Fuel Tank Temperature Profile
Development for Highway Driving
Final Report
&EPA
United States
Environmental Protection
Agency
-------
Fuel Tank Temperature Profile
Development for Highway Driving
Final Report
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
Prepared for EPA by
SGS-Environmental Testing Corporation (SGS-ETC)
Eastern Research Group (ERG)
EPA Contract No. EP-C-12-017
Work Assignment No. 1-05
NOTICE
This technical report does not necessarily represent final EPA decisions or
positions. It is intended to present technical analysis of issues using data
that are currently available. The purpose in the release of such reports is to
facilitate the exchange of technical information and to inform the public of
technical developments.
United States
Environmental Protection
Agency
EPA-420-R-14-026
October 2014
-------
Table of Contents
1.0 Objectives and Background 1
2.0 Study Equipment and Preparation 2
2.1 Test Vehicles 2
2.2 Laboratory and Test Equipment Overview 3
2.3 Fuel Procurement and Preparation 3
2.4 Vehicle Preparation 3
3.0 Test Program 4
3.1 Testing Overview 4
3.2 Data Collection Process 7
3.3 Data Validation and Analysis 7
3.4 Results 8
3.4.1 Honda Accord Test Results 9
3.4.2 Dodge Caravan Test Results 10
3.4.3 Toyota Corolla Test Results 11
3.4.4 Ford Focus Test Results 12
3.4.5 Chevrolet Silverado Test Results 13
3.4.6 Observations and Conclusions 14
4.0 References 16
5.0 Index of Appendices 17
List of Tables
Table 1. Test Vehicle Summary 2
Table 2. Test Vehicle Details 2
-------
Glossary of Terms and Acronyms
CDPHE Colorado Department of Public Health and the Environment
CFR Code of Federal Regulations
CRADA Cooperative Research and Development Agreement
CRC Coordinating Research Council
DAQ Data Acquisition System
DTP Ford's Driveability Test Facility
ECU Engine Control Unit
EPA US Environmental Protection Agency
ERG Eastern Research Group
FTTP Fuel Tank Temperature Profile
HC Hydrocarbon
LA-92 LA92 "Unified" Dynamometer Driving Schedule
MOVES Motor Vehicle Emissions Simulator
NREL National Renewable Energy Laboratory
OBDII Second-Generation On-Board Diagnostic System
OEM Original Equipment Manufacturer
PZEV Partial Zero-Emissions Vehicle
QAPP Quality Assurance Project Plan
RVP Reid Vapor Pressure
SGS-ETC SGS- Environmental Testing Corporation
US-06 US-06 Supplemental Federal Test Procedure
VECI Vehicle Emissions Control Information
VIN Vehicle Identification Number
-------
Executive Summary
This study, performed by Eastern Research Group (ERG) and subcontractor SGS-
Environmental Testing Corporation (SGS-ETC), under contract to the US Environmental
Protection Agency (EPA), was designed and conducted by EPA to develop fuel tank temperature
profiles that could be used for running loss emissions testing and/or modeling to represent
highway driving. Additional tests were performed to determine the repeatability of certification
temperature profiles developed by manufacturers and used in previous test programs. This work
builds on prior evaporative emissions test programs performed to determine the prevalence and
emission rates of diurnal and hot soak emissions in US vehicles1.
The testing for this program took place at Ford's Driveability Test Facility (DTF) in
Allen Park, Michigan under contract to ERG and subcontractor SGS-ETC. Five vehicles
supplied by EPA were tested using fuel with an RVP between 8.0 and 8.7 psi and an ethanol
volume content of 10 +/- 0.2. The test cell simulated the following conditions of ambient air
r\
temperature of 95 °F, solar loading of 1000 W/m , with a thermal mat to simulate pavement
temperature of 125 ° Fahrenheit. Each vehicle was driven over three different drive cycles of
similar length. The first cycle was the standard running loss cycle (FTTP) consisting of one
Urban Dynamometer Driving Schedule (UDDS), a 2-minute idle, two New York City Cycles,
another 2-minute idle, another UDDS, and then a final 2-minute idle (see §86.115). The second
cycle was three LA-92 cycles run back to back. The third sequence consisted of two US-06
cycles, a 70 mph highway cruise, and two final US-06 cycles. Fuel tank temperature, vapor
temperature, fuel tank pressure, and OBDII commanded purge (or purge solenoid voltage when
OBDII commanded purge was not available) were recorded throughout testing.
Results from the study showed the following:
• In general, higher-speed, more aggressive drive cycles result in lower final tank
temperatures, possibly due to higher wind speeds under the vehicle (and thus
higher amounts of convective cooling from the fuel tank)
• Fuel tank temperature profiles are strongly influenced by experimental conditions
and vehicle setup. Highly-controlled road surface temperatures with known set
points are necessary in order to replicate previously-developed tank temperature
profiles for specific drive cycles
• Although not performed for this study, it may be possible to use mathematical
models to "normalize" empirically-derived fuel tank temperature profiles in order
to approximate profiles developed under different, or unknown, conditions
in
-------
1.0 Objectives and Background
Over the past decade, the Environmental Protection Agency (EPA) and others have been
conducting studies to characterize real world evaporative emissions on aging newer technology
vehicles and their effects on the hydrocarbon (HC) inventory. The E-77 test program1 recently
performed by the Coordinating Research Council (CRC), the EPA and the Department of
Energy's National Renewable Energy Laboratory (NREL) collected emission rates of aging
enhanced evaporative emissions control and partial zero-emissions vehicle (PZEV) technology
vehicles at various temperatures using different fuels both with and without 0.020" diameter
leaks implanted at different locations over various evaporative test procedures including the
static permeation test, the diurnal test and the dynamic permeation test. A 0.020" leak diameter
was selected because this is the smallest size required to be detectable by vehicle OBDII
systems. In addition, recent studies of vehicles with high evaporative emissions in Denver,
Colorado were performed as part of a Cooperative Research and Development Agreement
(CRADA) between the EPA and the Colorado Department of Public Health and the Environment
(CDPHE) over three summers (2008-2010)2. These studies provided information on the
prevalence of evaporative emissions leaks and the emission rates of those leaks in the real world
fleet.
Work Assignment 1-08 under this contract also built on this work in order to simulate
real-world evaporative emissions leak rates during transient operation and static tests3. The leak
rates were collected using fuels with high and low RVPs on vehicles with various implanted leak
sizes and locations in order to determine running loss, hot soak and static emissions from
induced leaks on newer technology vehicles with enhanced evaporative emissions control
systems. Prior to that, EPA completed a multi-day diurnal test program to evaluate diurnal
emissions of vehicles which sit for more than the three days required in the certification test
cycle4. During this multi-day diurnal test program, nine vehicles were tested over a fourteen day
cycle with two fuels of varying RVPs.
EPA is accumulating this information to more accurately model evaporative emissions in
their Motor Vehicle Emissions Simulator (MOVES) emissions model. It has become apparent
through these studies that real world loading of the canister and purge strategies for on and off-
cycle emissions can have a significant impact on evaporative emissions. CRC's E-77-2
programs showed that running loss emissions with leaks can vary by orders of magnitude with
temperature, fuel volatility, and even leak location5. EPA is interested in appropriately applying
these large swings across the vehicle miles traveled in MOVES, and therefore is seeking
-------
temperature profiles for different modes of highway driving which tend to be cooler than the
certification running loss cycle for which emissions data were collected in WA 1-08.
This current work assignment (WA 1-05) was performed to develop evaporative
emissions testing tank temperature profiles for a different drive cycle than that used for
certification testing. In this work, the ERG / SGS team collected data in an environmentally
controlled test cell in order to develop fuel tank temperature profiles on a set of five vehicles
over non-standard drive cycles developed for this study. Internal tank pressure and OBDII
commanded purge (or purge solenoid voltage when OBDII commanded purge was not available)
were also collected during the drive cycle.
2.0 Study Equipment and Preparation
2.1 Test Vehicles
The EPA provided 5 vehicles to be tested in this study. The vehicles were chosen from
available vehicles that had previously been used in prior EPA test programs. The five vehicles
provided for the test program are listed in Table 1, and details regarding the selected vehicles and
test parameters are listed in Table 2.
Table 1. Test Vehicle Summary
Vehicle Make and
Model
Dodge Caravan
Toyota Corolla
Ford Focus PZEV
Honda Accord
Chevrolet Silverado
Model
Year
2007
2009
2010
2007
2006
Approx.
Odo
117k
121k
29k
124k
112k
Emissions
Standard
Tier 2 / Bin 5
Tier 2 / Bin 5
SULEV II PZEV
Tier 2 / Bin 5
Tier 2 / Bin 8
Canister
Cap1 (g)
177
115
110
140
177
Tank Vol
feal)
20.00
13.25
13.00
17.00
26.00
Canister/Tank
Ratio2
8.85
8.68
8.46
8.24
6.81
1 Canister Cap = canister working capacity, in grams
2 Canister working capacity (g) / Tank volume (gal)
Table 2. Test Vehicle Details
Vehicle
Caravan
Corolla
Focus
Accord
Silverado
Model
Year
2007
2009
2010
2007
2006
Inertia
Weight
4750
3250
3000
3500
5500
Tank
Capacity
20.00
13.25
13.00
17.00
26.00
Road
Load A
15.32
11.93
4.01
9.76
1.44
Road
LoadB
0.0948
0.0068
0.5575
0.2918
1.2678
Road
LoadC
0.02662
0.02276
0.01269
0.01602
0.02258
Engine Family
7CRXT03.8NEO
9TYXV01.8BEA
AFMXV02.0VZX
7HNXV02.4KKC
6GMXT05.3379
Evap Family
7CRXR0177GHA
9TYXR0115P12
AFMXR0110GCX
7HNXR0140BBA
6GMXRO 176820
-------
2.2 Laboratory and Test Equipment Overview
All testing was performed at Ford's DTP. The Ford DTP operates 24 hours per day, 7
days a week on a three shift per day schedule with reduced staffing on the weekends. The Ford
DTF has six vehicle labs equipped with a variety of features including chassis dynamometers,
extreme environmental and thermal simulation, wind simulation, heated road simulation and
barometric pressure simulation. The Ford DTF has the ability to operate light to medium-duty
two-wheel drive vehicles (2,000 to 32,000 Ibs. GVWR) under normal to extreme ambient
environmental conditions. All test cells are equipped with single 63" roll diameter electric
dynamometers. The test cell used to develop the fuel tank temperature profiles required for this
study can simulate a hot road surface. Ford DTF is not currently ISO accredited, but follows
quality standards developed during a former period of accreditation. To verify compliance with
the Tank Temperature Profile Development procedure defined in 40 CFR §86.129-94, SGS-ETC
provided an on-site technical lead at the DTF throughout the study to supervise the tests
performed by the DTF. The maintenance, calibration and verification of the measurement
equipment used in this study conformed to requirements defined in the work plan and quality
assurance project plan (QAPP) developed for this project.
2.3 Fuel Procurement and Preparation
SGS-ETC used a fuel provided by the Ford DTF which conformed with requirements set
forth in 40 CFR§86.113-04(a)(l) for fuel tank temperature profile development. This fuel had
an RVP between 8.0 and 8.7 psi and an ethanol volume content of 10 +/- 0.2. Additional fuel
specifications are provided in Appendix A.
2.4 Vehicle Preparation
SGS-ETC took the following steps to prepare each vehicle for tank temperature profile
development:
1. Create and prepare log books for recording and noting vehicle specifications and
process chronology
2. Check vehicles to verify safe operation on dynamometer
3. Document vehicle information (VTN, year, make, model, engine and evaporative
families); take pictures of the vehicles, VTN plate and the emissions control
system (VECI) labels
4. Read and record all OBDII diagnostic trouble codes and readiness codes
-------
5. Perform pressure test of evaporative emissions system using the Snap-On leak
detection unit (supplied by EPA) and record results
6. Install a fuel drain for changing fuel
7. Install two (2), type J, fuel tank thermocouples (one in the liquid and one in the
vapor space at a 40% fill)
8. Install tank pressure monitoring equipment and an induced pressurization port.
9. Fit ports on the fuel tank and in the vapor collection canister inlet for installing
induced leak orifices
10. Photograph the locations of the modifications to the vehicles' OEM configuration
11. Perform pressure test of evaporative emissions system using a Snap-On leak
detection unit (supplied by EPA) and record results
12. Check and adjust fluid levels and filters
13. Derive the appropriate vehicle road load for dynamometer testing
3.0 Test Program
3.1 Testing Overview
Test procedures conformed to the following steps.
Step 1) Drain and Refuel: An external pump connected to the fuel tank drain quick
connect was used to completely drain the tank. The tank was then fueled to 40% of tank
capacity with the specified fuel, and the vehicle was then placed into soak.
Step 2) Soak: The vehicle was soaked in a temperature-controlled environment to
stabilize fuel temperatures to 95 ±3 °F. Fuel temperatures were held at 95 ±3 °F for a minimum
of one hour before the beginning of a tank temperature profile development test.
Step 3) Tank Temperature Profile Development Test: A test was then performed
employing the following steps and procedures:
1. The test vehicle was moved onto the dynamometer. This process involved turning
the engine on and operating the vehicle for less than 60 seconds per the allowance
in §86.129-94(d)(4)(iv), "Fuel Temperature Profile, Profile Determination
Procedure". The vehicle engine compartment cover and any windows, doors, and
luggage compartments were closed.
-------
Fans were positioned as described in §86.135-90(b) and §86.107-96(d),
"Sampling and Analytical Systems; Evaporative Emissions, Fuel Temperature
Control System".
The vehicle air conditioning system (if so equipped) was set to the "normal" air
conditioning mode and adjusted to the minimum discharge air temperature and
high fan speed. Vehicles equipped with automatic temperature controlled air
conditioning systems were set to operate in "automatic" temperature and fan
modes with the system set at 72 °F as described in §86.129-94(d)(4)(iii).
The temperature of the liquid fuel was monitored and recorded at least every 1
second with the temperature recording system specified in §86.107-96(e). The
vapor temperature was monitored for reference only and was not used as a
process variable for controlling tank temperature.
When the ambient temperature was at least 95 °F (35 °C) and the fuel tank
temperature was 95±3 °F, the tank temperature profile development test was
started.
The tank temperature profile development test was conducted by operating the
test vehicle through two different drive cycles - repeating this entire procedure for
each cycle. The transmission was operated according to the specifications of
§86.128, "Transmissions", during the driving cycles. The driving cycles specified
were three LA92 cycles run back to back, then a US06, thirty minutes of 70 mph
highway cruise, followed by another US06. These are shown in the following
diagrams;
(a) Three LA92 cycles run back to back
374 748 1122 1496 1826 2156 2486 2816
Time(s)
3189 3562 3935 4308
-------
(b) US06, thirty minutes of highway cruise at 70 mph, another US06.
0 374 748 1122 1496 1826 2156 2486 2816 3189 3562 3935 4308
Time(s)
7. The ambient temperature was maintained at 95 °F during the tank temperature
profile development test and no more than a 2 °F drop was allowed during the
procedure.
8. The following parameters were measured and recorded during tank temperature
profile development
(a) Date and time of vehicle fueling
(b) Odometer reading at vehicle fueling
(c) Date and time car was parked, parking location
(d) Odometer reading at vehicle parking
(e) Date and time engine was started
(f) Time of initiation of selected cycle
(g) Time of completion of selected cycle
(h) Ambient temperatures throughout the period of the selected cycle
(i) Simulated wind speed throughout the period of the selected cycle
(j) Simulated surface temperature throughout the period of the selected cycle
(k) Trace Speed
6
-------
(1) Actual Speed
(m) Fuel Liquid Temperature
(n) Fuel Vapor Temperature
(o) Fuel Vapor Pressure
3.2 Data Collection Process
Various data files were collected during this study and have been provided to EPA as
separate deliverables to this study (Appendix D). All data files are in either tab delimited, or
comma separated variable format and consist of the following:
• Ambient temperatures, fuel liquid and vapor temperatures, and purge valve
voltages recorded by the Omega DAQ used during testing;
• Test dates, times, trace and vehicle speeds, wind speeds, surface temperatures and
fuel vapor pressures recorded by Ford DTF's data acquisition system; and
• OBD test data (including commanded purge, when available) recorded by the
HEM Data OBD Mini Logger
As described in the Appendix D data description, data files have been provided for the
FTTP, LA92 and US06 tests conducted during the study.
3.3 Data Validation and Analysis
Data was validated by ensuring the road surface mat temperatures did not fall beneath
125 °F and by ensuring the ambient temperatures did not fall beneath 93 °F during testing. The
trace was checked for driver violations.
Nonlinear regression was performed on the data to improve understanding of the
interaction of the final tank temperature with the mat temperatures, the proximity of the fuel tank
to the road surface, and the surface area of the fuel tank exposed to the road surface. Even though
a statistical model with a high coefficient of determination was developed with the
aforementioned variables, the influence of many other parameters, such as underbody wind,
exhaust system configuration, fuel tank materials and volume of fuel in the tank on tank
temperatures were not available for consideration. For these reasons, use of this model would
not be appropriate across broad ranges of inputs, and this model was, therefore, not included in
this report.
-------
Additional issues encountered during the study were investigated and resolved as
described in Appendix B, Issues Encountered and Solutions.
3.4 Results
The following subsections provide results for the LA-92/US-06 testing and the FTTP
testing. Results presented include separate graphs of fuel tank temperatures, fuel tank vacuum,
and commanded purge (based on a running 20-second average of the OBD datastream (or purge
solenoid voltage) versus time. Speed of each of the three traces is plotted on each of the graphs
for reference.
-------
3.4.1 Honda Accord Test Results
Original FTTP Acquired for WA 1-08
8/5/13 3 X LA-92 (8 psi RVP)
8/6/13 US - 06 + Cruise (8 psi RVP
8/27/13 FTTP (8 psi RVP)
Q.^.^^^
rflAT... fesBMi&S
118.9
117.0
114.5
1826 2156
Time (s)
2486 2816
0.40
0.35
0.30
0,25
0,20
0,15
0.10
0.05
jio.oo
£-0.05
1/25/13 FTTP (7 psi RVP) from WA 1-08
8/5/13 3 X LA - 92 (8 psi RVP)
8/6/13 US - OS + Cruise (8 psi RVP)
8/27/13 FTTP (8 psi RVP)
1496 1826 2156 24S6 2816
Time(s)
yu .
8/5/13 3 X LA - 92 (8 psi RVP)
8/6/13 US - 06 + Cruise (8 psi RVP)
8/27/13 FTTP (8 psi RVP)
*^^ • l
20 «
D S
1496 1826 2156 2485 2S16 3189 3562 3935 43 Of
Time (s)
-------
3.4.2 Dodge Caravan Test Results
130
125
120
115
110
105
"lOO
Original FTTP Acquired for WA 1-08
7/22/13 3 X LA - 92 (8 psi RVP)
8/5/13 VOID US -06 + Cruise (Spsi RVP)
- 8/6/13 FTTP (Spsi RVP)
8/6/13 US-06 +Cruise (Spsi RVP)
116.4
114.5
1496 1826 2156 2486
Time(s)
0.30
0.25
0,20
0.15
0,10
0.05
0.00
.
-0,05
1/29/13 FTTP (7 psi RVP) from WA 1-08
7/22/13 3 X LA - 92 (8 psi RVP)
8/5/13 VOID US-06 +Cruise (Spsi RVP)
8/6/13 FTTP (Spsi RVP)
8/6/13 US - 06 + Cruise (Spsi RVP)
20 g
0 £
1826 2156
Time(s)
3552 3935
8/5/13 VOID US-06 +Cruise (Spsi RVP)
8/6/13 FTTP (8psi RVP)
8/6/13 US - 06 + Cruise (Spsi RVP)
1122 1496 1826 2156 2486 2816 3189 3562 3935
10
-------
3.4.3 Toyota Corolla Test Results
-Original FTTP Acquired forWAl-QS
7/22/13 VOID 3 X LA-92 (8 psi RV
-7/23/13 3 X LA-92 (8 psi RVP)
8/5/13 US - 06 + Cruise (8 psi RVP)
--S/27/13 3 X LA-92 (8 psi RVP)
••9/24/13 FTTP (8 psi RVP)
r
'
^
0.20
I-/:.
B
-1/31/13 FTTP (7 psi RVP) from WA 1-08
7/22/13 VOID 3 X LA-92 (8 psi RVP)
-7/23/13 3 X LA-92 (8 psi RVP)
8/5/13 US-06 +Cruise (8 psi RVP)
-8/27/13 3 X LA-92 (8 psi RVP)
-9/24/13 FTTP (8 psi RVP)
A Ai
20 O
2156
Time (s)
2486 2816 3139 3562 3935 4308
- 7/23/13 3 X LA -92 (8 psi RVP)
8/5/13 US -06 + Cruise (8 psi RVP)
'"V rt^f
wjwUll LfT1
2155
Time(s)
11
-------
3.4.4 Ford Focus Test Results
130
125
120
115
110
105
Original FTTP Acquired for WA 1-08
7/22/13 VOID 3 X LA - 92 (S psi RVP)
7/23/13 3 X LA - 92 (8 psi RVP)
• 8/5/13 VOID US - 06 + Cruise (8 psi RVP)
• 8/6/13 FTTP (Spsi RVP)
8/6/13 US - 06 + Cruise (Spsi RVP)
9/24/13 FTTP (8 psi RVP)
119.Q
11S.S
Sr
&
2156
Time (s)
1/30/13 FTTP (7 psi RVP) from WA 1-08
7/22/13 VOID 3 X LA - 92 (8 psi RVP)
7/23/13 3XLA-92|8psiRVP
8/5/13 VOID US-06 +Cruise (8 psi RVP|
8/6/13 FTTP (Spsi RVP)
8/6/13 US- 06 + Cruise (Spsi RVP)
9/24/13 FTTP (8 psi RVP)
m/^W
-------
3.4.5 Chevrolet Silverado Test Results
- Original FTTP Acquired for WA 1-08
-7/23/13 3XLA-92(8psiRVP)
8/5/13 US-06 +Cruise (SpsiRVP)
-- 3/27/13 FTTP (8 psi RVP)
- 8/27/13 3 X LA - 92 (8 psi RVP)
114.3
113.6
112.8
^^
Ln
2156
Time(s)
1
|-0.05
S
Original FTTP Acquired for WA 1-08
7/23/13 3 X LA-92 (8 psi RVP)
8/5/13 US - 06 + Cruise (Spsi RVP)
-8/27/13 FTTP (Spsi RVP)
- 5/27/13 3 X LA - 92 (8 psi RVP)
A^
LJ v
' /""'n
"f~ --- \ ' f™< ..... ' ' J**""^ .................. /^ .....
20 «
0 ,E
1S26 2156
Time (s)
^
7/23/13 3 X LA - 92 (8 psi RVP)
8/5/13 US-06 +Cruise (Spsi RVP)
8/27/13 FTTP (8 psi RVP)
8/27/13 3 X LA - 92 (8 psi RVP)
1S26 2156
Time (s)
13
-------
3.4.6 Observations and Conclusions
Fuel tank temperature profile is influenced by a number of factors, such as the following:
• road surface temperature (or heated mat / simulated road surface temperature)
• distance from road surface to tank
• spatial arrangement, including exposed area of fuel tank to heated road surface
• fuel tank material and associated heat transfer properties
• volume of fuel in tank
• rate of heat generation from in-tank fuel pump
• proximity of exhaust to fuel tank
• arrangement of associated heat shielding, and
• volume and speed of air flowing under tank (which influences convective
cooling)
Although many of these parameters are fixed based on the vehicle type, the volume and
speed of air flowing under the tank will vary based on the drive cycle (speed of vehicle) while
the road surface temperature can vary from test to test.
This study demonstrated the difficulties in replicating the generation of OEM fuel tank
temperature profiles through an in-lab simulation, in particular without knowledge of actual road
surface temperatures and wind conditions from the original on-road temp profile development
work. It's also possible that the OEM fuel tank temperature profiles were numerically modified
(smoothed) before use in evaporative certification testing.
Further complications such as limited laboratory control of the road surface mat
temperature and potential temperature gradients in the laboratory flow tunnel resulted in
additional deviations between fuel tank temperature profiles developed during this study and
those originally developed by the original vehicle manufacturers.
As described in Appendix B, the Toyota Corolla showed increased temperatures for the
more aggressive drive cycles, an effect unique to this vehicle. All other vehicles demonstrated a
decrease in the tank temperature for the more aggressive cycles, likely due to the increased air
flow (and thus convective cooling) at the underside of the fuel tank. Further investigation
revealed the exhaust heat shields for the Corolla had been removed during a prior study, resulting
in exhaust system radiation heating of the fuel tank. This likely caused the tank temperature
profiles from the more aggressive traces to show increases above the original certification
profile. After new heat shields were acquired and installed, testing showed very good agreement
14
-------
between this study's temperature profile and the original certification temperature profile.
Figure 3.4.3 shows a corresponding increase in fuel tank pressure (reduction in tank vacuum) due
to the purge being overwhelmed from the excessive heating.
The Chevrolet Silverado (and to a lesser extent the Toyota Corolla) exhibited steep
temperature rate increases near the beginning of temperature development tests. This was seen
on all the Silverado tests, but was primarily limited to one of the LA-92 tests for the Corolla.
Supplemental measurements of the flow tunnel ambient temperatures conducted by SGS-ETC
suggested this could be due to ambient temperature gradients at the start of the tests (prior to
adequate ambient air mixing), although it's not entirely clear why this occurred with these two
vehicles and not the other test vehicles.
Exploratory testing conducted by SGS-ETC in order to determine the cause of differences
between the FTTP profiles developed for this study and those used for original vehicle
evaporative emissions certification testing did seem to indicate the original profiles could be
obtained by adjusting the road surface mat temperature set point. Since the OEM mat set point
was unknown in most cases, this would entail an iterative process in which the mat temperature
was adjusted based on its predicted effect in order to match the OEM FTTP profile. Based on
this testing, numerical strategies were developed to "normalize" drive trace temperatures to
match the OEM FTTP profiles. These strategies are described in Appendix B, and sample
calculations are provided electronically as Appendix C. However, the temperature profiles
provided for this study have not been normalized using any of these strategies. Future work
could be conducted to determine how best to apply adjustments to the profiles provided as part of
this study, although care must be taken to apply the data corrections only to deviations resulting
from systematic bias.
15
-------
4.0 References
1. CRC E-77 reports: Haskew, H., Liberty, T. (2008). Vehicle Evaporative Emission
Mechanisms: A Pilot study, CRC Project E-77; Haskew, H., Liberty, T. (2010),
Enhanced Evaporative Emission Vehicles (CRC E-77-2); Haskew, H., Liberty, T.
(2010), Evaporative Emissions from In-Use Vehicles: Test Fleet Expansion (CRC E-
77-2b); Haskew, H., Liberty, T. (2010), Study to Determine Evaporative Emission
Breakdown, Including Permeation Effects and Diurnal Emissions Using E20 Fuels on
Aging Enhanced Evaporative Emissions Certified Vehicles, CRC E-77-2c; DeFries,
T., Lindner, J., Kishan, S., Palacios, C. (2011), Investigation of Techniques for High
Evaporative Emissions Vehicle Detection: Denver Summer 2008 Pilot Study at Lipan
Street Station; DeFries, T., Palacios, C., Weatherby, M., Stanard, A., Kishan,
S.(2013) Estimated Summer Hot-Soak Distributions for Denver's Ken Caryl I/M
Station Fleet
2. DeFries, T., Lindner, J., Kishan, S., Palacios, C. (2011), Investigation of Techniques
for High Evaporative Emissions Vehicle Detection: Denver Summer 2008 Pilot Study
at Lipan Street Station; DeFries, T., Palacios, C., Weatherby, M., Stanard, A., Kishan,
S.(2013) Estimated Summer Hot-Soak Distributions for Denver's Ken Caryl I/M
Station Fleet.
3. Sabisch, M., Kishan, S, Stewart, J, Glinsky, G. (2014), Running Loss Testing with
Implanted Leaks.
4. Lindner, J., Sabisch, M., Glinsky, G, Steward, J., StDenis, M., Roeschen, J, (2012)
Multi-Day Diurnal Testing, Contract CP-C-06-080, WA 5-11.
5. Haskew, H., Liberty, T. (2010), Enhanced Evaporative Emission Vehicles (CRC E-
77-2); Haskew, H., Liberty, T. (2010), Evaporative Emissions from In-Use Vehicles:
Test Fleet Expansion (CRC E-77-2b); Haskew, H., Liberty, T. (2010), Study to
Determine Evaporative Emission Breakdown, Including Permeation Effects and
Diurnal Emissions Using E20 Fuels on Aging Enhanced Evaporative Emissions
Certified Vehicles, CRC E-77-2c.
16
-------
5.0 Index of Appendices
The following is a list of the appendices to be provided with this report. As noted below,
some appendices will be provided as separate electronic files.
Appendix A - Test Fuel Specifications
Appendix B - Issues Encountered and Solutions
Appendix C - Drive Trace Temperature Normalization Examples (electronic
appendix, *.xlsx format)
Appendix D - Descriptions of Study Data (the following files will be provided
electronically, by vehicle)
DAQ Recording (*.csv)
Ford DTF Data Recording (*.txt)
OBDII Data (*.csv)
Test Data (*.xlsx)
17
-------
Appendix A
Test Fuel Specifications
-------
Test fuel used for this study conformed to the following specifications:
Test Fuel Type
Test Fuel Specification Number
Fuel Description/ Application
RON
MON
Sulfur (ppm)
T10(°C/F)orE70(%v/v)
T50 ( °C/F) or E100 (%v/v)
T90 (°C/F) or El 50 (%v/v)
Vapor Pressure (kPa / psi)
Oxygenates (vol%) (e.g. ethanol,
methanol, MTBE, ETBE, etc.)
Additives
Lead (g/L)
General Purpose and other DV
WW XE-M4CX560-A
E10 GASOLINE, Worldwide Driveability Sign-off,
91 RON E10 Summer Nominal (0°C and above)
90-92
82-84
10 max
T10 = 50-60C(122- 140 F)
T50 = 90- 100 C (194- 212 F)
T90 = 160 - 170 C (320 - 338 F)
55 - 60 kPa (8.0 - 8.7 psi)
10+/- 0.2 %v/v Ethanol
Normal Commercial
0.0025 max
A-l
-------
Appendix B
Issues Encountered and Solutions
-------
This appendix provides a summary of issues that were encountered during this study and
a description of how each of those issues was addressed.
The study performed previously in Work Assignment 1-08, Running Loss Testing with
Implanted Leaks, demonstrated difficulties in collecting the OBD commanded evaporative purge
data stream. Only two of the vehicles from that study (which are the same as the vehicles in this
study, the 2009 Toyota Corolla and the 2010 Ford Focus PZEV) broadcasted the OBD
commanded evaporative purge. SGS-ETC measured and recorded the purge valve voltage to
determine the commanded purge signal for the remaining vehicles, using an Omega multichannel
portable data acquisition system (DAQ).
The Ford DTF did not have any viable means of measuring temperature data from the J-
Type thermocouples which had been previously installed in the vehicle's sending units for
vehicle fuel and fuel vapor temperature measurements, so SGS-ETC used the portable data
acquisition system previously described in order to measure temperatures of the fuel liquid and
vapor and also the ambient temperatures, in addition to the commanded evaporative purge.
Due to challenges with obtaining OBD data during Work Assignment 1-08, SGS-ETC
made stronger efforts to collect OBD data from all tests during this study. While SGS-ETC was
more successful in this endeavor than during Work Assignment 1-08, there are several instances
of missing data. This is most notable with the Corolla data, where OBD data is missing from
several tests. Since SGS-ETC expected to obtain the commanded evaporative purge from the
OBD datastream, purge voltage was not collected, so for these tests, no purge rate signal (voltage
or OBD) is available. Consequently, of the five tests performed on the Corolla, only two have
commanded evaporative purge.
The Caravan does not have commanded purge data available for the LA-92 trace. This
happened to be the first test performed using the DAQ system and it was not connected. This
test did not demonstrate any other problems so it was decided to accept this test despite missing
data.
Time alignment was handled by starting the DAQ recording as the test began. OBD data
and pressure data (as measured by the Ford DTF) were aligned using vehicle speed and
dynamometer roll speed, respectively.
During initial testing, the vehicle was moved into the wind tunnel while the wind tunnel
was heating up, before the SGS-ETC technical supervisor arrived on-site. This would cause the
liquid fuel temperature to fall below the 92 °F cutoff specified in 86.129-94(d)(4)(ii)(B), so test
B-l
-------
personnel would then wait until the fuel temperature reached 92 °F, plus an additional hour to
allow the fuel temperature to stabilize prior to the start of testing on that vehicle.
During the first day of testing (July 22, 2013), the SGS-ETC technical supervisor noticed
poor control of the thermal mat temperatures. The mat temperature was being controlled using a
simple step control method and the simulated air speed of the more aggressive trace was causing
large fluctuations in the mat temperature. Ford DTF improved the control by applying insulation
over the surface thermocouple. However, the mat temperatures continued to swing, so SGS-ETC
set the mat temperature to be greater than the minimum required 125 °F so that the temperature
swings would not cause the mat temperatures to fall beneath the minimum temperature specified
in the CFR.
At the beginning of testing, the air conditioning system in the Honda Accord was
inoperable. The temperature inside the vehicle exceeded 110 °F during the first test performed on
July 22, 2013 on this vehicle. This thermal extreme caused a problem with the data acquisition
system inside the car, resulting in the data having a severe noise issue. To correct this problem,
the vehicle was transported to SGS-ETC facilities in Jackson, MI, where the air conditioning
system was repaired. The original test was voided and later successfully repeated.
On August 5, 2013, a failure in the DTF wind tunnel resulted in wind speed not being
maintained for a portion of two of the US-06 tests performed on the Ford Focus and the Dodge
Caravan. SGS-ETC technical oversight elected to continue testing and performed a repeat of the
less aggressive 3 x LA-92 test cycle on the Accord to allow Ford DTF staff to make necessary
repairs to the wind tunnel. After repairs, two subsequent US-06 tests were successfully
completed on the Silverado and Corolla. The two tests on the Focus and Caravan were voided
and repeated the following night.
During the US-06 test conducted on August 5, 2013 the vent line on the Caravan was
clamped closed by one of the chain tie downs in the wind tunnel. This caused the amount of
vacuum on the system to increase well beyond the typical amount of vacuum. This test had
already been voided due to the previously-described wind tunnel malfunction. On subsequent
tests, as part of the test process SGS-ETC technical oversight verified that the vent line was not
pinched prior to test commencement.
The data acquisition computer crashed during the August 27, 2013 standard Fuel Tank
Temperature Profile (FTTP) testing on the Corolla. Since SGS-ETC was not able to recover data
from this experiment, this test was voided and repeated on September 24, 2012.
B-2
-------
The Chevrolet Silverado's brakes failed during the US-06 test on August 6, 2013, causing
a brief period of driver violations. The driver attempted to use the emergency brakes to slow the
vehicle during this period. This was discussed with EPA and ERG personnel and it was decided
that this was acceptable as the Silverado is an in-use vehicle.
Several vehicles, most notably the Silverado and the Corolla (prior to installing the heat
shield as later described in this section), displayed a very rapid increase in liquid fuel
temperature during the first ten minutes of testing, on occasion climbing as much as 10 °F during
this period of time. This behavior was rigorously investigated through several techniques. SGS-
ETC first performed a test in which the vehicle was parked in a shaded area and idled for at least
20 minutes. During this time, SGS-ETC measured the liquid fuel temperature to determine if the
rapid increase in liquid fuel temperature was being induced by the vehicle. The temperature
gains after 20 minutes were never greater than 5 °F. Results from additional data reviews
suggested this temperature increase was caused by the facility's wind tunnel air temperature. Air
temperature measurements collected in various locations by SGS-ETC showed the air
temperature dipped during the first minutes of testing and the following air temperature increase
coincided with the fuel temperature increase.
The Corolla showed greatly increased temperatures for the more aggressive cycles than
for the previously obtained tank temperature profile. This effect was unique to this vehicle. All
other vehicles demonstrated a decrease in the tank temperature for the newer cycles. During the
latter stages of the project, it was discovered that the Corolla was missing the original
manufacturer's fuel tank heat shield. A replacement shield was ordered and installed by SGS-
ETC. The FTTP test performed after installing the heat shield showed very good agreement
between the obtained temperature profile and the original certification temperature profile.
During the course of the study, the scope was expanded to include correlation testing to
verify the each vehicle manufacturer's original certification tank temperature profile. The
temperature profiles developed during this study using the FTTP drive trace did not match the
manufacturer's original FTTP temperature profiles. The tank temperature profiles developed at
the Ford DTF facility were consistently higher or lower than the OEM profiles. Additional
testing was performed on the Ford focus, increasing the mat temperature above the 125 °F set
point in an attempt to generate a matching temperature profile. This testing produced a good
match, with the final test performed on the Corolla (with the heat shield installed) obtaining a
superior match using the 125 °F mat temperature.
B-3
-------
Based on results of this verification testing, EPA, ERG and SGS-ETC discussed whether
or not the LA-92/US-06 tank temperature profiles that were developed during this study should
be "normalized" to correct for the temperature discrepancies seen between this study's FTTP
profiles and the vehicle manufacturer's original FTTP profiles. This would help ensure the LA-
92/US-06 temperature profiles developed during this study are the same as if they had been
developed by vehicle manufacturers using this study's drive cycles (manufacturer-equivalent,
with all vehicle and test conditions equivalent to those at the manufacturer's test facility). Using
"normalized" temperature profiles would help ensure running loss evaporative emissions
measured using this study's LA-92/US-06 drive traces would not be biased by drive trace
temperature discrepancies (which were shown in work assignment 1-08 to have a large influence
on running loss emissions for vehicles with induced leaks).
Three different candidate strategies were developed for drive trace temperature
"normalization". The first strategy (Final Temperature Correction of Original FTTP Profile)
would scale the original (vehicle manufacturers') FTTP profile by the ratio of the final
temperatures of the original FTTP profile to the recorded FTTP profile multiplied by the final
drive trace temperature. This strategy is shown mathematically as follows:
rri S-\ L/> lyiflUl I \JI lyillUl ™ fAQOR^ Q ^ I I Q ^
' C.nTTfrtfr} \^) 7^ ^ M n «^.x ,,. r- ' I 7^ / * n «^.x ' ' Rprnrrlprf I T'OUo J "D I ~r " J
The second candidate strategy (Final Temperature Correction of the Recorded FTTP
Profile) would scale the recorded FTTP profile by the ratio of the final temperatures of the
original FTTP profile to the recorded FTTP profile multiplied by the final drive trace
temperature. This strategy is shown mathematically as follows:
m_ ^corded (0-95 /7>rrPor.fl.nai(4308)
TcorrectedW ~ ~ - ^ - ' - ^corded (4308) -
- 95 T
yD \l FTTPRecorded
The third candidate strategy (Continuous Correction) would provide a continuous
correction of the recorded temperature by an ongoing ratio of the temperatures of the original
FTTP profile to the recorded FTTP profile. This strategy is shown mathematically as follows:
_ rgna _
1 Corrected W T /••% L Recorded v"v
lFTTPRecorded\l)
Examples of each of these strategies are provided electronically (Appendix C). None of
the temperature profiles developed during this study have been corrected using any of these
strategies.
B-4
-------
Appendix C
Drive Trace Temperature Normalization Examples
(Note: This appendix will be provided electronically as a *.xlsx file)
-------
Appendix D
Study Data
(Note: Descriptions of the data to be provided electronically follow)
-------
Several files were obtained for each test performed during this study. These are
described below, and the files are provided electronically as appendices. The file naming
convention for these files is as follows:
[Date]_[Vehicle]_[Test Type]_[DAQ/DTF/OBD Data SetJ.csv
The date stamp is compressed into a six-digit number, the first two digits represent the
year, the following two digits represent the month, and the last two digits are the day.
Date Stamp:YYMMDD so 130723 becomes 7/23/2013
The Test Type identifier designates the trace tested during a particular test. The following
table coordinates the test type with the specific trace performed.
Table 3. Running Loss Testing Sequence
Test Type
FTTP
LA92
US06
Specific Trace
Standard Running Loss Trace
3 X LA-92
2 X US-06 + 70 mph Cruise + 2 X US-06
Duration (s)
4308
4308
4308
Average
Speed (mph)
14.4
24.6
57.1
Maximum
Speed (mph)
56.7
67.2
80.3
DAQ Recording (*.csv)
The DAQ recording is parsed into a csv. During experimentation additional instruments
were added and removed from the configuration as necessary. The data presented in the DAQ
files doesn't conform to a single standard output. There was some consideration as to adding a
MFM to measure purge flow, however this instrument was never connected as it might interfere
with proper purge behavior on the vehicle. The existence of such a channel within the data is
erroneous, and reflects that nothing was measured using this device. The dataset contains the
following measurements.
Start Time: Time and date the data recording was initiated.
Elapsed Time: Measured in seconds since Start Time.
Purge Valve: Measured in voltage. When converting this to voltage, the minimum
voltage was assumed to correspond with a commanded purge of 0%, and the maximum voltage
was assumed to correspond with a commanded purge of 100%.
D-l
-------
Liquid Fuel Temperature: Measured in degrees Celsius, uses instrumentation installed for
WA 1-08 Running Loss Testing with Implanted Leaks.
Vapor Fuel Temperature: Measured in degrees Celsius, uses instrumentation installed for
WA 1-08 Running Loss Testing with Implanted Leaks.
Ambient Temperature/Ambient Top Temperature: Measured in degrees Celsius. A
thermocouple was installed on the roof of the vehicle to verify ambient temperature simulation.
Ambient Front Temperature: Measured in degrees Celsius. A thermocouple was installed
at the front of the vehicle extending downward so as to measure the temperature of the air before
it passes under the vehicle. This was done to check for thermal stratification.
Ford DTP Data Recording (*.txt)
This tab delimited file is produced by Ford DTP data recording systems. The file begins
with a row of header information containing the following data.
Vehicle: The test vehicle is identified.
Start Date: The date the test was performed.
Start Time: The time data recording began.
Operator: The technician responsible for controlling the wind tunnel during the test.
The file then produces several measurements of streaming data and test inputs. The
recording frequency for these data are 1 Hz. Only a few of these data were used in the post-
processing and quality analysis.
AID-Roll-Speed-F: Measured in miles per hour, this was used to align the individual data
files for front wheel drive vehicles.
AID-Roll-Speed-R: Measured in miles per hour, this was used to align the individual data
files for the one rear wheel drive vehicle, the Silverado.
AIRSPEED: Measured in miles per hour, this was checked against roll speed to verify
proper operation of the wind tunnel
D-2
-------
AIRHUMIDITY: Measured in percent relative humidity, this was checked to verify that
the wind tunnel temperature and humidity controls were operating properly.
AIRTEMP: Measured in degrees Fahrenheit, this was checked against the ambient
thermocouples installed on the vehicle and recorded using the DAQ.
MATS SUR: Measured in degrees Fahrenheit, this is the surface temperature of the
forward pavement simulation mat.
MAT4_SUR: Measured in degrees Fahrenheit, this is the surface temperature of the
middle pavement simulation mat.
MATS SUR: Measured in degrees Fahrenheit, this is the surface temperature of the rear
pavement simulation mat.
MAT6_SUR: Measured in degrees Fahrenheit, this is the surface temperature of the rear
pavement simulation mat used for the Silverado.
Fuel Vapor Pressure: Measured in inches of water, this is a measurement of positive
pressure on the fuel system.
OBD2 Data (*.csv)
OBDII data was collected using a HEM Data mini logger provided by the EPA during the
three driving portions (the FTP 72, the FTP 75, and the running loss test) of the test sequence.
The OBDII data collection system was problematic and didn't consistently record data for all
tests (as discussed above in the Issues Encountered and Solutions section). Data is stored in a
binary file that is processed using the HEM Data's DawnEdit software. The data available varies
for each vehicle as some vehicles use different communication standards and don't broadcast the
same types of data. The data description in below describes only the data that was used during
analysis. Also, per the discussion above in the Issues Encountered and Solutions section - OBDII
data is not available for all driving tests.
Time: When the drive cycle began
VIN: The vehicle identification number for the unit under test
Vehicle Speed: The vehicle speed measured in miles per hour
D-3
-------
Commanded Evaporative Purge: Measured in percent, this is how much purge was
commanded by the vehicle while the vehicle was in operation
Test Data.xlsx
The three data files are collated, and then added to a single Excel workbook containing
all test results and responsible for data visualization. Data from each vehicle is contained on a
single sheet for that particular vehicle.
D-4
------- |