SRI/USEPA-GHG-VR-29
August 2003
Environmental Technology
Verification Report
ConocoPhillips
Fuel-Efficient High-Performance (FEHP)
SAE 75W90 Rear Axle Gear Lubricant
Prepared by:
Greenhouse Gas Technology Center
Southern Research Institute
Under a Cooperative Agreement With
U.S. Environmental Protection Agency
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
EPA REVIEW NOTICE
This report has been peer and administratively reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
-------
SRI/USEPA-GHG-VR-29
August 2003
l lll ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM
V.l. ~
Protection Agency
Southern Research
INSTITUTE
ETV Joint Verification Statement
TECHNOLOGY TYPE:
Fuel Efficient Rear Axle Lubricant
APPLICATION:
Light-Duty Trucks and SUVs
TECHNOLOGY NAME:
ConocoPhillips Fuel-Efficient High-Performance
(FEHP) SAE 75W90 Rear Axle Gear Lubricant
COMPANY:
ConocoPhillips
ADDRESS:
1000 S. Pine St., Ponca City, OK 74602
E-MAIL:
kay.k.bjornen@conocophillips.com
The U.S. Environmental Protection Agency (EPA) has created the Environmental Technology
Verification (ETV) program to facilitate the deployment of innovative or improved
environmental technologies through performance verification and dissemination of information.
The goal of the ETV program is to further environmental protection by accelerating the
acceptance and use of improved and cost-effective technologies. ETV seeks to achieve this goal
by providing high-quality, peer-reviewed data on technology performance to those involved in
the purchase, design, distribution, financing, permitting, and use of environmental technologies.
ETV works in partnership with recognized standards and testing organizations, stakeholder
groups that consist of buyers, vendor organizations, and permitters, and with the full
participation of individual technology developers. The program evaluates the performance of
technologies by developing test plans that are responsive to the needs of stakeholders,
conducting field or laboratory tests, collecting and analyzing data, and preparing peer-reviewed
reports. All evaluations are conducted in accordance with rigorous quality assurance protocols to
ensure that data of known and adequate quality are generated and that the results are defensible.
The Greenhouse Gas Technology Center (GHG Center), one of six verification organizations
under the ETV program, is operated by Southern Research Institute in cooperation with EPA's
National Risk Management Research Laboratory. The GHG Center has completed the
performance verification of the ConocoPhillips Fuel-Efficient High-Performance (FEHP)
Society of Automotive Engineers (SAE) 75W90 Rear Axle Gear Lubricant. This verification
statement provides a summary of the test results for the lubricant.
S-l
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
TECHNOLOGY DESCRIPTION
The transportation sector accounted for approximately 32 percent of C02 emissions from fossil
fuel combustion during 2001. The US EPA reports that in 2001, automobiles and light-duty
trucks produced approximately 1.074 x 109 and 1.87 x 107 metric tons of carbon dioxide (C02)
equivalents from the combustion of gasoline and diesel fuel, respectively. Combustion of
gasoline and diesel fuel in automobiles and light-duty trucks was responsible for approximately
73 percent (57 percent from gasoline and 16 percent from diesel) of total transportation related
C02 emissions in the US during 2001. Small fuel efficiency or emission rate improvements are
expected to have a significant beneficial impact on nationwide greenhouse gas emissions because
of the large quantity of fuel consumed.
ConocoPhillips has developed the Fuel-Efficient High-Performance (FEHP) SAE 75W90 Rear
Axle Gear Lubricant in partnership with an axle manufacturer (Visteon Corporation) and an
additive supplier (Ethyl Petroleum Additives, Ltd.). The product is marketed as a fuel efficient,
high performance, multi-grade gear lubricant for light-duty trucks, automobiles, and sport utility
vehicles (SUVs). ConocoPhillips states that the product consists of a lower viscosity, synthetic
base lubricant with optimized fluidity and friction modifiers when compared to standard axle
lubricants. The developers report incremental fuel economy improvements of 0.1 to 0.2 miles
per gallon [mpg] with FEHP when compared to standard lubricants.
According to ConocoPhillips, the FEHP offers the following benefits:
¦ Improved axle efficiency,
¦ Reduced temperature under severe towing,
¦ Reduced spin losses, and
¦ Improved thermal and oxidative stability.
ConocoPhillips claims that the properties of the FEHP, including product durability, allow it to
be a replacement for standard SAE 75W140 rear axle gear lubricant typically specified by some
automobile manufacturers in light-duty trucks with FEHP rated at 75W90. Table 1 summarizes
typical FEHP physical properties as provided by ConocoPhillips.
S-2
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
Table 1: FEHP Fluid Properties3
Specified Test
Specified Method
Minimum
Value
Allowed
Maximum
Value
Allowed
Typical
Values
Kinematic Viscosity at 100 °C,
cSt
ASTM D445
17
18.5
17.65
Kinematic Viscosity at 40 °C,
cSt
ASTM D445
--
--
108.7
Viscosity Index
ASTM D2270
172
--
179.5
Pour Point, °C
ASTM D97
--
-42
-48
Sulfur, %
ASTM D1552
1.23
2.21
1.8
Phosphorus, %
ASTM D4951
0.07
0.123
0.09
Nitrogen, %
ASTM D4629
0.083
0.263
0.14
Boron, %
ASTM D4951
0.006
0.19
0.012
Moisture, %
Karl Fischer Titration,
ASTM D6304
--
0.10
0.04
Flash Point, °C
ASTM D92
150
--
193
Density (a), 60 °F, Kg/L
ASTM D4052
--
--
0.866
Copper corrosion
ASTMD130
--
2b
lb
"Provided by ConocoPhillips. Not verified by the GHG Center.
VERIFICATION DESCRIPTION
The goal of the performance verification testing for the ConocoPhillips FEHP rear axle gear
lubricant was the determination of a potential small change in fuel economy resulting from the
use of the FEHP lubricant when compared to a standard or reference lubricant. The test program
was completed in accordance with the requirements of the Test and Quality Assurance Plan for
ConocoPhillips Fuel-Efficient High-Performance SAE 75W90 Rear Axle Gear Lubricant
(SRI/USEPA-GHG-QAP-28), March, 2003. The sole verification parameter for testing of the
ConocoPhillips FEHP rear axle gear lubricant is the change in fuel economy (mpg). Emissions
of greenhouse gases and other pollutants were also determined.
Fuel economy testing was completed at Southwest Research Institute's (SwRI) Department of
Emissions Research (DER). The test site for the FEHP fuel economy change determination was
SwRTs light-duty vehicle Chassis Dynamometer #7. The dynamometer is equipped with a
constant volume sampling system, an array of emissions analyzers, a fuel supply cart, and
ambient monitoring and control equipment. Testing conditions (ambient conditions, test fuel,
vehicle driver, etc.) were consistent throughout the test period.
Testing was completed on a 2003 Ford F-150 Supercrew V8 with a straight beam axle. This
vehicle was determined to be representative of a large portion of straight beam axle vehicles in
current production, although a portion of vehicles in the future are likely to make use of
independent rear wheel suspensions. The vehicle was operated on the chassis dynamometer
over two test cycles for each test run using the Federal Test Procedure (FTP) (40 CFR 86.115)
and the Highway Fuel Economy Test (HFET) (40 CFR Part 600, Appendix I) to determine fuel
S-3
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
economy. Carbon monoxide (CO), carbon dioxide (CO2), nitrogen oxides (NOx), total
hydrocarbons (THC), and methane (CH4) emission rates for each test phase were determined
through analysis of exhaust samples using the Horiba VETS-9200 control system, emission
analyzers and a constant volume sampling system. Pollutant mass emission rates were calculated
in accordance with 40 CFR 86.144. Non-methane hydrocarbon (NMHC) emission rates were
calculated from THC and CH4 emissions in accordance with the same standard. Vehicle fuel
economy was calculated using the methods specified in 40 CFR 600.113. This method uses a
carbon balance based on the carbon content of test fuel used and carbon exhaust emissions
measured during each test phase to determine fuel economy.
The test period consisted of an initial set of five valid test runs using the reference lubricant
(75W140, as recommended by the manufacturer). Six runs using the 75W90 FEHP lubricant
were then completed. Six additional runs using fresh reference lubricant were completed after the
FEHP runs to determine if a change in fuel economy occurred as a result of mileage
accumulation effects and vehicle break-in. The mean fuel economies for each lubricant type
were compared to determine the fuel economy change. A statistical analysis was applied to the
data sets to determine the statistical significance of the measured fuel economy change. A
confidence interval was calculated for the observed fuel economy change.
The test vehicle was acquired on March 26, 2003, with vehicle setup, axle lubricant change, and
mileage accumulation occurring between March 26 and April 1, 2003. The fuel economy testing
verification period started on April 2, 2003. Testing was completed on May 31, 2003.
Quality assurance audits of the test facility laboratory were completed by the GHG Center field
team leader during testing. The GHG Center completed: (1) a technical systems audit to assure
the testing was in compliance with the test plan; (2) a performance evaluation audit to ensure that
measurement systems employed were adequate to produce reliable data; and (3) a data quality
audit of at least 10 percent of the test data to assure that the reported data and data reduction
procedures accurately represented the data generated during the test. In addition to the quality
assurance audits performed by the GHG Center, EPA QA personnel conducted a quality
assurance review of the Verification Report and a quality systems audit of the GHG Center's
Quality Management Program.
The GHG Center has made every attempt to obtain a reasonable and representative set of data to
examine fuel economy changes resulting from the use of the FEHP lubricant in light-duty trucks.
However, these results may not represent performance at significantly different operating
conditions or for different vehicle and axle types.
VERIFICATION OF PERFORMANCE
A total of seventeen valid fuel economy tests were completed on the test vehicle during the test
period. Fuel economy data was normalized to account for a slight upward drift in fuel economy
between the initial and final reference lubricant runs. Table 2 presents a summary of the
normalized mean fuel economy results and the standard deviation for each set of lubricant tests.
S-4
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
Table 2: Normalized Fuel Economy Test Results
Test Run ID
Normalized Fuel Economy (mpg)
Reference Lubricant- Initial
Mean
17.396
Standard Deviation
0.0414
FEHP Lubricant
Mean
17.566
Standard Deviation
0.0307
Reference Lubricant- Final
Mean
17.398
Standard Deviation
0.0447
Analysts completed a statistical analysis of the fuel economy data to determine whether a
statistically significant change in fuel economy had occurred. A confidence interval was also
calculated for the fuel economy change. The following summarizes the verification results:
The GHG Center's evaluation of the verification test results shows a statistically
significant improvement in overall fuel economy resulting from the use of the FEHP rear
axle lubricant on a 2003 Ford F-150 with beam axle.
The mean measured fuel economy improvement resulting from the use of the
ConocoPhillips FEHP 75W90 rear-axle lubricant is 0.169 mpg ± 0.0410 mpg. The error
specified represents the 95-percent confidence interval of the measured fuel economy
change data.
A 0.97 percent improvement in overall vehicle fuel economy occurred with the use of
the FEHP lubricant when compared to the mean vehicle fuel economy with the reference
lubricant.
Greenhouse gas and other pollutant emissions from the test vehicle were measured during
use of the reference lubricant and FEHP lubricant as part of the fuel economy test procedure.
The following tables present a summary of the mean pollutant emission rates observed for
both the FTP and HFET test cycles.
S-5
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
Table 3a: Greenhouse Gas and Ot
ler Pollutant Emissions - FTP
THC
CO
NOx
co2
NMHC
ch4
Test Run ID
g/mi
g/mi
g/mi
g/mi
g/mi
g/mi
Reference Lubricant-Initial
Mean
0.105
0.952
0.035
584.192
0.091
0.014
Standard Deviation
0.005
0.028
0.002
1.746
0.005
0.001
FEHP
Mean
0.106
0.964
0.035
575.927
0.092
0.014
Standard Deviation
0.005
0.071
0.003
1.199
0.004
0.000
Reference Lubricant-Final
Mean
0.111
0.990
0.036
580.072
0.095
0.015
SD
0.004
0.070
0.002
1.398
0.005
0.001
Table 3b: Greenhouse Gas and Other Pollutant Emissions - HFET
THC
CO
NOx
co2
NMHC
ch4
Test Run ID
g/mi
g/mi
g/mi
g/mi
g/mi
g/mi
Reference Lubricant-Initial
Mean
0.023
0.145
0.007
380.334
0.015
0.007
Standard Deviation
0.004
0.039
0.000
1.461
0.004
0.001
FEHP
Mean
0.025
0.164
0.008
376.320
0.017
0.008
Standard Deviation
0.003
0.018
0.001
0.880
0.002
0.001
Reference Lubricant-Final
Mean
0.025
0.168
0.008
378.109
0.017
0.008
Standard Deviation
0.003
0.030
0.001
2.469
0.002
0.001
Emissions are consistent throughout each group of test runs, with coefficients of variation below 0.3. A
comparison of mean gram per mile emission rates for the FEHP and reference lubricants indicates a
reduction in C02 emissions during the FEHP runs when compared to the reference lubricant runs for both
the FTP and HFET cycles. Carbon dioxide constitutes the majority of vehicle exhaust. Therefore, a
reduction in C02 emissions is expected as a result of the improvement in fuel economy attributed to the
use of the FEHP lubricant.
S-6
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
Signed by Hugh W. McKinnon, 9/2003
Hugh W. McKinnon, M.D., M.P.H.
Director
Signed by Stephen D. Piccot, 9/2003
Stephen D. Piccot
Director
National Risk Management Research Laboratory Greenhouse Gas Technology Center
Office of Research and Development Southern Research Institute
Notice: GHG Center verifications are based on an evaluation of technology performance under specific,
predetermined criteria and the appropriate quality assurance procedures. The EPA and Southern
Research Institute make no expressed or implied warranties as to the performance of the technology and
do not certify that a technology will always operate at the levels verified. The end user is solely
responsible for complying with any and all applicable Federal, State, and Local requirements. Mention
of commercial product names does not imply endorsement or recommendation.
EPA REVIEW NOTICE
This report has been peer and administratively reviewed by the U.S. Environmental Protection Agency,
and approved for publication. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
S-7
-------
This page has been left intentionally blank
SRI/USEPA-GHG-VR-29
August 20'0'3
S-8
-------
SRI/USEPA-GHG-VR-29
August 2003
Greenhouse Gas Technology Center
A U.S. EPA Sponsored Environmental Technology> Verification ( ) Organization
a
ft/
Environmental Technology Verification Report
ConocoPhillips
Fuel-Efficient High-Performance (FEHP) SAE 75W90
Rear Axle Gear Lubricant
Prepared By:
Greenhouse Gas Technology Center
Southern Research Institute
PO Box 13825
Research Triangle Park, NC 27709 USA
Telephone: 919/806-3456
Under EPA Cooperative Agreement CR 826311-01-0
U.S. Environmental Protection Agency
Office of Research and Development
National Risk Management Research Laboratory
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711 USA
EPA Project Officer: David A. Kirchgessner
i
-------
SRI/USEPA-GHG-VR-29
TABLE OF CONTENTS
Page
APPENDICES iii
LIST OF FIGURES iii
LIST OF TABLES iii
ACKNOWLEDGMENTS v
ACRONYMS/ABBREVIATIONS vi
1.0 INTRODUCTION 1-1
1.1. BACKGROUND 1-1
1.2. CONOCOPHILLIPS FEHP TECHNOLOGY DESCRIPTION 1-2
1.3. DESCRIPTION OF TEST FACILITY AND PROCEDURES 1-3
1.3.1. TestFacility 1-3
1.3.2. Test Vehicle 1-4
1.4. PERFORMANCE VERIFICATION OVERVIEW 1-5
1.4.1. Introduction and Verification Parameters 1-5
1.4.2 Fuel Economy Change Statistical Significance 1-6
1.4.3 Fuel Economy Change Confidence Interval 1-9
1.4.4 Calculation of Fuel Economy Improvement 1-10
1.4.5 Testing and Measurement Equipment 1-11
1.4.5.1 Chassis Dynamometer 1-11
1.4.5.2 Constant Volume Sampling System 1-11
1.4.5.3 Emission Analyzers 1-12
1.4.5.4 Ambient Monitoring Equipment 1-13
1.4.5.5 Fuel Cart 1-13
1.4.6 Testing Procedure and Sequence 1-14
1.4.6.1 Vehicle Receipt and Initial Preparation 1-14
1.4.6.2 Engine Oil Change and Driver Familiarity Runs 1-15
1.4.6.3 Rear Axle Lubricant Changes 1-15
1.4.6.4 Vehicle and Axle Inspection and Gear Rating 1-16
1.4.6.5 Mileage Accumulation 1-16
1.4.6.6 Test Fuel and Fuel System Preparation 1-16
1.4.6.7 Dynamometer Setup 1-17
1.4.6.8 Bag Cart and Emission Analyzer Setup 1-18
1.4.6.9 Vehicle Preconditioning 1-18
1.4.7 Reference Lubricant and FEHP Fuel Economy Test Procedure 1-19
1.4.8 Reference Lubricant and FEHP Fuel Economy Determination 1-20
1.4.9 Pollutant and GHG Emissions 1-22
2.0 VERIFICATION RESULTS 2-1
2.1. FUEL ECONOMY IMPROVEMENT 2-2
2.1.1. Fuel Economy Test Results 2-2
2.1.2. Fuel Economy Change 2-3
2.1.3. Reference Lubricant Fuel Economy 2-4
2.1.4. Fuel Economy Change 2-5
2.2. FUEL ECONOMY CHANGE STATISTICAL SIGNIFICANCE 2-10
2.3. FUEL ECONOMY SAVINGS CONFIDENCE INTERVAL 2-10
2.4. FUEL ECONOMY CHANGE CALCULATION CROSS-CHECKS 2-11
2.5. GREENHOUSE GAS AND OTHER POLLUTANT EMISSIONS 2-14
li
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
3.0 DATA QUALITY ASSESSMENT 3-1
3.1. DATA QUALITY OBJECTIVES 3-1
3.2. RECONCILIATION 01 DQOs AND DQIs 3-2
3.2.1. Dynamometer Specifications, Calibrations, and QA/QC Checks 3-2
3.2.2. CVS Sampling System Specifications, Calibrations, and QA/QC Checks 3-4
3.2.3. Emission Analyzer Specifications, Calibrations, and QA/QC Checks 3-5
3.2.4. Ambient Instrument Specifications, Calibrations, and QA/QC Checks 3-10
3.2.5. Test Fuel Specifications 3-11
3.2.6. Fuel Economy Volumetric and Gravimetric Cross-Checks 3-12
4.0 TECHNICAL AND PERFORMANCE DATA SUPPLIED BY CONOCOPHILLIPS 4-1
5.0 REFERENCES 5-1
APPENDICES
Page
APPENDIX A Engine Oil and Axle Lubricant Change Procedures and Records A-l
APPENDIX B Test Fuel Analysis B-l
APPENDIX C Daily Test Protocol and Checklist C-l
APPENDIX D Dynanometer Setup Data D-1
APPENDIX E Outlying Data Review and Analysis E-l
LIST OF FIGURES
Figure 1-1 CVS System Schematic 1-12
Figure 1-2 Instrumental Analyzer System 1-13
Figure 1-3 Fuel Cart Schematic 1-14
Figure 2-1 Reference Lubricant Fuel Economy Results vs. Mileage 2-6
Figure 2-2 Normalized Reference Lubricant and FEHP Fuel Economy Results vs. Mileage 2-8
LIST OF TABLES
Page
Table 1-1 FEHP Fluid Properties 1-3
Table 1-2 Estimated Vehicle Axle Production Quantities (2003) 1-4
Table 1-3 Test Vehicle Specifications 1-5
Table 1-4 Test Fuel Specifications 1-17
Table 1-5 Dynamometer Setup Coefficients 1-17
Table 1-6 Emission Analyzer Ranges 1-18
Table 1-7 Daily Test Procedure 1-20
Table 2-1 Fuel Economy Test Results 2-2
Table 2-2 F-test Evaluation of Reference Lubricant Fuel Economy Data Set Variances 2-4
Table 2-3 Statistical Analysis of Reference Lubricant Tests Fuel Economy Differences 2-4
Table 2-4 Reference Lubricant Data Regression Statistics 2-5
Table 2-5 Normalized Fuel Economy Test Results 2-7
Table 2-6 F-test Evaluation of Reference Lubricant Fuel Economy Data Set Variances 2-8
Table 2-7 Statistical Analysis of Reference Lubricant Fuel Economy Difference 2-9
in
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
LIST OF TABLES
Page
Table 2-8 Summary of Pooled Normalized Reference Lubricant Data 2-9
Table 2-9 Statistical Analysis of Fuel Economy Change 2-10
Table 2-10 F-test Evaluation of Fuel Economy Data Set Variances 2-11
Table 2-1 la Fuel Economy Change Cross-Check Calculations 2-12
Table 2-1 lb Fuel Economy Change Cross-Check Calculations 2-13
Table 2-12 Summary of Greenhouse Gas Emissions 2-15
Table 2-13 Greenhouse Gas and Other Pollutant Emissions - HFET 2-16
Table 3-1 Fuel Economy Change and Data Quality Objective 3-1
Table 3-2 Chassis Dynamometer Specifications and DQI Goals 3-3
Table 3-3 Chassis Dynamometer QA/QC Checks 3-3
Table 3-4 CVS Specifications and DQI Goals 3-4
Table 3-5 CVS System QA/QC Checks 3-5
Table 3-6 Emissions Analyzer Specifications and DQI Goals 3-6
Table 3-7 Emissions Analyzer QA/QC Checks 3-7
Table 3-8 Ambient Instrument Specifications and DQI Goals 3-10
Table 3-9 Ambient Instrument QA/QC Checks 3-10
Table 3-10 Test Fuel American Society for Testing and Materials Measurement Methods
and DQI Goals 3-11
Table 3-11 Volumetric and Gravimetric Cross-Checks - FEHP Lubricant Test Runs 3-12
Table 3-12 Volumetric and Gravimetric Cross-Checks - Reference Lubricant Test Runs 3-13
Table 3-13 Volumetric and Gravimetric Cross-Checks - Post FEHP 3-13
IV
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
ACKNOWLEDGMENTS
The Greenhouse Gas Technology Center wishes to thank Kay Bjornen of ConocoPhillips, Paul Schwarz
of Visteon, Arup Gangopadhyay of the Ford Research Laboratory, and Kevin Whitney of SwRI for
reviewing and providing input on the testing strategy and this Verification Report. Thanks are also
extended to Bill Olson of SwRI for his assistance in executing the verification testing.
v
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
ACRONYMS/ABBREVIATIONS
List of Acronyms and Abbreviations
ADQ
Audit of Data Quality
ANSI
American National Standards Institute
APPCD
Air Pollution Prevention and Control Division
ASME
American Society of Mechanical Engineers
ASTM
American Society for Testing and Materials
°C
degrees Centigrade
CFR
Code of Federal Regulations
CFO
critical flow orifice
CFV
critical flow venturi
CH4
methane
CO
carbon monoxide
co2
carbon dioxide
COV
coefficient of variation
cP
Centipoise
cSt
Centistoke
CVS
constant volume sampler
CWF
carbon weight fraction
DAS
data acquisition system
DDS
Durability Driving Schedule
DF
degrees of freedom
DOE
U.S. Department of Energy
DQI
data quality indicator
DQO
data quality objective
EPA-ORD
Environmental Protection Agency Office of Research and Development
EPA
Environmental Protection Agency
ETV
Environmental Technology Verification
°F
degrees Fahrenheit
FEHP
ConocoPhillips Fuel-Efficient High-Performance SAE 75W90 rear-axle gear
lubricant
FID
flame ionization detector
FRL
Ford Research Laboratory
FTP
Federal Test Procedure
ft3
cubic feet
gal
U.S. Imperial gallons
GC
gas chromatograph
GHG
greenhouse gas
GHG Center
Greenhouse Gas Technology Center
g/mi
grams per mile
HFET
Highway Fuel Economy Test
Hz
Herz
ISO
International Organization for Standardization
Kg/L
kilograms per liter
lb
pounds
lbf
pounds force
LHV
lower (or net) heating value
MAD
mileage accumulation dynamometer
VI
-------
SRI/USEPA-GHG-VR-29
August 20'3
mpg
miles per gallon
n2
Nitrogen
NIST
National Institute of Standards and Technology
NMHC
non-methane hydrocarbons
NO
nitric oxide
no2
nitrogen dioxide
NOx
blend of NO, N02, and other oxides of nitrogen
02
oxygen
03
ozone
ORD
Office of Research and Development
PEA
performance evaluation audit
ppmv
parts per million volume
psia
pounds per square inch absolute
psig
pounds per square inch gauge
QA
quality assurance
QA/QC
quality assurance / quality control
QMP
Quality Management Plan
Report
Environmental Technology Verification Report
RH
relative humidity
SAE
Society of Automotive Engineers
SAO
smooth approach orifice
SG
specific gravity
SOP
standard operating procedure
SRI
Southern Research Institute
SRM
standard reference material
SUV
sport utility vehicle
SwRI
Southwest Research Institute
SwRI DER
Southwest Research Institute Department of Emissions Research
Test Plan
Test and Quality Assurance Plan
THC
total hydrocarbons (as carbon)
TSA
technical systems audit
VETS
Vehicle Emissions Testing System
VEZ
vehicle emission zero (gas)
U.S.
United States
U.S. EPA
United States Environmental Protection Agency
-------
This page left intentionally blank
SRI/USEPA-GHG-VR-29
August 20'0'3
Vlll
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
1.0 INTRODUCTION
1.1. BACKGROUND
The U.S. Environmental Protection Agency's Office of Research and Development (EPA-ORD) operates
the Environmental Technology Verification (ETV) program to facilitate the deployment of innovative
technologies through performance verification and information dissemination. The goal of ETV is to
further environmental protection by accelerating the acceptance and use of improved and innovative
environmental technologies. Congress funds ETV in response to the belief that there are many viable
environmental technologies that are not being used for the lack of credible third-party performance data.
With performance data developed under this program, technology buyers, financiers, and permitters in the
United States and abroad will be better equipped to make informed decisions regarding environmental
technology purchase and use.
The Greenhouse Gas Technology Center (GHG Center) is one of six verification organizations operating
under the ETV program. The GHG Center is managed by EPA's partner verification organization,
Southern Research Institute (SRI), which conducts verification testing of promising GHG mitigation and
monitoring technologies. The GHG Center's verification process consists of developing verification
protocols, conducting field tests, collecting and interpreting field and other data, obtaining independent
peer-review input, and reporting findings. Performance evaluations are conducted according to externally
reviewed verification Test and Quality Assurance Plans (Test Plans) and established protocols for quality
assurance.
The GHG Center is guided by volunteer groups of stakeholders. These stakeholders guide the Center on
which technologies are most appropriate for testing, help disseminate results, and review Test Plans and
Technology Verification Reports (Reports). The GHG Center's Executive Stakeholder Group consists of
national and international experts in the areas of climate science and environmental policy, technology,
and regulation. It also includes industry trade organizations, environmental technology finance groups,
governmental organizations, and other interested groups. The GHG Center's activities are also guided by
industry specific stakeholders who provide guidance on the verification testing strategy related to their
area of expertise and peer-review key documents prepared by the GHG Center.
One sector of significant interest to GHG Center stakeholders is transportation - particularly technologies
that result in fuel economy improvements. The transportation sector accounted for approximately 32
percent of C02 emissions from fossil fuel combustion during 2001. The US EPA reports that in 2001,
automobiles and light-duty trucks produced approximately 1.074 x 109 and 1.87 x 107 metric tons of
carbon dioxide (C02) equivalent from the combustion of gasoline and diesel fuel, respectively.
Combustion of gasoline and diesel fuel in automobiles and light-duty trucks was responsible for
approximately 73 percent (57 percent from gasoline and 16 percent from diesel) of total transportation
related C02 emissions in the U.S. during 2001(1). Because of the large quantity of fuel consumed, small
fuel efficiency or emission rate improvements are expected to have a significant beneficial impact on
nationwide greenhouse gas emissions.
ConocoPhillips has developed the Fuel-Efficient High-Performance (FEHP) SAE 75W90 Rear-Axle Gear
Lubricant and has requested that the GHG Center independently verify its performance. ConocoPhillips
developed FEHP in partnership with an axle manufacturer (Visteon Corporation) and an additive supplier
(Ethyl Petroleum Additives, Ltd.), and markets it as a fuel-efficient high-performance multi-grade gear
lubricant for light-duty trucks, automobiles, and sport utility vehicles (SUVs). According to
1-1
-------
SRI/USEPA-GHG-VR-29
August 20'3
ConocoPhillips, the development process included durability tests on 43 vehicles operating over a total of
2.8 million fleet miles. The developers report incremental (0.1 to 0.2 miles per gallon [mpg]) fuel
economy improvements with FEHP as compared to standard lubricants. FEHP is a suitable verification
candidate considering its potentially significant beneficial environmental quality impacts (emission
reductions through reduced fuel consumption) and ETV stakeholder interest in verified transportation
sector emission reduction technologies.
The GHG Center completed verification testing from March 27 - May 31, 2003 to evaluate the fuel
economy performance attributable FEHP in a 2003 Ford Motor Company (Ford) F-150 light-duty truck.
Verification tests were conducted at Southwest Research Institute's (SwRI) Department of Emissions
Research (DER) in San Antonio, TX.
These tests were planned and executed by the GHG Center to independently verify the change in fuel
economy resulting from the use of FEHP. This report presents the results of these verification tests.
Exhaust emissions were also monitored during verification testing. Observed greenhouse gas emissions
are provided in this report.
Details on the verification test design, measurement test procedures, and Quality Assurance/Quality
Control (QA/QC) procedures can be found in the Test Plan titled Test and Quality Assurance Plan for the
ConocoPhillips Fuel-Efficient High-Performance SAE 75W90 Rear Axle Gear Lubricant (SRI/USEPA-
GHG) 2003, QAP-28). The Test Plan can be downloaded from the GHG Center's Web site (www.sri-
rtp.com) or the ETV program web site (www.epa.gov/etv'). The Test Plan describes the rationale for the
experimental design, the testing and instrument calibration procedures planned for use, and specific
QA/QC goals and procedures. The Test Plan was reviewed and revised based on comments received
from ConocoPhillips, Visteon, Ford Motor Company, SwRI, and the EPA Quality Assurance Team. The
Test Plan meets the requirements of the GHG Center's Quality Management Plan (QMP) and satisfies the
ETV QMP requirements. Deviations from the Test Plan were sometimes required. The rationale for
these deviations and their descriptions are discussed in this report.
The remainder of Section 1.0 describes the ConocoPhillips FEHP technology, the SwRI test facility, and
the performance verification procedures that were followed. Section 2.0 presents test results and Section
3.0 assesses the quality of the data obtained. Section 4.0, submitted by ConocoPhillips, presents
additional information regarding the FEHP lubricant. Information provided in Section 4.0 has not been
independently verified by the GHG Center.
1.2. CONOCOPHILLIPS FEHP TECHNOLOGY DESCRIPTION
ConocoPhillips states that the FEHP rear axle lubricant consists of a low viscosity, synthetic base
lubricant with proprietary optimized fluidity and friction modifiers when compared to standard axle
lubricants. ConocoPhillips states that the FEHP offers the following benefits:
¦ Improved axle efficiency,
¦ Reduced temperature under severe towing,
¦ Reduced spin losses, and
¦ Improved thermal and oxidative stability.
ConocoPhillips also claims that the FEHP characteristics, including durability and protection, allow for
the use of a SAE 75W90 FEHP product to replace standard SAE 75W140 lubricants typically specified in
light-duty trucks and SUVs. The FEHP lubricant is currently in use in Visteon axles and some model
1-2
-------
SRI/USEPA-GHG-VR-29
August 20'3
year 2003 production vehicles. Table 1-1, provided by ConocoPhillips, summarizes typical FEHP
physical properties.
Table 1-1. FEHP Fluid Properties
Parameter
Test Method
Minimum
Value Allowed
Maximum
Value Allowed
Typical
Values
Kinematic Viscosity at 100 °C, cSt
ASTM D445
17
18.5
17.65
Kinematic Viscosity at 40 °C, cSt
ASTM D445
--
--
108.7
Viscosity Index
ASTM D2270
172
--
179.5
Pour Point, °C
ASTMD97
--
-42
-48
Sulfur, %
ASTMD1552
1.23
2.21
1.8
Phosphorus, %
ASTMD4951
0.07
0.123
0.09
Nitrogen, %
ASTM D4629
0.083
0.263
0.14
Boron, %
ASTMD4951
0.006
0.19
0.012
Moisture, %
Karl Fischer Titration,
0.10
0.04
ASTM D6304
Flash Point, °C
ASTMD92
150
--
193
Density (t 60 °F, kg/L
ASTM D4052
--
--
0.866
Copper corrosion
ASTMD130
--
2b
lb
1.3. DESCRIPTION OF TEST FACILITY AND PROCEDURES
1.3.1. Test Facility
Fuel economy testing was completed at Southwest Research Institute's (SwRI) Department of Emissions
Research (DER). The SwRI DER maintains an International Organization for Standardization (ISO) 9002
"Model for Quality Assurance in Production and Installation" certification and ISO 17025 "General
Requirements for the Competency of Calibration and Testing Laboratories" accreditations. The terms of
these independently assessed quality systems allow SwRI to evaluate automotive fluids, fuels, emissions,
automotive components, engine/power-train performance, and equipment durability for regulatory
agencies, automobile manufacturers, and other clients. SwRI facilities include a wide variety of
stationary engine dynamometer test stands (light-duty, non-road, and heavy-duty), vehicle dynamometer
facilities, and associated state-of-the-art emissions test equipment. The GHG Center selected SwRI based
on its experience and capability in conducting fuel economy tests in accordance with the requirements of
the EPA city and highway driving cycles, 40 CFR Part 86 (Control of Emissions From New and In-Use
Highway Vehicles and Engines), 40 CFR Part 600 (Fuel Economy of Motor Vehicles), experience in
detecting small changes in fuel economy, and independence from the technology vendor and its
competitors or end users.
The test site for the FEHP fuel economy change determination was SwRI's light-duty vehicle Chassis
Dynamometer #7. The dynamometer is equipped with a constant volume sampling system, an array of
emissions analyzers, a fuel supply cart, and ambient monitoring and control equipment. The testing and
measurement equipment is described in detail in section 1.4.5.
1-3
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
1.3.2. Test Vehicle
ConocoPhillips developed the FEHP lubricant in conjunction with Visteon for use in Visteon axles. The
only axles currently using FEHP lubricant are those produced by Visteon for Ford Motor Company for
the current model year (2003) product line. The GHG Center selected the test vehicle to represent a
significant population of vehicles in which the FEHP lubricant is currently used or will be used in the
future. The GHG Center reviewed axle production numbers provided by Visteon for current model year
vehicles and axles that currently use or could potentially use the FEHP lubricant (see Table 1-2). The
GHG Center determined from this review that beam axles are dominant and the F-150 pickup truck was
the leader in production quantity. The majority of FEHP lubricant development and testing had
previously been completed for straight beam axles.
Table 1-2: Estimated Visteon Axle Production Quantities (2003)
Ford Vehicle Type
Estimated Visteon Axle Production (2003)
Beam Axle
F-150
750,000/yr
F-250/F-350
300,000/yr
Econoline
80,000/yr
Ranger
300,000/yr
Crown Vic/Grand Marquis
200,000/yr
Mustang
100,000/yr
TOTAL
1.7 million/yr
Independent Rear Suspension (IRS)
Explorer
350,000/yr
Expedition/Navigator
270,000/yr
TOTAL
620,000/yr
The Test Plan specified a current model year (2003) Lincoln Navigator SUV as the test vehicle.
However, the Navigator in current production did not have a rear beam axle. The Navigator and its
independent rear suspension represent a significantly smaller segment of the vehicle population than the
Ford F-150 with the beam axle. Independent rear suspensions will be used more in the future, specifically
in luxury sport utility vehicles. However, it is estimated that the light-duty truck market, as well as lower-
end vehicles, will continue to use straight beam axles.
The Ford F-150 was selected as the vehicle that is most representative of a significant population of
vehicles currently in production with axles that use the FEHP lubricant. The test vehicle selected for
evaluation of the ConocoPhillips FEHP rear axle lubricant was a 2003 Ford F-150 Supercrew with a 5.4L
V8 gasoline engine and a beam axle. A corrective action report (CAR #4) was completed for the change
in test vehicle from the Lincoln Navigator to the Ford F-150 and is on file at the GHG Center.
The test vehicle was selected in accordance with the requirements of the test plan. The test vehicle had
between 10,000 and 25,000 miles on the odometer and a beam axle with open differential. A vehicle is
expected to operate normally at this mileage with minimal aging effects resulting from mileage
accumulation during testing.
1-4
-------
SRI/USEPA-GHG-VR-29
August 20'3
The test vehicle was equipped with a 5.4L V8 multiport fuel-injected gasoline engine and an open, non-
limited slip gear differential on a straight beam axle. The initial mileage at vehicle receipt was 14,895. A
vehicle receipt form, documenting the vehicle condition and specifications at receipt is on file at the GHG
Center. The test vehicle specifications are summarized in Table 1-3.
Table 1-3: Test Vehicle Specifications
Model Y ear
2003
Model
Ford F-150 Supercrew
Engine
5.4LV8
Fuel
gasoline
VIN
1FTRW07L23KA13881
Engine Family ID
3FMXT05.4PFB
Transmission
4-speed automatic
Tire Size
P255/70R16
Air Conditioning
yes
Rear Axle Type
Straight beam axle w/ open non-limited slip differential
Rear Axle Gear Ratio
3.55:1
Rear Axle Diameter
9.75 inches
Rear Axle ID Tag No.
V942B 55 9 75 2H06
Initial Odometer Reading
14,895 mi
1.4. PERFORMANCE VERIFICATION OVERVIEW
1.4.1. Introduction and Verification Parameters
The goal of the ConocoPhillips FEHP rear axle lubricant performance verification testing is the
determination of a potential small change in fuel economy resulting from the use of the FEHP lubricant
when compared to a comparable standard replacement lubricant. The sole verification parameter for
testing of the ConocoPhillips FEHP rear axle lubricant is the change in fuel economy (mpg). Emissions
of greenhouse gases and other pollutants were also determined and reviewed for each axle lubricant as
part of the fuel economy test procedure.
Each fuel economy test run conformed to the widely accepted Federal Test Procedure (FTP) and Highway
Fuel Economy Test (HFET) for highway vehicles. Code of Federal Regulations (CFR) Title 40 Part 86,
"Control of Emissions from New and In-Use Highway Vehicles and Engines" 86.115(3), and Part 600,
"Fuel Economy of Motor Vehicles" 600.109(4), are the FTP and HFET source documents.
Verification of such small fuel economy changes is a multi-step process. First, appropriate test methods
must be selected and used to allow for repeatable tests of fuel economy, while minimizing variability in
testing conditions such that fuel economy changes can be attributed to the axle lubricant. Second,
assuming that appropriate test methods have been conducted, the difference between the reference
lubricant and FEHP mpg data must be statistically significant. Third, a confidence interval must be
determined for the fuel economy difference and must be refined as much as possible to ensure data
quality.
1-5
-------
SRI/USEPA-GHG-VR-29
August 20'3
The following summarizes the basic steps of the testing procedure:
Obtain and inspect representative vehicle and axle
Change vehicle engine oil in accordance with manufacturer specifications
Change vehicle axle lubricant to reference lubricant
Accumulate 1000 miles on mileage accumulation dynamometer
Precondition vehicle to test cycles
Perform fuel economy testing using FTP and HFET driving cycles
Evaluate results to determine required number of runs and minimize statistical error
Change vehicle engine oil in accordance with manufacturer specifications
Change vehicle axle lubricant to FEHP
Accumulate 1000 miles on mileage accumulation dynamometer
Precondition vehicle test cycles
Perform fuel economy testing using FTP and HFET driving cycles
Change vehicle engine oil in accordance with manufacturer specifications
Change vehicle axle lubricant back to reference lubricant
Accumulate 1000 miles on mileage accumulation dynamometer
Precondition vehicle test cycles
Perform fuel economy testing using FTP and HFET driving cycles
Determine fuel economy change, statistical significance, and confidence interval
The procedure specified in the Test Plan required the completion of the initial reference lubricant testing
and the FEHP testing. During testing, and after observation of a notable change in fuel economy using
the FEHP, the GHG Center decided that completion of a second round of reference lubricant tests was
necessary. This ensured that the observed fuel economy increase was attributable to the use of FEHP
lubricant and not the effects of mileage accumulation or additional vehicle and axle break-in during the
test period. Additional information regarding changes in calculation methods and tests runs as a result in
the addition of the post-FEHP reference lubricant runs is provided in Sections 1.4.2 through 1.4.4.
Section 1.4.6 provides details regarding the actual test procedures. The following sections discuss the
data analysis, statistical review, and testing equipment in detail.
1.4.2 Fuel Economy Change Statistical Significance
Fuel economy change is the difference between the reference lubricant and FEHP mean mpg results.
Each mean value is the result of a limited number of test runs. Statistical theory shows that the variability
between test runs determines how accurately the mean characterizes all possible fuel economy values
within a lubricant type (i.e. reference lubricant or FEHP). The mean can be sharply characterized if each
individual test run result is very close to the mean value, or if variability is small. The difference between
two such means would also be sharply characterized and the observed differences would be statistically
significant. Large run-to-run variabilities can, however, exist. This causes the mean to "spread out" over
a larger range of possible values. The difference between two such means may not be "statistically
significant", for example, if the reference lubricant mean fuel economy falls within the confidence
interval of the FEHP fuel economy. The statistical significance of the difference in mean fuel economics
is a measure of the probability or likelihood that the observed difference occurred by chance or is
representative of the sample population (for example, a series of test results).
1-6
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
The GHG Center evaluated the statistical significance of the difference between the reference lubricant
and FEHP fuel economies by the following hypothesis:
H0:
Hi:
|/"i - >"21 = 0
|/"i " ^2| >0
(Eqn. 1)
where:
H0 = Hypothesis that there is no statistically significant difference in fuel economy
Hi = Hypothesis that there is a statistically significant difference in fuel economy
jxi = Mean fuel economy for the population of vehicles treated with FEHP lubricant
jx2 = Mean fuel economy for the population of vehicles treated with reference lubricant
Essentially, the hypothesis is a comparison of the mean of the reference lubricant tests with the mean of
the FEHP tests. A statistical test is applied to the lubricant test data to evaluate whether there is a
statistically significant difference between the reference lubricant and FEHP lubricant means. If so, the
hypothesis, H0, is rejected, indicating that the fuel economy difference is significant. To evaluate the
statistical significance of the difference between the two lubricant fuel economy means, a test statistic,
ttest, is calculated for the fuel economy change test data. The t-statistic for the test data is:(5)
^test
(X^ X2) (//j ^2)
f
v
1 1
+
n, rir,
(Eqn. 2)
_ c,-l).y +c,-l).;,' 3)
P Hj+H2-2
where:
X1 = Mean fuel economy with FEHP lubricant
X2 = Mean fuel economy with reference lubricant
|ii - |i2 = Zero (H0 hypothesizes that there is no difference between the population means)
n, = Number of repeated test runs with FEHP lubricant
n2 = Number of repeated test runs with reference lubricant
Si2 = Sample standard deviation with FEHP lubricant, squared
s22 = Sample standard deviation with reference lubricant, squared
sp2 = Pooled standard deviation, squared
To determine whether a statistically significant change is observed, this calculated ttest statistic is
compared to a critical Student's t-distribution value, ta 2. df, for the same number of degrees of freedom
(DF) as the test runs with an acceptable uncertainty, a, of 0.05(5). This comparison evaluates whether the
fuel economy difference will be observed in similar tests at least 95 percent of the time, based on the test
data and the observed variance of the test data.
1-7
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
The decision rule for the hypothesis test is:
Do not reject Ha if ttest < 10.025,df- Conclude that the data cannot show a statistically significant
fuel economy difference between FEHP lubricant versus the reference lubricant.
Otherwise,
Reject H0 if ttest > t0.025,DF¦ Conclude that a significant fuel economy difference exists between the
FEHP lubricant versus reference lubricant.
Therefore, if the test data ttest value is greater than the critical t-distribution value, the observed fuel
economy change is statistically significant.
Sample variability for each lubricant provides an indication of the repeatability of the testing process.
The F-test statistic is a calculation that compares the variances of two data sets. An Ftest statistic is
calculated to determine the degree of similarity between the reference lubricant and FEHP sample
variances.
.2
where:
Fles,=^ (Eqn.4)
Ftest = F-test statistic
s2max = Larger of the reference lubricant or FEHP sample standard deviations, squared
s2mm = Smaller of the reference lubricant or FEHP sample standard deviations, squared
The calculated F-test statistic is compared to an F-statistic distribution value for the specified number of
test runs with an acceptable uncertainty (a; 0.05 for this verification).1-5-1
Analysts will conclude that the sample variances are substantially the same and the hypothesis test for
statistical significance and confidence interval calculations are valid approaches if the F-test statistic is
less than the corresponding distribution value. Analysts conclude that the sample variances are not the
same and will consequently modify the confidence interval calculation according to Satterthwaite's
approximation if the F-test statistic is equal to or greater than the specified distribution value.
Satterthwaite's approximation describes how to estimate the appropriate degrees of freedom for use in
calculating a modified critical t-distribution value and confidence interval.1-6-1
Satterthwaite's approximation is used to calculate the degrees of freedom for the critical t-distribution
value for data sets having unequal variances. Satterthwaite's approximation for degrees of freedom is:(6)
df =
/ 2 / -.2
(s n )
v 1 V
( 2 , 2
(s In)
v 2 2
(Eqn.5)
where s, = J 1- (Eqn. 5a)
M ll M _ 11
1-8
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
where:
Si = standard deviation of data set 1
s2 = standard deviation of data set 2
ni = number of tests in data set 1
n2 = number of tests in data set 2
sd = pooled standard deviation
The calculated degrees of freedom value is used in estimating the critical t-distribution value at the 95-
percent confidence level. The critical t distribution value is used in calculating the 95 percent confidence
interval as described in Section 1.4.3.
The approximate t-statistic specified in Equation 6, when compared to the critical t-distribution value, is
used to evaluate the statistical significance of the fuel economy change for data sets with unequal
variances.
(6)
where:
-X, )
(Eqn. 6)
t' = approximate t-statistic for test runs with unequal variances
Xx = mean fuel economy with FEHP lubricant
X2 = mean fuel economy with reference lubricant
Sd = pooled standard deviation (Eqn. 5a)
1.4.3 Fuel Economy Change Confidence Interval
It becomes meaningful to calculate the confidence interval of the fuel economy change if a statistically
significant change in fuel economy is determined as described in Section 1.4.2. The confidence interval
of the mean fuel economy change provides a range of values around the mean that indicate where the true
population of sample means can be expected to be located with a given level of certainty (95 percent for
this test). A narrow confidence interval implies that the fuel economy change is sharply characterized.
Conversely, a large confidence interval implies that the data was spread across a wide range and the
resulting mean fuel economy change could have limited utility.
The half-width (e) of the 95-percent confidence interval is:
.(5)
e ^.025,DF 11 ^p
1 1
1
\P\ n2 J
(Eqn. 7)
where:
to 025, df = the critical t-distribution value
sp2 = the pooled standard deviation squared
ni = the number of FEHP lubricant test runs
n2 = the number of reference lubricant test runs
1-9
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
If the variances of the two data sets are unequal, as determined by the F-test, and Satterthwaite's
approximation is used, the confidence interval is calculated as follows:
e = t
025,DF ¦
f S1
+
V«1
n
2 J
(Eqn. 8)
The mean fuel economy change is stated as A ire, where A is the change in fuel economy and e is the 95-
percent confidence interval.
1.4.4 Calculation of Fuel Economy Improvement
The statistical analysis of the fuel economy change, as described in sections 1.4.2 and 1.4.3, is based on a
comparison of the mean reference lubricant fuel economy to the mean FEHP fuel economy. The fuel
economy change would typically be the difference between the means of the reference lubricant and
FEHP tests, as described in the Test Plan. However, because a second set of reference lubricant test data
was collected after the FEHP test runs, analysts must determine how to evaluate the two sets of reference
lubricant data such that it can be compared to the FEHP fuel economy.
The first step is to determine whether the two sets of reference lubricant data are from the same
population and can therefore be pooled together as a single data set. It is possible that test and vehicle
conditions or vehicle break-in or wear could result in a drift in fuel economy vs. mileage, such that
significantly different means are observed between the two reference lubricant test data sets. A procedure
similar to the statistical significance test described in Section 1.4.2 is applied to evaluate the reference
lubricant data sets and determine if they are from the same population.
The mean and standard deviation are initially calculated for each set of data (initial reference lubricant,
FEHP lubricant, and final reference lubricant). The variances of the two reference lubricant data sets are
compared using the F-test described in Section 1.4.2 (Equation 4). If the two data sets have similar
variances, and the F-test is passed, then at-test is performed on the data sets as discussed in Section 1.4.2
(equation 2) to determine whether the data sets are statistically from the same or different populations. If
the data sets have differing variances, the reference lubricant data should not be combined and should be
reviewed independently.
There are several methods of evaluating the fuel economy improvement from the reference lubricant to
the FEHP lubricant based on the results of the analysis of the reference lubricant. The final fuel economy
improvement value will be evaluated by each method for statistical significance and confidence interval
using the statistical methods described in Sections 1.4.2 and 1.4.3.
The fuel economy improvement can initially be determined based on each separate group of reference
lubricant data. Analysts will compare the initial reference lubricant mean fuel economy with the FEHP
lubricant mean fuel economy, as well as the final reference lubricant fuel economy with the FEHP
lubricant fuel economy. A maximum and minimum fuel economy improvement will then be reported,
each with a specified confidence interval.
The second method, which is valid if it is shown that the reference lubricant data is all from the same
population, is to pool all of the reference lubricant fuel economy data, determine a mean reference
lubricant fuel economy and evaluate the fuel economy improvement, statistical significance, and
confidence interval based on this pooled data set. This method can also be used if the two reference
1-10
-------
SRI/USEPA-GHG-VR-29
August 20'3
lubricant mean fuel economies differ, assuming that the change in mean reference lubricant fuel economy
is the result of test variability.
The third method, which would be applicable if it is demonstrated that the two reference lubricant data
sets are from distinct populations, is to determine a mean reference lubricant fuel economy based on the
means of each reference lubricant data group. If analysts observe a statistically significant change in
reference lubricant fuel economy, with each data set having a low variance, it can be assumed that the
change in reference lubricant fuel economy is the result of vehicle fuel economy drift due to break-in,
mileage effects, or various other vehicle dynamics that cannot be controlled during the test procedure. All
test data is normalized to account for the observed vehicle drift under this assumption. The normalization
parameter is based on the rate of vehicle drift. Assuming similar variance, the normalized reference
lubricant data should then be pooled together and the mean compared to the mean normalized reference
lubricant fuel economy.
Sections 2.1 through 2.4 discuss the details and results of the data analysis and the calculation of the fuel
economy improvement.
1.4.5 Testing and Measurement Equipment
The equipment used in determining the fuel economy of the test vehicle was specified in the test plan.
The following subsections provide details regarding specific equipment used during testing.
1.4.5.1 Chassis Dynamometer
This verification used SwRI's Chassis Dynamometer #7 and its associated sampling and analysis system
for light-duty gasoline vehicles. The chassis dynamometer is a Power Converter 48-inch single-roll
electric dynamometer manufactured by Horiba Instruments. The chassis dynamometer consists of the 48"
single roll, power converter, power-exchange unit motor, bearing-drive motors, CDC-900 computerized
dynamometer controller, and a RTM 200 real-time dynamometer monitor. This chassis dynamometer
uses a feed-forward control system for inertia-and road-load simulation. The dynamometer electrically
simulates vehicle tire/road interface forces, including parasitic and aerodynamic drag. The vehicle
experiences the same speed, acceleration/deceleration, and distance traveled as it would on the road. A
preprogrammed road-load curve is the basis for the required force during each second of the driving
schedule. Observed road load and simulated inertia errors are less than ± 0.3 percent for light-duty trucks.
1.4.5.2 Constant Volume Sampling System
A Horiba Variable-Flow constant volume sampling (CVS) system was used to sample exhaust emissions.
Figure 1-1 is a CVS system schematic.1-1-1
The vehicle exhaust pipe is connected to the CVS inlet. An adjustable-speed turbine blower pulls
ambient air into the CVS while the vehicle operates on the dynamometer. The air is used to dilute the
exhaust stream to prevent the exhaust moisture from condensing and provide controllable sampling
conditions to the analyzers (specifically, sample flow rate). A sample pump and control system transfer
diluted exhaust to several different Tedlar bags during specific phases of each FTP and HFET test run. A
regulating needle valve maintains a constant sample flow rate into the bags.
The balance of the dilute exhaust passes through a Horiba smooth-approach orifice (SAO) which
measures the flow rate. The bag sampling rate must remain proportional to the total dilute exhaust
volume flow rate throughout each test run to ensure that the sample represents the entire volume. SAO
1-11
-------
SRI/USEPA-GHG-VR-29
August 2003
throat pressure and temperature measurements using calibrated pressure and temperature transducers,
correlated with the SAO's National Institute of Standards and Technology (NIST) - traceable calibration,
allow accurate dilute exhaust volume determinations. This determination generates a feedback signal that
adjusts the turbine blower speed. The continuous adjustment allows the blower to maintain constant
volumetric flow through the CVS system. The CVS both measures the dilute exhaust volumetric flow
and controls the sample dilution ratio to within ±0.5 percent.
1.4.5.3 Emission Analyzers
Technicians used a Horiba analytical bench equipped with instrumental analyzers to determine carbon
monoxide (CO), carbon dioxide (C02), total hydrocarbons (THC), methane (CH4) and nitrogen oxides
(NOx) concentrations in the dilute exhaust. Each analyzer is accurate to ±2 percent. Sample pumps
transfer the dilute exhaust from the sample bags to each analyzer as commanded by the control system.
Figure 1-2 is a schematic of the instrumental analyzer system.1-3-1
The Horiba triple analytical bench consists of feedgas, tailpipe and bag analytical benches, a sample-
conditioning unit, and various automated flow controls. The Horiba instrumental emission analyzers used
to analyze exhaust emissions using the CVS bag cart are:
AIA-210 Infrared Low-Low CO Analyzer (LLCO)
AIA-220 Infrared C02 and Low CO Analyzer (C02/LC0)
FIA-220 Flame Ionization Total Hydrocarbons (THC) Analyzer
CLA-220 Chemiluminescent NO/NOx Analyzer
GC-FIA Gas Chromatographic/Flame Ionization Methane Analyzer
Q "w*
Figure 1-1. CVS System Schematic
1-12
-------
SRI/USEPA-GHG-VR-29
August 2003
Sampling, analysis, dynamometer monitoring, and other equipment or processes, including bag leak
checks, calibrations, and analyzer zero/span checks are all controlled by a Horiba VETS-9200
computerized emissions testing control system. The VETS-9200 collects data from the test equipment,
calculates and reports test results, and facilitates system calibrations and quality control checks. The
VETS also records raw sensor outputs, applies the appropriate engineering conversion and averaging
algorithms, and flags data which are outside the permitted values.
Figure 1-2. Instrumental Analyzer System
1.4.5.4 Ambient Monitoring Equipment
Ambient conditions of the test area can affect test results and analyses. SwRI maintains the test site at 74
±2 °F with target humidity control of 70 ±10 grains of water per pound of dry air. Technicians measure
dry and wet bulb temperatures with an Industrial Instruments and Supplies "Psychro-dyne" wet and dry
bulb thermometer. Accuracy is ±0.5°F, as verified with a NIST-traceable calibration thermometer.
Temperature data is input into the VETS-9200 systems and actual humidity is calculated by the system.
Barometric pressure in the test site is uncontrolled. SwRI uses a barometer that is calibrated weekly to
±0.01" Hg with a NIST-traceable barometer to determine test site barometric pressure.
1.4.5.5 Fuel Cart
An external cart fueled the vehicle from a five-gallon fuel container during testing. The fuel container was
filled from a single certified batch of test fuel throughout the test period. Analysis of test fuel samples
was completed to ensure compliance with test fuel specifications in 40 CFR 86-113. The test fuel is
1-13
-------
SRI/USEPA-GHG-VR-29
August 2003
discussed further in Sections 1.4.6.6 and 3.2.5. The fuel cart was used to determine gravimetric and
volumetric fuel consumption for use in performing fuel economy cross-check calculations during testing.
Figure 1-3 shows a schematic of the equipment involved.
The fuel container rests on a Fairbanks Model SB12-0806-5 scale. The scale's range is from 0 to 60 lb,
±0.5 percent of reading. The fuel passes through a Max Machinery Model 213 positive-displacement
piston-type volumetric flow meter with maximum flow rate of 0.4 gal/min, ±0.75 percent of reading. A
day tank with a level controller maintains a constant circulating flow for vehicles equipped with a fuel
return system on the engine fuel rail, such as the test vehicle. This ensures that the fuel cart functions in a
manner similar to the vehicle's original fuel system. The flow meter records the make-up flow to the day
tank.
0.1 psi
Vafwe
Liquid Fuel Line and Flow Direction
Fuel Vapor Line and Flow Direction
Figure 1-3. Fuel Cart Schematic
1.4.6 Testing Procedure and Sequence
The test procedures and details regarding each phase of the test are described in the test plan and
summarized in the following sections.
1.4.6.1 Vehicle Receipt and Initial Preparation
The test vehicle was obtained from a local rental agency fleet on March 26, 2003. The vehicle selection
process was described in Section 1.3.2 of this report. SwRI completed an inspection of the vehicle.
Technicians verified proper vehicle function and documented all pertinent test vehicle information upon
receipt of the vehilcle. Copies of test vehicle receipt documents are on file with the GHG Center. The
Ford F-150 that was received was accepted as the test vehicle, pending inspection of the rear axle and
gears during the initial axle lubricant change.
1-14
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
1.4.6.2 Engine Oil Change and Driver Familiarity Runs
SwRI used a triplicate oil flush and fill procedure in combination with driver familiarity runs (the same
driver is used for all test runs) to ensure a proper flushing of the engine oil system and installation of fresh
oil prior to testing. SwRI technicians completed the engine oil flush and driver practice for the reference
lubricant runs on March 27, 2003. The engine oil change for the FEHP runs was completed on April 16,
2003. The engine oil change for the second set of reference lubricant runs was completed on May 15,
2003. The procedure for all engine oil changes consisted of installing the vehicle on the Chassis
dynamometer, running a FTP and single HFET driving cycle (both for driver familiarity and to ensure a
hot oil drain), draining the oil, changing the filter, and filling with oil. This series was completed three
times for each oil change. The engine oil used for all oil changes was Motorcraft brand SAE 5W20 motor
oil, as recommended by the vehicle manufacturer. The vehicle was filled with 6.675 quarts of oil during
each oil change. Technicians provided documentation of each oil change procedure. This entire
procedure was observed by the SRI field team leader or SwRI Project Manager.
1.4.6.3 Rear Axle Lubricant Changes
The test vehicle was delivered to the SwRI Fleet Lab for changing the rear axle gear lubricant to the
reference lubricant on March 27 and May 15, 2003. The change of FEHP lubricant was completed on
April 17, 2003. SwRI Fleet Lab personnel completed the rear axle lubricant change procedure
documented in the Test Plan Appendix A-3 with some exceptions noted by SRI. Documentation of the
reference and FEHP rear axle lubricant changes, as well as engine oil changes, is maintained by the GHG
Center.
The axle lubricant changes and initial axle inspection involved cleaning the exterior of the axle housing to
remove loose dirt and debris. Technicians removed the wheels, brakes and rotors, and rear cover of the
axle differential. The existing lubricant was allowed to drain. This used lubricant was collected in a
preweighed pan to determine the amount of lubricant in the axle system and for inspection of the lubricant
for wear, debris, etc. Technicians then removed the axle shafts. The shafts, gears, and accessible areas of
the axle were wiped clean with absorbent pads. Interiors of axle shafts were not wiped as described in the
test plan due to the interference with bearings and associated retention clips (see CAR #2). The axles,
gears, and associated parts were sprayed thoroughly with NAPA Max 4800 Brake Cleaner to remove
residual lubricant. Compressed air was used to force lubricant and solvent out of the axle tubes into the
gear box, as well as to ensure solvent was evaporated and no residual lubricant remained. The cleaning
was continued until the SRI field team leader or SwRI Project Manager observed no residual lubricant or
solvent.
The axle shafts were reinstalled (each shaft only fits in the vehicle on a specific side) after thorough
cleaning and inspection. The spider pin, securing the pinion and gears, was reinstalled and secured using
a torque of 22 ft lbs. A sealant (Permatex 34311 "Right Stuff for Imports") specified by the axle vendor
was installed on the rear cover using a 1/8-inch bead along the outer edge, the inner edge, and around bolt
holes. The rear axle cover was reinstalled. All bolts were tightened to 33 ft lbs of torque. The sealant
was then left to cure overnight (a minimum of four hours is recommended).
The appropriate amount of axle lubricant (5.5 pints), as recommended by the manufacturer, was measured
using a clean graduated cylinder after the rear cover sealant cured. Reference lubricant was installed in
the axle by pouring from the graduated cylinder back into the product bottles (reference lubricant) and
charging through the fill hole. Final quantities of reference lubricant were charged using a clean syringe.
The axle was charged with FEHP lubricant by pouring from the graduated cylinder into a funnel
connected to clean tygon tubing that was inserted into the fill hole. Technicians verified fill levels for
1-15
-------
SRI/USEPA-GHG-VR-29
August 20'3
both lubricants at approximately 'A inch below the bottom of the differential fill hole. The fill hole and
vent plugs were then replaced. The parking brake adjuster was backed off and brakes and wheels
reinstalled on the test vehicle.
1.4.6.4 Vehicle and Axle Inspection and Gear Rating
An SwRI certified gear rater inspected and rated the axle and gears during cleaning and installation of the
reference lubricant to ensure proper axle functioning for the testing process. SwRI retains certified
inspectors on staff for conducting such inspections. The axle, gears, seals, and bearings were inspected
visually and measurements taken to determine if any excess wear or damage to the axle was present.
SwRI completed and documented the following inspections:
Visual inspection of old lubricant
Verification of axle ratio
Visual inspection and rating of gears, axle shafts, bearings, and seals
Video bore scope of ring and pinion teeth and housing interior
Digital photographs of ring and pinion teeth
Video of all exposed shaft seals and bearings
Measurement of break and turn
Measurement of backlash
Documentation of these inspections is on file with the GHG Center. All inspections indicated that the
axles had appropriate wear for the vehicle age, no damage, and measurements within specifications.
Therefore, the field team leader approved the test vehicle and axle for use in the test protocol.
1.4.6.5 Mileage Accumulation
The vehicle was sent to SwRI's mileage accumulation facility after installation of each test lubricant to
accumulate 1000 miles to ensure proper break-in of the axle and engine lubricants. Technicians mounted
the vehicle on an eddy current-type mileage accumulation dynamometer (MAD) with 24-hour
capabilities. The MAD system incorporates a computer-based control system to operate the vehicle. The
control system maintains vehicle load and speed with a throttle actuator and electric motor. A large
blower provides airflow proportional to vehicle speed across the vehicle for cooling.
The dynamometer was operated over the Durability Driving Schedule (DDS) specified in 40 CFR 86,
Appendix IV. This test cycle is 4,960 seconds long at 29.5 mph average speed, including eleven 3.7-mile
"laps" at various speeds. Accumulation of 1000 miles required approximately two days. Detailed
mileage accumulation data logs are maintained by the GHG Center and SwRI.
1.4.6.6 Test Fuel and Fuel System Preparation
The vehicle fuel system was modified to accept fuel from the fuel cart system. SwRI technicians
completed modifications to provide quick connection to the fuel cart for the fuel feed and return fuel
lines. The fuel system was also drained and flushed prior to testing and only test fuel from a specified lot
was used during testing. The test fuel used met the requirements specified in 40 CFR 86.113. SwRI
completed analyses of the test fuel to verify fuel properties. Table 1-4 specifies the allowable and actual
test fuel properties for two fuel samples. Test fuel analyses provided by the manufacturer and analytical
results provided by SwRI are on file with the GHG Center. Test fuel quality and analytical data quality
are discussed further in Section 3.2.5.
1-16
-------
SRI/USEPA-GHG-VR-29
August 20'3
"able 1-4: Test Fuel Specifications
Parameter
Allowable Result
Actual Result
Sample ID
ETV1
ETV2
Octane, research
93 minimum
96.7
96.8
Sensitivity (research octane minus motor
octane)
7.5 minimum
7.8
7.8
Lead
0.050 g/U.S. gal maximum
<0.001g/gal
<0.001 g/gal
Distillation range
Initial Boiling Point
75 to 95°F
93.3°F
94.9°F
10 pet. Point
120 to 135°F
128.5°F
129.6°F
50 pet. Point
200 to 230°F
219.5°F
220.6°F
90 pet. Point
300 to 325°F
319.6°F
321.4°F
End Point
415 °F maximum
406.8°F
404.9°F
Sulfur
0.10 wt. percent maximum
0.0033%
0.0032%
Phosphorus
0.005 g/US gallon maximum
0.0001 g/gal
<0.0001g/gal
Reid Vapor Pressure
8.0 to 9.2 psi
9.15
9.11
Hydrocarbon composition
Olefins, max. pet
10% maximum
1.0
0.8
Aromatics, max. pet
35% maximum
30.9
31.2
Saturates
remainder
68.1
68
1.4.6.7 Dynamometer Setup
The chassis dynamometer requires appropriate setup to ensure proper road load and inertia simulation
specific to the test vehicle. SwRI technicians completed triplicate 65- to 15-mph coastdowns on each axle
with the reference lubricant initially installed in the rear axle and again with the FEHP in the rear axle.
This data was used to define the appropriate setup data. The coastdown data was used as input to the
Mears Model according to EPA-recognized least-square methods. The Mears Model calculates a three-
parameter road load force equation for dynamometer fuel economy tests.(7) This model incorporates
frictional coastdown data from drive and non-drive axles with wind and aerodynamic resistance
projections to yield the dynamometer force simulation equation "a", "b", and "c" coefficients.
Dynamometer setup coefficients were obtained for each lubricant to ensure that lubricant changes would
not improperly affect dynamometer simulation. The same front axle coastdown data was used in both
Mear's Model calculations to determine dynamometer coefficients because lubricants were not changed
in the front axle. SwRI used the triplicate coastdown data with the Mears Model to yield the following
dynamometer setup coefficients for each set of test runs:
Table 1-5: Dynamometer Setup Coefficients
Dynamometer
coefficient
Reference
Lubricant-Initial
FEHP
Lubricant
Reference
Lubricant-Final
A
19.37
19.37
19.37
B
0.31504
0.31504
0.31504
C
0.03248
0.03241
0.03227
1-17
-------
SRI/USEPA-GHG-VR-29
August 20'3
Proper dynamometer setup and simulation was verified by completing triplicate coastdown checks. These
data quality checks are discussed further in Section 3.2.1.
1.4.6.8 Bag Cart and Emission Analyzer Setup
Exhaust sampling is conducted during four phases of the fuel economy test driving schedule. This is
described further in Section 1.4.7. The exhaust is diluted with ambient air primarily to avoid
condensation and, in this case, to ensure that sample concentrations were within the calibrated span of the
analytical bench. Technicians determined estimated vehicle emission concentrations that were used to
estimate CVS bag flow rates for input into the VETS 9200 system. CVS flow rates used for each of the
four phases of fuel economy testing were 450 scfm, 350 scfm, 450 scfm, and 550 scfm, respectively
based on results of preconditioning runs. The flow rates were determined for each individual phase to
ensure that concentrations for each pollutant were within analyzer spans. This was also done to ensure
that concentrations from each phase were within the same span for the C02 analyzer for all phases. This
limited sampling and analytical variability during testing by limiting the number of calibration gases used
during analyzer spans.
Ambient air samples are simultaneously collected and analyzed in conjunction with exhaust sampling.
The VETS 9200 system automatically determines the dilution factor for the collected exhaust and ambient
pollutant concentrations for the dilution air. The system then calculates the actual exhaust concentration
based on the dilution rate, the dilute sample concentration, and ambient concentration for the bag pair for
each phase.
Technicians set the emission analyzer ranges (as shown in Table 1-6) based on the programmed CVS
flow rates. A higher CO range was required for phase 1 because it is a cold-start phase that typically
results in higher CO emissions.
Table 1-6: Emission Analyzer Ranges
Phase l(FTP)
Phase 2 (FTP)
Phase 3 (FTP)
Phase 4 (HFET)
CO
0-200 ppm
0-25 ppm
0-25 ppm
0-25 ppm
co2
0-4.0%
0-4.0%
0-4.0%
0-4.0%
NOx
0-10.0 ppm
0-10.0 ppm
0-10.0 ppm
0-10.0 ppm
THC
0-30 ppm C
0-30 ppm C
0-30 ppm C
0-30 ppm C
ch4
0-10 ppm
0-10 ppm
0-10 ppm
0-10 ppm
1.4.6.9 Vehicle Preconditioning
The test vehicle was "preconditioned" prior to beginning a series of test runs or after any soak period
greater than 24 hours. Preconditioning consists of running the vehicle through a complete fuel economy
test cycle (FTP and HFET) to condition the vehicle to the test cycle. Preconditioning is an attempt to
limit variability in testing by allowing the vehicle's adaptive controls to become familiar with the test
driving cycle.
Triplicate coastdown checks were run to verify proper dynamometer setup and road-load simulation
during each test period. This was performed either after a test run or after a preconditioning run. The
1-18
-------
SRI/USEPA-GHG-VR-29
August 20'3
driver ran the vehicle through a single HFET cycle after coastdowns to ensure consistency in the vehicle
driving patterns from test to test. This ensures that the last driving cycle the test vehicle saw prior to soak
and the next test was a HFET cycle.
1.4.7 Reference Lubricant and FEHP Fuel Economy Test Procedure
The vehicle was returned to the SwRI light-duty testing facility after mileage accumulation where it was
stored overnight in a shed inside the facility. The fuel economy testing procedure began after
preconditioning. The test vehicle was operated over two test cycles to determine fuel economy: (1) the
FTP as specified in 40 CFR 86.115 and (2) the HFET specified in 40 CFR 600 Appendix I. The FTP
simulates an 11-mile trip in an urban area. It includes stop-and-go driving, multiple vehicle starts, and a
short freeway driving segment. Average speed is about 20 miles per hour. It consists of four phases: (1)
a cold-start "transient" phase; (2) a stabilized phase; (3) a 10-minute soak period (vehicle off); and (4) a
hot-start transient phase.
The highway portion of the test (HFET) commenced immediately following the end of the FTP segment.
This dynamometer run employs a "hot" vehicle start and represents a 10-mile trip with an average speed
of 48 mph with little idling and no stops. The HFET cycle consists of a warm-up phase and the sampling
phase during which exhaust samples are collected.
Table 1-7 summarizes the daily test procedure. The Horiba VETS 9200 controlled most of the testing
procedure automatically based on user inputs. The sampling procedure for each test sampling phase
consisted of the completion of an initial screening test to verify the required analyzer span, an automated
zero and span check using system calibration gases, analysis of bag samples, recheck of analyzer zeroes,
and calculation of emission rates and fuel economy. Technicians completed a daily test checklist to
summarize the test procedures, test parameters, and some QA/QC checks.
The vehicle was stored in the same location overnight inside the light-duty testing area. Temperature and
humidity are controlled within the light-duty testing area. The test vehicle was stored in an open shed
inside the test area. The fuel cart and test fuel container were also stored in the same areas to ensure
consistency.
The Test Plan states that a maximum of seven reference lubricant tests were to be performed. SwRI
completed eight initial reference lubricant test runs, including three tests that were voided, resulting in a
total of five valid initial reference lubricant test runs. The axle lubricant was changed to the FEHP
lubricant and seven FEHP test runs were completed. One FEHP test run was voided, resulting in the
completion of six valid FEHP test runs. After completion of the FEHP testing, the axle lubricant was
changed back to the reference lubricant, and the entire test procedure repeated for six additional valid
tests (including mileage accumulation, engine oil changes, etc.). Test runs were voided based on results
of specific QA/QC checks or equipment error, as discussed in Section 2.1.1. This test procedure allowed
the GHG Center to verify that the fuel economy improvement observed was attributable solely to the use
of the FEHP lubricant and not to changes in vehicle performance as the result of additional mileage
accumulation and vehicle break-in over the course of the test period.
1-19
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
Table 1-7: Daily Test Procedure
1
Warm up dynamometer for a minimum of 15 minutes at average
speed of 50 mph
2
Run a parasitic friction curve on the dynamometer and verify no
losses > +/- 2.0 lbf
3
Setup Horiba VETS 9200 for test vehicle
4
Push vehicle from soak area onto dynamometer and tie down
5
Record vehicle odometer reading
6
Check vehicle tire pressure and inflate, if necessary
7
Install fuel cart system
8
Install fan
9
Determine ambient conditions and input into VETS 9200
10
Begin VETS 9200 testing procedure
11
Verily automatic bag leak check completed
12
Verily proper CVS flow rate
13
Start fuel cart and allow to circulate
14
Start Phase 1 (FTP) of test (samples automatically collected)
15
End Phase 1 - Begin Phase 2 (FTP)
16
Complete automatic analysis of Phase 1 samples
17
End Phase 2, Complete 10-minute soak. Begin Phase 3 (FTP)
18
Complete automatic analysis of Phase 2 samples
19
End Phase 3 - Begin Phase 4 (HFET)
20
Complete automatic analysis of Phase 3 samples
21
End Phase 4
22
Complete automatic analysis of Phase 4 samples
23
Disconnect vehicle and push to inside shed for overnight soak.
1.4.8 Reference Lubricant and FEHP Fuel Economy Determination
Composite fuel economy is determined from the quantity of carbon in the vehicle exhaust emissions
measured during the two driving cycles, the amount of carbon in the fuel, and the distance driven on the
dynamometer. This is the "carbon balance" method. This method generates a fuel economy value (in
mpg) by dividing the carbon mass in the fuel per unit volume by carbon mass in the emissions per mile:
S carbon, juel /
m/gal(or mpg) = / (Eqn- 9)
S carbon,emissions /
/mi
where:
mpg = vehicle fuel economy, miles per gallon
g = grams of carbon
mi = miles
1-20
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
The calculation relies on determination of carbon masses based on CO, C02, and THC mass emission
rates (in grams per mile or g/mi) as measured by the Horiba VETS 9200 emission testing system, the
measured test fuel carbon weight fraction, fuel specific gravity, and net heating value (as determined by
test fuel analyses). Emission rate determination procedures are specified in 40 CFR 80.144-94.
Weighted mass emissions are determined for the FTP cycle based on 40 CFR 86.114-94 criteria as
follows:
Ym =0.43
{Yct+Ysy
(Dct + Ds)
+ 0.57
(Yht+Ys)~
(Dht + Ds)
(Eqn.10)
where:
Ywm = weighted mass emissions of each pollutant, g/mi
Yct = mass emissions of each pollutant from the cold-start "transient" phase (Phase 1),
g/mi
Ys = mass emissions of each pollutant from the cold-start "stabilized" phase (Phase 2),
g/mi
Yht = mass emissions of each pollutant from the hot-start "transient" phase (Phase 3),
g/mi
Dct = Driving distance for the cold-start "transient" phase, mi
Ds = Driving distance for the cold-start "stabilized" phase, mi
Dht = Driving distance from the hot-start "transient" phase, mi
The FTP or HFET fuel economy is determined from 40 CFR 600.113 (e):
{5\1A*\04)*CWF*SG
mPg ~ [CWF * HC + (0.429 * CO) + (0.273 * C02)] * [0.6 * SG * LHV + 5471]
where:
mpg =
= miles per gallon
CWF =
= carbon weight fraction in the fuel
SG
= fuel specific gravity
HC
= total hydrocarbon emission rate, g/mi
CO
= carbon monoxide emission rate, g/mi
o
o
= carbon dioxide emission rate, g/mi
LHV =
= fuel lower (or net) heating value, Btu/lb
1-21
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
The composite fuel economy depends on the FTP- and HFET-cycle fuel economies. The composite fuel
economy is weighted based on the typical proportion of city (FTP) driving, 55 percent, to highway
(HFET) driving, 45 percent, for light-duty vehicles as specified by the US EPA. The equation for
composite fuel economy is:
mPgcomposite = ~^ 1 Q ^ (Eqn- 12)
mP§FTP mP§HFET
where:
mpgcomposite = composite fuel economy, mpg
mpgFTP = mean FTP fuel economy, mpg
mpgHFET = mean HFET fuel economy, mpg
The mean fuel economy (to be used as input to Equation 1) for either the reference lubricant or the FEHP
is:
Mean Fuel Economy (|-i) = 'V P^comP°slte (Eqn. 13)
11
1
where:
= average of all valid reference lubricant or FEHP test runs, mpg
n = number of test runs
Additional detailed calculations of emission rates and fuel economy are contained in 40 CFR 86.144.
Fuel economy was also determined for each test run by separate volumetric and gravimetric methods
using fuel cart data as a cross-check against the carbon balance method. The volumetric method
correlates the volume of gasoline (gallons) consumed during a test run with the dynamometer distance
traveled (miles) to yield mpg. The gravimetric method correlates the weight of gasoline consumed
(grams), its density (g/1), and the dynamometer distance traveled (miles) to yield mpg.
1.4.9 Pollutant and GHG Emissions
Each fuel economy test also provided emissions data for greenhouse gases (C02, CH4) and other
pollutants (NOx, CO, THC, and non-methane hydrocarbons [NMHC]). Emissions in g/mi are an
intermediate determination during the fuel economy testing and calculation procedure and are
automatically calculated by the Horiba VETS 9200 control system. Emission rates for CO, C02, NOx,
THC, and CH4 are determined using the analytical equipment and procedures described in Sections 1.4.5
and 1.4.6. NMHC emission rates are calculated by the Horiba VETS 9200 based on the THC and CH4
emission rates in accordance with 40 CFR 86.144. Section 2.5 summarizes the GHG emissions for the
test vehicle for both the reference and FEHP lubricants.
1-22
-------
SRI/USEPA-GHG-VR-29
August 20'3
2.0 VERIFICATION RESULTS
The test vehicle was acquired on March 26, 2003. Vehicle setup, axle lubricant change, mileage
accumulation, and preconditioning occurred between March 26 and April 1, 2003. The fuel economy
testing verification period started on April 2, 2003. Testing was completed on May 31, 2003.
The GHG Center acquired several types of data during the verification testing periods that represent the
basis of verification results presented here. The following types of data were collected and analyzed
during the verification:
Test vehicle fuel economy with reference and FEHP lubricants
Greenhouse gas and other pollutant emissions with reference and FEHP lubricants
Information was collected throughout the testing period to evaluate data quality and ensure the accuracy
of verification results. This information and associated review are discussed further in Section 3.0 - Data
Quality Assessment.
The field team leader reviewed, verified, and validated some data (test run results, statistical analysis,
QA/QC data) while on-site. He reviewed collected data for reasonableness and completeness. The data
from each of the fuel economy tests was reviewed on-site or within 24 hours, when possible, to verify that
test criteria were met. The field team leader validated emissions testing data by reviewing instrument and
system calibration data and ensuring that those and other reference method criteria were met. Calibration
and verification data for test equipment, including the dynamometer, CVS system, analyzers, ambient
monitoring equipment, and calibration gases were reviewed and verified prior to and during testing to
ensure proper function and accuracy. The field team leader classified all collected data as valid, suspect,
or invalid using the QA/QC criteria specified in the Test Plan. Review criteria were in the form of factory
and on-site calibrations, maximum calibration and other errors, audit gas analysis results, and lab
repeatability results. All results presented here are based on measurements that met the specified Data
Quality Indicators (DQIs) and QC checks as validated by the GHG Center during the testing period. DQI
goals were not completely satisfied for the entire test period (discussed in Section 3.0). However, this did
not result in a loss of data quality, as all QA/QC checks were satisfied during testing and data quality
objectives were met.
The observed fuel economy change resulting from use of the FEHP lubricant in the test vehicle (2003
Ford F-150 with beam axle) and the corresponding 95-percent confidence interval is:
A = 0.169 ± 0.0410 mpg
Section 2.1 discusses the evaluation of the change in fuel economy. Section 2.2 discusses the statistical
significance of the fuel economy change. Section 2.3 discusses the calculation of the 95-percent
confidence interval of the fuel economy change.
The verification test results provide the fuel economy change for a single representative vehicle and axle
setup (2003 Ford F-150 with beam axle) under the testing and driving conditions specified in the Test
Plan. This test vehicle and axle setup represent a large portion of the light-duty trucks currently in
production in the US. The test conditions and verification parameters were developed to obtain a
reasonable and representative set of data to examine fuel economy savings resulting from the use of the
FEHP lubricant in light-duty trucks. Performance at significantly different operating conditions or for
different vehicle and axle types can, however, affect the results from these types of test programs.
2-1
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
2.1. FUEL ECONOMY IMPROVEMENT
2.1.1. Fuel Economy Test Results
The GHG Center determined the change in fuel economy attributable to the use of the FEHP lubricant in
the rear axle of the test vehicle based on the means of several test runs of both the reference lubricant and
the FEHP lubricant. Table 2-1 presents the composite fuel economy results of each individual test run
and the mean fuel economy for each set of lubricant runs. The test data for runs Base-1 through Base-
5R2 have been corrected for a C02 emissions analyzer calibration error, as discussed in Section 3.2.3.
Table 2-1: Fuel Economy Test Results
Test Run ID
Date
Composite Fuel Economy (mpg)
Reference Lubricant
Base-1
4/2/03
18.070
Base-2
4/3/03
18.013
Base-3
4/4/03
(17.663) VOID-outlier
Base-4
4/8/03
17.994
Base-5
4/9/03
VOID - driver trace error
Base-5-R2
4/10/03
VOID - calibration gas error
Base-6
4/11/03
18.055
Base-7
4/12/03
17.973
Mean
18.021
Standard Deviation
0.0408
FEHP Lubricant
FEHP-1
4/23/03
18.272
FEHP-2
4/24/03
VOID - equipment error
FEHP-2-R2
4/25/03
18.272
FEHP-3
4/29/03
18.284
FEHP-4
4/30/03
18.233
FEHP-5
5/1/03
18.263
FEHP-6
5/2/03
18.206
Mean
18.255
Standard Deviation
0.0296
Reference Lubricant
Post Base-1
5/22/03
VOID - incorrect dyno settings
Post Base-1R2
5/23/03
18.208
Post Base-2
5/24/03
18.111
Post Base-3
5/28/03
18.143
Post Base-4
5/29/03
18.169
Post Base-5
5/30/03
18.121
Post Base-6
5/31/03
18.082
Mean
18.139
Standard Deviation
0.0448
2-2
-------
SRI/USEPA-GHG-VR-29
August 20'3
Five test runs were voided after review of data and crosschecks or during testing for various reasons. The
GHG Center voided reference lubricant run Base-5 because of a driver error. The driver is required to
follow or "trace" a specified route during the test period. The accuracy of the driving trace must meet
certain specifications set forth in the CFR. The driver compiled a deviation from the simulation speed
that exceeded the trace error limits for 2.0 seconds. Therefore, this run was invalidated in accordance
with the CFR requirements.
A review of test results and gravimetric and volumetric cross-checks for run Base-5R2 indicated
inconsistency in the data when compared to other runs. SwRI reviewed the data and determined that the
4-percent C02 calibration gas ran low during the test, causing error in the C02 measurement. The GHG
Center therefore voided run Base-5R2.
Technicians determined that the bags for sample collection during run FEHP-2 were not properly installed
on the CVS sampling system. Therefore, the run was invalidated due to "equipment error". The run was
completed to ensure consistency in the test pattern, but data was discarded. Test run Post Base-1 was
voided because the incorrect dynamometer settings were input into the Horiba VETS 9200 system prior to
testing. Test run Base-3 was voided because it was determined to be an outlying data point as determined
via the American Society for Testing Materials (ASTM) Standard Practice for Dealing with Outlying
Observations (E 178-02). An outlying data point, as defined by ASTM, is one that appears to deviate
markedly in value from other members of the sample in which it appears.(8) The analysis of run Base-3 as
an outlier using the ASTM procedure is presented in Appendix E.
2.1.2. Fuel Economy Change
The fuel economy change resulting from the use of FEHP, as presented in Section 2.0, is calculated as
discussed in Section 1.4 by comparing the FEHP fuel economy test results with the reference lubricant
fuel economy test results. An initial review of the data indicates that there is an observed increase in fuel
economy from each reference lubricant data set to the FEHP lubricant data set. There is also an observed
increase in fuel economy from the initial reference lubricant runs to the final reference lubricant runs.
Therefore, analysts must evaluate the reference lubricant data to determine the overall reference lubricant
mean fuel economy for comparison to the FEHP mean fuel economy.
There are three ways that the fuel economy change and reference lubricant results can be analyzed:
(1) Determine that there is no statistical difference in reference lubricant fuel economies from the
initial to final data sets. In this case, all reference lubricant data is pooled and compared to the
FEHP data.
(2) Compare each individual set of reference lubricant data to the FEHP data to obtain a range of fuel
economy changes based on the two data sets.
(3) Determine that the two reference lubricant data sets are statistically different and cannot be
directly pooled. Assume that the change in reference lubricant fuel economy from pre-FEHP to
post-FEHP is the result of a systematic drift in vehicle performance. In this case, all data can be
normalized to account for such systematic changes. The normalized reference lubricant data is
then pooled and compared to the normalized FEHP data.
Section 2.1.3 presents the statistical evaluation of the reference lubricant fuel economy to determine
which of the three fuel economy change calculations is most appropriate for data analysis. Section 2.1.4
applies the fuel economy change calculation. Section 2.2 discusses the evaluation of the statistical
significance of the fuel economy change. Section 2.3 presents the calculation of the confidence interval
2-3
-------
SRI/USEPA-GHG-VR-29
August 20'3
for the fuel economy change. Section 2.4 describes the application of the other two calculation methods
to determine fuel economy change as a cross-check.
2.1.3. Reference Lubricant Fuel Economy
Analysts evaluated the two sets of reference lubricant fuel economy data (discussed in Section 1.4.4) to
determine the statistical significance of the difference in mean fuel economy between the data sets. An
F-test (discussed in Section 1.4.2) was completed on the two reference lubricant data sets to compare the
data variance of the two groups. Table 2-2 presents the results of the F-test.
Table 2-2: F-test Evaluation of Reference Lubricant
Fuel Economy Data Set Variances
Parameter
Value
Standard Deviation, initial reference lubricant tests (mpg)
0.0408
Standard Deviation, final reference lubricant tests (mpg)
0.0448
F test
1.207
Fci.05
5.192
F test < F0.05 (variances statistically equivalent)?
Yes
Results of the F-test indicate that the two sets of reference lubricant data have equivalent variances at a
95-percent confidence level. Therefore, analysts applied the t-test to evaluate the statistical significance
of the change in fuel economy between the two reference lubricant data sets. Table 2-3 presents the
results of the t-test analysis for the two reference lubricant data sets.
The t-test results indicate that there is a statistically significant difference between the two reference
lubricant fuel economy data sets at a 95-percent confidence level. This analysis and SwRI's previous
experience indicate that it is likely that the change in fuel economy is the result of a systematic drift in
vehicle performance due to mileage effects or other phenomena. Therefore, analysts calculated the fuel
economy improvement using the method discussed in Section 2.1.4.
Table 2-3: Statistical Analysis Of Reference Lubricant Tests-Fuel Economy Difference
Parameter
Value
Initial reference lubricant standard deviation (mpg)
0.0408
Final reference lubricant standard deviation (mpg)
0.0448
Mean fuel economy - initial reference lubricant (mpg)
18.021
Mean fuel economy - final reference lubricant(mpg)
18.139
Change in fuel economy (mpg)
0.118
Change in fuel economy (%)
0.655
COV-Initial reference lubricant (%)
0.226
COV-Final reference lubricant (%)
0.247
Initial reference lubricant test count
5
Final reference lubricant test count
6
Total count
11
Degrees of freedom
9
(Pooled std. dev.)2
0.0019
(Pooled std. dev.)
0.043
Critical t-distribution value (t 0.025. df )
2.262
Calculated t-test value, ttest
4.525
ttest>t 0 025,df (Is the change statistically significant?)
YES
2-4
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
2.1.4. Fuel Economy Change
The two reference lubricant data sets are statistically independent based on the statistical analysis of the
reference lubricant fuel economy data presented in Section 2.1.3. Analysts must compare the complete
reference lubricant data set and FEHP lubricant test results to determine a representative fuel economy
change resulting from the use of FEHP lubricant. No viable explanation for the shift in reference
lubricant fuel economy was determined after review of test and QA/QC data. SwRI concluded that there
was a "drift" in vehicle performance associated with the mileage accumulation on the test vehicle. The
GHG Center evaluated the test data by making the assumption that, during this test period, vehicle drift
occurred and the drift exhibits a linear behavior with fuel economy improving with mileage accumulation.
The fuel economy data for all runs were normalized to remove the effects of the observed linear vehicle
performance drift. Any fuel economy change calculated for the normalized data set was attributable
solely to the FEHP lubricant and not mileage or other effects.
A linear regression was performed on the reference lubricant data (initial and final) to complete the
normalization. This provides the linear drift relationship. Table 2-4 presents the results of the linear
regression. Figure 2-1 presents the fuel economy results versus vehicle mileage with the linear regression
results.
Table 2-4: Reference Lubricant Data Regression Statistics
Parameter
Value
Intercept
17.397
Slope
3.86E-05
Standard error - intercept
0.163
Standard error - slope
9.10E-06
R-square
0.6664
Regression sum of squares
0.0364
Residual sum of squares
0.0182
Observations
11
2-5
-------
SRI/USEPA-GHG-VR-29
August 2003
18.5
18.3
o>
Q.
18.1
>
E
o
c
o
o
w 17.9
0)
3
y=0.0000386x+ 17.397
R2 = 0.6664
17.7
17.5
16000
17000 18000
Odometer (miles)
19000
Figure 2-1. Variation of Reference Lubricant Fuel Economy Results with Mileage
All test data (reference lubricant and FEHP) was normalized to a common point for comparison based on
the reference lubricant regression. The GHG Center normalized the test data to the y-intercept. Data was
normalized using the following equation:
FEnj = FEt
b
mx, + b
(Eqn. 14)
where:
FEN,i
FE,
m
b
Xi
normalized fuel economy for test run i
fuel economy for test run i
slope of "drift" line
intercept of "drift" line
vehicle odometer reading at beginning of test run i
2-6
-------
SRI/USEPA-GHG-VR-29
August 20'3
Table 2-5 presents the results of the normalization procedure. Figure 2-2 presents the normalized test
results as a function of mileage.
Table 2-5: Normalized Fuel Economy Test Results
Test Run
ID
Composite Fuel Economy
(mpg)
Normalized
Fuel Economy (mpg)
Reference Lubricant
Base-1
18.070
17.448
Base-2
18.013
17.392
Base-4
17.994
17.370
Base-6
18.055
17.425
Base-7
17.973
17.345
Mean
18.021
17.396
Standard Deviation
0.0408
0.0414
FEHP Lubricant
FEHP-1
18.272
17.588
FEHP-2-R2
18.272
17.584
FEHP-3
18.284
17.594
FEHP-4
18.233
17.543
FEHP-5
18.263
17.571
FEHP-6
18.206
17.515
Mean
18.255
17.566
Standard Deviation
0.0296
0.0307
Reference Lubricant
Post Base-1R2
18.208
17.468
Post Base-2
18.111
17.374
Post Base-3
18.143
17.402
Post Base-4
18.169
17.426
Post Base-5
18.121
17.379
Post Base-6
18.082
17.340
Mean
18.139
17.398
Standard Deviation
0.0448
0.0447
2-7
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
~ Reference Lubricant Fuel Economy
FEHP Fuel Economy
~
~
«
~ ~
y = 0.00x + 17.397
17.0-
16000 17000 18000 19000
Odometer Reading (miles)
Figure 2-2. Comparison of Normalized Reference Lubricant and
FEHP Fuel Economy Results
Analysts evaluated the normalized reference lubricant data to determine if the two data sets are from the
same population and can, therefore, be pooled to determine a mean reference fuel economy for
comparison to the normalized FEHP fuel economy. The normalized reference lubricant data was
evaluated as discussed in Section 2.1.3. An F-test was initially completed on the two normalized
reference lubricant data sets to compare the data variance of the two groups. Table 2-6 presents the
results of the F-test.
Table 2-6: F-test Evaluation of Reference Lubricant
Fuel Economy Data Set Variances
Parameter
Value
Standard deviation, initial reference lubricant tests (mpg)
0.0414
Standard deviation, final reference lubricant tests (mpg)
0.0447
F test
1.166
Fo.05
5.192
F test < F0 o5 (variances equal)?
Yes
18.0
O)
Q.
E
>
E
o
£
O
o
LU
a3
3
17.8
17.6
¦a
a>
N
re
E
L.
o
17.4
17.2
2-8
-------
SRI/USEPA-GHG-VR-29
August 20'3
Results of the F-test indicate that the two sets of normalized reference lubricant data have equivalent
variances at a 95-percent confidence level. Therefore, analysts applied the t-test to evaluate the statistical
significance of the change in fuel economy between the two normalized reference lubricant data sets.
Table 2-7 presents the results of the t-test analysis for the two normalized reference lubricant data sets.
Table 2-7: Statistical Analysis of Normalized Reference
Lubricant Fuel Economy Difference
Parameter
Value
Initial reference lubricant standard deviation (mpg)
0.0414
Final reference lubricant standard deviation (mpg)
0.0447
Mean fuel economy - initial reference lubricant (mpg)
17.396
Mean fuel economy - final reference lubricant (mpg)
17.398
Change in fuel economy (mpg)
0.002
Change in fuel economy (%)
0.011
COV-reference lubricant (%)
0.238
COV-FEHP lubricant (%)
0.257
Reference lubricant test count
5
FEHP test count
6
Total count
11
Degrees of freedom
9
(Pooled std dev)2
0.0019
(Pooled std dev)
0.043
Critical t distribution value (t 0.025. df )
2.262
Calculated t-test value, ttest
0.076
0 025,df (Is the change statistically significant?)
NO
The t-test results indicate that there is not a statistically significant difference between the two normalized
reference lubricant fuel economy data sets at a 95-percent confidence level. The two data sets have
statistically equivalent means and are from the same population. Therefore, the reference lubricant data
was pooled. Table 2-8 presents the results of the pooled reference lubricant data analysis.
Table 2-8: Summary of Pooled Normalized Reference Lubricant Data
Parameter
Value
Reference lubricant mean normalized fuel economy (mpg)
17.397
Standard deviation (mpg) - pooled normalize reference lubricant
0.0411
COV-pooled normalized reference lubricant (%)
0.236
The mean pooled, normalized reference lubricant fuel economy is compared to the mean normalized
FEHP fuel economy to determine the change in fuel economy resulting from the use of the FEHP
lubricant. The calculated fuel economy improvement attributable to the use of the FEHP lubricant in the
test vehicle is
A = 17.566 mpg - 17.397 mpg = 0.169 mpg
This represents a 0.97 percent improvement in fuel economy using the FEHP lubricant when compared
to the reference lubricant fuel economy.
2-9
-------
SRI/USEPA-GHG-VR-29
August 20'3
Sections 2.3 and 2.4 present the evaluation of the statistical significance and the determination of the 95-
percent confidence interval of the calculated fuel economy improvement. Section 3.0 provides a further
assessment of the quality of data collected throughout the verification period. The data quality
assessment is used to demonstrate whether or not the data quality objectives (DQOs) introduced in the
Test Plan were met for this verification.
2.2. FUEL ECONOMY CHANGE STATISTICAL SIGNIFICANCE
The GHG Center analyzed the calculated fuel economy change data as discussed in section 2.1 to
determine the statistical significance of the data. Table 2-9 summarizes the results of this analysis.
Table 2-9: Statistical Analysis of Normalized Fuel Economy Change
Parameter
Value
Normalized reference lubricant standard deviation (mpg)
0.0411
Normalized FEHP lubricant standard deviation (mpg)
0.0307
Mean fuel economy - normalized reference lubricant (mpg)
17.397
Mean fuel economy - normalized FEHP (mpg)
17.566
Change in fuel economy (mpg)
0.169
Change in fuel economy (%)
0.971
COV-reference lubricant (%)
0.236
COV-FEHP lubricant (%)
0.175
Reference lubricant test count
11
FEHP test count
6
Total count
17
Degrees of freedom
15
(Pooled standard deviation)2
0.0014
(Pooled standard deviation)
0.038
Critical t distribution value (t 0.025. df )
2.131
Calculated t-test value, ttest
8.777
ttest > t 0.025.DF ?
YES
Section 1.4.2 stated that if the t-test for the verification test data is greater than the t-distribution values
using a 95-percent confidence coefficient, the measured fuel economy change is deemed statistically
significant. A statistically significant fuel economy savings was observed based on the analysis shown in
Table 2-9. The confidence interval for the fuel economy savings was therefore calculated.
2.3. FUEL ECONOMY SAVINGS CONFIDENCE INTERVAL
The 95-percent confidence interval represents the range of values in which 95 percent of the fuel
economy data is expected to lie. A narrow confidence interval indicates a sharply characterized mean
fuel economy change. The method used to determine the 95-percent confidence interval depends upon
the relative variability, or variances, of the data set for each lubricant. Equation 7 can be applied to data
sets with similar variances to determine the 95 percent confidence interval half-width. The evaluation of
relative variance is completed using the F-test (Section 1.4.2). Results for the F-test evaluation are
presented in Table 2-10.
2-10
-------
SRI/USEPA-GHG-VR-29
August 20'3
Table 2-10: F-test Evaluation of Normalized Fuel Economy Data Set Variances
Parameter
Value
Standard deviation, normalized reference lubricant tests (mpg)
0.0411
Standard deviation, normalized FEHP tests (mpg)
0.0307
F test
1.785
Fci.05
4.735
F test < F0.05? (Variances statistically equivalent?)
Yes
The F-test evaluation indicates that the variances of the normalized reference lubricant data set and the
normalized FEHP data set are similar. Therefore, the confidence interval for the standard data set is
calculated using Equation 7 in Section 1.4.3. The 95-percent confidence interval for the fuel economy
change for this data set is 0.0410 mpg.
The fuel economy change resulting from the use of the FEHP lubricant is reported as:
A = 0.169 ± 0.0410 mpg
2.4. FUEL ECONOMY CHANGE CALCULATION CROSS-CHECKS
The fuel economy change was also evaluated using two additional methods. Fuel economy change was
initially calculated using the mean of each reference lubricant data set for comparison to the FEHP
lubricant fuel economy. Analysts calculated a minimum and maximum fuel economy change based on
the test data. The statistical analyses specified in the test plan and discussed in Sections 1.4.2 and 1.4.3
were applied to both fuel economy calculations. Table 2-1 la summarizes the results.
This method shows that the fuel economy change ranges from 0.116 ±0.0479 mpg to 0.234 ±0.0488 mpg.
This is equivalent to a 0.64-percent to 1.3-percent fuel economy improvement for the test vehicle.
A second method to evaluate the fuel economy improvement is to assume that the change in reference
lubricant fuel economy from initial to final testing is simply the result of test variability. Therefore, all
reference lubricant data is pooled together regardless of the t-test evaluation of statistical significance of
the difference between the two data sets. The mean and variance are then calculated for the entire data
set. Analysts compare the pooled mean to the FEHP fuel economy and calculate the fuel economy
change. The statistical significance of the calculated fuel economy change is again evaluated using the
methods described in Section 1.4.2 and the confidence interval determined using the methods discussed in
Section 1.4.3. Table 2-1 lb presents a summary of the results of this analysis.
2-11
-------
SRI/USEPA-GHG-VR-29
August 20'3
Table 2-lla: Fuel Economy Change Cross-Check Ca
culations
Parameter
Initial Reference
Lubricant Tests vs.
FEHP
Final Reference
Lubricant Tests vs.
FEHP
Reference lubricant standard deviation (mpg)
0.0408
0.0448
FEHP lubricant standard deviation (mpg)
0.0296
0.0296
Mean reference lubricant fuel economy (mpg)
18.021
18.139
Mean FEHP lubricant fuel economy (mpg)
18.255
18.255
Change in fuel economy, delta (mpg)
0.234
0.116
Change in fuel economy, delta (%)
1.298
0.640
COV-Reference lubricant (%)
0.226
0.247
COV-FEHP lubricant (%)
0.162
0.162
Reference lubricant test count
5
6
FEHP lubricant test count
6
6
Total test count
11
12
Degrees of freedom
9
10
Squared pooled standard deviation
0.0012
0.0014
Pooled standard deviation
0.035
0.038
Critical t-distribution Value (t 0.025, df)
2.262
2.228
Calculated t-test value, ttest
11.043
5.294
Is a statistically significant change (t test >
to.025,DF)?
YES
YES
Ftest
1.904
2.297
F0.05 from tables
5.192
5.050
Pass F-test (Ftest < F0,05)?
Yes
Yes
95% confidence interval (e) (mpg)
0.0479
0.0488
95% confidence interval as percentage of mean
fuel economy change (%)
20.5
42.1
Required confidence interval for data quality
objective (DQO)
0.1404
0.12
Meets CI DQO (95% CI < 60% of delta)?
Yes
Yes
2-12
-------
SRI/USEPA-GHG-VR-29
August 20'3
Table 2-llb: Fuel Economy Change Cross-Check Calculations
Parameter
Pooled Reference Lubricant vs.
FEHP Lubricant Fuel Economy
Reference lubricant standard deviation (mpg)
0.0739
FEHP lubricant standard deviation (mpg)
0.0296
Mean reference lubricant fuel economy (mpg)
18.085
Mean FEHP lubricant fuel economy (mpg)
18.255
Change in fuel economy, delta (mpg)
0.170
Change in fuel economy, delta (%)
0.938
COV-reference lubricant (%)
0.409
COV-FEHP lubricant (%)
0.162
Reference lubricant test count
11
FEHP lubricant test count
6
Total test count
17
Degrees of freedom
15
Squared pooled standard deviation
0.0039
Pooled standard deviation
0.063
Critical t distribution value (t 0.025, df)
2.131
Calculated t-test value, ttest
5.328
Is a statistically significant change (t test > to.025,df)?
YES
Ftest
6.257
F0.05 from tables
4.735
Pass F-test (Ftest < F0.05)? (If no,use Satterthwaite)
No
Satterthwaites Approximation
Approximate ttest statistic, t'
6.692
Degrees of freedom, DF (from Satterthwaite)
14.269
Critical t-distribution value (f 0.025, df)
2.145
Is a statistically significant change (ttest > t 0.025,df)?
YES
95% confidence interval (mpg)
0.0544
95% CI as percentage of mean FE change
32.0
Required confidence interval for data quality
objective (DQO)
0.12
Meets CI DQO (95% CI < 60% of Delta)?
Yes
2-13
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
The calculated fuel economy change using this method is 0.170 ± 0.0544 mpg. This equates to an
approximate 0.94 percent improvement in fuel economy as a result of the use of FEHP lubricant in the
test vehicle.
The fuel economy changes resulting from the two alternative methods presented here concur with the
calculated fuel economy and confidence interval presented in Section 2.3. All three methods yield an
average fuel economy improvement in the range of 0.94 percent to 0.97 percent.
2.5. GREENHOUSE GAS AND OTHER POLLUTANT EMISSIONS
Greenhouse gas and other pollutant emissions from the test vehicle were measured during use of the
reference lubricant and FEHP lubricant. Table 2-12 presents a summary of the individual and mean
greenhouse gas and other pollutant emission rates observed for the FTP test cycle. Table 2-13 presents a
summary of the individual and mean greenhouse gas and other pollutant emission rates for the HFET test
cycle. Pollutant concentrations (CH4, THC) and emission rates (THC, NMHC) were measured or
calculated by the Horiba VETS 9200 system using the equipment, methods, and analyzers described in
Sections 1.4.5 and 1.4.6. Methane emission rates are calculated from the THC and NMHC gram per mile
emission rates using the following equation:
ch4 =
1
THC
{P ^
ycHA
Pthc
-NMHC
{ P ^
yen,
\P~NMHC J
(Eqn.15)
where:
CH4 =
emission rate of methane, g/mi
THC =
emission rate of total hydrocarbons, g/mi
NMHC =
emission rate of non-methane hydrocarbons, g/mi
fcH4 =
flame ionization detector (FID) analyzer methane response factor, 1.205
for this test
PCH4 =
density of methane, 18.89 g/ft3
Pthc =
density of total hydrocarbons, 16.34468 g/ft3 for this test
Pnmhc =
density of non-methane hydrocarbons, 16.3433 g/ft3 for this test
This equation is derived from the hydrocarbon, methane, and non-methane hydrocarbon emission rate
calculations specified in 40 CFR 86.144. The density of total hydrocarbons and non-methane
hydrocarbons was calculated based on fuel properties for the test fuel used in this test procedure.
Technicians determined the FID methane response factor for the analyzer for this test period as part of the
emissions test procedure.
Emissions are consistent throughout each group of test runs with coefficients of variation below 0.3. A
comparison of mean emission rates for the FEHP and reference lubricants indicates a reduction in C02
emissions during the FEHP runs when compared to the reference lubricant runs for both the FTP and
HFET cycles. This is expected as a result of the improvement in fuel economy attributed to the use of the
FEHP lubricant.
2-14
-------
SRI/USEPA-GHG-VR-29
August 20'3
Table 2-12: Greenhouse Gas and Other Pollutant Emissions - FTP
THC
CO
NOx
co2
NMHC
ch4
Test Run ID
g/mi
g/mi
g/mi
g/mi
g/mi
g/mi
Reference Lubricant-
Initial
Base-1
0.103
0.982
0.034
581.996
0.088
0.014
Base-2
0.113
0.981
0.039
585.615
0.099
0.013
Base-4
0.099
0.939
0.034
585.939
0.085
0.013
Base-6
0.108
0.939
0.034
582.750
0.093
0.014
Base-7
0.103
0.919
0.036
584.659
0.089
0.013
Mean
0.105
0.952
0.035
584.192
0.091
0.014
SD
0.005
0.028
0.002
1.746
0.005
0.001
COV
0.051
0.030
0.062
0.003
0.060
0.038
FEHP
FEHP-1
0.102
0.911
0.037
575.606
0.088
0.013
FEHP-2-R2
0.103
0.925
0.032
575.518
0.088
0.014
FEHP-3
0.102
0.884
0.031
574.004
0.088
0.013
FEHP-4
0.105
1.019
0.040
577.401
0.091
0.013
FEHP-5
0.110
0.974
0.034
576.096
0.096
0.013
FEHP-6
0.113
1.072
0.036
576.935
0.098
0.014
Mean
0.106
0.964
0.035
575.927
0.092
0.014
SD
0.005
0.071
0.003
1.199
0.004
0.000
COV
0.044
0.074
0.096
0.002
0.049
0.036
Reference Lubricant-
Final
Post Base-1R2
0.112
1.064
0.037
579.196
0.096
0.015
Post Base-2
0.111
0.893
0.033
582.510
0.097
0.013
Post Base-3
0.108
0.977
0.037
580.764
0.092
0.015
Post Base-4
0.105
0.936
0.036
579.183
0.090
0.014
Post Base-5
0.118
1.068
0.035
578.733
0.103
0.014
Post Base-6
0.109
1.001
0.035
580.048
0.094
0.014
Mean
0.111
0.990
0.036
580.072
0.095
0.015
SD
0.004
0.070
0.002
1.398
0.005
0.001
COV
0.040
0.070
0.043
0.002
0.048
0.050
2-15
-------
SRI/USEPA-GHG-VR-29
August 20'3
Table 2-13: Greenhouse Gas and Other Pollutant Emissions - HFET
THC
CO
NOx
co2
NMHC
ch4
Test Run ID
g/mi
g/mi
g/mi
g/mi
g/mi
g/mi
Reference Lubricant-
Initial
Base-1
0.022
0.107
0.007
380.053
0.015
0.007
Base-2
0.020
0.143
0.007
379.006
0.013
0.007
Base-4
0.028
0.206
0.008
379.772
0.020
0.008
Base-6
0.026
0.154
0.007
380.002
0.018
0.008
Base-7
0.017
0.117
0.007
382.838
0.011
0.006
Mean
0.023
0.145
0.007
380.334
0.015
0.007
SD
0.004
0.039
0.000
1.461
0.004
0.001
COV
0.197
0.267
0.062
0.004
0.237
0.116
FEHP
FEHP-1
0.021
0.143
0.008
375.861
0.014
0.007
FEHP-2-R2
0.028
0.184
0.008
375.870
0.020
0.008
FEHP-3
0.028
0.171
0.007
377.078
0.019
0.009
FEHP-4
0.023
0.145
0.007
375.724
0.015
0.008
FEHP-5
0.022
0.157
0.007
375.634
0.015
0.007
FEHP-6
0.027
0.181
0.010
377.755
0.018
0.009
Mean
0.025
0.164
0.008
376.320
0.017
0.008
SD
0.003
0.018
0.001
0.880
0.002
0.001
COV
0.128
0.109
0.149
0.002
0.148
0.112
Reference Lubricant-
Final
Post Base-1R2
0.024
0.167
0.006
374.912
0.016
0.008
Post Base-2
0.025
0.169
0.007
376.973
0.017
0.008
Post Base-3
0.021
0.129
0.007
377.152
0.013
0.008
Post Base-4
0.026
0.157
0.008
377.529
0.018
0.008
Post Base-5
0.030
0.220
0.010
380.536
0.020
0.010
Post Base-6
0.025
0.164
0.009
381.549
0.016
0.009
Mean
0.025
0.168
0.008
378.109
0.017
0.008
SD
0.003
0.030
0.001
2.469
0.002
0.001
COV
0.116
0.176
0.188
0.007
0.140
0.098
2-16
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
3.0 DATA QUALITY ASSESSMENT
3.1. DATA QUALITY OBJECTIVES
The GHG Center selects methods and instruments for all verifications to ensure a stated level of data
quality in the final results. The GHG Center specifies data quality objectives (DQOs) for each
verification parameter before testing commences. Each test measurement that contributes to the
determination of a verification parameter has stated data quality indicators (DQIs) which, if met, ensure
achievement of that verification parameter's DQO.
The establishment of DQOs begins with the determination of the desired level of confidence in the
verification parameters. The next step is to identify all measured values which affect the verification
parameter and determine the levels of error which can be tolerated. The DQIs, most often stated in terms
of measurement accuracy, precision, and completeness, are used to determine if the stated DQOs are
satisfied. This verification's DQO is the fuel economy change's desired confidence level. The DQO
statement for this verification is:
The 95-percent confidence interval of the fuel economy change (A) will be less than 60 percent of
the mean A for A values as low as 0.2 mpg. For mean values of A less than 0.2 mpg, the
confidence interval will be less than or equal to ±0.12 mpg.
Table 3-1 summarizes the data quality objective, DQO goal, and results achieved for the valid test runs.
Table 3-1: Fuel Economy Change And Data Quality Objective
Mean Fuel Economy Change
0.169 mpg
95% Confidence Interval
0.0410 mpg
DQO Confidence Interval
0.12 mpg
Meets DQO Goal?
yes
Each testing, sampling, and analytical method produces results that contribute to the overall fuel economy
change determination. The GHG Center concludes that the data and the resulting confidence interval
calculation are valid if each contributing measurement conforms to the applicable method specifications.
These quantitative or qualitative protocols constitute this verification's DQI goals. The DQIs, goals, and
achieved results are summarized and discussed in the following sections. Achievement of the DQI goals
implies that the contributing measurement conforms to the applicable method specifications and its use in
calculating the achieved DQO is valid. The field team leader also used several QA/QC checks to verify
that test conditions were appropriate prior to and after testing, minimizing the number of potential invalid
test runs. SwRI made some adjustments to test equipment to ensure data quality because of the QA/QC
checks.
The following DQIs contain accuracy, precision, and completeness levels that must be achieved to ensure
that DQOs can be met. Reconciliation of DQIs is conducted by: (1) performing independent performance
checks in the field with certified reference materials; (2) following approved reference methods; (3)
performing factory calibration of instruments prior to use; and (4) conducting QA/QC procedures in the
field to ensure that instrument installation and operation are verified. The Test Plan stated that in some
instances, reconciliation of DQIs was performed by completion of specific QA/QC checks that infer
3-1
-------
SRI/USEPA-GHG-VR-29
August 20'3
proper functioning of equipment. Achievement of such QA/QC checks for these DQIs, in conjunction
with original equipment installation calibrations or certifications, indicated that DQI goals were met at the
time of testing.
The field team leader voided test runs that did not meet the QA/QC checks or DQIs specified in the test
plan. He also addressed DQI failure by requiring test runs to be repeated the next available test day.
Some data quality indicator goals were not met during the test period. However, failure to meet these
data quality goals did not adversely impact the achievement of the specified data quality objective
because supporting data QA/QC goals were met.
3.2. RECONCILIATION OF DQOs AND DQIs
The following sections discuss and summarize the range of measurements observed in the field and the
DQI and completeness goals. The GHG Center completed the majority of tasks specified in the test plan
for DQI measurements and determinations for all valid test runs used in data analysis. Therefore, the
completeness for the majority of these DQI and QA/QC checks is 100 percent. SwRI did not complete
the specified tasks for two DQI checks according to the schedule specified in the Test Plan. This is
discussed in the following sections. The following sections also include accuracy goals for measurement
instruments. Actual measurement accuracy achieved is reported for each item based on instrument
calibrations conducted by manufacturers, field calibrations, reasonableness checks, and/or independent
performance checks with a second instrument. The QA/QC procedures conducted for key measurements
and the procedures used to establish DQIs are also included. The accuracy results for each measurement
and their effects on the DQIs are discussed. Accuracy goals were met for the majority of QA/QC checks.
Some accuracy checks were not met in certain instances.
3.2.1. Dynamometer Specifications, Calibrations, and QA/QC Checks
Table 3-2 summarizes the dynamometer specifications and the associated data quality indicator goals.
The field team leader verified all DQIs during the test period and completed daily QA/QC checks that
were used to assure that DQI goals were met for the dynamometer. The field team leader also used the
QA/QC checks to ensure that test conditions were within the specified parameters to minimize the
number of invalid test runs.
The field team leader verified the dynamometer specifications and DQI goals by reviewing: (1) the
original installation calibration data, (2) updated load simulation calibration data, and (3) additional daily
QA/QC checks. SwRI indicated that the dynamometer was capable of 0-120 mph speed measurements
and 0-1750 lbf load measurements with accuracy of +0.02 percent full-scale for speed and ±0.1 percent
full-scale for load. Review of the original installation data for the dynamometer enabled the field team
leader to determine that the accuracy of the system as installed was within the allowable error.
3-2
-------
SRI/USEPA-GHG-VR-29
August 20'3
Table 3-2. Chassis Dynamometer Specifications and DQI Goals
Data Quality Indicator Goals
Measurement
Variable
Date
Completed
Instrument Type
/ Manufacturer
Instrument
Range
Measurement
Frequency
Accuracy
Goal Actual
How Verified /
Determined
Complete-
ness
Speed
12-15-95
Horiba LDV-48-
86-125HP-AC
Light-Duty
Chassis
Dynamometer
0 to 120 mph
10 Hz
(10/sec)
+
0.02%
FS
Max. Error =
-0.0016% FS
Sensors
calibrated and
verified during
original
installation.
Calibration
records
reviewed.
100%
Load
3-19-03
Oto 1,750 lbf
+ 0.1%
FS
Max. Error =
-0.1%
100%
The results of the data quality QA/QC checks for the valid test runs are presented in Table 3-3. The field
team leader recorded results of all QA/QC checks. Documentation is on file at the GHG Center.
Table 3-3. Chassis Dynamometer QA/QC Checks
QA/QC Check
How Verified /
Determined
Date
Completed
Goal
Actual Results
Source
Complete-
ness
Road load horsepower
calibration
Triplicate coastdown
checks completed
during testing for each
lubricant
4/1/03;
4/24/03;
5/20/03
+ 1.5 lbf of target
+0.73 lbf (Ref.
Lube)
+0.24 lbf
(FEHP)
Coastdown
Run Data
Sheets
100%
Dyno calibration
certificate inspection
Once during the test
campaign
4/2/03
Sensor accuracies
conform to
specifications
See Table 3-2
Dyno
Calibration
Records
100%
Parasitic friction
verification
Daily, prior to testing
Each test
+2 lbf from
existing settings
Maximum
change +2 lbf*
Daily
Parasitic
Loss Check
Records
100%
Dyno warmup
verification
Before each test run
Each test
>15 minutes of
operation; at least
50 mph within 2
hours of the start
of testing
Daily dyno
warmup time
minimum 15:29
at 50 mph
Dyno
Warmup
Records
100%
Roadload and inertia
simulation check
End of each test run
Each test
+ 0.3%) average
over the entire
driving sequence
Max. simulation
error 0.17%)
Dyno Test
Run Record
100%
Valid driver's trace
End of each test run
Each test
No deviation from
tolerances given in
40CFR 86.115
None**
Dyno Test
Run Record
100%
*A new parasitic loss curve was accepted before run Base-4 due to verifications exceeding + 2 lbf.
** Test run Base-5 was voided due to an invalid driver's trace.
3-3
-------
SRI/USEPA-GHG-VR-29
August 20'3
There were two instances when the dynamometer or test simulation did not meet the QA/QC checks. The
first occurred prior to test run Base-4 when SwRI checked the parasitic friction losses of the
dynamometer. Friction loss results were out of range of the specifications. Technicians accepted a new
parasitic friction loss curve to meet the specifications. All tests completed prior to and after the
acceptance of the new parasitic loss curve met the data QA/QC requirements. The second occurred
during test run Base-5 when the test driver deviated from the required driving cycle for 2.0 seconds,
exceeding the tolerances specified in the test plan. Therefore, the field team leader voided test run Base-
5. Test run Base-5R2 was completed in its place the next testing day.
The remaining valid test runs met all required dynamometer QA/QC checks, indicating that DQI goals for
the chassis dynamometer were achieved and completeness for valid test runs was 100 percent.
3.2.2. CVS Sampling System Specifications, Calibrations, and QA/QC Checks
Table 3-4 summarizes the Horiba CVS system specifications and DQI goals. The field team leader
reviewed the calibration records for the initial installation of the CVS system to verify that the CVS
system met the specified DQI goals. This included pressure, temperature, and flow measurement devices.
The actual accuracy of the equipment is summarized in Table 3-4. All CVS instrumentation met DQI
goals based on a review of available calibration data.
Table 3-4. CVS Specifications and DQI Goals
Data Quality Indicator Goals
Measurement
Variable
Date DQI Check
Completed
Instrument
Description
Range
Measurement
Frequency
Accuracy
Goal Actual
Complete
-ness
How Verified /
Determined
Pressure
11/94
Horiba Variable-
Flow Constant
Volume Sampler
0 to 1500 millibar
10/sec
±2%
reading
Max. Error =
2% of reading
100%
Sensors
calibrated and
verified during
installation.
Temperature
8/94
Oto 100°C
±2%
reading
Max Error =
0.1% of reading
Volumetric
Flow Rate
1/17/95; 11/7/94
0 to 700 ft3/min
±0.5%
reading
±0.5% reading
Daily QA/QC checks were required to ensure continued proper CVS function although initial installation
records indicate acceptable functioning of the pressure, temperature, and flow measurement equipment.
The CVS measurement variables achieved the specified DQI accuracy during the test period if the daily
QA/QC checks conformed to specifications. The field team leader and SwRI technicians performed the
daily QA/QC checks of the CVS system as specified. Results of the checks are summarized in Table 3-5.
The CVS system and associated equipment met all daily QA/QC check requirements during the entire test
period. The sample bag leak check is an automated procedure that was completed by the Horiba VETS
9200 system prior to each test. The specifications of the automated leak check met the QA/QC check
requirements. It was not necessary to change the propane cylinder during testing. Therefore, technicians
did not need to complete the new propane tank verification check. The field team leader completed a log
form for all CVS system QA/QC checks. Copies of the QA/QC log are on file with the GHG Center.
3-4
-------
SRI/USEPA-GHG-VR-29
August 20'3
Table 3-5. CVS System QA/QC Checks
QA/QC Check
Date(s) Performed
(Required Frequency)
Expected or Allowable
Result
Actual
Result
How Verified
New propane tank composition verification
NA (prior to placing new
propane tank in service)
<0.35% difference in reading
from previously verified tank
NA - not
changed
during
testing
NA - not required
CVS and propane critical flow orifice
calibration certificate inspection
4/2/03 (once during the
test campaign)
Sensor accuracies conform to
specifications
see Table
3-4
Review of initial
calibration certificates
Propane injection check
3/28/03; 4/3/03; 4/10/03;
4/14/03; 4/23/03; 4/30/03;
5/14/03; 5/20/03; 5/28/03
(weekly)
Difference between injected
and recovered propane
< + 2.0%.
max error
= -1.96%
Review of weekly
Propane Injection Check
records
Flow rate verification
Each test (before each test
run)
+ 5 cfm of appropriate
nominal set point
all test runs
pass
Review of daily test log
forms and intermittent
visual verification
Sample bag leak check
Each test (before each test
run)
Maintain 10" Hg vacuum for
10 seconds
all test runs
pass
Review of daily test
output and log forms
3.2.3. Emission Analyzer Specifications, Calibrations, and QA/QC Checks
Table 3-6 lists the emission analyzers used during the test campaign, the expected values, and associated
DQI goals. SwRI technicians calibrate the emission analyzers monthly using calibrated gas dividers and
calibration gases verified against NIST-traceable reference gases. Technicians calibrate analyzers at 11
calibration points throughout the range of the analyzer in accordance with SwRI standard operating
procedures. Technicians accept a revised calibration curve for the emission analyzers if calibration
checks indicate error exceeding the accuracy goals. The field team leader reviewed all analyzer
calibrations for the expected range of operation of the analyzers prior to commencement of testing to
ensure analyzers would meet the specifications during the test period. SwRI also completed several
analyzer calibrations during the test period. The CO analyzer with the 0-3000 ppm range originally
identified in Table 3-5 of the Test Plan was not used during the test period because the higher range limit
was not needed. Therefore, calibration data was not obtained for this CO analyzer range. The field team
leader reviewed the calibration results during testing to ensure QA/QC goals would be met. The
maximum actual calibration error for each analyzer is summarized in Table 3-6.
Table 3-7 summarizes the applicable QA/QC checks for the emission analyzers, calibration gases, and
associated equipment. Since calibrating analyzers prior to each test would be cumbersome and time-
consuming, the test plan specified that if all calibration gases and QA/QC checks met their specifications,
SwRI and the GHG Center would infer that the emission analyzers met Table 3-7 accuracy specifications.
The LLCO (200 ppm) and C02 (4 percent) accuracies observed during the analyzer calibration checks
exceeded the accuracy specification of one percent of full scale for two data points each. The C02 (4
percent) analyzer also exceeded the one percent of reading for one point and two percent of full scale for
two points for the May 2003 calibrations. SWRI performs an 11-point calibration on the analyzers,
although 40 CFR 86.122 and .124 only require an 8-point calibration. SwRI sometimes observes one or
two points out of acceptable range on a specific analyzer. Technicians then use their judgment to evaluate
whether or not the analyzer should be recalibrated so the analyzer will still meet the 8-point calibration
requirement. SwRI did not recalibrate the analyzers in these two cases because the remainder of
calibration data points met the allowable error requirements. The C02 analyzer was also not recalibrated
to allow for the direct comparison between the calibration curves for the incorrectly input calibration gas
value and the replacement calibration gas. This allowed analysts to develop a correction factor as
3-5
-------
SRI/USEPA-GHG-VR-29
August 20'3
discussed below. Completeness for these DQI goals is calculated as the percentage of valid data points
for a given set of calibration data for each analyzer.
Table 3-6. Emission Analyzer Specifications and DQI Goals
Data Quality Indicator Goals
Measure-
ment
Variable
Operating
Range
During
Tests
Instrument Mfg.,
Model / Type
Date
Completed
Measurement
Frequency
Accuracy3
How Verified /
Determined
Complete-
ness
Goal
Maximum
Actual
Low CO
0-200
ppm
Horiba AIA-210 /
NDIR
3/6/03;
4/8/03;
5/5/03
1 analysis per bag,
8 bags (4 dilute
exhaust, 4 ambient
air) per run. 45-
second purge
period, then 10-
second analysis
period per bag.
Analyzer output to
VETS @ 10/sec
+ 1.0% FS or
+ 2.0% of the
calibration
point
1.81%
reading;
1.07% FS
Gas divider
with protocol
calibration
gases at 11
points evenly
spaced
throughout span
(including zero)
95.5%
co2
0-4.0%
(vol)
Horiba AIA-220 /
NDIR
3/6/03;
4/8/03;
5/6/03
2.10%
reading;
2.10% FS
92.4%
NOx
0-10 ppm
Horiba CLA-220 /
Chemiluminescence
3/6/03;
4/8/03;
5/5/03
1.85%
reading;
0.32% FS
100%
THC
0-30 ppm
(carbon)
Horiba FIA-220 /
HFID
3/5/03;
4/8/03;
5/5/03
1.23%
reading;
0.74% FS
100%
aThe most stringent accuracy specification applies for each calibration point.
SwRI verifies all new Standard Reference Material (SRM) or other calibration and reference gas
concentrations with an emissions analyzer that has been calibrated within the last 30 days. The operator
zeroes the analyzer with a certified zero-grade gas and then spans it with a NIST SRM (or equivalent)
three times to ensure stability and minimal analyzer drift. The operator then introduces the new reference
gas into the analyzer and records the concentration, followed by reintroduction of the NIST SRM to
ensure that the analyzer span point does not drift more than ±0.1 meter divisions. The operator repeats
these last two steps until three consistent values are obtained for the NIST SRM and the new candidate
reference gas. The mean of the three NIST SRM concentrations must be within one percent of the
certified NIST SRM concentration. SwRI then considers the reference gas suitable for emissions analyzer
calibrations. SwRI refers to this process as calibration gas "naming." The field team leader reviewed all
calibration gas naming records to ensure compliance with the specified protocol and quality standards.
All calibration gases met the specified QA/QC standards.
3-6
-------
SRI/USEPA-GHG-VR-29
August 20'3
Table 3-7. Emission Analyzer QA/QC Checks
QA/QC Check
Date(s) Performed
(Frequency)
Expected or Allowable
Result
Actual Result
NIST-traceable calibration gas verifications
C02 (4%): 1/15/02;
4/10/03
THC (30 ppm): 2/21/01
CO (25 ppm): 6/25/02
NOx (10 ppm): 6/25/02
CO (200 ppm): 4/10/03
(prior to being put into
service)
Average of three readings
must be within +1% of
verified NIST SRM
concentration
Maximum Error Observed
=0.042%) (see discussion)
Zero-gas verification
3/25/03; 4/7/2003;
4/25/2003
(prior to being put into
service)
HC < 1 ppmC
CO < 1 ppm
C02 < 400 ppm
NOx <0.1 ppm
02 between 18 and 21 %
HC = 0 ppmC
CO = 0 ppm
C02 = 0 ppm
NOx = 0 ppm
02=21.0-21.64%
Gas divider linearity verification
3/13/03;
5/2/03
(monthly)
All points within + 2% of
linear fit
FS within + 0.5% of known
value
+ 0.5% of point
+ 0.1% FS
Analyzer calibrations
See Table 3-6
(monthly)
All values within + 2% of
point or + 1% of FS;
Zero point within + 0.2% of
FS
See Table 3-6
Wet C02 interference check
3/7/03;
4/23/03;
5/6/03
(monthly)
CO-O to 300 ppm,
interference < 3 ppm
CO > 300 ppm, interference <
1% FS
Interference <0.1 ppm
Interference < 0.00373%
NOx analyzer interference check
4/30/03
(monthly)
C02 interference < 3%
Interference < 0.32%o
NOx analyzer converter efficiency check
3/7/03;
4/23/03;
5/6/03
(monthly)
NOx converter efficiency >
95%
Efficiency > 98.95%o
CO and C02 PEAs
C02: 4/24/03;
CO: 5/2/03
(once during testing)
+ 2% of analyzer span
+ 0.42% of span (C02)
+ 0.25% of span (CO)
Calibration gas certificate inspection
4/3/03
(once during testing)
Certificates must be current;
concentrations consistent with
cylinder tags
Concentrations match tags
& naming sheets. All
current.
Bag cart operation
Each Test
(prior to analyzing each
bag)
Post-test zero or span drift
shall not exceed +2% full-
scale
All pass.
The test plan specifies verification of the new calibration gas concentration to within one percent of the
NIST-traceable reference gas. However, the standard method SwRI used for gas "naming" requires that
repeated analyses of a NIST SRM on the analyzers have an error of less than one percent when compared
to the certified NIST concentration. The NIST gas is used as a reference gas to ensure the analyzer
accuracy, but is not meant as a direct comparison to the calibration gas. Error reported in Table 3-7 is the
maximum error for the NIST SRM gas readings.
SwRI also verifies each new working zero air (or N2) cylinder's impurities to ensure that it is suitable for
emission analyzer zero checks. Comparisons between a certified Vehicle Emission Zero (VEZ) gas (or
3-7
-------
SRI/USEPA-GHG-VR-29
August 20'3
equivalent) and the candidate zero gas serve this purpose. SwRI employs an emissions cart (or suite of
instruments) that has been calibrated within the last 30 days for this procedure. The operator zeroes the
analyzers with certified VEZ gas and spans them with NIST-traceable reference gases to ensure stability
and minimal analyzer drift. The operator then introduces the candidate cylinder's zero gas to the sample
train and records the THC, CO, C02, and NOx values. The field team leader reviewed zero gas
verification records for all zero gases used during the test period. This review showed that the oxygen
content of one of the zero gases exceeded the specified criteria. This zero gas cylinder was installed
without proper verification so testing could begin on schedule. It was verified during the test period. The
pollutant concentrations met the specified criteria and the gases were deemed suitable for instrumental
analyzer calibrations although the oxygen concentration exceeded the specified limit.
SwRI uses a gas divider to obtain a range of concentrations from a single reference gas to complete
analyzer calibrations across the appropriate instrument range. SwRI verified the calibration gas divider
linearity using an HC analyzer known to have a linear response and an HC span gas. The field team
leader reviewed the gas divider linearity verification records for the divider used during analyzer
calibrations during the test period. The maximum error observed for the gas divider was 0.5 percent of
reading or 0.1 percent of full scale, which is within the specifications of the Test Plan.
The NIST-traceable calibration gases, in conjunction with the verified gas divider and zero gas, were used
to create individual gas concentrations to verify the calibration of each instrumental analyzer. Eleven gas
concentrations were generated in ten-percent increments from 0 to 100 percent of each analyzer's span for
calibration verification. The Horiba VETS 9200 records analyzer response at each point and determines
associated error. The field team leader reviewed the calibration verification records for each analyzer
range used during the test period. The LLCO analyzer calibration completed on April 1 indicated errors
in excess of one percent full scale. Therefore, a new calibration curve was accepted for the LLCO 0-200
ppm range on April 1, 2003. The C02 analyzer calibration completed on March 6, 2003, indicated an
error in excess of one-percent of full scale for the four percent C02 range. A calibration was run again
after the four percent C02 calibration gas was replaced on April 8. The new calibration curve was
accepted based on the correct calibration gas concentration. The field team leader did not identify any
other calibration errors that were outside of the specified allowable error.
The four-percent C02 range calibration gas ran low during the Base 5-R2 test period. Technicians
changed the tank immediately after the test, with a new verified calibration gas taking its place. However,
SwRI identified a shift in C02 analyzer response after replacing the four-percent C02 gas. SwRI
determined after further review that the concentration of the previously used four-percent C02 reference
gas was input incorrectly into the Horiba VETS 9200 during analyzer calibrations. Therefore, the Horiba
analyzer incorrectly determined the C02 concentration calibration curve used for runs Base-1 through
Base-5. Previous calibration data for three months prior to testing showed consistent analyzer response
for the C02 calibration verification. The maximum observed error among three of the calibration curves
was 0.885 percent. Technicians performed repeated analyzer calibration verifications after replacement
of the four-percent C02 gas. The maximum observed error among resultant calibration curves was 0.651
percent.
Comparison of the two sets of calibration curves for the old (incorrect concentration) and new gas bottles
allowed analysts to determine that a shift in C02 concentration of 2.37 to 2.40 percent occurred as a result
of the incorrect calibration gas concentration input for the range of exhaust C02 concentrations
encountered during testing (0.9070 to 1.1452 percent C02). A shift in C02 concentration of 2.55 percent
was also observed for C02 in the range of ambient C02 concentrations encountered during testing (0.0394
percent to 0.0449 percent C02). SwRI applied a correction factor to the test data equivalent to the
average percent offset observed between the correct calibration gas analyzer response and the incorrect
3-8
-------
SRI/USEPA-GHG-VR-29
August 20'3
analyzer response to account for the incorrect calibration curve. A Corrective Action Report (CAR #3)
for this is on file at the GHG Center.
A sensitivity analysis was completed to demonstrate that the error incurred by using an average value for
the offset, as opposed to a specific offset value based on the calibration curves, was minimal. The C02
concentrations for one test run (Base-1) were corrected using the maximum, minimum, and average offset
correction factors. The resulting differences in fuel economy were calculated. The differences in fuel
economy calculated using the average C02 correction factor and the maximum and minimum correction
factor were 0.017 percent and 0.011 percent. The correction method, supporting documentation,
sensitivity analysis, and sample calculations were documented and are on file with SwRI and the GHG
center.
The CO analyzer wet C02 interference check was completed in conjunction with the monthly
calibrations. This procedure determines the CO analyzer's response to water vapor and C02. The field
team leader reviewed documentation of the C02 interference checks completed for the test period.
Analyzer response to the interference gas was <0.1 ppm for spans below 300 ppm and < 0.00373 percent
of span for higher ranges. This is well within the allowable error specified in the Test Plan.
The NOx analyzer C02 interference check was not completed monthly as scheduled in the Test Plan. The
NOx analyzer C02 interference (quench) check is normally completed once every six months in
accordance with SwRI's SOP and was completed in conjunction with one of the monthly calibrations.
This does not meet the schedule specified by the Test Plan but it does meet the schedule specified in
SwRI's SOP. This check is not required by EPA regulations nor specified in the CFR for light-duty
vehicle fuel economy testing. A verified gas divider was used to dilute NIST-traceable C02 by 50 percent
with NIST-traceable NO. The operator then calculated the expected dilute NO concentration and
recorded the analyzer's actual response to this challenge. The difference between the calculated NO and
measured NO concentrations was < 0.32 percent.
The field team leader reviewed documentation of NOx analyzer converter efficiency checks for the test
period. The check procedure uses a NOx generator that dilutes NIST-traceable NO with air. An ozone
generator then converts a quantitative portion of the air's oxygen to 03 that converts the same proportion
of NO to N02. This creates a NOx blend (NO plus N02) of known concentration. The difference
between the analyzer's NO response and NOx response will be the measure of the NOx to NO converter
efficiency. SwRI determined the NOx converter efficiency to be > 98.95 percent for the test period. The
allowable minimum NOx converter efficiency is 95 percent.
The field team leader introduced NIST-traceable CO and C02 audit gases to the analyzer at the analyzer's
external ports as an independent performance evaluation audit (PEA). The audit gas concentrations used
were within the analyzer ranges used during testing. The C02 audit gas was C02 in N2 gas with a
certified concentration of 1.003 percent C02 and an accuracy of +2 percent. The CO audit gas was a 49.9
ppm CO in air mixture with an accuracy of +5 percent. Analyzer audits yielded analyzer accuracies of
0.42 percent and 0.25 percent, considered acceptable according to the test plan specification of ±2 percent
of span. The CO audit gas used for the PEA did not meet the audit gas accuracy specification of +2
percent indicated in the Test Plan. A CO audit gas with a +2 percent accuracy specification was not
available from gas suppliers in a reasonable time frame or at a reasonable cost necessary to ensure use
during the test period. Therefore, the CO audit gas with an accuracy specification of +5 percent was used.
The field team leader also reviewed certificates for all calibration and zero gases used during the test
campaign. All certificates were current and the cylinder tag concentrations matched those on the
applicable certificate and the calibration gas naming records. Records of analyzer calibrations and
QA/QC checks, as compiled by the field team leader, are on file with the GHG Center.
3-9
-------
SRI/USEPA-GHG-VR-29
August 20'3
3.2.4. Ambient Instrument Specifications, Calibrations, and QA/QC Checks
Meteorological parameters collected or calculated during the test period include ambient air temperature,
relative humidity, and barometric pressure. These values are used in a variety of calculations. SwRI
acquired these data prior to each test with the instruments listed in Table 3-8. The DQI goals for the
ambient monitoring instruments are also specified in Table 3-8.
Table 3-8. Ambient Instrument Specifications and DQI Goals
Data Quality Indicator Goals
Measurement
Variable
Expected
Operating
Range
Manufacturer
Instrument
Range
Measurement
Frequency
Acc
Goal
uracy
Actual
Complete
-ness
How Verified /
Determined
Wet- and Dry-
Bulb Temperature
68 to 86 °F
Psychro-Dyne
10 to 110 °F
Prior to each
test
±1.0 °F
± 0.4 °F
100%
Regular verification
checks with NIST-
traceable standards
Barometric
Pressure
28 to 31"
Hg
Heise 901A
pressure
transducer
20 to 35" Hg
±0.1"
Hg
±0.004"
Hg
The barometric pressure transducer measures test site pressure directly. Wet-bulb and dry-bulb
temperatures are used to estimate relative and absolute humidity. Relative humidity and temperature are
also recorded continuously to verify test site conditions. SwRI verified meteorological instrument
performance with the QA/QC checks outlined in Table 3-9. The field team leader reviewed records of
these QA/QC checks for the testing period.
Table 3-9. Ambient Instrument QA/QC Checks
QA/QC Check
When Performed
(Required Frequency)
Expected or Allowable
Result
Actual Result
Test site barometer calibration
verification
Weekly (Prior to each
set of lubricant tests)
±0.1" Hg of NIST-
traceable standard
±0.004" Hg of NIST-traceable standard
Wet-bulb and dry-bulb
temperature calibration
verification
Monthly - 3/31/03;
4/24/03; 5/21/03 (Prior
to each set of lubricant
tests)
±1.0 °F of NIST-traceable
standard
±0.4 °F of NIST-traceable standard
Test site dry-bulb temperature
verification
Prior to each test run
68 to 86 °F
71-74 °F
SwRI maintains separate NIST-traceable primary standard and secondary standard barometers. Operators
compare the primary and secondary standards with each other to ensure the primary standard's accuracy.
SwRI requires the primary standard to be within ±0.05" Hg of the secondary standard. SwRI also
requires the test site barometer readout to be within ±0.1" Hg of the primary standard. The Test Plan
incorrectly specified that the test site barometer should be within ±0.01" Hg of the primary standard. A
Corrective Action Report (CAR #5) was issued to revise the Test Plan requirement to meet the SwRI
SOP. Although not a requirement of SwRI's standard operating procedures, barometer verification data
demonstrates that both the existing and revised data quality checks are satisfied.
Verification of the wet-bulb and dry-bulb thermometers occurs monthly. Comparisons with NIST-
traceable thermometers (accurate to ±0.1 °F) ensure that test site temperature measurements are within
±1.0 °F. The field team leader reviewed thermometer calibrations for the test period and verified that the
psychrometer meets the specified accuracy requirement of 1.0 °F, achieving an accuracy of ±0.4 °F. 40
CFR 86.130 specifies that test site temperatures must be between 68 and 86 °F during vehicle testing.
Operators monitored temperatures prior to the start of every test run to ensure that this specification is
3-10
-------
SRI/USEPA-GHG-VR-29
August 20'3
met. A review of the temperature records indicates that all test runs were completed under appropriate
temperature conditions as specified in the test plan.
The field team leader monitored and documented SwRI's QA/QC check performance for ambient
monitoring equipment. Documentation of QA/QC checks is on file with the GHG Center.
3.2.5. Test Fuel Specifications
SwRI received certification-grade test fuel in 55-gallon drums to use throughout testing. All drums were
obtained from a single batch of gasoline maintained by the supplier. The Test Plan specifies analysis of
duplicate samples from each drum of test fuel, but the GHG center felt it unnecessary to analyze each
drum because all test fuel was from the same fuel batch. Therefore, SwRI's analytical laboratory
completed analysis of two samples from the test fuel batch received by SwRI. Samples were analyzed in
accordance with the methods listed in Table 3-10. Table 3-10 also presents the DQI goals, allowable test
fuel specifications, analytical results, and accuracy of analyses. The field team leader also summarized
test fuel analyses on the required log forms and obtained copies of analytical reports from SwRI. All
documentation is on file with the GHG Center. All sample results meet the DQI goals for accuracy and
are within the required specifications for the test fuel.
Table 3-10. Test Fuel ASTM Measurement Methods and DQI Goals
Parameter
ASTM Test
Method
Sample Results
Required
Spec
Method
Accuracy
Accuracy Goal (2x
Method Accuracy)
Measured
Accuracy
Sample ID
ETV1
ETV2
Octane - Research
D2699
96.7
96.6
>93
+0.32
+0.64
+0.1
Octane - Motor
D2700
88.9
88.8
NA
NA
NA
NA
Sensitivity (Octane)
D2699, D2700
7.8
7.8
>7.5
NA
NA
NA
Lead (g/gal)
D3237
<0.001
<0.001
<0.05
+0.0004
+0.0008
NA
Distillation Range
D86
Initial
93.3 °F
94.9 °F
75-95 °F
+2.54 °F
+5.08 °F
+1.6
10%
128.5 °F
129.6 °F
120-135 °F
+2.36 °F
+4.72 °F
+1.1
50%
219.5 °F
220.6 °F
200-230 °F
+1.96 °F
+3.92 °F
+1.1
90%
319.6 °F
321.4 °F
300-325 °F
+1.57 °F
+3.14 °F
+1.8
End
406.8 °F
404.9 °F
415 °F
+5.11 °F
+10.22 °F
+1.9
Sulfur (wt%)
D1266
0.0033
0.0032
<0.1
+0.00042
+0.00084
+0.0001
Phosphorus (g/gal)
D3231
0.0001
<0.0001
<0.005
+0.0007
+0.0014
NA
Reid Vapor Pressure
(psia)
D5191
9.15
9.11
8.0-9.2
+0.07
+0.14
+0.04
Hydrocarbons( wt%)
D1319
Olefins
0.8
0.9
<10
+0.64
+1.28
+0.1
Aromatics
30.4
30
<35
+0.54
+1.08
+0.4
Saturates
68.8
69.1
-
+0.59
+1.18
+0.3
Hydrocarbons
Duplicate
D1319
Olefins
1
0.8
<10
+0.64
+1.28
+0.2
Aromatics
30.9
31.2
<35
+0.54
+1.08
+0.3
Saturates
68.1
68
-
+0.59
+1.18
+0.1
Specific Gravity
D1298
0.7423
0.7424
-
+0.5
+1
+1E-04
3-11
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
3.2.6. Fuel Economy Volumetric and Gravimetric Cross-Checks
SwRI and the GHG Center performed cross-checks of the carbon balance method fuel economy results
with separate volumetric and gravimetric fuel economy determinations. An external fuel cart provided
fuel for the vehicle from a five-gallon fuel container during each test run. The fuel cart used a flow meter
to monitor volumetric fuel flow to the vehicle and return fuel. The fuel container was placed on a
Fairbanks scale to monitor weight of fuel consumed during each test phase. Fuel economy was then
calculated from both the volumetric and gravimetric fuel consumption data and vehicle miles traveled
during the test phase for comparison to the carbon balance fuel economy result. SwRI technicians
calibrated the fuel cart prior to beginning the test period to ensure proper fuel cart function. The error in
volumetric fuel cart readings during calibration averaged 0.27 percent, with a maximum of 0.36 percent.
The acceptable error specified by SwRI's SOP is +2%.
The Test Plan for this verification specified a ±0.3 difference in COVs for each fuel economy calculation
method as an indicator of potential data bias for the carbon balance calculation method. Observed fuel
economy differences of greater than 0.2 mpg were also to be investigated for evidence of bias.
Volumetric and gravimetric cross-checks were completed for each test run instead of the maximum of 10
test runs specified in the Test Plan. The field team leader and SwRI project manager reviewed the
volumetric and gravimetric cross-check data after each test run. Tables 3-11, 3-12, and 3-13 summarize
the volumetric and gravimetric cross-check results for all test runs.
Table 3-11: Volumetric and Gravimetric Cross-Checks - FEHP Lubricant Test Runs
Run ID
Date
Carbon Balance
mpg
Gravimetric
mpg
cov
Difference
Mean mpg
Difference
FEHP 1
4/23/03
18.272
18.77
FEHP 2R2
4/25/03
18.272
18.72
FEHP 3
4/29/03
18.284
18.74
0.096
0.467
FEHP 4
4/30/03
18.233
18.69
0.058
0.465
FEHP 5
5/1/03
18.263
18.6
0.242
0.449
FEHP 6
5/2/03
Date
18.206
18.69
0.151
0.447
Run
ID
Carbon
Balance mpg
Volumetric
mpg
COV
Difference
Mean mpg
Difference
FEHP 1
4/23/03
18.272
18.46
FEHP 2R2
4/25/03
18.272
18.42
FEHP 3
4/29/03
18.284
18.50
0.179
0.184
FEHP 4
4/30/03
18.233
18.40
0.119
0.180
FEHP 5
5/1/03
18.263
18.32
0.206
0.165
FEHP 6
5/2/03
18.206
18.39
0.336
0.160
3-12
-------
SRI/USEPA-GHG-VR-29
August 20'3
Table 3-12: Volumetric and Gravimetric Cross-Checks - Initial Reference
Lubricant Test Runs
Run
ID
Date
Carbon
Balance mpg
Gravimetric
mpg
COV
Difference
Mean mpg
Difference
Base 1
4/2/2003
18.070
18.624
Base 2
4/3/2003
18.013
18.487
0474
Base 4
4/8/2003
17.994
18.389
0.419
Base 6
4/11/2003
18.055
18.478
0.328
0.462
Base 7
4/12/2003
17.973
18.403
0.279
0.455
Run ID
Date
Carbon Balance
mpg
Volumetric mpg
COV Difference
Mean mpg
Difference
Base 1
4/2/2003
18.070
18.280
Base 2
4/3/2003
18.013
18.202
Base 4
4/8/2003
17.994
18.108
0.254
0.171
Base 6
4/11/2003
18.055
18.247
0.213
0.176
Base 7
4/12/2003
17.973
18.206
0.129
0.188
Table 3-13: Volumetric and Gravimetric Cross-Checks - Post FEHP Reference
Lubricant Test Runs
Run
ID
Date
Carbon
Balance mpg
Gravimetric
mpg
COV
Difference
Mean mpg
Difference
PostBaselR2
5/23/03
18.208
18.601
Post Base 2
5/24/03
18.111
18.649
Post Base 3
5/28/03
18.143
18.712
0.025
0.500
Post Base 4
5/29/03
18.169
18.690
0.034
0.505
Post Base 5
5/30/03
18.121
18.629
0.024
0.506
Post Base 6
5/31/03
18.082
18.673
-0.029
0.520
Run ID
Date
Carbon Balance
mpg
Volumetric mpg
COV Difference
Mean mpg
Difference
PostBaselR2
5/23/03
18.208
18.328
Post Base 2
5/24/03
18.111
18.336
Post Base 3
5/28/03
18.143
18.449
0.094
0.217
Post Base 4
5/29/03
18.169
18.375
0.073
0.214
Post Base 5
5/30/03
18.121
18.375
0.044
0.222
Post Base 6
5/31/03
18.082
18.350
-0.010
0.230
Tables 3-11, 3-12, and 3-13 indicate that the COV difference for the cross-checks was typically less than
0.3. The field team leader and SwRI project manager reviewed test run data to determine if any bias or
error may have been introduced for those instances where the COVs differed by greater than 0.3. No bias
or error was identified in these tests. Continued cross-checks indicated that the ratio of volumetric or
gravimetric fuel economy to the carbon balance fuel economy remained consistent throughout the test
runs. Differences in mean fuel economy followed the same trend for each set of cross-checks. Test
personnel observed a difference greater than 0.2 mpg (0.447 - 0.520 mpg) between the carbon balance
3-13
-------
SRI/USEPA-GHG-VR-29
August 20'3
and gravimetric cross-checks, but the difference remained stable throughout the test runs. This indicated
that the gravimetric method consistently overestimated the fuel economy and did not warrant further
investigation. The volumetric method also consistently overestimated the fuel economy when compared
to the carbon balance method by 0.160 to 0.230 mpg. The ratio of volumetric to carbon balance fuel
economy was consistent throughout the test period. Therefore, no further investigation was warranted.
3-14
-------
SRI/USEPA-GHG-VR-29
August 20'3
4.0 TECHNICAL AND PERFORMANCE DATA SUPPLIED BY CONOCOPHILLIPS
The following data is supplied by ConocoPhillips for informational purposes only. The data has not been
verified by the GHG Center.
ConocoPhillips and Visteon state that FEHP provides excellent fuel economy, extreme pressure
lubrication, and antiwear protection under severe service. The FEHP development process included
extensive bench, dynamometer, and vehicle tests. In addition, developers used proprietary axle efficiency
and spin-loss tests to evaluate frictional losses and to optimize axle efficiency while maintaining low
temperatures. ConocoPhillips' controlled test results found FEHP lubricant properties to be better than
synthetic reference fluids under most conditions. Subsequent fuel economy testing by the Ford Research
Laboratory (FRL) confirmed this by showing a 1.5-percent increase in fuel economy over the reference
lubricant normally installed in light truck rear axles. These tests were completed using 1999 Lincoln
Navigators.
The FEHP's unique fluid properties include high lubricant film strength under heavy loads and high
temperatures. This is said to provide excellent component surface protection. FEHP minimizes frictional
drag at low temperatures with a characteristic viscosity of 90,000 cP at -40°C. These fluid properties
allow the FEHP 75W90 lubricant to be used in lieu of 75W140 standard axle lubricants
Projects to certify the FEHP for use in limited slip differentials have been completed successfully. The
FEHP is in use in current production vehicles.
4-1
-------
This page intentionally left blank
SRI/USEPA-GHG-VR-29
August 20'0'3
4-2
-------
SRI/USEPA-GHG-VR-29
August 20'3
5.0 REFERENCES
1. Inventory of Greenhouse Gas Emissions and Sinks. US Environmental Protection Agency.
Washington, DC. April, 2003.
2. Correspondence from Paul Schwarz (Visteon) re: Visteon Beam vs. Independent Rear Suspension
Vehicle Volume. March 20, 2003.
3. 40 CFR Part 86, Control of Emissions from New and In-Use Highway Vehicles and Engines. U.S.
Environmental Protection Agency Code of Federal Regulations. Washington, DC. Feb. 18, 2000.
4. 40 CFR Part 600, Fuel Economy of Motor Vehicles. U.S. Environmental Protection Agency Code of
Federal Regulations. Washington, DC. Aug. 3, 1994.
5. Statistics Concepts and Applications. D.R Anderson, D.J. Sweeney, and T.A. Williams. West
Publishing Company. St. Paul, MN. 1986.
6. A Modern Approach to Statistics. R.L. Iman and W.J. Conover. John Wiley & Sons. New York,
NY. 1983.
7. Proposed A-, B-, C-, Coefficient Estimation Procedure, including Appendix A ~ Calculated
Dynamometer Coefficients (W. Mears, April 24, 1995). Personal communication from Gerald A.
Esper, American Automobile Manufacturer's Association, to Phil Lorang of the U.S. EPA and K.D.
Drachand of the California Air Resources Board. Detroit, MI. September 28, 1995.
8. Standard Practice for Dealing with Outlying Observations, ASTM Standard E178-02. ASTM
International. West Conshohocken, PA. 2002.
SwRI Standard Operating Procedures
SOPs#
Revision date:
06-002
06-003
06-007
06-010
06-011
06-013
06-014
06-016
06-021
06-023
06-036
06-041
06-042
NOx converter efficiency determination
Linearity verification of gas dividers
Naming monthly calibration gas
Barometric pressure verification
Propane recovery check
Temperature calibration and verification
CVS tunnel stratification check
Wet C02 interference check for CO analyzers
FID response for methane
Calibration of analyzers using digital readout
Verification of zero gases
NOx analyzer C02 quench check
Verification of SRM or NIST-traceable gases
01-13-1998
01-19-1998
10-16-1997
04-10-2000
01-22-1999
06-17-1996
11-03-1995
09-09-1996
10-20-1995
03-04-1999
08-11-1997
04-05-1999
06-25-1998
5-1
-------
SRI/USEPA-GHG-VR-29
August 20'3
06-043
Verification of pure propane gas
06-02-1999
06-044
Hydrocarbon analyzer optimization
04-04-2002
06-048
48" dyno coastdown procedure
01-23-2002
06-049
Load cell calibration check
03-23-2001
07-013
Light-duty FTP
08-07-1998
07-027
Light-duty HFET
11-16-1995
08-004
Verification of driver's trace
02-14-1996
12-001
Quality system and process audits
02-16-2001
5-2
-------
SRI/USEPA-GHG-VR-29
August 20'3
APPENDIX A
Engine Oil and Axle Lubricant Change Procedures and Records
A-l
-------
SRI/USEPA-GHG-VR-29
August 2003
Vinton L2~ February 2L MB DRAFT
Do not au, quote, use, or dint. mission mm GHG Center
' jipt'.lu""',
k ¦ s ,rr Is i iv ",-jl.; \. ¦ r o" ;i m,]-
Notes;
Ford Research Laboratory has developed this procedure to yield the most consistent possible run - to - run
test results and submitted it for use in this Test Plan,
The same technician will perform ail rear axJe lubricant changes.
Change engine lubricant and filter at the same time,
1, Remove the axle cover and drain plug; allow tie lubricant to drain,
2, Remove rear wheels and brakes ami soak up any lubricant in the axle tube (for solid beam axlesi with
A-2
-------
SRI/USEPA-GHG-VR-29
August 2003
A-3
-------
SRI/USEPA-GHG-VR-29
August 2003
-------
SRI/USEPA-GHG-VR-29
August 2003
A-5
-------
SRI/USEPA-GHG-VR-29
August 2003
m DRAFT
quote, use, or distribute without written permission from GHO Center
Appssciix a-3
Rear-Axle Lubricant Orange Procedure and Observations
-------
SRI/USEPA-GHG-VR-29
August 2003
x cover and measure volume of oil that comes out. Save
A-7
-------
SRI/USEPA-GHG-VR-29
August 2003
n i-1 M uir11 tc
11_
I )r»in ni
dition dyno setui
S + single HFET
ih.
nil Ti} tit a rip
puct ana nil with fc
-------
SRI/USEPA-GHG-VR-29
August 2003
-------
SRI/USEPA-GHG-VR-29
August 2003
-------
SRI/USEPA-GHG-VR-29
August 2003
A-ll
-------
This page left intentionally blank
SRI/USEPA-GHG-VR-29
August 20'0'3
A-12
-------
SRI/USEPA-GHG-VR-29
August 20'3
APPENDIX B
Test Fuel Analyses
B-l
-------
SRI/USEPA-GHG-VR-29
August 2003
B-2
-------
SRI/USEPA-GHG-VR-29
August 20'3
APPENDIX C
Daily Test Protocol and Checklist
C-l
-------
SRI/USEPA-GHG-VR-29
August 2003
initial
' ,i.is : ¦:'vnir.ii- < up v-i -i
i' - ? Tititc rv'wivn.v -ur : .--il-'V
rii c *jti runiiirv lv ih- >.!d .1 v ' ' V ^
m ¦ v rn,:it:" vr iK;
ehicle on chassis dyno and note time; ,
v 1" i' iv nj " i ion*; i.'T, m ii! ir r - r
ib 1 cart and scsls in corrsct positions sncl cc
¦ us \ r ' 'i\ii¦¦ i, 'Mi' v '/» hi
ii i , i, i ¦ 1» ii>. ii . rii i". ) v 1 -i ,,i
ran
if U
i option
How:
)U cm
test information on dyno computer s
v h ; i " rPv i' v '"-in' 'VMi'
novo now rait
C-2
-------
SRI/USEPA-GHG-VR-29
August 2003
ainqs at ©no
and scale readinas
C-3
-------
This page left intentionally blank
SRI/USEPA-GHG-VR-29
August 20'0'3
C-4
-------
SRI/USEPA-GHG-VR-29
August 20'3
APPENDIX D
Dynanometer Setup Data
D-l
-------
SRI/USEPA-GHG-VR-29
August 2003
-------
SRI/USEPA-GHG-VR-29
August 2003
8 3 £
w. O
iggss
<
£
5 g 6 8 «
» <0 -*¦
C*:5 CO U> N3
6 6 6 6
O C O CJ
3383s
CO <5 tn it
6 66 6
111
o
a
w
-------
O) O
CVS o
is
> *J
6 "
I
o
<
Q.
UJ
£0
ro
0£
m
-------
SRI/USEPA-GHG-VR-29
August 20'3
APPENDIX E
Outlying Data Review & Analysis
E-l
-------
SRI/USEPA-GHG-VR-29
August 20'0'3
OUTLYING DATA REVIEW & ANALYSIS
The GHG Center reviewed all valid test run data (four test runs were excluded and explanations are
provided) for consistency and completed statistical analyses to determine the statistical significance of
observed fuel economy changes. The Center noted during the data review that the fuel economy for the
reference lubricant for run Base-3 was considerably lower than the fuel economy for the other five test
runs. The repeatability of the six FEHP test runs, as measured by the sample standard deviation and
COV, compared favorably to the standard deviation and COV of the reference lubricant data without the
Base-3 run data included in the data set. Run Base-3's inclusion in the reference lubricant data set made
the calculated COV and standard deviation notably larger than the COV and deviation for the FEHP data
set. This indicated that the data for run Base-3 could be suspect. Table E-l summarizes this observation.
Table E-l: Evaluation of Run Base-3 Impact on Variability
Parameter
FEHP Runs3
Reference Lubricant Runs (Base
1,2,3,4,6,7)
Reference Lubricant Runs (Base
1,2,4,6,7)
Mean Fuel Economy
(mpg)
18.255
17.961
18.021
Standard Deviation
(mpg)
0.0296
0.151
0.0408
COV
0.162
0.839
0.226
A Six FEHP runs.
The GHG Center evaluated the data for run Base-3 using ASTM standard El87-02: Standard Practice for
Dealing With Outlying Observations*''1 This standard presents methods to test the statistical significance
of outlying observations in sample data sets.
The standard suggests a review of test data for inconsistencies or error in the test procedure. A review of
all test data for test run Base-3 did not indicate any problems or errors in the test procedure, equipment,
calibrations, or other potential sources of error. Gravimetric and volumetric fuel economy cross checks
also correspond to the observed carbon balance fuel economy result for run Base-3. Therefore, no
physical reason could be identified for the suspect data. The ASTM standard describes a statistical test
that can be used to determine whether a data point lies outside the distribution exhibited by the remainder
of the data. The standard recommends the following criteria for a single outlying sample in a series of n
tests:
Tn =(X-Xl )/s
(Eqn. E-l)
where:
Tn = sample test criterion
xi = low outlying data point
x = mean of sample series for all values
s = sample standard deviation for all values
The sample test criterion for the initial reference lubricant test data is calculated as follows:
E-2
-------
SRI/USEPA-GHG-VR-29
August 20'3
T6 = (17.961 mpg- 17.663 mpg)/0.151 =1.980 (Eqn. E-2)
where:
T6 = test criteria for n=6 test runs
17.961 = mean fuel economy for initial reference lubricant runs (mpg)
17.663 = fuel economy for test run Base 3 (mpg)
0.151 = standard deviation for initial reference lubricant test runs (mpg)
This value is compared to the calculated critical T values for six tests found in the standard. The critical
T values at a significance level of 0.1 percent and 0.5 percent are 2.011 and 1.973, respectively.1-8-1 Based
on a comparison to these values, a large T-value for a data point of this low magnitude would occur by
chance no more often than 0.5 percent of the time because 1.98 exceeds 1.973 but is less than 2.011. This
is significantly less than the 95 percent confidence interval that has been specified as an interval of
concern for data obtained for this test. Therefore, based on the analyst's application of the ASTM
standard test, the data from run Base-3 can be identified as an outlier.
SRI sought further verification that run Base-3 is the sole potential outlier in the reference lubricant data
set. The next lowest fuel economy value for the reference lubricant was evaluated according to the same
procedure (with run Base-3 eliminated). The sample test criterion for this value is:
T5 = (18.021 mpg -17.973 mpg)/0.0408 = 1.176 (Eqn. E-3)
where:
T5 = test criteria for n=5 test runs
18.021 = mean fuel economy for initial reference lubricant runs not including
Base 3 (mpg)
17.973 = fuel economy for run Base 7
The smallest critical T value specified in the ASTM Method (corresponding to a 10 percent significance
level) is 1.602,(8) much larger than the calculated T value. Therefore, the likelihood of the calculated T5
value occurring is much greater than 10 percent, indicating that the subject data is likely from the same
sample set as the remainder of the data and is valid.
E-3
------- |