional emissions depending on the source of the grid power. 31
Studies on the emissions impacts of EVs can be divided into disaggregate and aggregate 32
approaches. The disaggregate approach focuses on the emissions performance of small sample of 33
EVs compared to conventional gasoline-powered vehicles (13,14, 15, 16, 17). The methods used 34
to estimate EV emissions are mainly in-use testing using portable emissions measurement 35
systems (PEMS) and microsimulation models of traffic flow. The aggregated approach provides 36
estimates of the percentage of emissions reductions due to various EV market penetration 37
scenarios at regional or national levels (18,19, 20, 21, 22, 23). These studies are predominantly 38
top-down, meaning emissions benefits are solely related to the number of EVs without 39
considering how those EVs are used differently from their internal combustion counterparts. The 40
following is a sample of studies that researchers included in their literature review for this 41
research. The research team used this information to develop an approach to collect field data 42
and characterize tailpipe emissions of EVs. 43
Farzaneh, Johnson, Lee, Gu, and Ramani
5
Disaggregate Approach 1 Graver et al. developed a framework to estimate real-world emissions of a plug-in hybrid electric 2
vehicle (24). PHEV running exhaust emissions were measured by PEMS, and grid emissions 3
were based on the electricity consumption data and regional power grid resource mix. The study 4
reported distance-based emission rates of CO2, NOx, sulfur dioxide (SO2), and particulate matter 5
(PM) for running exhaust emissions only and running exhaust plus grid emissions. The results 6
showed the tested PHEV’s running exhaust emissions rates were 3 percent to 140 percent lower 7
than the corresponding regulatory limits. However, when grid emissions were included, the 8
distance-based emissions rates were similar or higher than the regulatory limits. 9
Robinson and Holmen compared second-by-second particle numbers from a 2010 hybrid electric 10
vehicle and a comparable conventional vehicle under the same real-world driving conditions 11
(25). The particle number concentrations were recorded using PEMS. The researchers found that 12
the average particle number per trip for the HEV was two times higher than the conventional 13
vehicle. The high particle number emissions from the HEV were mainly due to the restart 14
behavior at low or stop-and-go driving conditions, which resulted in air quality concerns in areas 15
such as intersections. 16
Karabasoglu and Michalek investigated the potential of HEVs, PHEVs, and BEVs in reducing 17
lifetime GHG emissions under various driving cycles and charging scenarios (26). The driving 18
cycles considered were the Urban Dynamometer Driving Schedule (UDDS), Highway Fuel 19
Economy Test, US06, and LA92. The results showed that EVs could achieve most CO2 20
emissions reductions under urban driving conditions (such as UDDS). At aggressive driving 21
conditions, such as US06, the benefits of EVs in reducing CO2 emissions diminished 22
significantly. 23
Millo et al. analyzed the CO2 emissions benefits and operating cost reduction of a PHEV and a 24
comparable conventional vehicle under real-world driving conditions (27). They developed an 25
optimal control model that optimized the PHEV’s driving and charging activities. The running 26
exhaust emissions and electricity consumptions of the PHEV under the New European Driving 27
Cycle (NEDC) were simulated using the MATLAB program, and grid emissions were later 28
calculated based on electricity consumptions. The results indicated the PHEV could reduce 10–29
30 percent of CO2 emissions depending on the electricity generation resource mix. 30
In a study by EPRI and NRDC, the authors concluded that a PHEV could reduce 40–65 percent 31
or more GHG emissions than a conventional vehicle and 7–46 percent or more than a HEV in 32
2050 if a large number of PHEVs enter the vehicle fleet from 2010 to 2050 (28). Findings of a 33
study by Thomas suggest that BEVs with a 300-mi range will have higher GHG emissions 34
compared with conventional vehicles if electricity generation is based on the current coal 35
technology (29). 36
Aggregate Approach 37 Doucette and McCulloch conducted a study that simulated BEVs’ CO2 emissions due to 38
electricity consumptions in four countries (US, France, China, and India) and compared those 39
CO2 emissions with similarly configured conventional ICE vehicles (30). They found that in 40
countries that have high CO2 emissions per unit of electricity (such as China and India), driving a 41
BEV could potentially lead to higher CO2 emissions compared with a conventional ICE vehicle. 42
Depending on factors such as BEV battery range and charging infrastructure locations, the 43
Farzaneh, Johnson, Lee, Gu, and Ramani
6
differences in CO2 emissions in driving a BEV could range from −2 percent to 30 percent 1
compared with a conventional ICE vehicle. 2
In another study conducted by Doucette and McCulloch, the authors developed a model that 3
evaluated the prospects of plug-in hybrid electric vehicles in reducing CO2 emissions. Compared 4
with a conventional ICE vehicle, a plug-in hybrid electric vehicle uses both an ICE and electric 5
motor as power sources (31). Hence, driving a plug-in hybrid electric vehicle could reduce fuel 6
demand and achieve higher energy efficiency. However, if the CO2 intensity in electricity is 7
higher than the transportation fuel, driving a plug-in electric vehicle could actually result in 8
higher CO2 emissions compared with driving a conventional ICE vehicle. 9
Ma et al. investigated the true ability of BEVs to reduce GHG emissions through a life cycle 10
assessment based on various BEV driving patterns in the United Kingdom and California (32). 11
The driving patterns included different profiles in speed, loading, accessory usage, and more. 12
The results showed that BEVs can deliver significant driving GHG emissions savings compared 13
with conventional vehicles under conditions in which the grid GHG intensity used to charge the 14
batteries is sufficiently low. The study also showed that BEVs perform best relative to ICE 15
vehicles in terms of driving GHG emissions at low speeds and lightly loaded driving. However, 16
the overall vehicle life cycle emissions are higher for BEVs than ICE vehicles due to the GHG 17
emissions associated with battery manufacturing. 18
Sharma et al. quantified the economic and greenhouse gas emission performance of 19
conventional, hybrid, and battery electric vehicles in Australia (33). The evaluations were based 20
on Australian specific driving conditions. They used the Powertrain System Analysis Toolkit 21
simulation package to simulate economic and greenhouse gas emission performance of various 22
types of vehicles. The results showed that for large vehicles such as light commercial vehicles, 23
the BEV had higher life cycle GHG emissions than an equivalent conventional vehicle. For 24
passenger cars, hybrid electric vehicles were the most effective in terms of cost and GHG 25
emissions under life cycle assessment. 26
Silver studied the impacts of introducing HEVs, PHEVs, and BEVs in terms of criteria 27
pollutants, such as CO, HC, NOx, PM, and CO2 in Portugal (34). The study assumed that the 28
emissions from each vehicle followed a probability distribution and executed a Monte Carlo 29
simulation to estimate emissions as a result of various increasing patterns of the EVs’ market 30
shares. The results indicated that 10 percent to 53 percent reductions in various criteria pollutants 31
could be achieved with a scenario of 50 percent fleet replaced with EVs. In addition, a 23 percent 32
reduction in CO2 could be achieved. 33
STUDY APPROACH AND RESULTS 34
To establish the MOVES-compatible Texas-specific emissions rates for HEVs and PHEVs, the 35
local distance-based emissions rates are required at each average speed bin and each road type. A 36
series of field data collections and a MOVES-based data analysis approach were developed and 37
executed to characterize the tailpipe emissions from a sample of HEVs and PHEVs. The field 38
data collections included an activity data collection using GPS loggers and an in-use emission 39
data collection using PEMS. Researchers developed a series of vehicle type specific drive cycles 40
based on the GPS records. The emissions measurements were translated into operating-mode-41
based emissions rates for each individual test vehicle. The research team then combined the 42
Farzaneh, Johnson, Lee, Gu, and Ramani
7
operating mode distributions from the drive schedules with operating-mode-based emission rates 1
to generate distance-based emission rates (g/mi) for individual vehicles by average speed and 2
road type. Market shares (within each category, i.e., HEVs/PHEVs) were estimated from 3
historical vehicle sales data and used to calculate the market-share weighted average emission 4
rates (g/mi) for each vehicle type category by average speed and road type. These emission rates 5
were then organized in MOVES-compatible emission rate lookup tables. The following provides 6
a summary of these steps. 7
Development of EV Drive Schedules 8 Researchers developed and executed a data collection protocol based on related literature and the 9
research team’s previous experiences, specifically Texas Department of Transportation (TxDOT) 10
Report 0-6629-1, Texas-Specific Drive Cycles and Idle Emissions Rates for Using with EPA’s 11
MOVES Model (35). The test protocol included an unsupervised GPS data collection for 12
developing the representative drive cycles and a supervised on-road emissions testing to develop 13
basic emission rates for EVs. 14
Data were collected from the following three categories of EVs: HEVs, PHEVs, and BEVs. 15
Vehicles from these categories were identified, recruited separately for each data collection 16
effort (drive cycles and emissions), and equipped with data loggers. The data were processed and 17
analyzed following the MOVES model format for vehicle activity characterization. 18
The data collection plan consisted of the following major items: 19
Vehicle sample sizes for each vehicle category. 20
Technology, methodology, and installation procedures for data collection. 21
Required duration of the data collection. 22
Procedures for protecting participants’ privacy. 23
The focus of the data collection effort was on the nonattainment and near-nonattainment urban 24
areas of Texas: Houston, Dallas–Fort Worth, Austin, and San Antonio. The research team used 25
the GPS units that were identified and obtained in TxDOT Project 0-6629, i.e., the QStarz BT-26
Q1000eX Xtreme Recorder (36). Based on the findings of TxDOT Project 0-6629 (37), 27
researchers determined that a two-week period would provide the necessary amount of data. 28
Each vehicle was equipped with three GPS data loggers to ensure completeness and continuity in 29
case one unit malfunctioned or provided erroneous observation. Information from each of the 30
three data loggers was downloaded and merged into one table that was labeled with variables 31
describing unit number, date of initial activation, and type of vehicle observed. 32
A total of 21 vehicles, 10 HEVs and 11 PHEVs, were recruited for the GPS data collection. GPS 33
data from each vehicle were collected for a period of one to two weeks, depending on the level 34
of driving activity. Table 1 shows the distribution of vehicles observed by vehicle type and area. 35
Farzaneh, Johnson, Lee, Gu, and Ramani
8
Table 1. Recruited Vehicles by Vehicle Type and Region. 1
Vehicle Type Austin
Dallas–
Fort Worth Houston San Antonio Total
HEV 2 6 – 2 10
PHEV 4 – 7 – 11
Total 6 6 7 2 21
Vehicle Type Ford Escape Toyota Prius Chevy Volt Total
HEV 6 4 – 10
PHEV – 7 4 11
Total 6 11 4 21
2 Potential participants for the research project were recruited through a variety of methods 3
including e-mail messages distributed to local electric vehicle enthusiast groups, web-based 4
classified listing Craigslist, and Facebook social media account maintained by the Texas A&M 5
Transportation Institute (TTI). 6
Data collection followed an unsupervised procedure in which drivers were instructed to follow 7
their normal driving activities for a period of two weeks. The result of this effort was a database 8
of second-by-second speed for all participating vehicles. The data processing and analysis were 9
conducted according to the methodology developed in TxDOT Project 0-6629 consisting of the 10
following general steps. Details on these steps are documented in the final research report (37). 11
1. Raw data quality control and validation. 12
2. Data processing. 13
3. Data analysis and drive schedule development. 14
4. MOVES default drive schedule comparison. 15
The goal of this effort was to develop drive schedules for each of the EV types according to 16
MOVES road types; i.e. rural and urban arterials and freeways/highways. Therefore, the location 17
information of each second of the data needed to have study area and road type information 18
assigned. The research team processed the data in a geographic information system (GIS) 19
environment for this purpose. The research team then developed drive schedules based on the 20
methodology documented in detail in the TxDOT Research Project 0-6629 final report. 21
An ideal drive schedule for a given driving condition is the one that has the maximum amount of 22
information regarding that condition, i.e., in the context of this study, the one that has all the 23
observations corresponding to EV driving conditions. However, using an entire database of 24
second-by-second speed data is impractical. Therefore, a sub-ideal solution is a continuous short 25
drive schedule constructed from a limited number of micro-trips, which will closely represent the 26
ideal solution. This suboptimal solution is easy to implement and is currently used in MOVES in 27
the form of default drive schedules. 28
In addition to second-by-second speed data, MOVES lets the user input vehicle activity 29
information in terms of the equivalent operating mode distribution. This method provides an 30
Farzaneh, Johnson, Lee, Gu, and Ramani
9
opportunity to implement the ideal solution (i.e., all observations) in a practical way. The 1
research team analyzed the data and developed both ideal and sub-ideal solutions. The final drive 2
schedules and target opMode distributions are submitted to TxDOT in a database format. 3
Researchers compared the resulting target opMode distributions (i.e., target) and drive schedules 4
(i.e., cycles) and to the default drive schedules of MOVES in terms of their corresponding 5
distribution of modal operating bins. Figure 1 shows an example of the results of this simple 6
comparison effort. Figure 1 show opMode distributions for the 50 mph speed bin on urban 7
arterial and freeway/highway from plug-in hybrid vehicles in urban area. 8
Plug-in Hybrid Vehicles—Arterial, Statewide Urban, Speed 50 mph
Plug-in Hybrid Vehicles—Freeway/Highway, Statewide Urban, Speed Bin 50 mph
Figure 1. OpMode Distribution Comparisons for Plug-in Hybrid Vehicles. 9 10
A visual review of the comparison charts suggested noticeable differences between the EV drive 11
cycles and MOVES default drive cycles which are based on data from combustion engine 12
0%
5%
10%
15%
20%
25%
30%
35%
40%
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
Fre
qu
ence
y P
erce
nta
ge
Operating Mode
Local_Target
Local_Cycle
MOVES
0%
5%
10%
15%
20%
25%
30%
35%
40%
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
Fre
qu
ence
y P
erce
nta
ge
Operating Mode
Local_Target
Local_Cycle
MOVES
Farzaneh, Johnson, Lee, Gu, and Ramani
10
vehicles. For example, the case shown in Figure 1 suggests that PHEVs have lower percentage of 1
high speed and high power instances (opMode bins 35 and higher) compared to their internal 2
combustion counterparts. The research team is planning to conduct a more in-depth analysis of 3
these differences and hope to publish the findings in a future paper. 4
Researchers also created the overall average drive schedules for each vehicle type, i.e., combined 5
all the valid observations into a single drive schedule for a certain scenario. The purpose of 6
generating the average drive schedule was to review the driving behavior at a macroscopic level. 7
Figure 2Error! Reference source not found. shows an example of the results of this effort. The 8
basic data statistics of each drive schedule and opMode distribution were also calculated. 9
10 Figure 2. OpMode Distribution Comparisons for Plug-in Hybrid Vehicles, Hybrid Vehicles, 11
and Pure Electric Vehicles on Urban Arterial Roads. 12 13
To ensure the representativeness of the results, the research team decided to discard the drive 14
cycle and opMode distributions that had less than 500 seconds of observations for a speed bin. In 15
such instances, a substitute opMode distribution from other road types would then be used to 16
calculate the emission rates for that speed bin. The following criteria were used to assign a 17
substitute for such cases: 18
Step 1. Use opMode distribution for the same speed bin and road class from the 19
alternative area type, e.g., if it is rural restricted, then use urban restricted. 20
Step 2. If Step 1 did not yield a valid opMode distribution, use opMode distribution 21
for the same speed bin and area type from the alternative road class, e.g., if it is rural 22
restricted, then use rural unrestricted. 23
0%
5%
10%
15%
20%
25%
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
Fre
qu
ence
y P
erce
nta
ge
Operating Mode
HEV_TARGET
PHEV_TARGET
Farzaneh, Johnson, Lee, Gu, and Ramani
11
Step 3. If Steps 1 and 2 did not yield a valid opMode distribution, use opMode 1
distribution for the same speed bin from the alternative road class and area type, e.g., 2
if it is rural restricted, then use urban unrestricted. 3
Step 4. If none of the above steps work, use the opMode distribution from the 4
MOVES database for the corresponding speed and road type. 5
Tailpipe Emissions Profiles of HEVs and PHEVs 6 The driving schedules described above provide an operating mode distribution for each road type 7
and average speed bin. Operating-mode-based emissions rates are required in order to estimate 8
distance-based emissions rates (i.e. grams/mile) for HEVs and PHEVs. Researchers developed 9
estimates of these operating-mode-based emissions rates based on in-use testing of HEVs and 10
PHEVs using PEMS equipment. The observations from in-use testing of EVs were used to 11
obtain emissions rates for each operating mode bin (in units of g/s of emissions). These rates 12
were then combined with operating mode distributions obtained from the drive schedules to 13
calculate distance-based emissions rates (in units of g/mi) at each speed bin for each road type. 14
The data collection protocol for developing emissions rates for EVs consisted of the following 15
major components: 16
Test procedures for driving activities. 17
Test procedures for idling. 18
Data collection equipment and installation procedures. 19
Vehicle samples from each vehicle category. 20
To ensure that sufficient data were collected for each vehicle, the driving tests were conducted as 21
a supervised data collection effort at TTI and surrounding areas in Bryan–College Station. This 22
allowed the team to collect data from each vehicle under real-world driving conditions while 23
giving the flexibility to collect significant amount of data in a relatively short period of time. 24
Each vehicle was transported to TTI’s Environmental and Emissions Research Facility (EERF) 25
at the Texas A&M Riverside Campus in Bryan, Texas, for approximately one week of testing. 26
Both the idling and driving tests were conducted during this period. 27
The procedures for driving activities were developed to ensure that the test vehicles operated in 28
each MOVES opMode bin in order to establish the emission rates. The total distance for each 29
test was approximately 31 mi. Each vehicle was driven on the test route for a minimum of three 30
runs in order to collect sufficient data for establishing the rates. 31
Twelve vehicles were recruited for data collection of in-use testing, including four vehicles from 32
each target vehicle category. The test vehicles were selected based on availability and because 33
they were considered to be as representative as possible of the existing overall fleet. In some 34
vehicle categories, such as the BEVs, where one make of vehicle is more prevalent than others, 35
multiple vehicles of the same make were recruited for testing. Table 2 shows the vehicle 36
specifications of the EV units that were tested. A Tesla was also considered for testing, but the 37
Farzaneh, Johnson, Lee, Gu, and Ramani
12
available data loggers needed to collect the information on the battery performance did not 1
support Tesla vehicles. 2
Table 2. Tested Vehicle Specifications. 3
PHEV
Vehicle Parameter 2012 Chevy Volt 2011 Chevy Volt 2012 Prius Plug-in
Electric Motor 111 kW, 149 HP 111 kW, 149 HP 60 kW, 80 HP
Battery 16 kWh 16 kWh 4.4 kWh
All Electric Range 35 mi 35 mi 15 mi
Internal Combustion
Engine 1.4 L, 84 HP 1.4 L, 84 HP 1.8 L, 98 HP
2013 Units Sold 23,094 12,088
HEV
Vehicle Parameter
2012
Ford Fusion
Hybrid
2012
Toyota Camry
Hybrid
2012
Ford Escape
Hybrid
2011
Toyota Prius
Engine 2.5 L, 156 HP 2.5 L, 156 HP 2.5 L, 155 HP 1.8 L, 98 HP
Electric Motor 106 HP 141 HP 94 HP 80 HP
Combined HP (electric
motor and ICE engine) 188 HP 200 HP 177 HP 134 HP
4
In-use testing provided second-by-second emissions, speed, and acceleration of tested vehicles 5
under real-world driving conditions. Each second of data was categorized into a operating mode 6
bin according to definitions used in MOVES. The calculations in this process use a series of 7
vehicle parameters including vehicle’s mass; i.e. sourceMass. The sourceMass of each test 8
vehicle was calculated by taking the vehicle’s curb weight and adding the weight of the PEMS 9
unit and the driver of the vehicle. 10
11
Farzaneh, Johnson, Lee, Gu, and Ramani
13
Table 3 shows the values of the vehicle parameters used in the calculations. By averaging 1
emissions under each operating mode, time-based EV emissions rates per operating mode bin 2
were estimated. 3
The PEMS units used in testing included a vehicle interface connection and GPS unit that 4
allowed for the collection of second-by-second speed data during all testing. These second-by-5
second speed profiles were then used to calculate a second-by-second vehicle-specific power 6
(VSP), which was then used to determine the opMode that the vehicle was in during a given 7
second of operation. The emissions data from each second were then averaged for each opMode 8
bin. The results were documented in a series of opMode-based emissions tables for each 9
individual vehicle. Figure 3 shows opMode-based CO emissions rates for HEVs as an example 10
of the results of this step. 11
12
Farzaneh, Johnson, Lee, Gu, and Ramani
14
Table 3. MOVES Variables for VSP Calculation. 1 Test Vehicle MOVES Variables
rollingTermA rotatingTermB dragTermC sourceMass
2011 and 2012
Chevy Volt 0.156461 0.00200193 0.0049265 1.71503275
2012
Toyota Prius Plug-In 0.156461 0.00200193 0.0049265 1.63250049
2012
Ford Escape Hybrid 0.221120 0.00283757 0.00698282 1.81346230
2012
Ford Fusion Hybrid 0.156461 0.00200193 0.0049265 1.86880056
2012
Toyota Camry Hybrid 0.156461 0.00200193 0.0049265 1.77309257
2011
Toyota Prius Hybrid 0.156461 0.00200193 0.0049265 1.56126494
2
3 Figure 3. HEV CO Emissions Rates. 4
5
The opMode emissions information were then combined with the opMode distributions 6
developed in the previous section to establish distance-based (i.e. grams/mile) emissions rates for 7
each average speed. Because the actual average speeds of the opMode distributions were slightly 8
different from the assigned speed of MOVES speed bins, and because not all bins had a local 9
opMode distribution, the team used a linear interpolation to calculate emission rates at the 10
assigned speeds of MOVES speed bins. Figure 4 shows a sample of the resulting distance-based 11
emission rates for individual vehicles. 12
< 25 mph <50 mph
Farzaneh, Johnson, Lee, Gu, and Ramani
15
1 Figure 4. CO2 Emission Rates for HEVs on Rural, Arterial Road. 2
Adjustment for PHEVs Charge Depleting Mode 3 The distance-based emissions rates calculated for PHEVs in the previous section were done only 4
for the charge sustaining mode where the vehicle switched between its internal combustion 5
engine and electric motor. PHEVs produce no tailpipe emissions when driving in the CD mode. 6
Understanding percentages of PHEV miles driven on electricity is important for estimating 7
PHEVs’ overall average distance-based tailpipe emissions rates (i.e. combined CD and CS 8
modes). 9
A study conducted by the US Department of Energy’s Clean Cities Program estimated that 10
70 percent of PHEV miles are driven on electricity. This study was based on travel behavior data 11
from a 2001 National Household Travel Survey (38). Several online electric vehicle consumer 12
forums let electric vehicle users report their EV mile reading from the odometer of their vehicles. 13
These data can be used to develop a relatively accurate estimate of the overall electric mode 14
percentage for PHEVs in the US market. Volt Stats.com collects data from electric vehicle users 15
and reports the total miles driven, electricity miles driven, and overall fuel economy (39). 16
Researchers obtained volt usage data for the entire US and three states (Texas, California, and 17
New York) and analyzed them to estimate the percentage of PHEVs’ electric mode operation. 18
Table 4 shows summary statistics of PHEV mileage and percentage of miles driven on electricity 19
in 2013, respectively. 20
The data indicate that the percentages of miles driven on electricity ranged between 74 percent 21
and 80 percent in the three states, and the national average was at 76.0 percent. Those real-world 22
observations are consistent with the 70 percent electric miles driven estimated by the National 23
Renewable Energy Laboratory, whose study was based on nationwide transportation survey data 24
and considered both electric vehicle battery technology evolvement and users’ driving behaviors 25
(38). 26
0
100
200
300
400
500
600
700
800
2.5 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
CO
2 (
g/m
i)
Average Speed (mph)
Camry 2012 (HEV)
Escape 2012 (HEV)
Fusion 2012 (HEV)
Prius 2011 (HEV)
Farzaneh, Johnson, Lee, Gu, and Ramani
16
1
Table 4. Summary of Miles Driven by PHEVs. 2
Average per Vehicle VMT US TX CA NY
Min 7 626 193 88
Max 149,090 69,271 78,399 57,500
Mean 16,670 16,467 15,270 16,985
Standard Deviation 12,322 12,707 13,245 12,770
Electric Mode VMT
Percentage US TX CA NY
Min 1.4 31.2 20.9 40.4
Max 100 99.8 99.0 97.4
Mean 76.0 74.2 79.8 75.7
Standard Deviation 14.8 18 14.2 16.7
3 Based on literature and analysis of PHEV user reported data, 76 percent (value of national 4
average) was selected as the share of PHEVs’ VMT on electricity. This means that 24 percent of 5
PHEVs’ VMT are driven using the internal combustion engine and thus producing running 6
exhaust emissions, while the other 76 percent of PHEVs’ VMT do not produce exhaust 7
emissions. Therefore, after obtaining the distance-based running exhaust emissions rates for 8
PHEVs, a reduction factor equaling the share of total PHEV miles driven using the ICE was 9
applied. This factor allowed equivalent distance-based emissions rates for PHEVs to be 10
established per VMT as a whole, instead of only for ICE engine operation. 11
SUMMARY AND CONCLUSIONS 12
The increasing presence of hybrid electric and plug-in hybrid electric vehicles the U.S. vehicle 13
fleet is expected to results in lower on-road mobile source emissions and energy consumption. 14
However, currently there is limited research on the energy consumption and emissions 15
implications of increased HEVs and PHEVs based on fine-resolution field data. This paper 16
presents the methodology and results of an investigation of in-use activity and tailpipe emissions 17
from HEVs and PHEVs. The work presented in paper is part of a broad study that examined the 18
implications of HEVs and PHEVs in terms of regional emissions inventories. 19
The study approach for the work presented in this paper consisted of two major steps; - 20
developing drive cycles for HEVs and PHEVs based on GPS data collected in Texas, and - 21
developing MOVES-compatible emission rate tables for HEVs and PHEV based on a series of 22
in-use emissions measurements. Using self-powered GPS units, researchers conducted an 23
extensive vehicle activity data collection exercise from a sample of EVs in major Texas 24
metropolitan areas. These second-by-second speed and location data from GPS units were 25
processed and analyzed according to a micro-trip-based methodology. More than 60 drive 26
schedules and their equivalent opMode distributions were developed each representing an 27
average speeds on different road type as defined by in the MOVES model. 28
In-use emissions testing of EVs was then conducted using a portable emissions measurements 29
system (PEMS) and engine control unit (ECU) loggers to obtain second-by-second emissions 30
and electricity consumption rates. Eight vehicles, 4 HEVs and 4 PHEVs, were recruited for data 31
Farzaneh, Johnson, Lee, Gu, and Ramani
17
collection of in-use testing. The test vehicles were selected based on availability and because 1
they were considered to be a good representative of the existing U.S. fleet of HEVs and PHEVs. 2
The in-use emissions measurements were conducted as a supervised data collection effort at 3
Texas A&M Riverside Campus and surrounding areas in Bryan–College Station. This allowed 4
the team to collect data from each vehicle under real-world driving conditions on various types 5
of roads while giving the flexibility to collect significant amount of emissions data in a relatively 6
short period of time. 7
The emissions data were analyzed following a MOVES-oriented approach; i.e. second-by-second 8
observations from in-use testing grouped according to MOVES’ operating mode bin (in units of 9
g/s of emissions). These rates were then combined with operating mode distributions obtained 10
from the drive schedules to calculate distance-based emissions rates (in units of g/mi) at each 11
speed bin for each road type. The opMode-based emissions rates were then combined with the 12
opMode distributions representing different road types and average speeds for HEVs and PHEVs 13
to establish distance-based (g/mi) emissions rates. To account for PHEVs’ operation on battery 14
power, i.e. when there are no tailpipe emissions, an adjustment was made based on a sample of 15
user-reported mileage information and literature. 16
ACKNOWLEDGEMENTS 17
The study reported in this paper was funded by the Texas Department of Transportation 18
(TxDOT) research program as part of TxDOT project 6763 (40). The authors would like to thank 19
Bill Knowles, Kevin Pete, and Wade Odell of TxDOT for their support on all stages of the study. 20
REFERENCES 21
1 Retail New Passenger Car Sales. Bureau of Transportation Statistics.
http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statist
ics/html/table_01_16.html. Accessed August 31, 2014.
2 US HEV and PHEV Sales by Model. Alternative Fuel Data Center.
http://www.afdc.energy.gov/data/10301. Accessed August 31, 2014.
3 U.S. Department of Energy (DOE). History of Electric Vehicles: The Early Years (1890–1930).
4 Kazimi, C. “Evaluating the Environmental Impact of Alternative-Fuel Vehicles.” Journal of
Environmental Economics and Management, 1997, Vol. 33(2), Page 163-185.
5 Chan, C. “The State of Art of Electric Hybrid, and Fuel Cell Vehicles.” Proceedings of the IEEE, 2007,
Page 704-718.
6 Garner, J. “Electric Car Driving Range.” http://www.plugincars.com/vehicle-driving-range.html,
accessed September 2013.
7 DOE Environmental and Emissions Research Facility (EERE). Plug-In Electric Vehicle Handbook—for
Fleet Managers, 2011.DOE/GO-102012-3273.
8 Kezunovic, M., S. Waller, and I. Damnjanovic. “Framework for Studying Emerging Policy Issues
Associated with PHEVs and Managing Coupled Power and Transportation Systems.” Green
Technologies Conference, 2010.
Farzaneh, Johnson, Lee, Gu, and Ramani
18
9 Pike Research. “Electric Vehicle Charging Equipment—Level 1 and Level 2, DC Fast Charging
(CHAdeMO), and Wireless EVSE: Market Analysis and Forecasts.”
http://www.navigantresearch.com/research/electric-vehicle-charging-equipment.
10 Institute for Energy Research. “Obama Administration Pushes Electric Vehicles.”
http://www.instituteforenergyresearch.org/2011/03/10/obama-administration-pushes-electric-vehicles/,
accessed August 2012.
11 Michalek, J., M. Chester, P. Jaramillo, C. Samaras, C.-S. N. Shiau, and L. B. Lave, 2011. “Valuation
of Plug-In Vehicle Life-Cycle Air Emissions and Oil Displacement Benefits.” Proceedings of the
National Academy Sciences of the United States of America, Vol. 108, No. 40 16554-16558.
12 Energy, U. S. D. O. One Million Electric Vehicles By 2015.
https://www1.eere.energy.gov/vehiclesandfuels/pdfs/1_million_electric_vehicles_rpt.pdf, accessed
May 11, 2014.
13 Robinson, M.K., and B.A. Holmen. “On-Board, Real-World Second-by-Second Particle Number
Emissions from a 2010 Hybrid and Comparable Conventional Vehicle.” Transportation Research
Record, 2009.
14 Karabasoglu, O., and J. Michalek. “Influence of driving patterns on life cycle cost and emissions of
hybrid and plug-in electric vehicle powertrains.” Energy Policy, 2013, Vol. 60, 445-461.
15 Millo, F., L. Rolando, R. Fuso, and F. Mallamo. “Real CO2 emissions benefits of end user’s operating
costs of a plug-in hybrid electric vehicle.” Applied Energy, 2014, Vol. 114, 563-571.
16 Strecker, B., A. Hausmann, and C. Depcik. “Well to wheels energy and emissions analysis of a
recycled 1974 VW Super Beetle converted into a plug-in series hybrid electric vehicle.” Journal of
Cleaner Production, 2014, Vol. 68, Pages 93-103.
17 Sonntag, D.B., H.O. Gao, and B.A. Holmen. “Modeling On-Road Particle Number Emissions from a
Hybrid Diesel-Electric Bus.” Transportation Research Record, 2007, pp. 40-48.
18 Electric Power Research Institute (EPRI) and National Resources Defense Council (NRDC).
Environmental Assessment of Plug-In Hybrid Electric Vehicles, Volume I: Nationwide Greenhouse
Gas Emissions. EPRI Report # 1015325, 2007.
19 Thomas, C. E. “Cost-Benefit Analyses of Alternative Light-Duty Transportation Options for the 21st
Century.” National Hydrogen Association Conference, Columbia, SC, March 31, 2009.
20 Doucette, R.T., and M.D. McCulloch. “Modeling the CO2 emissions from battery electric vehicles
given the power generation mixes of different countries.” Energy Policy, 2010, Vol. 39, 803-811.
21 Doucette, R.T., and M.D. McCulloch. “Modeling the prospects of plug-in hybrid electric vehicles to
reduce CO2 emissions.” Applied Energy, 2011, Vol. 88, 2315-2323.
22 Jansen, K.H., T.M. Brown, and G.S. Samuelsen. “Emissions impacts of plug-in hybrid electric vehicle
deployment on the U.S. western grid.” Journal of Power Sources. 2010, Vol. 195, 5409-5416.
23 Kim, J.D., and M. Rahimi. “Future energy loads for a large-scale adoption of electric vehicles in the
city of Los Angeles: Impacts on greenhouse gas (GHG) emissions.” Energy Policy, 2014, Vol. 73,
620-630.
24 Graver, B.; Frey, H. C.; Choi, H. In-use Measurement of Activity, Energy Use, and
Emissions of a Plug-in Hybrid Electric Vehicle. Environmental Science & Technology. 2011, 45
(20), 9044-9051.
Farzaneh, Johnson, Lee, Gu, and Ramani
19
25
Robinson, M.K., B.A. Holmen. On-Board, Real-World Second-by-Second Particle
Number Emissions from a 2010 Hybrid and Comparable Conventional Vehicle. Transportation
Research Record, 2009.
26
Karabasoglu, O., J. Michalek. Influence of driving patterns on life cycle cost and
emissions of hybrid and plug-in electric vehicle powertrains. Energy Policy, 2013, Vol. 60, 445-
461.
27
Millo, F., L. Rolando, R. Fuso, F. Mallamo. Real CO2 emissions benefits of end user’s
operating costs of a plug-in hybrid electric vehicle. Applied Energy, 2014, Vol. 114, 563-571.
28
EPRI and NRDC. Environmental Assessment of Plug-In Hybrid Electric Vehicles,
Volume I: Nationwide Greenhouse Gas Emissions, EPRI Report # 1015325, 2007.
29
Thomas, C. E. Cost-Benefit Analyses of Alternative Light-Duty Transportation Options
for the 21st Century. National Hydrogen Association Conference, Columbia, SC, March 31,
2009.
30
Doucette, R.T., M.D. McCulloch. Modeling the CO2 emissions from battery electric
vehicles given the power generation mixes of different countries. Energy Policy, 2010, Vol. 39,
803-811.
31
Doucette, R.T., M.D. McCulloch. Modeling the prospects of plug-in hybrid electric
vehicles to reduce CO2 emissions. Applied Energy, 2011, Vol. 88, 2315-2323.
32
Ma, H., F. Balthasar, N. Tait, X. Riera-Palou, A. Harrison. A new comparison between
the life cycle greenhouse gas emissions of battery electric vehicles and internal combustion
vehicles. Energy Policy. 2012, Vol. 44, 160-173.
33
Sharma, R., C. Manzie, M. Bessede, R.H. Crawford, M.J. Brear. Conventional, hybrid
and electric vehicles for Australian driving conditions. Part 2: Life cycle CO2-e emissions.
Transportation Research Part C. 2013, Vol. 28, 63-73.
34
Silve, C. Electric and plug-in hybrid vehicles influence on CO2 and water vapour
emissions. International Journal of Hydrogen Energy. 2011, Vol. 36, 13225-13232.
35
M. Farzaneh, J.A. Zietsman, D. Lee, J.D. Johnson, N. Wood, T.L. Ramani, C. Gu. Texas-
Specific Drive Cycles and Idle Emissions Rates for Using with EPA's MOVES Model—Final
Report. 0-6629-1. Texas A&M Transportation Institute, College Station, TX. May 2014.
36
BT-Q100eX Users Manual, http://www.qstarz.com/download/BT-Q1000eX Users
Manual.pdf.
Farzaneh, Johnson, Lee, Gu, and Ramani
20
37
M. Farzaneh, J.A. Zietsman, D. Lee, J.D. Johnson, N. Wood, T.L. Ramani, C. Gu. Texas-
Specific Drive Cycles and Idle Emissions Rates for Using with EPA's MOVES Model—Final
Report. 0-6629-1. Texas A&M Transportation Institute, College Station, TX. May 2014.
38
Alternative Fuels Data Center, http://www.afdc.energy.gov/vehicles/search/light/.
Accessed January 31, 2014.
39
Chevy Volts online forum. http://www.voltstats.net/. Accessed May 4, 2014.
40
Farzaneh, Reza, Yuche Chen, Jeremy Johnson, Josias Zietsman, Chaoyi Gu, Tara Ramani, L.
D. White, Megan Kenney, and Yue Zhang. Accounting for Electric Vehicles in Air Quality
Conformity. No. FHWA/TX-15/0-6763-1. 2014.