United States May 11,
Environmental Protection 1993
Agency \ EPA-AA-AQAB-93-01
Air
» EPA Evaluation of a Four-Mode
Steady-State Test With
Acceleration Simulation Modes
As An Alternative Inspection and
Maintenance Test for Enhanced
l/M Programs
William M. Pidgeon
Daniel J. Sampson
PaulRBuibagelV
Lany C. I -anrfman
William B. Clemmens
Erik Herzog
David J. Bizezinski
David Sosnowski
-------
1. Executive Summary 1
1.1 Purpose 1
1.2 Findings 2
1.2.1 Ability to Correctly Identify Vehicles Needing Repair ..2
1.2.2 Ability to Distinguish Sufficiently Repaired Vehicles
From Insufficiently Repaired Vehicles 4
1.2.3 Ability to Distinguish Between Functioning and
Malfunctioning Evaporative Canister Purge Systems 4
1.2.4 Test Costs 5
1.2.5 Adequacy of the ASM for Enhanced I/M Programs 5
2. Background 6
3. Test Procedures 12
4. Data Description .13
4.1 Data Listings 13
4.2 Database Statistics 14
4.3 Quality Control (QC) Protocol 17
5. Analyses/Discussion 19
5.1 Introduction 19
5.2 Analyses Techniques 19
5.2.1 Reducing Four Steady-State Modes to a Single Score per
Pollutant For Comparison to One Cutpoint per
Pollutant 19
5.2.1.1 Reporting Overall ASM Results Versus Reporting
Individual Mode Results 20
5.2.1.2 Determination of Individual Mode Scores 21
5.2.2 Multiple Linear Regressions to Find ASM Coefficients ...21
5.2.3 Applying ASM Coefficients 22
S.2.4 ASM Concentration Measurements 23
J— v
5.2%5 Explanation of the Criteria Used To Compare I/M Tests ..23
5.2.5.1 Excess Emission Identification Rate (IDR) 23
5.2.5.2 Failure Rate 24
5.2.5.3 Error-of-Commission (Ec) Rate 24
5.2.5.4 Two-Ways-To-Pass Criteria 25
5.2.5.5 Discrepant Failures (DFs) 26
5.2.5.6 Unproductive Failure (UF) Rate 27
-------
5.2.5.7 Vehicles with Malfunctions That Were Not
Counted as Ecs and DFs 27
5.2.5.8 Weighting Factors to Correct Biased
Recruiting 29
5.3 Comparison of IM240 Versus ASM Using Cutpoint Tables 31
5.4 Comparison Using Scatter Plots and Regression Tables 40
5.4.1 Using the Coefficient of Determination (R2 ) and
Standard Error of the Estimate for Objective
Comparisons 40
5.4.2 Advantage of Using Weighted Modes 42
5.4.3 Observations of Scatterplots 43
5.4.4 Poor ASM HC Correlation 43
5.5 Derivation of ASM Coefficients and Mode Contribution
Variations From Sample to Sample 45
5.5.1 ASM Versus IM240 As The Dependent Variable For
Determining ASM Coefficients 45
5.5.2 Variability of ASM Coefficients 51
5.5.3 Significance of Mode Contributions 53
5.5.4 Conclusions on ASM Mode Contributions 57
5.6 Repair Analyses 57
5.6.1 Contractor Repairs 57
5.6.2 Commercial repairs 71
5.6.2.1 Introduction 71
5.6.2.2 Database/Analysis 71
5.6.2.3 Results/Conclusions 73
5.6.3 In-Use Emission Reductions from Real World Repairs 77
5.6.4 One-Mode Repairs on ASM 90
5.7 Purge Analyses 99
5.7.1 Introduction 99
5.7.2 The Database 100
5.7.3 The Results 100
5.8 IM240 Improvements and the Four-Mode IM240 104
5.8.1 Reduce Test-to-Test variability 105
5.8.2 Statistical Techniques to Improve the IM240's
Correlation With the FTP 107
ii
-------
6. Test Programs by Other Organizations 110
6.1 Colorado Test Program 110
6.2 California Test Program 110
7 . Test Costs Comparison 112
8. Evaluation of the Adequacy of the ASM for Enhanced I/M Programs 115
8.1 Introduction 115
8.2 MOBILESa Analysis 116
9. Appendices Table of Contents 120
iii
-------
1 . Executive Summary
1 . 1 Purpose
On November 5, 1992, the U.S. Environmental Protection Agency (EPA)
promulgated a regulation1 for state-operated enhanced Inspection and
Maintenance (I/M) programs. This regulation established the IM2402 as the
benchmark I/M test, against which any alternative test must be found
equivalent, or nearly so but with compensating improvements in other program
aspects.
EPA performed tests on over 1500 vehicles in Mesa, Arizona to evaluate a
four-mode, steady-state procedure utilizing two Acceleration Simulation
Modes3. (This four-mode test procedure will herein be referred to as the
"ASM" test, although only the first two modes are strictly ASM modes.) This
evaluation was designed for determining whether the ASM is a suitable
alternative to the IM240 for enhanced I/M testing.
The ASM test utilizes equipment costing about half of the anticipated cost
of the equipment required for IM240 testing. This equipment is less expensive
because the ASM does not involve transient driving and the equipment only
approximates mass emissions via pollutant concentration measurements. In
contrast, the IM240 is a transient test requiring more expensive equipment
measuring true mass emissions during typical driving.
The purpose of this document is to provide:
EPA's evaluation regarding the effectiveness of the ASM test;
- a description of the analysis techniques EPA used;
- the data used in the evaluation; and
1 Inspection/Maintenance Program Requirements; Final Rule 40 CFR Part 51,
Federal Register, November 5,1992
2 William M. Pidgeon, and Natalie Dobie, "The IM240 Transient I/M Dynamometer
Driving Schedule and The Composite I/M Test Procedure," EPA-AA-TSS-91-1,
January 1991
3 Thomas C. Austin and Larry Sherwood, "Development of Improved Loaded-Mode
Test Procedures for Inspection and Maintenance Programs," Sierra Research,
Inc. and California Bureau of Automotive Repair, SAB Technical Paper No.
891120, May 1989.
-------
- a description of the test program.
This is the only ASM study conducted in an official I/M station. The
vehicles were randomly selected and tested under the widely varying ambient
conditions and preconditioning that normally attend official I/M tests. Many
more cars were tested than in any other ASM study. Also, this is the only
study to use one sample to develop the ASM mode weighting factors and an
independent sample to evaluate their effectiveness. EPA strongly believes that
this study should be given far more weight than all previous ASM studies.
1. 2 Findings
EPA's findings are based on performance comparisons between the ASM and
the IM240 regarding five important considerations:
- their relative ability to fail malfunctioning vehicles (needing
exhaust emission control system repairs) and to avoid failing
properly functioning vehicles;
- their relative ability to distinguish repaired vehicles (exhaust-
repairs) that are sufficiently repaired from those that are
insufficiently repaired;
- their relative ability to distinguish between functioning and
malfunctioning evaporative canister purge systems;
their relative costs; and
the adequacy of the ASM for Enhanced I/M Programs using MOBILE5a.
1.2.1 Ability to Correctly Identify Vehicles Needing
Repair
EPA commonly uses the rate of excess emissions identified during an I/M
test to objectively and quantitatively compare I/M test procedures. Excess
emissions are those FTP-measured emissions that exceed the certification
emission standards for the vehicle under consideration. For example, a
vehicle certified to the 0.41 g/mi HC standard whose FTP result was 2.00 g/mi,
would have excess emissions equalling 1.59 g/mi HC (i.e., 2.00 - 0.41 =» 1.59).
The excess emissions identification rate (IDR) equals the sum of the
excess emissions for the vehicles failing the I/M test divided by the total
-------
excess emissions.
better the test.
The more excess emissions an I/M test identifies, the
EPA uses IDR instead of merely comparing the number of vehicles that
correctly fail and correctly pass. The IDR better contrasts the relative
merits of competing- I/M test procedures because failing vehicles with high
emissions is more important than failing those that are only slightly above
their certification standards. For example, take two I/M procedures that
correctly failed 100 of the 500 vehicles that had FTP emissions greater than
their certification standards, but only 50 cars failed both tests. If the
fifty cars that failed Test A were high FTP emitters, and the other 50 cars
that failed Test B had FTP emissions only slightly above their standards,
obviously Test A would be preferred, and its IDR would reflect its better
performance. Test A's better performance is not evident in comparing the
number of vehicles that correctly fail.
The ASM does not find high emitting vehicles as well as the IM240. Some
high emitters which could be caught with the IM240 give low ASM scores. Table
1.2.1 shows the percent decrease in the excess emissions identification rate
that would accompany substituting the ASM for the IM240. For example, an
IM240-based I/M program's HC and NOx IDRs will suffer nearly a 20% decrease by
substituting the ASM test at the same failure rate (18%) that is produced by
EPA'a recommended cutpoints for biennial I/M programs.
Table 1.2.1 Loas in Identification Bffectiveness With ASM Test
Scenario
Failure Rate Held at 18%
(0.8/15/2.0 + 0.50/12.0 IM240
Cutpoints)
Best IDRS' with ECS Held
Below 5%
BC CO NOz
19.0% 9.5% 18.5%
14.0% 14.3% 17.5%
(92 2-74 7)
These valuaa are % differences. For example: J—' - ' ' * 100 - 19.0%
Source: Table 5.3.1, Section 5.3
An aggressive I/M program, tolerating both higher failure rates and higher
false-failure rates would relinquish about 15% of its inspection effectiveness
by substituting the ASM test.
Additional related findings are listed below:
-------
• The ASM fails cars that are actually clean more often than the IM240.
About 1 in 10 cars failed by the ASM did not appear to need repair,
compared to about 1 in 30 for the IM240. EPA knows from other testing
that more preconditioning can eliminate IM240 errors; we are not sure
whether it can for ASM failures.
• Making ASM cutpoints more stringent in an attempt to get the same
effectiveness as the IM240 increases the failure rate and/or the error
rate beyond what EPA believes any I/M program would want or is willing
to commit to in binding regulation form.
The comparative ability to identify vehicles needing repair is fully
discussed in Section 5.3. Why IDRs and associated criteria are important, how
the criteria are derived, and the tradeoffs associated with increasing
cutpoint stringency to increase IDRs are discussed in Section 5.2.
1.2.2 Ability to Distinguish Sufficiently Repaired
Vehicles From Insufficiently Repaired. Vehicles
Vehicles that do fail the ASM test and get repaired, can pass ASM
cutpoints with repairs that are not as effective as the repairs needed to pass
IM240 cutpoints, even when repaired in good faith. Also, the ASM modes are
prone to "adjust to pass/readjust after" strategies like the idle and
2500/idle tests.
Several of the 17 cars which failed the Arizona test and the ASM were
repaired in local shops, after which they passed the Arizona and ASM test but
still had high IM240 emissions. This is the same pattern seen in 2500/idle I/M
programs. Repair analyses are discussed in Section 5.6.
1.2.3 Ability to Distinguish Between Functioning and
Malfunctioning Evaporative Canister Purge
Systems
In purge testing, the ASM and the IM240 should do equally well in
identifying malfunctioning purge systems, so their comparative ability to fail
vehicles with malfunctioning purge systems has not been an issue. Therefore,
the research issue has been whether, and how many properly functioning
vehicles would fail. That is, EPA is more concerned with errors-of-commission
than with errors-of-omission. About 4-6% of the vehicles failed the ASM
evaporative canister purge system test but were actually properly functioning.
This is about 38% to 52% of all cars that failed the ASM purge.
-------
About 1% of the vehicles failed the IM240 purge system test, but were actually
properly functioning. This is about 12% to 18% of all cars that failed the
IM240 purge.
Unlike transient IM240 testing, which requires vehicles to operate through
a wide range of speeds and loads, the four steady-state modes of the ASM do
not provide a purge opportunity for a significant portion of the fleet. The
purge system test results are discussed in Section 5.7.
1.2.4 Test Costa
The 180 seconds required for this four-mode ASM test is the same as would
be needed for the IM240 if special algorithms are used to pass obviously clean
cars and fail obviously dirty cars early in the cycle. So, the ASM does not
save test time or reduce the number of lanes required. A shorter test based on
fewer than four modes would have even less benefit.
The only cost advantage for this ASM test is that up to about half the
equipment cost can be avoided by not having variable inertia capability in the
dynamometer and low-concentration measurement capability in the gas analysis
instruments. This savings works out to about 75 cents per test in a
centralized program. Test costs are discussed in Section 7.
1.2.5 Adequacy of th« ASM for Enhanced I/M Programs
The MOBILESa analysis results show that even in a maximum annual program,
covering all weight classes, with ASM, purge, and pressure testing of all
model years and comprehensive anti-tampering inspections, the ASM test yields
insufficient benefits to meet the performance standard for HC, CO, or NOx.
-------
2. Background
EPA began development of a transient I/M test procedure, named the IM240,
during 1989. EPA published a Notice of Proposed Rulemaking on July 13, 1992
which proposed a performance standard for enhanced I/M programs that assumed
the use of the IM240 test procedure.
On May 8, 1992, ARCO Products Company released a report4 recommending that
an alternative to the IM240 be allowed for enhanced I/M programs. The
operating modes for ARCO's alternative procedure were not conclusively
determined, but the modes were based on "Acceleration Simulation Mode"
procedures developed by the California Bureau of Automotive Repair and Sierra
Research, Inc.
In contrast to ARCO's report which was somewhat ambiguous on which modes
should be included in.the alternative test, the earlier BAR/Sierra report5 had
recommended an ASM I/M test that included the traditional 2500/idle test (2
modes) with two loaded dynamometer modes, at 15 mph and at 25 mph. The
authors concluded that these two dynamometer modes were needed for NOx
correlation with the FTP*, and that the 2500/idle modes were necessary for
good HC/CO correlation.
ARCO's report, which reached different conclusions, was based on test
results from five newer vehicles that were tested with and without implanted
malfunctions, resulting in 30 tests. ARCO's conclusions are directly quoted
below:
4 Kenneth L. Boekhaus, Brian K. Sullivan, and Charles E. Gang, "Evaluation of
Enhanced Inspection Techniques on State-of-the-Art Automobiles," ARCO Products
Company, May 8, 1992.
5 Austin and Sherwood.
* The Federal Test Procedure (FTP) is a mass emissions test created to
determine whether prototype vehicles comply with EPA standards, thus allowing
production vehicles to be certified for sale in the United States. The FTP
has become the "gold standard" for determining vehicle emission levels, so it
is also used to determine the emission levels of "in-use" vehicles. The FTP
is too costly to use for I/M because vehicles must be maintained in a closely
controlled environment for over 13 hours. The FTP driving cycle includes 31
minutes of actual driving which takes 41 minutes to complete due to a 10
minute engine shut-off between the second and third modes of this 3 mode test.
-------
1. An enhanced IM program utilizing steady-state exhaust emission testing is
as effective in identifying cars needing repair as is the EPA's proposed
IM240 test Because the cost of the IM240 equipment is four times that of
an enhanced I/M test, the enhanced I/M test is far more cost effective.
2. Canister purging can be tested as effectively in an enhanced I/M test as in
anIM240test
3. The current BAR90 exhaust emissions test conditions of idle and 2500
rpm/no load are not effective in identifying most malfunctions in state-of-
the-art automobiles.
4. The ASM501S steady-state test condition is effective in identifying
malfunctions for HC, CO and NOx and should be included in any
enhanced I/M test developed.
S. Further development work is needed to develop one or more other steady-
state test conditions to complement the ASM501S test
6. The IM240 test correlates better with the FTP test in predicting absolute
emissions levels than does the ASM501S test With one or more
additional steady-state test conditions, steady-state testing would likely
correlate as well as the IM240 test
7. The BAR 90 Test Analyzer System, with NOx analyzer, can be used for
enhanced I/M testing incorporating a steady-state dynamometer.^
On November 5, 1992, EPA promulgated the final I/M rule establishing the
IM240 as the benchmark I/M test, against which any alternative test must be
found equivalent. The IM240 is a transient test which measures true mass
emissions during typical driving. In contrast, the Acceleration Simulation
Mode procedures only approximate mass emissions via pollutant concentration
measurements during several steady-state modes. A BAR90 HC and CO analyzer
with an NO analyzer is sufficient. Emissions measurements are not made during
the accelerations and decelerations between the steady-state driving modes
because such measurements require more expensive equipment including a
constant volume sampler to dilute the exhaust and measure flow, and analyzers
capable of accurately measuring the resulting low concentration samples.
The purpose, of EPA's alternative I/M test procedure study is to evaluate
whether the IM240's performance as an I/M test can be attained, or nearly so,
by a multi-mode, steady-state procedure (including two ASM modes) that
utilizes equipment costing about half of the anticipated cost of the equipment
required for IM240 testing.
6 Boekhaus, et al.
-------
Due to widespread interest and the need for states to move forward with
specific testing plans, EPA prepared to initiate a test program to evaluate a
steady-state loaded I/M test as a potential alternative to the IM240 for
enhanced I/M. EPA's alternative test study focused on the Acceleration
Simulation Mode procedures.
EPA needed to select a practicable number of steady-state operating modes,
but there was disagreement among the proponents of steady-state testing for
enhanced I/M. ARCO concluded that the 2500 rpm & idle modes are not effective
for identifying most malfunctions. In contrast, BAR/Sierra recommended an I/M
test consisting of the following modes: 5015, 2525, 2500 rpm, and idle.
EPA's desire to evaluate a single steady-state procedure agreeable to all
interested parties led to a conference call with the interested parties on
July 27, 1992. The participants included: ARCO, Sierra Research, California
Bureau of Automotive Repair, Allen Test Products/SAVER, EPA's Testing
Contractor (Automotive Testing Labs), and EPA.
-------
The parties reached a consensus that the steady-state test to be evaluated
should include four modes (a fifth mode was only to be performed on the first
50 cars with automatic transmissions):
15 mph (ASM 5015)*
25 mph (ASM 2525)**
50 mph at road load horsepower***
idle (automatic transmissions using Drive rather than Neutral)****
idle (first 50 vehicles with automatic transmissions using Neutral
rather than Drive)****
This test procedure will herein be referred to as the "ASM" procedure,
although only the first two modes are strictly ASM modes.
* This is a steady-state 15 mph mode (5015). The dynamometer load is set to
simulate 50% (5015) of the power required to accelerate the particular
vehicle being tested at 3.3 mph/second at 15 mph. The ASM does not include a
true speed changing acceleration during emissions measurement, instead the
speed is held constant while the dynamometer load is set to simulate the power
required to accelerate the car. The 3.3 mph/second acceleration rate is the
maximum acceleration rate during the Federal Test Procedure (FTP). The FTP is
the transient (accelerations and decelerations) procedure used to certify that
vehicles comply with Federal emissions standards, which is required before the
manufacturer can offer them for sale. The IM240, for the most part, is taken
directly from the FTP. The 5015 mode usually requires a higher load setting
than the 2525 or the 50 mph road load modes.
** This is a steady-state 25 mph mode (2525). It is analogous to the ASM5015
mode in that the dynamometer load set to simulate 25% (2525) of the power
required to accelerate the particular vehicle being tested at 3.3 mph/second
at 25 mph.
*** This is a 50 mph mode with the dynamometer set to the power required for a
vehicle to maintain 50 mph on level road talcing into account air resistance,
tire losses, bearing friction in the drivetrain, etc. Air drag is the major
resistance at 50 mph.
**** Because the vehicles were to be operated on the dynamometer, it was
judged that the vehicles could be safely tested at idle in Drive. Because
automatic-transmission-equipped vehicles idle in drive during the FTP, and
some ECM algorithms for the emission control system change with transmission
selector position, idle in Drive is expected to yield better correlation with
the FTP than idle in Neutral. Since all known idle emissions tests had been
run in Neutral prior to EPA's ASM evaluation, the first 50 vehicles, or more,
were also run in Neutral to allow comparison with other databases.
-------
Having reached a consensus on the procedure to be evaluated, EPA issued a
work assignment7 on July 30, 1992, directing EPA'3 testing contractor to
implement the new procedure. Shakedown testing began in August and the first
official ASM test was performed on September 10, 1992. The last as-received
ASM test was performed at the I/M lane on March 19, 1993. This analysis
includes tests that were performed through February 17, 1993.
Another issue EPA must consider is the impact of approving alternative I/M
procedures on the automobile manufacturers. The Motor Vehicle Manufacturers
Association, in written comments on the I/M NPRM, said:
6.0.0 ALTERNATIVE TEST METHODS
MVMA agrees that EPA should not allow enhanced I/M areas to
implement alternative tests until they have submitted substantial data supporting
the quality of the test, and showing that the test produces emission reductions
equivalent to those of the IM240 Test
One such report by ARCO describes an acceleration simulation mode
(ASM) that was compared to the IM240 test A substantial cost advantage with
this alternative test is that it does not require the use of a constant volume
sample (CVS) sampling system. The report references an earlier study by Sierra
Research that calculates mass emissions by multiplying a "constant" times
"emissions concentration" times "inertia weight". Yet during the comparison of
the two test methods, the mass emissions for the ASM were measured utilizing a
CVS. For a more accurate comparison, the ASM data should have been
calculated in the method prescribed for use in the field, i.e., with BAR-90
readings and without the use of CVS equipment
Probably of greater concern, however, is the outpoints selected for each
test process. Since outpoints are an important criteria in comparing and
evaluating test processes, realistic outpoints have to be determined before an
accurate comparison can be made. The IM240 Test outpoints selected for this
[ARCO's] comparison are extremely low and thus "create" false failures. In
contrast, the [ARCO] selection process for the ASM outpoints is not well
explained and remains ambiguous, making IM240 Test versus ASM Test
comparison speculative at best
ARCO used only five vehicles in the study. Their objective was to
"evaluate die viability of an alternative enhanced I/M test" It appears much more
work is required before such an alternative could be properly defined and
evaluated. In the NPRM preamble. EPA stated that if this ASM test can be
shown to be as effective as the IM240 Test, it could be permitted as a
7 Statement of Work Change 1, July 30, 1992; Work Assignment 0-2, Contract
No. 68-CZ-0055, "Teat Procedure to Evaluate the Acceleration Simulation Mode
and the Emissions Measurement Capabilities of a BAR90 Certified Analyzer with
An Integrated Fuel Cell Type NO Analyzer."
10
-------
"substitute". MVMA is concerned that "substitute" tests could lead to several
alternative tests with varying degrees of effectiveness. MVMA requests that
EPA continue to critically assess any alternative tests proposed by enhanced VM
areas. This review process will help assure that any alternative tests are able to
properly identify failing vehicles.
EPA is also puzzled by ARCO'a conclusions which seem contradictory.
ARCO's first conclusion states:
An enhanced IM program utilizing steady-state exhaust emission testing
is as effective in identifying cars needing repair as is the EPA's proposed IM240
test
Their fourth conclusion states that:
The ASM501S steady-state test condition is effective in identifying
malfunctions for HC, CO and NOx and should be included in any enhanced I/M
test developed.
ARCO's fifth conclusion states that:
Further development work is needed to develop one or more other steady*
state test conditions to complement the ASMS01S test
These statements suggest that ARCO's testing indicated that the ASM5015
was not as effective as the IM240, that the other modes they evaluated were
not helpful, and additional work was required to identify better alternatives,
This report will document additional work performed by EPA to evaluate the
ASM5015 and three additional steady-state modes.
11
-------
3. Test Procedures
The beat way to compare I/M tests is to utilize actual results from an I/M
station in conjunction with FTPs run on a subset of the vehicles also tested
at the I/M station. A highly inferior method is to compare the procedures
based only on test data collected in a laboratory which is not subject to the
range of vehicle operating conditions which normally precede actual I/M tests,
nor the range of ambient conditions encountered during actual I/M tests.
It is widely acknowledged that a given vehicle's emissions can vary widely
with changes in vehicle operating conditions that precede emissions tests, and
a given vehicle's emissions can vary widely with ambient conditions
encountered during an emissions test. So, in contrast to laboratory test
results, the results from pilot tests run in an official I/M station provide
significantly more confidence that the pilot test results will accurately
represent future results when the procedure is mandated for official I/M
testing.
For these reasons, EPA carried out IM240 and ASM testing (through a
contractor) in an I/M station in Mesa, Arizona, with FTP testing in a
contractor-owned laboratory also in Mesa. In this respect, EPA's results have
much greater applicability to the real world than results from recent WASM"
testing by Environment Canada8, ARCO, California Air Resources Board, and the
Colorado Department of Health.
The test procedures are discussed in detail in Appendix A.
8 Vera F. Ballantyne, Draft. Steady State Testing Report and Data.
Environment Canada, August 28, 1992.
12
-------
4. Data Description
From September 10, 1992 through March 19, 1993, EPA'a contractor,
Automotive Testing Laboratories (ATL), conducted a vehicle testing program in
Mesa, Arizona, a suburb of Phoenix, on mostly 1983 and newer vehicles.
This program included several tasks designed to produce data for an
analysis comparing the ASM test to the IM240 test as predictors of actual FTP
emissions. These tasks included the operation of an Arizona I/M inspection
lane. Vehicles at this lane received XM240s with a functional test of the
evaporative canister purge system (referred to as the purge test in the
remainder of this report), ASMs with the purge test, Arizona I/M tests, and
fuel system pressure tests under real-world I/M testing conditions. In
addition, vehicles were recruited from the I/M lane for additional tasks,
which included:
• FTP laboratory testing
• IM240 laboratory testing
• Contractor IM240-targeted repairs
• Commercial repairs obtained by vehicle owners to pass the official
Arizona I/M test.
Choosing vehicles for laboratory testing was driven by the importance of
testing and assessing emissions from—and the impact of repair on—dirty in-
use vehicles. A random sample of vehicles visiting the I/M station would
result in the contractor recruiting mostly clean vehicles (see Section
5.2.5.8). But most excess emissions come from a relatively small percentage of
vehicles known as high to super emitters. To avoid the problem and cost of
evaluating a majority of vehicles that will ultimately be assessed as clean, a
stratified recruitment plan was employed to deliberately over-recruit dirty
cars, based on lane-IM240 results at the Mesa lane. A nominally 50/50 mix of
IM240-clean and IM240-dirty vehicles were to be recruited for FTP exhaust
testing. In actual practice, more dirty cars than clean have been recruited
which is shown in Table 4.2.2.
Specifics^concerning the recruitment criteria and the test procedures for
these tasks ar* discussed in Appendix A.
4.1 Data Listings
Appendix B provides a listing of the data used for the cutpoint
effectiveness analysis, the contractor repair analysis, and the commercial
repair analysis, which are all discussed in Section 5.
13
-------
Data for the over 1400 vehicles that only received one set of lane tests
(no laboratory tests and no after-repair lane tests) are only available on
disk. These data include the purge analysis data, and the lane data used to
calculate ASM coefficients. The available disk(s) will include all IM240 and
ASM lane data including lane data for the laboratory tested vehicles. These
can be requested by contacting:
William M. Pidgeon
U.S. EPA
National Vehicle and Fuel Emissions Laboratory
2565 Plymouth Road
Ann Arbor, Michigan 48105-2425
Tel. No. 313-668-4416
Fax. No. 313-668-4497
Fax requests for data disks are preferred and a form is provided at the
end of Appendix B; questions can be addressed by phone.
4.2 Database Statistics
The first official ASM/IM240 test series was run on September 9, 1992.
Data collected up to March 17, 1993 were considered for these analyses.
During that period, 1574 vehicles received 1758 ASM/IM240 test series at the
Arizona Z/M lane. Priority for testing was given to 1983 and newer model year
fuel injected vehicles. The following table illustrates the model year and
fuel metering distribution of the tested fleet:
14
-------
Table 4.2.1
Lane Data By Modal Year and Fuel Metering
MYR I
81
82
83
84
85
86
87
88
89
90
91
92
Totals
PFI
-
-
12
31
38
94
105
100
119
133
150
104
885
1 TBI
.
2
6
30
42
53
50
61
77
48
35
11
415
ICARB
3
3
26
46
49
45
46
36
17
2
-
-
273
Totals
3
5
44
107
129
192
201
197
213
183
185
115
1574
Of the 1574 vehicles tested 27 were recruited for the commercial repair
program and 127 vehicles were recruited to the laboratory for additional tests
and for contractor repairs when the repair criteria were met (Section 5.6).
The following list summarizes the criteria used for recruiting laboratory
vehicles and for data completeness:
• The IM240 and the ASM were designed to distinguish between
malfunctioning and properly functioning newer technology cars, so
only 1983 and newer fuel-injected (no carbureted) cars were used.
• One-half of the laboratory vehicles were to exceed 0.80/15.0/2.0
(HC/CO/NOx) on their lane-IM240.
• One-half of the recruited vehicles were to have the lane-IM240
performed prior to the ASM.
• Only vehicles having an as-received FTP, an as-received lane-IM240,
and an as-received lane ASM test were used. Vehicles missing any one
of these three tests were not included in the analysis.
The resulting database consisted of 106 fuel-injected. Table 4.2.2 lists
actual distribution statistics for these laboratory vehicles.
15
-------
Tabla 4.2.2
Distribution of Laboratory Recruited Vehicles
Lane-IM240
Passed: 20.80 / 15.0 / 2.0
Failed: >0.80 / 15.0 / 2.0
Totals
Fuel
Metering
PFI
TBI
PFI
TBI
ASM Prior
to IM240
18
5
18
14
55
IM240 Prior
to ASM
14
3
23
11
51
Totals
32
8
41
25
106
Table 4.2.3 shows the model year and fuel metering distribution for the
106 laboratory recruited vehicles.
Table 4.2.3
Lab Data by Modal Year and Fuel Metering
MYR I
83
84
85
86
87
88
89
90
91
92
Totals
PR
6
12
7
13
9
5
7
6
7
1
73
1 TBI
2
8
7
6
1
4
2
2
1
-
33
Totals
8
20
14
19
10
9
9
8
8
1
106
Table 4.2.4 provides FTP HC/CO emitter group statistics for these
recruited vehicles. FTP emitter groups are defined based on FTP emissions as
follows:
16
-------
I Normals; HCX0.82 and CCX10.2 I
Highs: 0.82£HC<1.64
HC<1.64
and
or
and
CO<13 . 6
10.2SCCX13.6
Very Highs: 1.64SHC<10.2
HC<10.2
and
or
and
CCX150
13.6£CO<150
Supers: HCS10.2
or
C02150
Table 4.2.4
Lab Vehicle FTP HC/CO Emitter Category Distribution
Normals
67
HiRhs
13
VeryHiRh
25
Supers
1
For more detailed information on the data used for these analyses refer to
Section 5. For details on the data excluded from these analyses refer to
Section 4.3 and Appendix C.
4. 3 Quality Control (QC) Protocol
This Section provides a general description of the QC process. For more
detailed descriptions of the QC criteria and data excluded from these analyses
see Appendix C, which lists the QC criteria in detail and the vehicles removed
due to the QC protocol.
Data were received from ATL in two forms. Calculated cycle-composite
values for all tests (lab and lane), except the ASM tests/ and second-by-
second data for lane-IM240s and ASMs were provided. The calculated data and
the raw second-by-second data followed separate but similar QC processes. The
calculated datfr were processed using a program which performed checks on FTP
data and IM24&- (lab and lane) data. These checks included bag result
comparisons, fuel economy checks, test distance checks, dynamometer setting
checks, and test-to-test comparisons. For details on these checks see
Appendix C.
The second-by-second data were processed by a separate program which
performed similar checks for the raw data. The QC checks for the second-by-
second data included checks for acceptable speed, correct test/mode duration,
17
-------
sampling continuity, reasonable ambient background concentrations, acceptable
purge flow, and reasonable fuel economy. Again details concerning these QC
criteria are included in Appendix C. The second-by-second QC program also
calculated composite values for the IM240 and ASM tests.
In addition to the QC program comparisons, the calculated results reported
by ATL were compared to those results calculated from the second-by-second
data. Significant differences were investigated. All lab vehicles violating
the QC criteria were hand checked by EPA staff and the data were corrected or
removed, as warranted. Due to the volume of lane data, lane vehicles that
violated QC tolerances were removed from all pertinent analyses, without
further attempts to "save" the data unless solutions were obvious. These
unutilized data will be checked, as time permits, for future use. In
contrast, because the vehicles that received FTPs were relatively precious,
significant effort was expended to correct data that were identified by the QC
process.
Vehicles removed from the sample are discussed in Appendix C on page C-4
through C-8.
18
-------
5. Analyses/Discussion
5 . 1 Introduction
The purpose of Section 5 is to present the analysis EPA used to assess
whether the ASM test is sufficiently effective in identifying high emitting
cars needing repair when compared to the IM240 test, and the findings of that
analysis. Additionally, it provides a comparison of the repair issues for
those vehicles that were identified as needing repairs.
Section 5.3 compares the ability of the ASM and the IM240 to identify
vehicles needing repair, and presents EPA's major findings regarding the
effectiveness of the ASM as an alternative to the IM240 for enhanced I/M
programs. It discusses comparisons of IM240 versus ASM using information from
cutpoint tables. Section 5.2 provides information needed to understand how
the cutpoint tables were derived.
Section 5.4 compares the correlation of the IM240 and ASM with the FTP
using traditional regression analysis. Section 5.5 discusses the somewhat
specialized issue of how four ASM scores are combined in one score and the
uncertainties and sensitivities in this process.
Section 5.6 discusses the repairs performed by the contractor and repairs
performed by commercial repair shops.
Section 5.7 discusses canister purge system test results and Section 5.8
discusses methods that will be explored to improve the power of the IM240.
5 . 2 Analyses Techniques
This section discusses the methodology and criteria EPA used to compare
the ability of the ASM and the IM240 to identify vehicles needing repair.
This section explains why the criteria are important, how the criteria are
derived, and indicates the tradeoffs associated with these interrelated
criteria. Then, Section 5.3 contrasts the ASM and the ZM240 using the
criteria explained in Section 5.2.
5.2.1 Reducing Four Steady-State Hodes to a Single
Score per Pollutant For Comparison to One
Cutpoint per Pollutant
This section explains how the final ASM score is computed. Two questions
will be answered in this section:
19
-------
1. Should the four mode scores for each pollutant be combined to
calculate a single result or score for each pollutant, or should a
separate score be reported for each of the four modes, and apply
those separate scores to separate cutpoints for each of the four
modes?
2. How is the score computed for each ASM mode?
5.2.1.1 Reporting Overall ASM Results Versus Reporting
Individual Mod* Results
There are three alternatives for reporting overall ASM test results. The
first alternative does not combine the scores from the separate modes, so this
alternative is analogous to the way 2500/idle test results are reported. The
HC and CO scores are reported for the 2500 mode and separate HC and CO scores
are reported for the idle mode for a total of four scores and up to four
cutpoints. For the four mode ASM test, this is too complicated. With NOx
added, three cutpoints are needed for the 3 dynamometer modes and two
cutpoints (HC and CO only) for the idle mode, necessitating 11 separate
cutpoints. (Because NOx emissions are insignificant during an idle test, NOx
is only considered for the 3 dynamometer modes.) This first alternative is
too unwieldy for a four mode test.
The second and third alternatives are two different ways to report a
single score for each pollutant by combining one pollutant's scores from all
the modes.
For the second alternative, the single score would be the sum of the
scores from each mode, using a weighting of 25% for each mode. For example,
to calculate the single ASM score for HC, the equation would be as follows:
ASM HC = (0.25 * 5015 HQ + (0.25 * 2525 HQ + (0.25 * 50RL HQ + (0.25 * idle HC)
In the third alternative, which EPA used, a single score is determined
from the sum. off the individual mode scores, but the weighting or coefficient
for each was determined by regression techniques. A multiple regression was
performed wherein all four of the mode scores are independent variables that
were regressed against FTP scores. The regression produced coefficients for
each mode, plus a constant. These coefficients weight each mode more
appropriately than the second alternative's method of just assigning each mode
a weighting of 25%. BAR/Sierra used this regression method, and likewise
EPA's analyses for this report also used this regression method, with one
difference that is discussed in Section 5.5. This yields an equation to
20
-------
calculate a single ASM score for each pollutant. For example, the equation
for calculating a single ASM HC score is as follows:
ASM HC = (x * 5015 HQ + (y * 2525 HC) + (z * 50RL HC) + (t * idle HC) + Constant
the x, y, z, t, and constant terms are listed in Table 5.2.2.
While ASM advocates have used this concept, none have proposed specific
coefficients for EPA to evaluate. Thus, EPA had to develop coefficients.
The remaining question is: "How were each of the individual mode scores
determined?"
5.2.1.2 Determination of Individual Mode Scores
Emission concentrations were measured on each of the four ASM modes (see
Section 5.2.3 for more details). These concentration measurements were then
converted to simulated grams/mile emissions, because concentration
measurements do not provide a reliable indication of the magnitude of
pollutants emitted per mile traveled. At the same exhaust concentration
level, a heavy vehicle will emit more per mile than a light vehicle.
To calculate simulated g/mi results, EPA followed BAR/Sierra's method,
which was also followed by ARCO, wherein the measured concentration values are
multiplied by the Inertia Weight (engine displacement for the idle mode) of
the vehicle. The Idle Mode was not considered for NOx since it is a no load
test. EPA also divided these simulated g/mi results by the scaling factors
listed in Table 5.2.1. Using these factors yield overall ASM scores that have
magnitudes similar to FTP and IM240 magnitudes.
Table 5.2.1: Scaling Factors Used to Keep
Regression Coefficients of Kqual Magnitude
Pollutant
HC
CO
NOx
Modes 1-3
[COMC1 * IW / x
105
102
106
Mode 4
[CONG] * Diso(L) / x
103
10°
NA
5.2.2 Multiple Linear Regressions to rind ASM
Coefficients
As previously discussed, multiple linear regressions were performed using
the four modes (three for NOx) of the ASM test as the independent variables
21
-------
(XI,...,X4) The one difference mentioned in Section 5.2.1.1 above is that the
IM240 (rather than the FTP) was used as the dependent (Y) variable*. This was
done for testa on which the ASM was run first only, because the corresponding
IM240s are pre-conditioned, and thus more closely resemble an FTP.
Vehicles that were recruited to the lab or received commercial repairs
were not included in the database used to develop coefficients, because these
are the cars to which the coefficients were applied. EPA determined that
including these would cause the developed coefficients to mask the test
variability of the ASM. (This is also discussed in Section 5.5.)
The multiple linear regressions were run on a database of 608 lane ASM
tests versus pre-conditioned lane-IM240s, giving the following coefficients
for each mode.
Table 5.2.2
Coefficients Developed from Multiple Regression ASM Versus
(see Table 5.2.1 for scaling factors)
ZM240
Mode
Constant
5015
2525
50MPH
Idle
Adjusted R2
HC
0.083
0.025
0.059
0.136
0.124
29.0%
CO
2.936
0.040
0.043
0.356
1.350
50.1%
NOx
0.258
0.061
0.219
0.352
NA
59.1%
5.2.3 Applying ASM Coefficients
The coefficients were then used to calculate composite ASM scores for all
lab vehicles and commercially repaired vehicles. These are the ASM scores
that are reported in the ensuing cutpoint tables, scatterplots, and
regressions.
* Why the IM240 was used as the dependent variable, rather than the FTP, is
explained in Section 5.5. This is not discussed here because the purpose of
this section is to explain how, rather than why. Also, this issue requires a
lengthy discussion and relies on information presented in Section 5.5, so
repetition is also avoided.
22
-------
5.2.4 ASM Concentration Measurements
The ASM concentrations were measured over a 40 second period. Because the
exhaust sample delay to the most downstream analyzer cell is almost 10
seconds, the first 10 seconds of data were ignored. The concentrations that
are used in the composite ASM score calculations are actually reported
averages over the last 30 second period. For various reasons, the time
allotted for measured concentrations was occasionally less than 30 seconds.
In these few cases, EPA calculated the average concentrations over this
shortened period and reported these values. No ASM tests were accepted with
concentrations averaged over a period of less than 20 seconds.
5.2.5 Explanation of the Criteria Used To Compare I/M
Tests
In assessing the overall effectiveness of the ASM relative to the IM240,
it is important to determine their effectiveness in measuring and determining
a variety of factors, including the excess emissions identified, the failure
rate, the error-of-commission rate, the two-ways-to-pass criteria, the
discrepant failures, and the unproductive failure rate. Each of these is
discussed below. These criteria are used in Section 5.3 to compare the
effectiveness of the two procedures.
5.2.5.1 Excess Emission Identification Rate (IDR)
EPA commonly uses the rate of excess emissions identified during an I/M
test to objectively and quantitatively compare I/M test procedures. Excess
emissions are those FTP-measured emissions that exceed the certification
emission standards for the vehicle under consideration. For example, a
vehicle certified to the 0.41 g/mi HC standard whose FTP result was 2.00 g/mi,
would have excess emissions equalling 1.59 g/mi HC (i.e., 2.00 - 0.41 - 1.59).
The excess emissions identification rate (IDR) equals the sum of the
excess emissions for the vehicles failing the I/M test divided by the total
excess emissions (because of imperfect correlation between I/M tests and the
FTP, some I/Jfrrpissing vehicles also have excess emissions which are used for
calculating that total excess emissions). Thus, assuming an I/M area that
tests 1000 vehicles, 100 of which are emitting 1.59 g/mi excess emissions
each, while the I/M test fails (identifies) 80 of the excess emitting
vehicles, the excess emission identification rate can be calculated as
follows:
80 failing vehicles * 1.59 q/mi excess per vehicle ^ QQ m afl%
100 vehicles * 1.59 g/mi excess per vehicle
23
-------
EPA uses IDR instead of merely comparing the number of vehicles that
correctly fail and correctly pass. The IDR better contrasts the relative
merits of competing I/M test procedures because failing vehicles with high
emissions is more important than failing those that are only slightly above
their certification standards. For example/ take two I/M procedures that
correctly failed 100 of the 500 vehicles that had FTP emissions greater than
their certification standards, but only 50 cars failed both tests. If the
fifty cars that failed Test A were high FTP emitters, and the other 50 cars
that failed Test B had FTP emissions only slightly above their standards,
obviously Test A would be preferred, and its IDR would reflect its better
performance. Test A's better performance is not evident in comparing the
number of vehicles that correctly fail.
5.2.5.2 Failure Rat*
As the IDR increases with different test procedures or different
cutpoints, the opportunity to identify vehicles for emission repairs also
increases. However, this measure is not sufficient for determining which is
the more efficient and cost-effective I/M test. Other criteria must also be
addressed before such an assessment can be made. One such criterion is the
failure rate, which is calculated by dividing the number of failing vehicles
by the number of vehicles tested. For example:
50 vehicles failed I/M . .... _. „.„ _ .,
,.., ——: r— * 100 - 5% I/M failure rate
1000 vehicles tested
The ideal I/M test is one that fails all of the dirtiest vehicles
while passing those below the FTP standard or close to it but still above it.
The potential emission reduction benefit decreases as emission levels from a
vehicle approach the standard, because the prospect for effective repair
diminishes. Thus, achieving a high IDR in conjunction with a low failure rate
(as a result of identifying fewer vehicles passing or close to the standard)
efficiently utilizes resources.
5.2.5.3 Krror-of-Conmission (Be) Rat*
Properly functioning vehicles which pass FTP standards sometimes fail the
I/M test; these are referred to as false failures or errors-of-commission
(Ecs). When error-of-commission vehicles are sent to repair shops, no
emission control system malfunctions exist. Often, the repair shop finds that
the vehicle now passes the test without any changes. These false failures
waste resources, annoy vehicle owners, and may lead to emissions increases as
a result of unnecessary and possibly detrimental "repairs." Automobile
manufacturers see this as a significant problem, since it can contribute to
24
-------
customer dissatisfaction and increased warranty costs. An I/M program seeking
larger emission reductions through more stringent emission test standards may
actually increase the number of false failures. The error-of-commission rate
is, therefore, an important measure for evaluating the accuracy of I/M tests.
To see how an error-of-commission rate is calculated, assume an I/M area
which tests 1000 vehicles, of which 100 fail the I/M test, although only 50 of
those 100 failing vehicles also exceed the FTP standards. The error-of-
commission rate equals the number of vehicles that fail the I/M test while
passing the FTP divided by the total number of vehicles which were I/M tested:
50 vehicles failed I/M but passed FTP . ,AA _. „
100Q venlcles te3t@pd * 100 - 5% EC rate
As the error-of-commission rate decreases, vehicle owner satisfaction and
acceptance of the I/M program increases. Thus, while it is relatively easy to
improve the IDR by making the I/M test standards more stringent, this
"improvement" comes at the cost of potential increases in the error-of-
commission rate.
5.2.5.4 Two-Ways-To-Pass Criteria
The theory behind the two-ways-to-pass criteria for the IM240 is as
follows. Assuming that the IM240 test was correctly performed in the first
place, the most likely reason that a properly functioning vehicle would fail
an IM240 is that the evaporative canister was highly loaded with HC molecules
and that they were being purged into the engine during the test. This has
been a significant cause of false failures in existing I/M programs and it has
been shown that highly loaded canisters can cause both high HC and CO
emissions, even though the feedback fuel metering system is functioning
properly.
Since the canister is being purged during the IM240, the fuel vapor
concentration from the canister continually decreases during IM240 operation.
The decreasing fuel vapor concentration results in decreasing HC and CO
emissions. 3o^ emissions during Mode-2 (the last 136 seconds of the 239
second cycle}- should be lower than the composite results. On the other hand,
if the vehicle is actually malfunctioning, Mode-2 emissions should remain
high.
Catalyst temperature can also affect test outcome. Emissions are
generally highest after a cold start, before the catalyst has had a chance to
warm up. If a vehicle is standing in line for a prolonged period of time, or
was not sufficiently warmed up before arriving at the test lane, this can
25
-------
cause the vehicle to fail, when, in fact, it should be passed. Under the two-
ways-to-pass criteria, Mode-1 acts as a preconditioning mode, thus providing
insurance against this particular variety of false failure.
NOx outpoint criteria are not included in EPA's two-ways-to-pass
algorithm. So a vehicle which meets the IM240 NOx cutpoint (i.e., composite
NOx £2.0) only fails if both its composite emissions exceed the HC or CO
composite cutpoints, and its Mode-2 emissions exceed the HC or CO Mode-2
outpoints. In other words, a vehicle can pass by having low HC/CO emissions
in Mode-2 even if its Mode-1 HC/CO emissions were high. EPA is mandating this
approach to IM240 cutpoints.
The IM240 cutpoint tables, in Appendix E and Table 5.3.1 in the next
section, were calculated using the two-ways-to-pass-criteria.
The two-ways-to-pass criteria were optimized only at the cutpoints EPA
recommends for biennial enhanced Z/M programs, which are referred to as
"standard" or *recommended" IM240 cutpoints; For composite emissions, the
standard cutpoints are 0.80 g/mi HC, 15.0 g/mi CO and 2.0 g/mi NOx. The Mode-
2 criteria for the standard cutpoints are 0.50 g/mi HC and 12.0 g/mi CO. The
Mode-2 cutpoints were carefully selected from EPA's ZM240 data collected in
Indiana, to pass properly functioning vehicles while continuing to fail
malfunctioning vehicles. (The Mode-2 criteria were not redetermined for this
new Arizona sample.) The Mode-2 criteria, listed in the cutpoint tables in
Appendix F and Table 5.3.1, simply increase proportionally with increasing
composite cutpoints (i.e., become less stringent) and decrease proportionally
with decreasing composite cutpoints (i.e., become more stringent). The point
is that these Mode-2 criteria have not been optimized at every stringency
level to provide the best tradeoff among IDR, failure rate, and Ecs, so it is
probable that the effectiveness of the IM240 Mode-2 cutpoints can be improved.
5.2.5.5 Discrepant Failure* (DF«)
Discrepant failures are vehicles that fail an Z/M test for HC and/or CO
and pass th« FT? for HC and CO, but fail the FTP for NOz, or vice versa. The
table below illustrates one possible discrepant failure scenario:
Test
Short Test
FTP
HC or CO
Pass
Fail
NOx
Fail
Pass
Zn this example, a false failure for NOx happens to occur on a vehicle
which is a false pass for HC/CO.
26
-------
Repair diagnostic routines are frequently selected on the basis of which
pollutant caused the I/M test failure. Given that HC/CO and NOx move in
opposite directions with changes to the A/F ratio, there is not much reason to
expect that fixing a NOx problem will reduce HC/CO emissions. Therefore,
these scenarios represent an error of sorts for the short test . If a vehicle
was to fail the short test for NOx, whereas the only high FTP pollutant was
CO, chances are the mechanic will be looking for a problem that causes high
NOx. In this case, the problem that is causing high CO emissions is not
likely to be found.
5.2.5.6 Unproductive Failure (X7F) Rate
The unproductive failure rate represents the percentage of vehicles that
will be identified as needing repair, but either repair is not needed (Ecs),
or it is not likely the reason for repair will be found (DFs) . The
unproductive failure rate is calculated by adding errors-of-commission to
discrepant failures, and dividing the quantity by the total number of vehicles
which were I/M tested. Keeping with the same example as above, take an I/M
area which tests 1000 vehicles. 100 fail the I/M test, 50 of those 100
failing vehicles are Errors-of-Commission, and 5 are Discrepant Failures:
- »•» »• »«•
'Unproductive Failure
5.2.5.7 Vehicles with Malfunctions That Were Not
Counted as Bcs and DFs
Errors-of-commission in I/M programs have been most often caused by test-
to-test variability or incompatibility between the I/M test procedure and
vehicle emission control systems (e.g., air pump switching), so attempting to
repair EC vehicles were fruitless, with the IM240, however, EPA has found
that some vehicles that had failed the IM240 and passed the FTP actually did
have malfunctions, so they were correctly identified and air quality would
suffer by ignoring them. By the strict definition of Ecs, the IM240 is
penalized despite its successfully identifying malfunctioning vehicles.
A likely reason for vehicles passing the FTP despite a malfunction is
that malfunctions are sometimes intermittent. Vehicle 3172 provides a good
example. This vehicle had a number of IM240s performed, some with high NOx
and others with low NOx. The mechanic indicated that the vehicle had a sticky
EGR valve. The mechanic's diagnosis was not influenced by the FTP result
because the contractor had been instructed to report only IM240 scores to the
27
-------
mechanics, not the FTP score. This has become standard practice to allow the
contractor-repairs to simulate commercial repairs/ where mechanics will not
have access to FTP results.
For this analysis, EPA did not count vehicles as Ecs when they had a
malfunction that would logically explain the IM240 test failure. To
facilitate a fair comparison between the ASM and the IM240, the ASM failing
vehicles that passed the FTP, but had malfunctions, also were not counted as
Ecs when their malfunction would logically explain the ASM failure.
EPA was very conservative in that a vehicle was counted as an EC unless
the malfunction clearly explained the ASM or IM240 test failure. For example,
vehicle 3239 failed the IM240 with 2.4 g/mi NOx yet passed the FTP. The
vehicle was diagnosed as having a slow responding 02 sensor and it was
replaced. Because (1) a report of a slow-responding O2 sensor does not
indicate that objective criteria were used, (2) NOx failures are not strongly
associated with defective 02 sensors, and (3) all of its other IM240 tests
had passing NOx, the car waa counted as an EC despite the mechanic's judgement
the the O2 sensor should be replaced, which he did.
Using the similar logic, some vehicles with discrepant failures were also
not counted as DFs when their malfunctions could logically explain the I/M
test failure and a proper repair could be expected to reduce FTP emissions of
the affected pollutant even though FTP emissions of that pollutant were
initially below FTP standards. For example, the vacuum leaks on vehicle 3154
could cause a lean air/fuel ratio which can lower the catalyst's NOx
conversion efficiency and cause higher combustion temperatures, both of which
can cause high NOx on the ZM240 and ASM. FTP NOx emissions should also be
affected but perhaps not enough to cause an FTP failure. Because it is
logical for a mechanic to check for vacuum leaks on a car that fails NOx, and
this vehicle did have vacuum leaks, the I/M tests shouldn't be penalized for
correctly identifying the malfunction. On the other hand, if this vehicle had
failed CO on an I/M test and NOx on the FTP, the mechanic would look for
problems causing a rich air/fuel ratio, which would probably preclude looking
for vacuum leaks.
Table 5.2 lists the five vehicles the met the strict definitions for Ecs
or DFs, but were not counted for the reasons discussed. Note that while these
vehicles were not counted as Ecs or DFs In the outpoint tables, they still do
count toward the Failure Rate.
28
-------
Table 5.2.3.1:
Cara not Counted as Bca or DFa
Vehicle
Original Status
Malfunctions Explaining
I/M Test Failure
3154
3172
3200
3216
3244
Discrepant failure; failed
IM24Q and ASM NOx, but
failed FTP CO only.
Error-of-Commission/ IM240
NOx.
Discrepant Failure; failed
IM240 and ASM NOx, but
failed FTP HC only.
Injector seals leak at
intake manifold;
distributor advance vacuum
hose broken.
EGR valve sticks, EGR valve
vacuum line plugged.
EGR position sensor out of
range.
Error-of-Commission/ ASM HC ECM malfunction
Error-of-Commission/ IM240 Injector malfunctions
and ASM NOx. intermittently.
5.2.5.8 Weighting Factors to Correct Biased Recruiting
The criteria used to recruit vehicles for laboratory testing heavily
biased this laboratory sample toward IM240 failing vehicles. Sixty-two
percent of the 106 laboratory vehicles had failed the lane-IM240 criteria
(>0.80/15.0/2.0), whereas only 19% of 2,070 cars tested at the lane failed the
IM240. This resulted in a laboratory sample that was highly biased toward
failing vehicles. (Two-ways-to-pass criteria was not considered for
laboratory recruiting.)
Using this biased database results in unrealistically high excess emission
identification rates, and unrealistically low error-of-commission rates. So
the laboratory database must be corrected to represent the pass/fail vehicle
ratio in therin-use fleet to correctly determine IDRs, failure rates, and Ecs.
The database waa corrected using the weighting factors presented in Table
5.2.5.2.
Weighting factors are used as follows: If the 66 failing vehicles that
received FTP tests had excess HC emissions which totaled 100 g/mi, the
database would be corrected in this case by multiplying 100 by the 5.97
weighting factor, resulting in corrected total excess emissions of 597 g/mi
for the dirty vehicles. In comparison, the total excess emissions of the
29
-------
IM240-clean vehicles have to be multiplied by 41.9 to make their excess
emissions representative. The total simulated excess emissions are the sum of
the simulated excess emissions from the clean and dirty vehicles in the I/M
lane sample. The number of vehicles tested was similarly adjusted with the
factors for the purpose of calculating failure rates. The sample of 40 clean
vehicles provides confidence in conclusions about a test's relative tendency
to avoid failing clean cars.
Table 5.2.5.2
Weighting Factors Used To Adjust the Laboratory Database
IM240 at Lane * at Lane i at Lab Weighting Factor
Pass:
Fail:
£0.80/15.0/2.0
>0. 80/15. 0/2.0
1676
394
40
66
41.90
5.97
The resulting weighted database was used to produce the realistic
estimates of IDRs, failure rates, and Ecs that are listed as outpoint tables
in Appendices D & E. These cutpoint tables are sorted by failure rates. For
the outpoints that produce the same failure rate, the results are sorted first
by HC IDRs (in descending order) and then by NOx IDRs.
30
-------
5.3 Comparison of IM240 Versus ASM Using Outpoint Tables
In assessing the overall effectiveness of I/M test procedures, as
discussed in Section 5.2, it is important to determine the test's
effectiveness in terms of IDR, the failure rate, discrepant failures and
unproductive failure rate.
Appendices D and E list the same criteria for many different cutpoints.
Table 5.3.1 provides a summary of these criteria to compare the ASM with the
IM240 for the following three important scenarios:
- ASM cutpoints selected to achieve the same 18% failure rate (using
the cutpoint tables that are reweighted to correct the lab sample
bias) that result from EPA's recommended IM240 two-ways-to-pass
cutpoints of .80/15.0/2.0 + 0.50/12.0. Among the ASM cutpoint
combinations with this failure rate (see Appendix E), a combination
was selected that produced the maximum IDRs for all the pollutants
simultaneously, so there was no need to set priorities among
pollutants.
- ASM cutpoints selected to achieve IDRs similar to the IDRs that
result from EPA's recommended IM240 two-ways-to-pass cutpoints of
.80 / 15.0 / 2.0 + 0.50 / 12.0. Because ASM CO and NOx IDRs could
more favorably be presented by excluding HC, two ASM cutpoint sets
are presented, one to provide matching ASM and IM240 HC IDRs
(resulting in better IDRs for CO and NOx), and the second to provide
matching ASM and IM240 CO & NOx IDRs.
- ASM and IM24O cutpoints selected to achieve the highest IDRs possible
while keeping the unproductive failure rate below 5%. This case was
addressed on the possibility that an aggressive I/M program might be
willing to operate with such a high EC rate.
31
-------
Table 5.3.1
Comparison of the Ability of IM240 and ASM to Identify Vehicles
Whose Emissions Exceed Certification Standards Based on 106 Lab
Vehicles Weighted to Represent 1676 Lane Vehicles
Test
IM240
ASM
ASM
ASM
ASM
IM240
Excess Emissions
Failure Identified
Scenario Rate HC CO NOx
% % % %
Standard 18 92.2 67.S 83.4
Ctpu.
Same Fail 18 74.7 61.1 68.0
Rate
Similar HC 42 92.4 78.1 95.0
IDR
Similar CO & 24 80.4 66.2 89.4
NOx IDR*
Best IDR* 28 82.S 67.0 80.1
w/UF@<5%
Best IDR. 33 9S.9 78.2 97.1
w/UF@<5%
Weighted *
of Vehicles = 1676
Unproductive
Discrepant Failure
EC* Failures Rate** Cutpotnts
* # %
0 12 0.6 .80/15.0/2.0
+
0.50 / 12.0
42 6 2.3 1.00/8.0/2.0
174 180 17.1 1.00/8.0/1.0
84 48 6.4 1.00/11.0/1.4
48 48 4.6 .40/8.0/1.5
60 12 3.5 JO/ 9.0/1.7 +
.19 / 7.0
* Excludes Be vehicles that had malfunctions that caused an Z/M test failure,
but because they were intermittent malfunctions, did not fail the FTP. FTPa
were always performed on a different day. Since they were correctly
identified by the I/M test, they are not vehicles that will "ping-pong".
** The Unproductive Failure Rate includes the traditional EC vehicles and the
discrepant failures, without including the traditional EC vehicles that were
found to have intermittent malfunctions that were not identified by the FTP
test.
32
-------
For the first scenario with an 18% failure rate for both tests, the ASM
statistics in Table 5.3.1 show that the IM240 identifies 18 percentage points
more of the excess HC emissions and 15 percentage points more of the excess
NOx emissions than the ASM identifies, with a significantly lower unproductive
failure rate. Expressed differently, an IM240-based program would relinquish
about 19% of its HC effectiveness and 18.5% of its NOx effectiveness by
substituting the ASM test at the same failure rate. Some relatively dirty
vehicles are missed by the ASM and replaced by relatively clean vehicles.
This scenario is illustrated in Figure 5.3.1.
The second scenario shows that in order to achieve HC IDRs similar to the
IM240's at an 18% failure rate, the ASM'3 failure rate must be increased to
42%, resulting in an unacceptable EC rate of 17%. To achieve similar CO and
NOx IDRs with the ASM, an ASM failure rate of 24% is necessary, and that also
results in an unacceptable unproductive failure rate of 6.4%. This scenario
is illustrated in Figures 5.3.2 and 5.3.3.
The last scenario compares the tests at the maximum IDRs achievable while
limiting the unproductive failure rate to less than five percent. Again, the
IM240 IDRs are significantly higher than the ASM's, with a lower unproductive
failure rate. The IM240 HC ZDR is 14% higher, the CO IDR is 14.3% higher, and
the NOx ZDR is 17.5% higher, with an EC rate that is 1% lower. Expressed
differently, an aggressive IM240-based program with a 3.5% unproductive
failure rate would relinquish about 14% of its HC and 17.5% of its NOx
effectiveness by substituting the ASM test at at an even higher unproductive
failure rate. This scenario is illustrated in Figure 5.3.4.
These statistics indicate that the ASM test is significantly less
effective than the IM240 as an I/M test.
The second scenario, wherein the ASM's HC IDR is raised to match the
IM240's HC IDR of 92%, is anticipated to raise the following question:
Why didn't EPA make the ASM's HC cutpoint more stringent to increase
the ASM's IDR without increasing the stringency of the ASM's NOx cutpoint,
thereby allowing a lower ASM failure rate?
The answ«r^±jr that eight vehicles (see Table 5.3.2) have a major impact on
the ASM HC IDR, but their ASM HC scores are less than 0.3 g/mi. Although
their ASM HC scores are very low, they account for roughly 10.5% of the total
excess FTP HC emissions. These eight vehicles also have ASM CO scores below
8.0 g/mi. While developing the ASM cutpoint tables, EPA found that ASM
outpoints below 0.3/8.0 caused failure rates and Ecs to increase excessively,
so the final cutpoint tables did not include tighter outpoints. So to achieve
33
-------
the IM240's HC IDR, the only "practical" way to identify these cars ia through
the NOx outpoint.
Five of the eight cars with high FTP HC that pass the ASM HC outpoints are
failed by a NOx cutpoint of 1.5 or less. These five cars account for 7.6% of
the total excess HC emissions. So the 1.0 g/mi ASM NOx cutpoint achieves an
HC IDR comparable to the 92.2% achieved by the IM240 at EPA's standard
cutpoints.
Table 5.3.3 summarizes the ASM cutpoint table in Appendix E to show that
the only way for the ASM to achieve the IM240's HC IDR of 92.2% at the
recommended cutpoints for biennial programs is to lower the ASM NOx cutpoint
to 1.0.
Table 5.3.2: Vehicles that Pass a
0.30 g/mi ASM HC Cutpoint While Failing FTP HC
Ve hicle HC FTP CO FTP NOx FTP HC ASM CO ASM NOx ASM
3180
3192
3195
3199
3201
3254
3257
3259
0.96
0.49
0.51
0.53
0.94
1.87
1.26
1.94
9.75
6.31
5.80
10.90
19.73
35.87
8.57
14.95
1.22
0.53
0.66
1.53
1.72
1.16
0.90
0.53
0.15
0.20
0.18
0.29
0.17
0.29
0.25
0.23
3.64
6.42
3.26
3.89
4.73
7.37
4.65
4.61
0.68
0.92
1.27
1.53
1.22
1.13
1.03
1.22
Table 5.3.3: Alternative ASM Cutpoints For High HC IDRs
ASM
Outpoints 1
0.30 /
0.30 /
0.30 /
0.30 /
0.30 /
0.30 /
0.30 /
0.40 /
1.00 /
1.00 /
1.00 /
8.0 / 2.0
8.0 / l.a
8.0 / 1.3
8.0 / 1.4
8.0 / 1.3
8.0 / 1.2
8.0 / 1.0
8.0 / 1.0
8.0 / 1.2
9.0 / 1.0
8.0 / 1.0
Identification
HC I CO
88.0%
88.2%
89.0%
90.3%
90.3%
91.4%
96.6%
92.4%
84.1%
91.3%
92.4%
69.2%
69.3%
71.0%
75.1%
75.2%
76.4%
82.1%
78.8%
71.9%
74.9%
78.1%
Rates
1 NOx
74.9%
78.2%
82.7%
89.5%
89.5%
89.9%
95.0%
95.0%
89.8%
95.0%
95.0%
Failure
Rate
29%
30%
33%
38%
40%
42%
48%
45%
32%
40%
42%
EC Discrepant UF
1 Rate* | Failures I Rate I
4.3%
4.3%
4.3%
6.4%
6.4%
6.4%
8.7%
8.7%
4.0%
8.4%
8.4%
0
0
42
42
84
126
132
138
132
221
180
4.3%
4.3%
6.4%
8.4%
10.4%
12.4%
15.0%
15.3%
10.4%
19.1%
17.1%
34
-------
To achieve an HC IDR rate greater than 89% a NOx outpoint of less than 1.5
is necessary. To achieve an HC IDR rate greater than 91.4% a NOx cutpoint of
less than 1.2 is necessary, and to achieve an HC IDR rate greater than 92% a
NOx cutpoint of 1.0 is necessary. Once a tight NOx cutpoint is used to fail
these cars with excess HC, the HC cutpoint no longer determines the result, at
least in this sample. So, ASM outpoints of 1.00/8.0/1.0 are the least
stringent ASM cutpoints that can achieve a 92% HC IDR.
Another consideration is that the ASM cutpoints have been optimized for
this database. In contrast, the IM240 recommended cutpoints were optimized
for the Indiana IM240 database. Because of sample to sample differences, the
optimum cutpoints are expected to vary slightly from one database to another.
So the optimum ASM cutpoints are being compared to standard IM240 cutpoints,
which while optimum for the Indiana data, are not the optimum cutpoints for
this database. Applying ASM cutpoints optimized for this data base, to a
different database, is expected to further lower the ASM's performance.
35
-------
Figure 53.1
Comparison of ASM to IM240
Using IM240 Standard Outpoints
With Maximum ASM IDRs at Equivalent Failure Rates
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
83%
Note Higher
ASM UF Rate
HCIDR
COIDR
NOxIDR Failure Rate Unproductive
Failures
BIM24O at Recommended Cutpoints D ASM at Same Failure Rate
36
-------
Figure 5.3 J
Comparison of ASM to IM240
With ASM CO & NOx IDRs Matching EV1240 IDRs
100% -r-
90%
80%
89%
10%
0%
HCIDR COIDR
NOx IDR Failure Rate Unproductive
Failures
BIM24O at Recommended Outpoints D ASM at Matching CO & NOx IDRs
38
-------
Figure 5.3.4
Comparison of ASM & IM240
Maximum IDRs @ EC <5%
100% -r
90% - J
80%
70%
96%
97%
HCIDR
COIDR NOxIDR
Failure Rate Unproductive
Failures
HIM240 Best IDRs @<5% EC D ASM Best IDRs @ <5% EC
39
-------
5.4 Comparison Using Scatter Plots and Regression Tables
The objective of this analysis was to check the correlation of both the
IM240 and the ASM test with the FTP. The correlations are illustrated in
Figures F-l through F-9, in Appendix F. Appendix F includes regression tables
along with these scatterplots. The regressions show similar R2 over the
entire data range, but the IM240 correlates much better to the FTP for
vehicles emitting closer to the FTP standards.
5.4.1 Using the Coefficient of Determination (R2 ) and
Standard Error of the Estimate for Objective
Comparisons
R2 represents the percentage of variability in the dependent variable
(FTP result) that is explained by the independent variable (I/M test result)
and is often used to compare one I/M test's effectiveness with another's, but
R2 can often be misleading. Since R2 is often used in correlation studies,
it does provide an indication of comparative test accuracy that would be of
interest to readers accustomed to seeing such comparisons. More important,
however, is how well these tests discriminate between malfunctioning and
properly functioning vehicles at an I/M station, which is best measured using
the techniques discussed in Section 5.2.
For a vehicle to fail an IM240, it must fail the two-ways-to-pass-criteria
developed by EPA (see Section 5.2.2). The R2 values presented in this section
are for composite IM240 scores only and do not account for this. Two-ways-to-
pass affects the quantitative correlation between IM240 and FTP significantly
because the Mode-2 HC and CO values are often more representative of vehicles'
actual FTP emissions. However, EPA believes it is not appropriate to mix and
match composite and Mode-2 scores into one quantitative correlation analysis.
Additionally, the R2 comparisons presented here do not account for the
sample's bias toward high emitters (discussed in Section 5.2.5.8). The 106
vehicles that were recruited to the lab for FTP testing were purposely biased
to include a high number of dirty vehicles. When regressing the I/M test
scores versus the FTP to determine R2 values, these high emission values
disproportionately influence some regression statistics, given the typical
distribution of in-use emissions data. Thus the emission values close to the
FTP standards (where comparing I/M tests is most important), have less
influence on the R2 statistic than desirable for determining the actual
merits of these tests. Cutpoint tables account for this sample bias by
weighting each vehicles' emissions according to the population of vehicles
tested at the I/M lane.
40
-------
To account for these limitations the sample was divided into the following
three groups:
• All Vehicles. This database is not very useful for comparing
correlation because the cleanest and dirtiest vehicles dominate the
R2 statistic. Both tests correctly differentiate these. More
pertinent are the vehicles with emissions closer to the FTP standard,
where the capability of I/M tests is not masked by the very clean and
very dirty vehicles. Also, vehicle 3211 is a CO outlier for both
tests. It has a major effect on the regression equation and the R2 ,
thus masking the typical capability for both procedures.
• Vehicle 3211 Removed for HC, CO, NOx. This database better
characterizes the correlation of both short tests with the FTP, but
for the reasons discussed, it is not the most relevant for comparing
the effectiveness of the tests.
• Marginal Emitters: Only vehicles that are not very clean or not
very dirty on FTP using following criteria:
HC 20.30 g/mi and <1.5 g/mi on the FTP
CO 22.5 g/mi and <25.0 g/mi on the FTP
NOx 20.5 g/mi and <2.25 g/mi on the FTP
Also, Vehicle 3211 was excluded as an outlier.
All vehicles with FTP emissions less than 0.30 HC, 2.5 CO, and 0.5
NOx passed the ASM and IM240 tests, for all the cutpoint sets
evaluated in Section 5.3.
The standard error is an objective measurement of test variability
expressed in the units (g/mi. in this case) of the variables used in the
regression. Because R2 are expressed as percents, standard errors have an
advantage of being less abstract.
Table 5.4.1 provides a summary of R2 and standard errors for Figures F-l
through F-9 in Appendix F, divided into the 3 groups just discussed. The
"Marginal Emitters" group indicates that the R2 for the IM240 are
considerably higher for HC and NOx, and somewhat higher for CO. Likewise, all
the standard errors are lower for the IM240, most notably for HC.
41
-------
Table 5.4.1 Summary of R2 and Standard Brrors
Data Set :
n :
Procedure :
R2 -
HC
Std. Error =
R* =
CO
Std. Error »
R2 =
NOx
Std. Error =
All Vehicles
106
IM240
82%
0.62
54%
13.4
70%
0.65
ASM
73%
0.76
68%
11.2
71%
0.64
Vehicle 3211
Removed
105
IM240
83%
0.61
75%
10.0
70%
0.66
ASM
74%
0.75
80%
8.9
71%
0.64
Vehicles Near
Standards
43
IM240
63%
0.19
25%
4.3
46%
0.34
ASM
17%
0.28
13%
4.6
26%
0.39
The standard errors listed in Table 5.4.1 can be used to estimate the
lowest FTP value that would confidently predict a dirty vehicle. For example,
the HC standard error is 0.28 g/mi for the "Marginal Emitters" group. Since
95% of the time, a vehicle's result will be within ±2 standard errors, this
suggests that the lowest ASM HC score that confidently predicts an HC-dirty
vehicle (i.e., FTP HC > 0.41) is the ASM HC score that yields (using the
regression equation) an FTP HC of 0.97 g/mi [0.41 + (2 * 0.28)]. In contrast,
using the IM240 error of 0.19 g/mi means the lowest IM240 score that
confidently predicts a HC-dirty vehicle is 0.79 g/mi, over 18% less than the
score needed to confidently predict an ASM HC dirty vehicle.
5.4.2 Advantage of Using Weighted Modes
The ASM teat; is given a big advantage in the way the regressions are
performed because each mode is weighted separately according to the IM240. On
the other hand, the ZM240 score is a non-weighted score. EPA developed the
IM240 to contain similar driving conditions as the FTP. However, the
frequency of each condition is not proportional to the FTP. By weighting
different modes of the IM240 to the FTP similar to the way EPA has weighted
the 4 modes of the ASM test, EPA has found the R2 for the ZM240 to improve
immensely. The current score reported for the IM240 is something like
42
-------
weighting each mode of the ASM test 25%. This would hurt the correlation of
the ASM with the FTP, because, as is shown in Section 5.5, the 50 mph mode
accounts for roughly half of the composite ASM scores for each pollutant.
5.4.3 Observations of Scatterplots
The scatterplots for the first two sets of data (Figures F-l through F-6)
do not appear much different for either test, mainly because the high emitters
cause the emissions close to the standards to appear as a tight pack of data.
The plots for vehicles near the standard only (Figures F-7 through F-9),
however, suggest the following:
HC - The IM240 identifies the dirtier cars much better. Notice on the ASM
HC plot how many high emitters (FTP HO0.82 g/mi) still score
relatively low on the composite ASM score. Six vehicles pass the
very tight ASM HC cutpoint of 0.3 predicted g/mi, yet have FTP
emissions greater than twice the FTP standard (0.82 g/mi).
CO - Neither test appears to correlate very well over this emission range
for CO. Two issues come into play that explain why this is. First,
cars with loaded canisters will have high IM240 Mode-1 CO emissions
at the lane, causing the short test to have a high score while the
FTP at the lab is relatively low. The second scenario is cold start
problems. Two vehicles in the database (Vehicles 3175 and 3227)
appear to have cold start problems, with high Mode-1 FTP CO
emissions, and low Bag-3 FTP CO emissions. Since the lane test is a
hot start test, these vehicles will show up clean at the lane, and
the cold start FTPs will be significantly dirtier.
NOx - The IM240 has a slightly tighter fit to the regression line, and more
of the FTP dirty cars fall to the upper right of the scatterplot
(i.e., fail the test properly).
5.4.4 Poor ASM BC Correlation
As discussed in Section 5.3, ASM HC scores do not correlate very well with
FTP HC scores. This section briefly discusses theoretically why some of these
vehicles had very low ASM HC emissions, yet failed the IM240 and FTP for HC
emissions. Because the contractor's mechanics were not aware of the ASM
scores, vehicles were not diagnosed with the objective of determining the
cause of the performance differences on these Z/M tests.
The first four vehicles in Table 5.4.2 were found to have ignition
problems. This is logical considering that misfire, which causes high HC, is
43
-------
sometimes related to load. As load increases the voltage required to jump the
spark plug gap also increases. Some portions of the IM240 load the vehicle
more heavily than any of the ASM modes, so a marginal ignition system that
only causes misfire during the higher IM240 loads will not be identified by
ASM HC.
Vehicles 3180 and 3264 were found to have bad 02 sensors and other
malfunctions. Slow responding O2 sensors are more likely to be identified
during a transient test, because changing throttle position tends to cause
air/fuel ratio excursions that will cause high emissions unless the fuel
induction system rapidly compensates to maintain the optimum air/fuel ratio.
So slow responding 02 sensor might explain the high HC on the IM240 and low HC
on the ASM.
The disconnected vacuum lines on vehicle 3201 could have caused lean-
misfire during accelerations on the IM240 that would not be apparent on the
steady-state modes of the ASM.
These explanations can not be proven with the existing data, but they
should indicate that a steady-state test suffers from known disadvantages in
identifying vehicles with these types of malfunctions.
Table 5.4.2 Vehicles with Poor ASM BC Correlation
VBH ASM HC
FTP HC IM240 HC Problem Found
3259
3257
3155
3210
0.23
0.25
0.34
0.35
1.94
1.26
3.25
1.40
3180
3264
3254
3201
3165
0.15
0.49
0.29
0.17
0.39
0.96
1.36
1.87
0.94
1.96
1.50 Ignition Module
1.92 Plug Wires, Plugs Transducer,
Ignition Coil Transistor
2.77 Incorrect Plugs, Torn wire boot
1.04 02 Sensor, Spark Plugs, Fuel Hose,
Catalyst
1.33 O2 Sensor and Injectors
2.16 02 Sensor, Vacuum Switching Valve
2.26 ECU Intermittent, Catalyst
1.15 Vacuum Lines Disconnected
1.59 Dirty Injectors
44
-------
5.5 Derivation of ASM Coefficients and Mod* Contribution
Variations From Sample to Sample
This section discusses why the ASM coefficients that EPA based its
analyses on were developed using the IM240 as the dependent variable rather
than the FTP. BAR/Sierra and ARCO used the FTP to develop ASM Coefficients.
Also discussed are the rather large variations in the ASM coefficients
with different samples, and the variation in the contribution of each ASM mode
to the final ASM score (expressed as percent contribution).
5.5.1 ASM Versus ZM240 As The Dependent Variable For
Determining ASM Coefficients
EPA faced a dilemma in determining the best method for developing the ASM
equation coefficients. No ASM advocate has recommended specific coefficients,
on which, EPA should accept or reject the ASM approach. So two options were
to: 1) perform the multiple regressions on all the lab recruited vehicles for
ASM versus FTP. Or, 2) perform the multiple regressions for ASM versus IM240
on a subset of the lane sample, excluding all lab recruited cars.
Obviously, the ideal method is to regress the ASMs versus FTPs (i.e.,
option 1), but this raises a problem.' To evaluate how well the ASM identifies
FTP failing vehicles, the coefficients must be applied to the lab recruited
vehicles to calculate simulated grams/mile scores. However, good statistical
practice mandates applying the coefficients to a different sample than those
from which they were developed.
This interlinking method, wherein the coefficients are applied to the same
vehicles from which they were developed, would minimize the effects of the
test's variability. This improper interlinking is illustrated using results
from Vehicle 3211. This vehicle's lane-IM240 CO score was 93 g/mi and its ASM
CO score was 65 g/mi using the coefficients from the 608 vehicle sample listed
in Table 5.5.1 (The relevance of the other samples in this table will be
discussed latex). Its FTP CO score was only 10.8 g/mi.
45
-------
Table 5.5.1
CO Coefficients Developed from Different Samples
Mode
Constant
5015
2525
50MPH
Idle
Adjusted R2
CO Sample 1
2.814
0.035
0.072
0.425
0.891
34.2%
CO Sample 2
2.836
0.116
-0.058
0.391
2.014
58.4%
CO All 608
2.936
0.040
0.043
0.356
1.350
50.1%
COvsFTP
5.533
-0.047
0.565
0.050
1.968
80.6%
The scatter plots below show that using the IM240-developed coefficients
cause this vehicle to be easily identified as an outlier. In marked contrast,
using the FTP-developed coefficients make it look like this vehicle's ASM
score highly correlates to its FTP score. The ASM mode scores are weighted
differently, so the high scoring mode(s) are de-emphasized. But these same-
sample FTP-based ASM coefficients are obviously peculiar to this sample, and
highly dependent on it containing this one particular car. (See Tables 5.5.1
and 5.5.6.)
Still not answered is which ASM coefficients better indicate whether the
vehicle is malfunctioning or not. Some could argue that this vehicle should
not be an outlier. Instead, the IM240-developed coefficients inappropriately
make it appear as an outlier. Attempting to resolve this, the raw ASM
concentration measurements were checked. This vehicle's 50 mph mode CO
concentration was 4.96%, which is higher than 97% of vehicles recruited to the
lab (103 of the 106 vehicles).
Using the same-sample, FTP-based ASM coefficients prevents this vehicle
from being an outlier because they adjust themselves to minimize the effect of
the 50 mph mode score from all cars. Additionally, the very high concentration
measurement (4.96%) proves that the vehicle had a malfunction causing very
high emissions that had been inappropriately minimized. (This was also
verified during the mechanic's inspection which found a defective O2 sensor
and that an ECM PROM update was required.) This evidence strongly supports
EPA's properly uaing the preconditioned IM240s as the dependent variable for
developing ASM"coefficients to compare the ASM and IM240 correlations with the
FTP.
This evidence also casts doubt on conclusions developed from test programs
that used interlinked coefficients. Interlinking makes the correlation
between the ASM, or any other test, and the FTP significantly better than
could be expected in an official I/M program. Since I/M programs will apply
ASM coefficients to vehicles that were never FTP tested, the opportunity for
46
-------
interlinking will not exist, so ASM performance should not be evaluated using
interlinked coefficients.
Another reason for EPA's not using FTP-based coefficients is because some
are negative, which means that as ASMS015 emissions increase, FTP emissions
decrease. This is counter-intuitive.
47
-------
ASM Coefficients Developed
from
608 Lane ASM vs Pre-Conditioned IM240s
120 -r
100 ••
80 ••
60 ••
40 ••
20 ••
20
40 60
ASM CO Score
Vehicle 3211
80
100
ASM Coefficients Developed
from
106 Lane ASM vs Lab FTP
8
48
-------
EPA decided that the best method, using the Arizona data, was to regress
the four steady-state modes (three for NOx) against lane-only, preconditioned
IM240s. There were three major factors leading to this decision. First, this
allowed applying the coefficients to a different subset of data (the lab
recruited vehicles). Second, the sample size was considerably larger (608 vs.
106 tests) . EPA'3 FTP sample was too small to divide and use half to
determine coefficients and the other half to evaluate ASM effectiveness, which
is supported by the negative coefficient yielded for the ASM5015 listed in
Table 5.5.1. Third, only preconditioned IM240s were used because they
correlate better with the FTP than non-preconditioned IM240s. The only
significant compromises in using IM240s instead of FTPs is that the composite
ASM score does not include a cold start excess (which would be independent of
warmed-up ASM mode concentrations anyway) and that the mix of speeds and loads
in IM240 is not exactly like that in the full FTP driving cycle (a hardship
borne by the IM240 in its own correlation to the FTP).
Figure 5.5.1 illustrates that preconditioned IM240s strongly correlate
with the FTP. These data are from the 106 lab recruited vehicles, but are
restricted to IM240s that were performed following the ASM at the lane, making
them preconditioned IM240s.
NOx has the worst correlation because of a few outliers at the high end,
but this is not a concern for the ASM since the NOx coefficients are
relatively stable, which is discussed in the next section.
49
-------
5-S.I High Correlation of Preeondit-innod Lane—IM24Qa with FTPa
IM240vsFTP
Preconditioned IM240s
0.0
0.0
0.5
•
1.0
1.5
2.0
IS
3.0
3.5
I
4.0
4.5
IM240HC
IM240vsFTP
Preconditioned IM240s
10
20
30
40
50
60
70
80
IM240CO
IM240vsFTP
Preconditioned IM240s
0.0
50
-------
5.5.2 Variability of ASM Coefficients
The objective of the following analysis was to investigate the stability
of the coefficients used to calculate composite ASM simulated grams/mile
scores. The database was divided into four samples for comparison.
Sample 1 was developed by using a random number generator to select 304
vehicles from the lane-only fleet of 608. The remaining 304 vehicles became
Sample 2. Sample 3 was all 608 vehicles, and the fourth sample was the 106
laboratory vehicles. The ASM coefficients for the first three samples were
developed using the IM240 as the dependent variable and the FTP sample used
the FTP for the dependent variable. The resulting coefficients are listed in
the following tables. (Table 5.5.3 is a duplicate of Table 5.5.1.).
Table 5.5.2
HC Coefficients Developed from Different Samples
Mode
Constant
5015
2525
50MPH
Idle
Adjusted R2
HC Sample 1
0.080
0.045
0.047
0.147
0.084
21.7%
HC Sample 2
0.073
0.008
0.059
0.123
0.585
38.9%
HC All 608
0.083
0.025
0.059
0.136
0.124
29.0%
HCvsFTP
0.291
-0261
0.507
0.238
0.154
79.4%
Table 5.5.3
CO Coefficients Developed from Different Samples
Mode
Constant
5015
2525
50MPH
Idle
Adjusted R2
CO Sample 1
2.814
0.035
: 0.072
--'• 0.425
0.891
34.2%
CO Sample 2
2.836
0.116
•0.058
0.391
2.014
58.4%
CO All 608
2.936
0.040
0.043
0.356
1.350
50.1%
COvsFTP
5.533
•0.047
0.565
0.050
1.968
80.6%
51
-------
Tabla 5.5.4
NO* Coefficients Developed from Different Samples
Mode
Constant
5015
2525
50MPH
Adjusted R2
NOX Sample 1
0.230
0.088
0.206
0.386
60.2%
NOX Sample 2
0.279
0.045
0.212
0.333
57.9%
NOX All 608
0.258
0.061
0.219
0.352
59.1%
NOX vs FTP
0.190
0.148
0.093
0.291
71.1%
The negative coefficients are highlighted in bold. One could infer from
the negative coefficients that increasing the emissions during that mode of
the ASM would lower the composite score.
These coefficients were used with the ASM data, from each of the 106 lab-
recruited vehicles, to calculate the emissions for each mode and the percent
of the total emissions that each mode contributed. These mode contributions
give a better indication of each modes importance in the final ASM score, than
the coefficients, which are more difficult to interpret. The results are
listed in the following tables:
Tabla 5.5.5
Average Contribution of Total HC Emissions by Mod*
Mode
Constant
5015
2525
50MPH
Idle
HC Sample 1
17%
20%
17%
43%
2%
HC Sample 2
17%
4%
22%
39%
18%
HC All 608
19%
12%
22%
43%
4%
HCvsFTP
11%
-75%
116%
45%
3%
52
-------
Table 5.5.6
Average Contribution of Total CO Emissions by Mode
Mode
Constant
5015
2525
50MPH
Idle
CO Sample 1
26%
4%
7%
57%
6%
CO Sample 2
27%
12%
-6%
53%
15%
CO All 608
29%
4%
4%
51%
11%
COvsFTP
28%
-5%
55%
7%
15%
Table 5.5.7
Average Contribution of Total NOx Emissions by Mode
Mode
Constant
5015
2525
50MPH
NOX Sample 1
15%
11%
24%
50%
NOX Sample 2
20%
6%
27%
47%
NOX All 608
17%
8%
27%
48%
NOX vs FTP
20%
22%
13%
45%
The HC coefficients in particular are very volatile, and that the negative
FTP-developed coefficients are counter-intuitive. When applying the
coefficients from Sample 1, the idle mode, on average, only contributes 2% to
the total score. This contribution jumps to 18% when the coefficients from
Sample 2 are applied. Similarly the ASM5015 contribution drops from 20% to
4%. These examples indicate that the largest sample (608 vehicles) with
preconditioned IM240s was the best sample available for developing ASM
coefficients.
5.5.3 Significance of Mode Contributions
The ASM node contributions also vary as the composite ASM score moves from
low values (fo* which the constant term will be the primary contributor to the
composite score) to relatively high values (for which the constant term will
be a relatively small contributor to the composite emission). This is
illustrated in Figures 5.5.2 to 5.5.4. For CO, the ASM5015 and ASM2525 are
combined, because of the negative contributions of 2525 and the small
contribution of the ASMS015 in relation to the 50 mph mode.
53
-------
Figure 5.5.2
Mod* Contributions for HC
Using Sample 1 Coefficients
100% -r
80% ••
60% ••
40% ••
20% -•
0% •-
Constant • Idle DsOMPH • 2525
5015
FTP HC S
0.82 < HC S 1.64
FTP HC > 1.64
Using Sample 2 Coefficients
I Constant • Idle
50 MPH • 2525
El 5015
FTP HC £
0.82 < HC £ 1.64
FTP HC > 1.64
54
-------
5.5.3
Moda Contributions for CO
Using Sample 1 Coefficients
FTP CO s 6.80
6.80 < CO S 13.60
13.6 < CO s ISO
Using Sample 2 Coefficients
FTP CO s 6.80
6.80 < CO s 13.60
55
-------
Figure 5.5.4
Moda Contributions for NOx
100% j
80% • •
60% ••
40% ••
20% ••
0% -I-
Using Sample 1 Coefficients
(Constant DSOMPH B 2525
5015
0.60
0.6 4.00
Using Sample 2 Coefficients
100% T
(Constant D 50 MPH • 2525 U 5015
.
0.6 4.00
56
-------
The fact that the 50 mph mode contributes so much to the composite score
for each pollutant is also reason for concern. This opens the opportunity for
mechanics to adjust vehicles to lower emissions for just one mode (namely the
50 mph), which will be further discussed in Sections 5.6.3 and 5.6.4.
5.5.4 Conclusions on ASM Mode Contributions
While not a mode that was recommended by the ASM developers, the 50 mph
mode at road load horsepower appears to be more important for identifying
dirty vehicles than the lower speed, acceleration simulation modes (ASM5015
and ASM2525). Surprising was the small contribution of the ASM5015 (mode 1)
for identifying dirty vehicles considering that BAR/Sierra and ARCO both found
this mode to be the most effective. This suggests that the first mode in a
four mode sequence serves mainly to precondition vehicles for the following
modes. Randomizing the order of the modes may be useful in determining the
best sequence.
For the cutpoint analysis in Section 5.3 and the regression analysis in
Section 5.4, the ASM scores used were those calculated from the coefficients
developed from the 608 ASM versus preconditioned lane-IM240s. However, the
variability of the HC coefficients between the two random subsets of 304 tests
suggest that a different sample of 608 tests might produce substantially
different equation coefficients. The resulting change in HC (and in some
cases CO and NOx) ASM scores would produce different failure rates, IDRs, and
EC rates in the cutpoint tables, and different R2 values in the regressions
of ASM versus FTP. So the volatile coefficients may vary from sample to
sample, or worse yet region to region, resulting in disparate I/M programs
which would be hard to evaluate on a consistent basis.
5.6 Repair Analyses
5.6.1 Contractor Repairs
The objective of thia analysis was to investigate the performance of both
the IM240 and the ASM tests as predictors of changes (i.e., decreases or
increases) in FTP emissions following contractor-performed, IM240-targeted
repairs.
Of the 106 vehicles used in the cutpoint analysis (Table 4.2.2), 56
exceeded the lane-IM240 0.80/15.0/2.0 + 0.50/12.0 cutpoint and were repaired
by the contractor. Of these, 52 received each of the three following tests
both prior to repairs (i.e., as-received) and following repairs:
57
-------
- a Lane-IM240,
- a Lane-ASM, and
an FTP.
These 52 were used in this analysis and are included with the data listed in
Appendix B. The resulting database of those 52 fuel-injected vehicles has the
following distribution:
Fuel
Maturing
PFI
TBI
Order of
ASM Prior
to XM240
14
11
Lane Testing
IM240 Prior
to ASM
16
11
The contractor was instructed to perform the minimum repairs necessary in
order that each vehicle's IM240 emissions after repair (as tested at the
contractor's laboratory) meet the following criteria:
- composite IM240 HC S 0.80 g/mi,
- composite IM240 CO £ 15.00 g/mi, and
- composite IM240 NOx £ 2.0 g/mi.
The contractor was allowed multiple repair attempts if the first set of
repairs did not reduce the IM240 emission levels enough. The repairs were
limited to $1,000 per car. And, the contractor was instructed that "the
mechanic should only be aware of the IM240 scores for the IM240-targeted
repairs." Because ASM outpoints, that could distinguish malfunctioning
vehicles from properly functioning vehicles, were not yet developed, only
IM240-targeted repairs were performed.
These IM240 emission repair criteria were met at the contractor's
laboratory foe all cars prior to the second and final FTP, with the highest
after-repair laboratory ZM240 composite HC emission score of 0.56, CO of
10.82, and NOar of 1.93 (g/mi). The effects of those IM240-targeted repairs on
FTP emissions are illustrated in the following table:
58
-------
Table 5.6.1.1
FTP Emissions Prior to and Following
XM240-Targeted Repairs
FTP Emissions
HC As-Received
After Repair
CO As-Received
After Repair
NOx As -Received
After Repair
Mean
1.458
0.326
19.707
3.331
1.649
0.739
Rang<
Emiai
Minimum
0.16
0.10
0.28
0.63
0.20
0.05
a of
lions
Maximum
13.07
0.75
113.40
8.82
7.56
1.81
The resulting FTP emissions after the IM240-targeted repairs were essentially
independent of the as-received FTP emissions. (That is, the R-squares
associated with before and after HC, CO, and NOX were only 0.1%, 1.2%, and
1.0%, respectively.)
The data from these repaired vehicles can give insight into the question
of whether the IM240 test and outpoints cause repairs to be made which also
reduce FTP emissions. In other words, does the IM240 and the FTP correlate
well on a single vehicle? This correlation is to be expected based on the
realistic nature of the IM240 driving cycle, and the good correlation found in
samples of vehicles not repaired.
For each of those 52 vehicles (all 1983 and newer fuel-injected cars), the
change in each pollutant (HC, CO, and NOx), following contractor repairs, was
calculated for each of those three test cycles. Regressing the reductions in
the lane emissions against the reductions in FTP emissions produced Tables
5.6.1.2 and 5,.6.1.3. The six graphs (Figures 5.6.1.1 through 5.6.1.3) that
follow those regression tables illustrate the results of this analysis.
59
-------
•Pablo 5.6.1.2
Regression of Chang** in Lane-XM240 Bmiaaiona Following
Contractor Repairs Versus Corresponding Changes in FTP Bmiaaiona
Dependent variable
R2 =81.9%
is: A(FTP HC)
s » 0.8320 with 52-2-50 degrees of freedom
Source
Regression
Residual
Variable
Constant
AUM240 HC)
Sum off Squares df
156.693 1
34.611 50
Coefficient s.e. off Coefff
-0.365173 0.1524
1.4106 0.0938
Mean Square
157
0.69222
t-ratlo
-2.4
15
F-ratio
226
Dependent variable is: A(FTP CO)
R2 -47.5%
S- 18.50 with 52
Source
Regression
Residual
Variable
Constant
AHM240 CO)
- 2 - 50 degrees of freedom
Sum of Squares df
15469.4 1
17110.1 50
Coefficient s.e. off Coefff
4.64057 3.103
0.846373 0.1259
Mean Square
15469
342.203
t-ratio
1.5
6.72
F-ratlo
45.2
Dependent variable is: A(FTP NOx)
R2 -64.5%
S- 0.8846 with
Sou re* -r
Regression
Residual
Variable
Constant
AOM240 NOx)
52-2-50 degrees of freedom
Sum of Squares df
71.1008 1
39.1265 50
Coefficient s.e. of Coefff
-0.275563 0.1747
0.738523 0.0775
Mean Square
71.1
0.78253
t-ratlo
-1.58
9.53
F-ratlo
90.9
60
-------
Tahlo S.K 1 3
Regression of Changaa in Lana-ASM Emissions Following
Contractor Repairs Versus Corresponding Changes in FTP Emissions
Dependent variable
R2 -71.7%
s- 1.040 with 52
Source
Regression
Residual
Variable
Constant
A(ASM HC)
is: A(FTP HC)
-2-50 degrees of freedom
Sum of Squares df
137.187 1
54.1169 50
Coefficient s.e. of Coeff
0.270944 0.1633
2.18967 0.1945
Mean Square
137
1.08234
t-ratlo
1.66
11.3
F-ratlo
127
Dependent variable
R2 -79.5%
is:
A(FTP CO)
s - 1 1 .55 with 52-2-50 degrees of freedom
Source
Regression
Residual
Variable
Constant
AfASM CO)
Sum of Squares
25906.3
6673.29
Coefficient
5.71775
1.08685
df
1
50
s.e. of Coeff
1.775
0.078
Mean Square
25906
133.466
t-ratlo
3.22
13.9
F-ratio
194
Dependent variable
R2 -70.8%
is: A(FTP NOx)
s » 0.8016. with 52-2-50 degrees of freedom
Source - ~
Regression
Residual
Variable
Constant
A(ASM NOx)
Sum of Squares df
78.0956 1
32.1317 50
Coefficient s.e. of Coeff
-0.013624 0.1392
0.829714 0.0753
Mean Square
78.1
0.642635
t-ratlo
-0.098
1 1
F-ratlo
122
61
-------
Figux* 5.6.1.1
Decreases in HC Emissions Following Repairs
AFTPvsAIM240HC
Regression Line
It-Squared = 81.9%
135
AIM240 HC Emissions (g/mi)
AFTPvsAASMHC
f
15
10
Regression Line
R-Squared = 7L7%
135
A ASM HC Emissions (g/mi)
62
-------
Figure 5.6.1.2
Decreases in CO Emissions Following Repairs
AFTP vs AIM240 CO
8
120
100
60
20
0
•20
Regression Line
It-Squared = 47.5%
. \
•20
20 40 60
AIM240 CO Emissions (g/mi)
80
100
-20
AFTPvsAASMCO
Regression Line
R-Sqoared = 79.5%
20 40 60
A ASM CO Emissions (g/mi)
80
100
63
-------
Figur* 5.6.1.3
Decreases in NOx Emissions Following Repairs
AFTPvsAIM240NOx
Regression Line
R-Squared = 64.5%
-1
135
AIM240 NOx Emissions (g/mi)
AFTPvsAASMNOx
\Regression Line
R-Squared = 704%
135
A ASM NOx Emissions (g/mi)
64
-------
Comparing the two graphs that examine the changes in HC emissions (Figure
5.6.1.1), it is apparent that one vehicle (vehicle number 3190) exhibited a
reduction in FTP HC emissions substantially greater than any of the other 51
vehicles (12.88 g/mi HC reduction compared to only 3.86 for the next larger
FTP HC reduction). Since it is possible that one such vehicle could
substantially affect the regression analysis, a second set of regressions were
performed on the remaining 51 cars (i.e., with vehicle number 3190 deleted) to
determine the effect. The effects on the slopes of the regression lines are
given in Tables 5.6.1.4 and 5.6.1.5.
Table 5.6.1.4
Effect on IM240 Regression Line
For BC Emissions
Of Deleting Vehicle 3190
Constant
Coefficient
R2
Based on Al
52 Vehicles
-0.365173
1.4106
81.9%
Based on 51
Vehicles
0.018227
0.93431
74.7%
Table 5.6.1.5
Effect on ASM Regression Line
For BC Emissions
Of Deleting Vehicle 3190
Constant
Coefficient
R2
Based on Al
52 Vehicles
0.270944
2.18967
71 .7%
Based on 51
Vehicles
0.452113
1.35391
67.8%
From Tables 5.6.1.4 and 5.6.1.5, we see that deleting that potential HC
outlier (vehicle number 3190) has a similar effect on each regression line.
The slope of the IM240 regression line decreases 11.6 degrees, and the slope
of the ASM regression line decreases 11.9 degrees. Since deleting the change
in HC emissions of vehicle 3190 from the sample has the same effect on both
regression lines, it would be advisable to use the equations based on 51 cars
to estimate changes in FTP HC emissions based on IM240 and/or ASM HC changes,
for IM240 and/or ASM HC changes between -1.0 and +4.0 g/mi.
65
-------
Comparing the two graphs that examine the changes in CO emissions (Figure
5.6.1.2), it appears, at first glance, that the composite IM240 tends to over
predict the repair benefit to CO emissions for some vehicles with relatively
small FTP CO repair benefits. However, the actual situation is that several
relatively cleaner vehicle* (though still exceeding FTP standards) had
unusually high IM240 results on their first test. IM240 CO was a lot lower
after repair, but the FTP emissions had comparatively little room to improve.
The five vehicles in Figure 5.6.1.2 that exhibit this problem (vehicles
numbered: 3157, 3175, 3211, 3213, and 3214) all have as-received composite
FTP CO less than 15 g/mi. For two of those five, most of the high composite
IM240 emissions resulted from the first mode (i.e., the first 93 seconds) of
the IM240. For this reason, EPA has recommended that vehicles which fail the
composite HC or CO cutpoint be given a second chance to pass by examining the
Mode-2 emissions (see "Two-ways-to-pass" in Section 5.3). A similar situation
cannot happen for the ASMs as analyzed in this report because the weighting
factors, in effect, cause the CO scores on the first mode (5015) to be
ignored. One vehicle that deserves special note is vehicle number 3211. That
vehicle exhibited the largest IM240 CO reduction (91.70 g/mi), but an FTP CO
reduction of only 8.00 g/mi. This high lane-IM240 CO reduction resulted from
a high initial (i.e., as-received) lane test score of 93.07 g/mi, but an
initial FTP CO score of 10.79. (However, the lane score was confirmed by an
indolene-fueled lab-IM240 following the FTP which had a CO result of 52.48
g/mi.) The ASM tests on this vehicle did not exhibit a large CO reduction
following repairs because both the initial ASM and the ASM following repairs
exhibited very high CO emissions (more than 5%) during the 50 mph cruise mode.
(Thus, the ASM did not over estimate the CO repair benefit on vehicle 3211
because the ASM over estimated both the initial FTP CO emissions, as well as,
the FTP CO emissions following repair.) In spite of the few over predictions
of emission benefits from repairs, it should be noted (as illustrated in Table
5.6.1.1) that following the IM240-targeted repairs, no vehicle was left with
high unrepaired FTP emissions.
Most of the vehicles, which exhibited very little if any HC or CO
improvement following the IM240-targeted repairs, had been recruited for
repairs because they exhibited, on the lane-IM240 test, low HC and CO, but
high NOx> Therefore, no significant improvement in either FTP HC or CO was to
be expected.
Comparing the two graphs that examine the changes in NOx emissions (Figure
5.6.1.3), it is apparent that one vehicle (vehicle number 3202) exhibited a
reduction in FTP NOx emissions greater than any of the other 51 vehicles (6.31
g/mi NOx reduction compared to 4.98 for the next larger FTP NOx reduction).
Since it is possible that one such vehicle could substantially affect the
regression analysis, a second set of regressions were performed on the
remaining 51 cars (i.e., with vehicle number 3202 deleted) to determine the
66
-------
effect. The effects on the slopes of the regression lines are given in Tables
5.6.1.6 and 5.6.1.7. From Tables 5.6.1.6 and 5.6.1.7, we see that deleting
that potential NOx outlier (vehicle number 3202) has virtually no effect on
either regression line. The slope of the IM240 regression line decreases only
3.3 degrees/ and the slope of the ASM regression line decreases less than half
a degree.
Table 5.6.1.6
Effect on IM240 Regression Line
For NOx Emissions
Of Deleting Vehicle 3202
Constant
Coefficient
R2
Based on Al
52 Vehicles
-0.275563
0.738523
64.5%
Based on 51
Vehicles
-0.183932
0.652265
57.4%
Table 5.6.1.7
Effect on ASM Regression Line
For NOx Emissions
Of Deleting Vehicle 3202
Constant
Coefficient
R2
Based on Al
52 Vehicles
-0.013624
0.829714
70.8%
Based on 51
Vehicles
-0.003336
0.816328
60.1%
Six vehicles (vehicle numbers: 3172, 3200, 3212, 3239, 3240, and 3244)
exhibited large decreases in lane NOx emissions, but little if any change in
FTP NOx emissions. These six had a number of factors in common:
- *_'
- All six- had low as-received FTP HC (for five of the six HC £ 0.37,
and HC - 0.59 for the sixth), CO (CO S 3.47), and NOX (NOx * 2.34) .
All six had low as-received lane-IM240 HC (HC £ 0.29) and CO
(CO £ 3.33), but high lane-IM240 NOx (NOx * 1.14).
67
-------
- All six had low as-received ASM composite HC (HC S 0.24) and CO
0.80 and Mode-2 IM240 HC > 0.50.
2) the 16 vehicles that were recruited (and repaired) because their
initial lane-IM240 exceeded the CO cutpoint of:
Composite IM240 CO > 15.00 and Mode-2 IM240 CO > 12.00.
3) the 30 vehicles that were recruited (and repaired) because their
initial lane-IM240 exceeded the NOx cutpoint of:
Composite IM240 NOx > 2.00.
As previously-discussed, two vehicles (vehicles numbered 3211 and 3190) could
be deleted from the "HC-Repaired* and from the "CO-Repaired" data bases due to
questionable test results. Additionally, vehicle number 3202 could be deleted
from the "NOx-Repaired* data base for similar reasons. Thus, in addition to
performing regression analyses on the entire 52 car data base, we can also
perform regressions on the 32/16/30 (HC/CO/NOx) subsets, as well as, (after
deleting the questionable vehicles) on the 30/14/29 car subsets. Within these
various data sets, we performed 16 linear regressions, the results of which
are summarized in Tables 5.6.1.8 through 5.6.1.10.
68
-------
56 1.8
Ragreaaion Lines of AHC for Short Teata Versus FTP
Constant
Coefficient
R-Squared
......... !••«-•*»
Based on
All 52
Vehicles
-.365173
1.41060
81.9%
Based on
32
Exceeding
Initial HC
-.932339
1.63036
82.7%
Based on
32 Minus
Two
0.014741
0.958254
635%
r-r-,-,. A«S««
Based on
AB52
Vehicles
.270944
2.18967
71.7%
Based on
32
Exceeding
Initial HC
0.371352
2.14853
672%
Based on
32 Minus
Two
0.7929
1.11862
605%
Tabla 5.6.1.9
Regression Lines of AGO for Short Teata Versus FTP
Constant
Coefficient
R-Squared
......... !•««.««
Based on
All 52
Vehicles
4.64057
0.846373
47.5%
Based on
16
Exceeding
Initial CO
17.5097
0.611959
18.7%
Based on
16 Minus
Two
-0.36712
1.28527
55.1%
......... *«••
Based on
AI52
vehicles
5.71775
1.08685
79.5%
Based on
16
Exceeding
Initial CO
13.9744
0.95755
73.0%
Based on
16 Minus
Two
13.4061
0.962594
74.1%
69
-------
Table S.6.1.10
Regression Lines of ANOx for Short Tests Versus FTP
Constant
Coefficient
R-Squared
Based on
All 52
Vehicles
-.275563
0.738523
64.5%
__ |M24fl ..
Based on
X
Exceeding
Initial NOx
-.778908
0.886525
54.2%
Based on
30 Minus
One
-0.45849
0.733666
39.9%
Based on
AH52
Vehicles
-0.013624
0.829714
70.8%
.... AIM ...
Based on
X
Exceeding
Initial NOx
0.209571
0.763686
62.3%
Based on
30 Minus
One
0.273927
0.711124
45.7%
Examining the slopes and y-intercepts (i.e., the "coefficient" and
"constants" in Tables 5.6.1.8 through 5.6.1.10) of the 18 regression lines, we
make the following observations:
- Limiting the analysis to only those vehicles whose initial lane-IM240
test exceeded the outpoint for the pollutant being examined:
— had virtually no effect on the regression line predicting FTP
HC changes based on ASM HC changes, and only a relatively small
effect on the line predicting FTP CO changes based on ASM CO
changes;
— had moderate effects on the two regression lines predicting FTP
HC and CO changes based on IM240 HC and CO changes; and
— had only relatively small effects on the regression lines
predicting FTP NOx changes based on ASM or IM240 NOx changes.
Again, the effect was larger for the IM240 case.
- Deleting the one or two questionable vehicles prior to performing the
regression analysis:
— produced only small effects in the two NOx cases (IM240 and
ASM) and in the ASM CO case and
— produced substantial effects in the two HC cases and in the
IM240 CO case.
In summary, this analysis indicates that the change in ASM scores before
and after repairs correlates with changes in FTP emissions, about as well as
70
-------
for the IM240. However, because ASM outpoints were not recommended by ASM
proponents and EPA did not have outpoints to use as repair targets, the
contractor repairs were performed to attain IM240 scores that complied with
the standard IM240 cutpoints. So the contractor repairs offer little insight
into the primary question of whether vehicles repaired to pass an ASM test
will be as effective as vehicles repaired to pass the IM240 test. The next
two sections will further discuss repair issues.
5.6.2 Commercial repairs
5.6.2.1 Introduction
The purpose of this analysis was to compare the effects of commercial
repairs, for vehicles that failed the Arizona I/M test, on IM240 and ASM
after-repair test results. Experience has shown that commercial repairs
geared to steady-state I/M tests have not met expectations for in-use emission
reductions. Because vehicles are operated only at steady-state, repairs have
been geared to reducing emissions at those operating conditions. As a result,
emissions over the full range of operating conditions are often not
effectively reduced, even when vehicles are repaired to pass a steady-state
I/M test. This is one reason EPA has established a transient test for
enhanced I/M. Since the IM240 requires vehicles to perform over a wide
variety of real-world operating conditions, IM240-successful repairs must be
effective in reducing emissions over a wide range of operating conditions.
By comparing the effects of commercial repairs on ASM and IM240 test
results at selected cutpoints, an evaluation of the comparative repair
effectiveness can be made. As discussed above, EPA analyzed the results of
repairs performed to pass the Arizona I/M test to determine whether such
repairs would significantly reduce ASM emissions without significantly
reducing FTP emissions. Since the ASM test and the Arizona I/M test are
somewhat similar in that they are steady-state tests, repairs for the Arizona
I/M test may provide information on whether ASM-successful repair are as
effective as- ZM240-effective repairs. The data show that successful repairs
for the Arizon* I/M test are more likely to be successful for the ASM test
than for th* IM240.
5.6.2.2 Database/Analysis
EPA1s commercial repair program in Mesa consisted of offering incentives
to owners of 1983 and newer vehicles that failed the Arizona I/M test, but
were not needed or declined to participate in laboratory testing, to return to
EPA1s IM240 lane for after-repair ASM and IM240 tests. To receive their
incentive, they were told to return with a receipt for commercial repairs. No
71
-------
instructions were given to owners regarding where or how to get their cars
repaired, and owners were not compensated for the actual repair itself.
As of April 1, 1993, before- and after-repair data were available for 23
of these vehicles. One vehicle, #13239 (CR# 24) was removed from the database
due to unacceptable speed deviations on its initial ASM test, leaving 22
vehicles available for analysis. For this analysis, five other vehicles were
excluded because they continued to fail the Arizona test after repairs. The
resulting database consisted of 17 .successfully repaired, 1983 and newer
vehicles.
Outpoints were applied to the IM240 and ASM data to determine pass/fail
status. The pass/fail determinations were then compared to evaluate the
effects of commercial repairs. Three different cutpoint sets were used to
make the comparisons. Since the Arizona test measures HC/CO only, the first
comparisons involved only HC and CO criteria. Two additional comparisons were
made which included NOx outpoints. All three are listed below (Section 5.3
discusses the relevance of these cutpoints.):
• IM240 recommended cutpoints for HC/CO with ASM cutpoints that produce
the highest IDRs at the same failure rate as the IM240 recommended
cutpoints:
IM240 - 0.80 / 15.0 + 0.50 / 12.0
ASM - 1.00 / 8.0
• EPA recommended IM240 cutpoints including NOx with ASM cutpoints that
produce the same 18% failure rate. These cutpoints are listed below:
IM240 - 0.80 / 15.0 / 2.0 + 0.50 / 12.0
ASM - 1.00 / 8.0 / 2.0
• ASM and IM24O cutpoints selected to achieve the highest IDRs possible
while keeping the probable EC rate below 5%. These cutpoints are
listed below:
IM241T - 0.30 / 9.Q/ 1.7 + 0.19 / 7.0
ASM - 0.40 / 8.0 / 1.5
For each set of cutpoints, a comparison of the initial and final test
results were made. To evaluate the effects of repairs on a specific I/M test
a vehicle must be identified by the I/M test for repairs. Thus, while the
initial test result comparison allowed the identification ability of these two
I/M tests to be compared, the final test result allows an evaluation of the
relative repair effectiveness of the I/M tests. The data were restricted to
72
-------
vehicles which were identified by all teats for the comparison of final test
results. Using these common vehicles allows the comparison of repair effects
to be clearly illustrated.
The results, which are discussed in the next section, indicate that the
IM240 is superior at identifying vehicles requiring repair and that for
vehicles which initially fail both the IM240 and ASM, steady-state repairs are
more likely to result in ASM passing scores than in IM240 passing scores.
5.6.2.3 Results/Conclusions
Initially, all 17 vehicles used in this analysis failed their initial
Arizona I/M test. However, the IM240 and ASM identified slightly different
sets of vehicles as needing repairs. Vehicles of interest are those that pass
the initial IM240 and fail the initial ASM and those that fail the initial
IM240 and pass the initial ASM.
As shown in Table 5.6.2-1, for the initial HC/CO only comparison, one car
passed the IM240 and failed the ASM and four cars passed the ASM and failed
the IM240 (see Appendix B for data listings). These errors-of-omission
support the assertion made in Section 5.3 that the ASM is weaker than the
IM240 at identifying malfunctioning vehicles with HC and/or CO emission
problems.
Table 5.6.2-1
Initial Pass/Fail Status Comparison
HC/CO only
Cutpointa
IM240 ASM
PASS PASS 3
IN240 ASM
FAIL PASS 4
IM240 ASM
PASS FAIL 1
XM24Q*: ASM
FAH»"L FAIL 9
Common Failure
Rate Cutpoints
IM240 ASM
PASS PASS 2
ZM240 ASM
FAIL PASS 2
ZM240 ASM
PASS FAIL 0
ZM240 ASM
FAIL FAIL 13
Optimal IDR/Max
Be Cutpointa
IM240 ASM
PASS PASS 1
ZM240 ASM
FAIL PASS 2
IM240 ASM
PASS FAIL 0
IM240 ASM
FAIL FAIL 14
Vehicle 13504 (CR# 25) failed the ASM and Arizona test due to a CO problem
which appears to occur only at idle operation. Because the IM240 driving
cycle includes little idle operation, this vehicle was not identified by the
IM240 HC/CO only outpoints. An air/fuel mixture adjustment reduced emissions
sufficiently to pass the HC/CO outpoints for the ASM and Arizona tests.
However, this vehicle did exhibit excessive NOx emissions that were identified
by the addition of a NOx outpoint. Incidentally, the fuel mixture adjustment
73
-------
did little to address or reduce this vehicle's NOx emissions on either the
IM240 or the ASM.
In contrast to the IM240, which failed to identify only one vehicle, four
vehicles passed the ASM HC/CO only cutpoints and failed the IM240 and Arizona
cutpoints. Three of these vehicles are examples of ASM errors-of-omission and
illustrate the superior identification ability of the IM240. The fourth
vehicle failed NOx and will be discussed after the three that passed.
Vehicle 13471 (CRt 27) failed the Arizona and IM240 tests because of high
CO emissions, but passed the ASM test. Vehicle 13125 (CRt 12) failed HC on
both the IM240 and Arizona tests and was not identified by the ASM cutpoints.
Vehicle 13202 (CRt 15) failed the HC and CO idle modes of the Arizona test.
On the IM240, vehicle 13202 failed HC and NOx but passed CO due to the two-
ways-to-pass algorithm. The ASM identified this vehicle for NOx emissions
only.
The fourth vehicle that initially passed only the ASM test was vehicle
12771 (CRt 8). This vehicle exemplifies the weakness of steady-state I/M
tests and is discussed in detail in Section 5.6.3. Vehicle 12771 failed CO
on the loaded mode of the Arizona test but passed the CO outpoint on both the
IM240 and the ASM. However, the car failed NOx and HC on the IM240 and failed
only NOx on the ASM. After repair, this car passed both the ASM and Arizona
tests even when ASM cutpoints were tightened. These repairs did not
sufficiently reduce emissions over the full operating range of the vehicle,
demonstrated by the vehicle continuing to fail both HC (1.01 g/mi) and NOx
(3.01 g/mi) on the IM240. This supports the assertion made in the
introduction that repairs to pass a steady-state test may not be effective in
reducing emissions over normal driving conditions and, therefore, do not
effectively reduce in-use emissions.
To illustrate the effects of commercial repairs on ASM and IM240 after-
repair test results, data were restricted to vehicles that failed both the
initial ASM and ZM240 (see Table 5.6.2-1). The results of these comparisons
are graphically depicted in Figures 5.6.2-1 thru 5.6.2-3.
74
-------
Figure 5.6.2-1
Commercial Repairs Passing ASM and IM240 Outpoint,
HC/CO only Comparison
9
Pass ASM
But Fail
IM240
Commercial
Figure 5.6.2-2
R.P.lt8 Pa8aing ASJf
Common Failure Rate
Comparison
Pass ASM
But Fail
IM240
75
-------
Figure 5.6.2-3
ixcial Repairs Passing ASM and IM240 Outpoints
Bigh EC Comparison
Pass ASM
But Fail
IM240
These graphs show that vehicles can and will be repaired to pass the ASM
test but will continue to fail the IM240.
For the first comparison using only HC and CO cutpoints, three vehicles
passed the ASM but continued to fail the IM240. The second comparison added
the NOx outpoint which in combination with the HC/CO cutpoints produced the
same failure rates for the IM240 and ASM. Again, three vehicles passed the
ASM but continued to fail the IM240. For the comparison using the most
stringent cutpoints for the IM240 and ASM, five vehicles passed the ASM but
continued to fail the IM240. For all of these comparisons, there were no
vehicles that failed the ASM and passed the IM240 after commercial repairs.
This indicates that repairs which are sufficient to pass the ASM test are not
necessarily sufficient to pass the IM240, indicating that the repair
effectiveness of the IM240 is superior to that of the ASM.
Based on these results, repairs to pass the steady-state Arizona I/M test
are significantly more effective at reducing ASM emission scores than IM240
emission scores. Although the sample of successful commercial repairs is
small, these results indicate that the ASM test, if implemented, will result
in significantly lower identification rates and emission reduction benefits
than those of the IM240. A more detailed investigation of ASM emission
reduction benefits is discussed in Section 5.6.3.
76
-------
5.6.3 Xn-Use Emission Reductions from Real World
Repairs
One of the concerns with any test is the ability of an observed reduction
on the test to reflect real and permanent in-use reductions. Two particular
concerns are (1) can unscrupulous mechanics find repair strategies that would
allow a vehicle to temporarily pass the I/M test without resulting in
permanent in-use reductions (i.e., temporary repairs would be undone after
passing the test), and (2) is the test sufficiently imprecise such that
honest, but insufficient, repairs would not be detected by the I/M retest.
last Defeating strategies
It is common knowledge that the current idle test can be, and is being,
defeated by a variety of methods. Most can be used only in the privacy of a
test-and-repair station. Some of the common ones that can be used in a test-
only station include creating a vacuum leak to lean out the air-fuel ratio for
CO failures, and raising the idle speed to create a similar effect. A logical
question is, what is the likelihood that test defeating strategies can be
developed by unscrupulous mechanics for the ASM or for the IM240 I/M tests.
On the surface, the ASM test appears easier to beat than the IM240 because
of its steady-state nature and number of limited operating modes. In theory
at least, the mechanic could employ a similar method to the idle test for ASM
CO failures. The process would include creating a vacuum leak and disabling
the feedback control system. Since it is assumed that most shops would have a
dynamometer in an ASM I/M scenario, the mechanic would simply need to operate
the vehicle on the dynamometer and adjust the leak until the car was under the
cutpoints. Most likely the driveability of the car would be quite poor;
however, it would only need sufficient driveability to drive to the test
center and return, where the test beating repairs could b« undone.
If the vehicle could drive to the test center, then it could certainly
drive the steady-state test, since driveability is not required on the ASM,
and emission*- ar« not recorded during the transitions between ASM test modes.
Conversely, tht> emissions are measured during driving transitions on the
IM240, and the? lack of driveability would require more throttle movement with
a likely substantial increase in CO emissions. If misfire occurred during
driving transitions because of the lean condition, the HC, and possibly NOx,
would increase on the IM240, but would not on the ASM (because emissions are
not measured during driving transitions).
Another potential test defeating strategy that could occur on the ASM for
NOx failures deals with ignition timing. Retarding ignition timing has long
been an approach to reducing NOx. Retarding the ignition timing excessively,
77
-------
however, reduces driveability. Once again, however, driveability is not
required on the ASM. A severe loss in driveability on the IM240 would be
expected to increase CO significantly, but would be expected to have little
effect on the ASM CO levels.
Some may point out that many new cars do not have adjustable distributors,
and others do not have distributors at all. Therefore, it would not be
possible to retard the timing, so such a test defeating strategy would not
exist on these cars. What many may not know is that all cars with non-
adjustable distributors and those without distributors have a base timing mode
that can be activated. Activation of base timing will severely retard the
timing in most cases, and could be used to lower NOx emissions.
Since these are only a few of the less creative methods that might be
attempted to defeat an ASM or IM240 test, it would be useful to verify if the
theoretical potential really could occur. Currently there is no data on
purposefully test defeating repairs. However, data from vehicles tested in
Arizona and sent for commercial repair may shed some light on the potential
for test defeating or improper repairs to be identified by either the ASM or
the IM240.
In our commercial repair data base, twenty-two vehicles failed the Arizona
I/M test which includes a steady-state loaded mode and an idle mode. All of
these vehicles received a 4-mode ASM test and an IM240 test. The vehicle
owners took the vehicles for commercial repair, and volunteered for repeat ASM
and IM240 tests when they returned for their Arizona retest.
Five of the 22 vehicles were excluded from this analysis because their
four-mode ASM emissions did not exceed the cutpoints of 1.0/8.0/2.0
(HC/CO/NOx). The resulting 17 vehicles represent the portion of the 22 car
sample that would have failed a four-mode ASM test if that had been the
official test. Note that this group of 17 vehicles represents a different
portion of the sample of 22 commercially repaired vehicles than the 17
vehicles used for analysis in Section 5.6.2. The analysis in Section 5.6.2
excluded fiv* vehicles that ultimately did not pass the Arizona I/M test after
repairs. TO* analysis in this section excluded five vehicles that passed the
initial ASM tcat, but included those vehicles that did not ultimately pass the
Arizona I/M test after repairs.
The repairs conducted on the 17 vehicles are listed in Table 5.6.3-1.
From these repairs and the resulting ASM and IM240 scores, the possibility of
test defeating strategies can be evaluated. Not* that the multiple repairs
represent retest failures on the Arizona I/M test. Also, four of the 17
vehicles that initially failed the ASM test did not ultimately pass the
Arizona I/M test.
78
-------
The before and after repair emission results are graphically represented
in Figures 5.6.3-1 though 5.6.3-3. From the repair data reported by the
commercial garages (in Table 5.6.3-1), it is clear that many vehicles had the
air-fuel ratio adjusted or received repairs that would likely affect air-fuel
ratio. Many of the vehicles were feedback carbureted; however, this should
make no difference for the purposes of evaluating the effect of air-fuel ratio
on the test type. Only vehicles CR-07 and CR-16 had reported commercial
repairs that would not likely affect air-fuel ratio (it was assumed that the
"tune-up" repairs in Table.5.6.3-1, in some cases, could have involved
adjustment of air-fuel ratio).
On these other vehicles, the degree of the effect on air-fuel ratio is
unknown. But, from the CO emission results in Figure 5.6.3-2, it is clear
that in general, a repair that resulted in reduced CO on the IM240 also
reduced CO on the ASM. However, there are some exceptions. These are
vehicles CR-10, and CR-25. Vehicle CR-10 failed the before and after IM240,
failed the before-ASM, but passed the after-ASM. Whereas vehicle CR-25 passed
the before and after IM240, failed the before-ASM, and passed the after-ASM.
Since CO is primarily a function of air-fuel ratio, the observation from
these two vehicles is that the air-fuel ratio during the steady-state test can
be different than the average over the transient test. To some extent, this
observation also appears to be evident in the CO results for vehicles CR-03,
CR-06, and CR-22 (see Figure 5.6.3-2). In the case of vehicle CR-10, the air-
fuel ratio during steady-state operation is sufficiently lean after repairs to
allow the vehicle to pass the ASM, but rich enough overall during transient
driving to cause an IM240 failure. The opposite is apparently true for
vehicle CR-25, where the before repair air-fuel ratio during steady-state is
apparently sufficiently rich to cause an ASM failure, but lean enough during
average driving to allow the vehicle to pass the IM240.
Certainly, the CO level can also be affected by the catalyst. But the
catalyst was the same in all of these tests, so the catalyst effect should
wash out. Also, catalyst efficiency can be somewhat gauged by HC levels as
seen in Figux*5.6.3-1. The after repair HC levels on vehicle CR-10 clearly
pass the ASK outpoint. The after repair IM240 HC status parallels the CO
status. In other words, based on the IM240 this vehicle was still broken, but
was passed on the ASM. The HC levels on vehicle CR-25 were low for all IM240
and ASM teats. Based on IM240 results, this vehicle should not have been
failed for HC or CO. However, vehicle CR-25 did have serious problems as
evidenced by the NOx emissions in Figure 5.6.3-3.
The emission results on these two vehicles, reinforce the following point.
Air-fuel ratio can affect the CO levels on both tests. In particular, the
79
-------
air-fuel ratio during a steady-state mode can be different than the overall
ratio during transient operation. Therefore, it is likely that with willful
intent, a mechanic could purposefully create a vacuum leak, and adjust it so
that a car could pass the ASM, but not the IM240. Whereas the amount of
leanness in vehicle CR-10 was not sufficient to pass CO on the IM240, it was
sufficient to pass the ASM. Furthermore, the amount of leanness was not
sufficient to cause vehicle CR-10 to fail either the IM240 or the ASM NOx
cutpoints. Therefore, the results on vehicle CR-10 support the theoretical
possibility that unscrupulous mechanics could, with proper adjustment of
vacuum leaks, be able to adjust vehicles to temporarily pass the ASM CO
without increasing the NOx emissions sufficiently to cause an ASM NOx failure.
As indicated previously, the likelihood of such improper and temporary repairs
would be exacerbated in a program where the ASM was the official test, because
unscrupulous repair centers could conveniently maladjust a vehicle on a
dynamometer to pass the steady-state modes of the ASM test.
80
-------
VEH.NQ
CR-01
CR-02
CR-03
CR-04
CR-05
CR-06
CR-07
CR-08
CR-09
CR-10
CR-13
CR-15
CR-16
CR-21
Table 5.6.3-1
Vehicle Repairs
1st Repair
Adjusted air/fuel mixture on
carburetor.
Adjusted an/fuel mixture on
carburetor.
Adjusted air/fuel mixture on
carburetor. Replaced heat valve.
Repaired vacuum leak and adjust
ignition timing.
Replaced O2 Sensor and performed
TuneUp.
Rpl O2 .plugs, cap and rotor,
cleaned fuel injector
Rpl fen belt, plugs, fuel flt Adj
tuning. Changed oiL
Tune-up, replaced fuel filter,
replaced air filter.
Adjusted emissions. Scoped and
adjusted an/fuel mixture.
Adjusted air/fuel mixture and idle
speed.
Adjusted air/fud mixture, idle speed.
Replaced Oxygen sensor.
Set Ignition timing to manufacturer's
Tune-upjlpl plugs, wires,
QISulOluQf CflD flOu fOIOf•
tion of choke
Adjodedan/rad mixture and idle
2nd Repair
Rpred electrical short in
harness from ECU to mix
control
Adjusted Idle, an/fuel
injectors.
A one year waiver was
granted for this vehicle.
Adjusted air/fuel mixture
and idle speed.
Performed Tune-up.
Scoped engine and
adjusted carburetor.
3rd Repair
Overhauled
Carburetor.
CR-26 Performed basic tune up.
81
-------
Commercial Repair Effects
Clung* h HC EmlMtan* en Cm
Frifeg ASH HC. CO, or NOB
IU240
Am,
Boloni .. X.
ASH
i® m
im m
MSI m m m \\
§ I I I I I 3 I 5 I I I I I
V«hld« Number
8
Commercial Repair Effects
Ctang* to CO EnlMtorv on C*n
FUlng ASH HC, CO, Of MO«
11 i11 i \ i i i
Vtttcto NumlMr
82
-------
Commercial Repair Effects
Clung* ki NO* Cmmicn* on CM
FaUbig AMI HC, CO, «r NOB
I § I I II I 3 ? 33 33 3 2 5 3
Vchteto Number
Insufficient Repairs
Another concern with an Z/M test is the ability for the test to cause
'proper and sufficient repairs to be performed if the vehicle fails the I/M
test. For this analysis, proper and sufficient repairs are considered to be
repairs sufficient to pass the IM240 outpoints. The commercial repair data
used in the preceding section on test defeating strategies can also provide
some insight into this issue.
Of interest- is the comparison of test modes between the 4 -mode ASM and the
Arizona I/It teeft. Both have an idle mode, and both have a steady-state loaded
mode. The *Trif*«M loaded mode is similar to the ASM 2525 mode.
Using the> general similarity of the test (i.e., idle and loaded modes ),
the general sufficiency of ASM repairs can be approximated by observing the
results from vehicles used in the previous section that failed the initial ASM
test and the initial Arizona Z/M test. A case history on vehicle number CR-
08, which initially failed the ZM240 HC and NOx outpoints (as well as the
Arizona CO outpoint and the ASM NOx cutpoint) , illustrates the concern about
the ability of the ASM teat to cause proper and sufficient repairs to occur
in-use .
83
-------
After vehicle number CR-08 had failed the Arizona I/M test for CO on
January 4, and had received initial tests on the IM240 and ASM, it was
enlisted in the commercial repair program. Two weeks after the commercial
repairs (January 14), the vehicle returned for its after-repair IM240, ASM,
and Arizona I/M retest (and for the owner to obtain the recruitment incentive
payment) . At that time, it was discovered that two days after the initial I/M
test (which was conducted on January 4), and following repairs (listed in
Table 5.6.3-1), the repair center had taken the vehicle to another Arizona
test lane for an I/M retest. At this other I/M lane, on January 6 the vehicle
easily passed the Arizona outpoints of 1.2% CO and 220 ppm HC. However, when
retested on January 14 at the IM240 test lane, this vehicle failed the Arizona
HC cutpoint by a wide margin (see Table 5.6.3-2). The owner was demons tr ably
upset (even though a valid Arizona passing certificate had been issued) , and
left the test center abruptly. However, the owner returned again in another
two weeks (January 26) . At this time, the vehicle passed all of the Arizona
cutpoints. The owner did not divulge any information on corrections or
repairs that may have occurred between January 14 and January 26.
Table 5.63-2
Test Data - Vehicle No. CR-08
I — StattTMt — I
Loaded kfl« I— IM240— I I— ASM —1
HC CO HC CO HC CO NOx HC CO NOx
-ft Afifl& -A fuU gQ^ UDL XttB SuBL
1/0403 LaaeIM240 2771 86 121 87 0.38 121 12.2 2JU 0.46 5.7 2.19
1/0603 StataTe* — 40 0.84 116 0.78 — — — — — —
1/1403 LawIM240 2977 38 0.63 8U 0.07 13S 4.1 2JJ O33 3.3 1.63
1/26/93 LueIM240 316S 73 O38 41 0.06 Ifll 4.4 Ifll 0.16 3.4 1.51
Several important aspects should be noted. First, while this vehicle
failed the Arizona CO cutpoints, it passed CO on all ASM and IM240 tests.
Also, while this vehicle passed HC in all of ASM tests, it failed HC on all of
the IM240 testa. Further, after the first repair, this vehicle passed the ASM
NOx cutpoint*^ for all subsequent ASM tests, even though it failed the IM240
NOx cutpointokfor all of the subsequent IM240 tests (aa well as the initial
IM240) . Tber ZM240 NOx actually increased slightly from the first test to the
last.
The most pessimistic scenario on this vehicle is that once the vehicle
failed the Arizona test for CO, the mechanic maladjusted the vehicle, took it
to an I/M lane, where it passed, and then undid the maladjustments. These
undid maladjustments were then observed on the January 14 Arizona retest. A
more benign conclusion is that the mechanic performed incomplete repairs, but
84
-------
the repairs were ultimately sufficient to pass the Arizona test. Also, the
repairs were sufficient to pass the ASM NOx outpoints (HC and CO were always
below the ASM outpoints), while they were not sufficient to pass the IM240 NOx
outpoint, nor were they sufficient to pass the IM240 HC cutpoint. In fact,
the ASM would not even have identified this vehicle as a high HC emitter.
Clearly, on this particular vehicle, commercial repairs were not
sufficient to pass the IM240, but were sufficient to pass the ASM test.
Reviewing the data for all 17 vehicles that failed the initial ASM test,
all of these vehicles, except CR-07, CR-21, and CR-22 eventually passed the
Arizona HC Z/M retest. Also, all vehicles, except these three vehicles and
CR-06 eventually passed the Arizona CO I/M retest.
However, vehicles CR-03, CR-04, CR-06, CR-Q8, CR-09, CR-10, CR-15, and CR-
16 which initially failed the IM240, continued to fail HC on the IM240 after
all commercial repairs (see Figure 5.6.3-1). Further, after all commercial
repairs, these same vehicles passed the ASM HC cutpoint (note CR-8, CR-9, and
CR-15 passed the initial ASM test, see Figure 5.6.3-1). Given the
similarities of the ASM and the Arizona test, these data suggest that the
level of HC repair on the ASM would be similar to the current Arizona Z/M
test. This assumption on test similarity and stringency of repair
effectiveness is further supported by the fact that the three vehicles that
failed to pass the Arizona test after repairs (CR-7, CR-21, and CR-22) were
also the only vehicles that failed the after-repair ASM test (see Table Figure
5.6.3-1).
Another method of looking at the ability of the ASM to enforce proper and
sufficient repairs is to look at the test status of the ASM results before and
after repairs relative to the before and after ZM240 status. The test status
for the 17 vehicles initially failing the ASM for at least one pollutant is
listed in a truth-table format in Table 5.6.3-3 by pollutant (i.e., HC, CO,
and NOx). The roughly square boxes in a diagonal row represent test results
where the IM240 and ASM status before and after repair were identical.
Deviations fro*-the diagonal row, obviously represent results where the status
differs between the ZM240 and ASM. The fuzzy horizontal rectangular box in
Table 5.6.3-3. highlights those vehicles which passed the ASM test after
repair, but wer* still failing the ZM240 for HC, CO, or NOx.
From the Table, a total of 11 vehicles continued to fail the ZM240 HC
cutpoint after repair. Of these 11 vehicles, 8 vehicles (or 73%) passed the
ASM after repair (three of the eight also passed the initial ASM test). All
eight vehicles also passed the Arizona HC cutpoint after repair. Aa
previously mentioned, in the 11 vehicle sample that continued to fail the
ZM240 HC cutpoint, 100% of the vehicles that failed the Arizona HC cutpoint
85
-------
(CR-07, CR-21, and CR-22), also failed the ASM HC outpoint. Thus, every
vehicle that continued to fail the Arizona HC outpoint also continued to fail
the ASM HC outpoint. In other words, in this sample of commercial repairs,
the ASM did not fail anymore reteat vehicles than the Arizona I/M test.
CO test status represents somewhat of a mixed bag. Seven vehicles
continued to fail the IM240 for CO after repair. Only one vehicle in this
group of seven (or 14%) passed the ASM for CO after repair. This vehicle also
passed the Arizona retest. In this seven vehicle sample that continued to
fail the IM240 CO cutpoint, six vehicles continued to fail the ASM, and four
vehicles (CR-06, CR-07, CR-21, and CR-22) failed the Arizona retest. In this
case, the ASM found 2 more vehicles than the Arizona I/M retest after
commercial repairs.
However, it should be noted, that four other vehicles had anomalous CO
results. Two vehicles, CR-15 and CR-16 failed the initial IM240 for CO,
passed the initial ASM, and subsequently passed both the IM240 and ASM for CO.
Two other vehicles (CR-09 and CR-25), passed the initial IM240, failed the ASM
for CO, and also subsequently passed both the IM240 and ASM retests. If the
ASM was as good as the IM240 in identifying vehicles that should fail a retest
(at the same overall failure rate), one might expect a random scatter on each
side of the diagonal boxes, particularly for vehicles just marginally failing
or passing (which all of these were, except CR-25). Even so, all of these
vehicles also passed the Arizona CO retest. So they do not represent any
additional retest failures following commercial repairs that the ASM would
have found over the standard Arizona test. Also note, that these four
vehicles initially failed and continued to fail the IM240 for HC or NOx, and
•that the commercial repairs reduced the IM240 CO levels in all cases.
A total of 5 vehicles in this sample of 17 continued to fail IM240 NOx
after commercial repairs. Two of the five (or 40%) passed the ASM after
repairs. The Arizona I/M test does not test for NOx, therefore, it is more
difficult to judge the effectiveness of the ASM (using the Arizona test as a
surrogate) to force proper and sufficient commercial repairs.
This analysis began with a concern about the ability of the ASM test to
foster proper-and sufficient commercial repairs following an I/M failure.
Because of the general similarity of the Arizona I/M test to the ASM, it was
expected that repairs targeted by the commercial repair industry towards the
Arizona test would be similar to those that would be targeted towards the ASM,
at least for HC and CO. Thus, if the ASM were more effective in forcing
better repairs than the Arizona test, the ASM retest should fail more cars for
a given pollutant than the Arizona retest. Further, if the ASM were very
effective in forcing proper repairs, it would fail as many cars, for a given
pollutant, as an IM240 retest.
86
-------
The analysis shows that for this small sample, the ASM fails no more cars
for HC after commercial repairs than the Arizona retest, and fails only
twenty-seven percent of those that failed HC on the IM240 after repair. These
results imply that the ASM test would not force the repair industry to make
any more.repairs for high HC emissions than the current Arizona test, and
obviously, not as many HC repairs as the IM240. Therefore, the data from this
sample suggest that the repair effectiveness credits for HC in the MOBILE
model for commercial repairs on the ASM should be no greater than that
currently given for existing basic I/M programs.
The analysis for CO retest failures, indicates that the ASM found two more
vehicles than the Arizona test. The Arizona I/M retest found about 57 percent
of the IM240 retest failures, and the ASM found about 86 percent of the retest
failures. Thus in this small sample, it appears that an ASM retest would have
forced the repair industry to make additional CO repairs over and above those
that would have been required to pass the Arizona outpoints, but again, not as
many as the IM240 cutpoints would require. These results suggest that the
repair effectiveness credits for CO in the MOBILE model for commercial repairs
on the ASM should probably be given additional credit over that currently
given for existing I/M programs. The additional credit would be approximately
equal to 60 percent of the difference between that currently given for
existing I/M programs and that given to I/M programs employing the IM240.
However, given the potential ease that unscrupulous mechanics could defeat the
CO portion of the ASM retest, assigning additional CO repair effectiveness
credits in the model for ASM over those currently given for existing I/M
programs would be difficult to rationalize at this time.
The analysis for NOx retest failures is somewhat hampered by the fact that
the Arizona test only fails vehicles for HC and CO. Even though the repair
industry was not repairing vehicles to an NOx standard, the IM240 and ASM
retests after commercial repairs can be used to determine whether either
retest would have forced the repair industry to make additional repairs.
Clearly, both teats would have required some vehicles to get additional
repairs for high NOx emissions. However, the results from this sample
indicate that- •** ASM retest would only require 60 percent of the vehicles that
failed the OC240 retest to get additional NOx repairs. Therefore, this result
would suggest that the repair effectiveness credits for NOx in the MOBILE
model for commercial repairs on the ASM should be about only sixty percent of
that given for the IM240.
87
-------
IM240
Status
Fafl-Fai
HC
CO
NOx
Pass-Fail
HC
CO
NOx
HC
CO
NOX
HC
CO
NOx
Table 5.63-3
Effect of Commercial Repairs on Test Status
ASM Status
(Status Before Repair - Status After Repair)
Fafl-Fal Pass-Fail FaB-Pasa Pagg.Pflftfl
CR-07, CR-21,
CR-22
CR-03.CR-06.
CR-04.CR-07,
CR-21, CR-22
CR-9, CR-15,
CR-25
CR-03.CR-04,
CR-06, CR-10,
CR-16
CR-10
CR-08, CR-16
CR-06, CR-13
CR-01, CR-08,
CR-13.CR-26
CR-02
CR-09, CR-25
CR-08, CR-09,
CR-15
CR-01, CR-26
CR-15, CR-16
CR-02, CR-25
CR-02, CR-8
CR-01, CR-03.
CR-04.CR-05.
CR-06.CR-07.
CR-10, CR-13,
CR-21, CR-22.
CR-26
88
-------
5.6.4 On* -Mod* Repairs on ASM
The objective of this analysis was to investigate the theoretical effects
of targeting the ASM repairs to a single mode. That is, if it were possible
for a mechanic to reduce the emissions sufficiently on a single mode while
leaving the remaining three modes unaffected:
• Could an ASM failing vehicle, with such a repair, be made to pass the
ASM composite cutpoint?
• What are the emission characteristics of such passing vehicles?
Examining the 106 laboratory test vehicles, we can determine whether the
as-received NOx emissions met or exceeded an FTP NOx standard of 2.0 g/mi.
Also, we can determine FTP HC/CO emission range. That is:
Pass
FTP HC
0.41 and CO
3.40 (g/mi)
Marginal (Failing) Emitters
FTP HC > 0.41 or CO > 3.40 (g/mi)
and
FTP HC S 0.82 and CO £ 10.20 (g/mi)
High Emitters
FTP HC > 0.82 or CO > 10.20 (g/mi)
and
FTP HC S 1.64 and CO S 13.60 (g/mi)
Very High Emitters
FTP HC > 1.64 or CO > 13.60 (g/mi)
and
FTP HC £ 10.00 and CO £ 150.00 (g/mi)
taitters
FTP HC > 10.00 or CO > 150.00 (g/mi)
Classifying the laboratory vehicles in this way produces ten strata;
however, two of those strata are empty, and one stratum has only a single test
vehicle. Using the weighting factors (Table 5.2.5.2), we can model the lane
vehicles and characterize the emissions of that simulated lane sample of 2,071
1983 and new fuel-injected passenger cars. (Actually, the lane sample was
2,070 cars, the additional vehicle resulted from rounding off the estimated
89
-------
number of vehicles in those eight strata.)
following table.
The distribution is given in the
N OXS2.0
NOx>2.0
Flaafc Platglbufcion
Laboratory Sample
FTP HC/CO Emission*
Pass
27
0
Marginal
36
4
Hlgne
10
3
V.HIgh
18
7
Super
1
0
Simulated Lana Reat
FTP HC/CO Emlsslona
Paaa
808
0
Mirglnal
934
24
High
96
18
V. High
143
42
Super
6
0
An ASM outpoint of 1.00/8.0/2.0 (i.e., composite ASM HC £ 1.00, composite
ASM CO £ 8.0, and composite ASM NOx £ 2.0) will fail 372 vehicles in that
simulated lane fleet. The distribution of those 372 vehicles is given in the
following table.
NOx>2.0
«*fc nf •fr.g
galling 1,00/8.0/2.0
Laboratory Sample
FTP HC/CO Emission*
Pase
3
0
Marginal
6
3
|_|fi^i»
rtifln*
5
2
V.HIgh
12
6
Super
1
0
Simulated Lane Fleet
FTP HC/CO Emission*
Paae
54
0
Marginal
144
18
High
30
12
V. High
72
36
Super
6
0
If mechanics were able to repair those 372 vehicles so that the emissions
on the 2525 mode, the 50 mph mode, and the idle mode remained unchanged, but
.the emissions (HC, CO, and NOx) on the 5015 mode were reduced by 80 or 90
percent (the model yields the same result for each), then only 42 of those 372
would be able to pass the 1.00/8.0/2.0 outpoint. Thus, a repair strategy that
targeted only the 5015 mode would result in "successfully" repairing only
about 11 percent of the originally failing vehicles. The distribution of
those 42 passing vehicles is given in the following table.
90
-------
an gailin 1.00 /a fl/9 0
Passing after Reducing 5015 Mod* by 80 or 90%
NOx S 2.0
NOx > 2.0
Laboratory Sample
FTP HC/CO Emlaalona
Pass
0
0
Marginal
1
0
High*
0
0
V. High
0
0
Super
0
0
Simulated Lane Reel
Pasa
0
0
FTP HC/CO Emissions
Marginal
42
0
High
0
0
V. High
0
0
Super
0
0
Rather than attempting to reduce the emissions on the 5015 mode by a flat
percentage, the mechanic could target the typical emissions on the 5015 mode
of vehicles whose ASM composite emissions pass the 1.00/8.0/2.0 outpoint.
Such a repair strategy would not change a single failing vehicle into a
passing vehicle in our model.
If mechanics were able to repair those 372 vehicles so that the emissions
(HC, CO, and NOx) on the 2525 mode were reduced by 80 percent (while the
emissions on the other three modes remained unchanged), then only 30 of those
372 would be able to pass the 1.00/8.0/2.0 outpoint. The distribution of
those 30 "successfully" repaired vehicles is given in the following table.
Flaafc Plgfeglhiifcion galling 1.00/9.0/2.0
Passing after Reducing 2525 Mode) by 80%
NOx S 2.0
NOx>24
Laboratory Sample
FTP HC/CO Emissions
Pass
2
0
Marginal
1
1
UlMlM
niytW
0
0
V. Htgh
1
0
Super
0
0
Simulated Lane Reel
Paae
12
0
FTP HC/CO Emleelone
Mtrgbwl
6
6
High
0
0
V. High
6
0
Super
0
0
Reducing the 2525 mod* emissions by 90 percent would add 48 vehicles (42
marginal HC/CO emitters with NOx £ 2.0 and 6 very high HC/CO emitters with
NOx > 2.0) to th« 30 whose estimated ASM composite score would pass the
cutpoint. TtaM» * repair strategy that targeted only the 2525 mode would be
successful o«£oaly 21 percent of the originally failing vehicles.
;^rr
Rather thatt attempting to reduce the emissions on the 2525 mode by a flat
percentage, the* mechanic could target the typical emissions on the 2525 mode
of vehicles whose ASM composite emissions pass the 1.00/8.0/2.0 cutpoint.
Such a repair strategy would not change a single failing vehicle into a
passing vehicle in our model.
If mechanics were able to repair those 372 vehicles so that the emissions
(HC, CO, and NOx) on the 50 mph cruise mode were reduced by 80 percent (while
91
-------
the emissions on the other three modes remained unchanged), then 264 of those
372 would be able to pass the 1.00/8.0/2.0 outpoint. The distribution of
those 264 passing vehicles is given in the following table.
Fl««t Piafcribufeinn galling- 1.00/8.0/2 0
Passing after Reducing SO mph Cruise Mode by 80%
NOx £ 2.0
NOx > 2.0
Laboratory Sample
FTP HC/CO Emissions
Pass
3
0
Marginal
5
2
Highs
3
2
V. High
3
2
Super
0
0
Simulated Lane Fleet
FTP HC/CO Emissions
Pass
54
0
Marginal
138
12
High
18
12
V. High
18
12
Super
0
0
Reducing the 50 mph cruise emissions by 90 percent would add 6 vehicles (all
with very high HC/CO emitters and NOx > 2.0) to the 264 whose estimated ASM
composite score would pass the outpoint. Thus, a repair strategy that
targeted only the 50 mph cruise mode would result in "successfully" repairing
about 73 percent of the originally failing vehicles.
Rather than attempting to reduce the emissions on the 50 mph cruise mode
by a flat percentage, the mechanic could target the typical emissions on the
50 mph cruise mode of vehicles whose ASM composite emissions pass the
1.00/8.0/2.0 cutpoint. Such a repair strategy would result in "successfully"
repairing 162 (44%) of the originally failing vehicles. The distribution of
those 162 passing vehicles is given in the following table.
gluat Ptatrthtifcton Failin l.QQ/fl.Q/2.0
Passing after Reducing 50 mph Cruise Mod* to Nominal Score
NOx $24
NOx > 24
Luorttory sa
impto
FTP HC/CO Emissions
Pass
1
0
• •••nlnal
MewgWwl
2
0
Highs
4
0
V. Htah
2
0
Super
0
0
Simulated Lane Rest
Pas*
42
0
FTP HC/CO Emission*
Marginal
84
0
High
24
0
V. High
12
0
Super
0
0
If mechaafe* were able to repair those 372 vehicles so that the emissions
(only HC and CO) on the idle mode were reduced by 80 or 90 percent (the model
yields the same result for each) while the emissions on the other three modes
remained unchanged, then only 96 of those 372 would be able to pass the
1.00/8.0/2.0 cutpoint. Thus, a repair strategy that targeted only the idle
mode would result in "successfully* repairing only about one-fourth of the
originally failing vehicles. The distribution of those 96 passing vehicles is
given in the following table.
92
-------
Flaafc Distribution Falling* 1.00/8.0/2.0
Parsing after Reducing Idle Mode, by 80 or 90%
NOX32.0
NOx>2.0
Laboratory Sample
FTP HC/CO Emissions
Pass
0
0
Marginal
2
0
Highs
1
0
V. Hkjh
1
0
Super
0
0
Simulated Lane Fleet
Paee
0
0
FTP HC/CO Emission*
Marginal
84
0
High
6
0
V. High
6
0
Super
0
0
Rather than attempting to reduce the emissions on the idle mode by a flat
percentage, the mechanic could target the typical emissions on the idle mode
of vehicles whose ASM composite emissions pass the 1.00/8.0/2.0 cutpoint.
Such a repair strategy would result in "successfully" repairing only 60 (16%)
of the originally failing vehicles. The distribution of those 60 passing
vehicles is given in the following table.
F!«»«»» Plafcrlbufclon Pallingr 1 00/9 0/2 0
Passing after Reducing Idle Mode to Nominal Score
NOxs&O
Laboratory Sample
FTP HC/CO Emissions
Pass
0
0
Marginal
1
0
Highs
1
0
V.HIgh
2
0
Super
0
0
Simulated Lane Fleet
FTP HC/CO Emissions
Pass
0
0
Marginal
42
0
High
6
0
V. High
12
0
Super
0
0
Tightening the ASM cutpoint from 1.00/8.0/2.0 to a more stringent cutpoint
of 0.40/8.0/1.5 produces similar results in our model.
An ASM cutpoint of 0.40/8.0/1.5 will fail 587 vehicles in that simulated
lane fleet. The distribution of those 587 vehicles is given in the following
table.
Plat rthtit--ton Falling 0.40/8.0/1.S
NOXS2J
NOx>24
te
t:
PSSS)
6
0
Laboratory Sample
FTP HC/CO Emissions
A^BMMBeWeit
Mewyww
16
4
IMi^tm
raBne
7
3
V.HIgh
13
6
Super
1
0
Simulated Lane Fleet
FTP HOCO Emissions
Pass
72
0
Marginal
311
24
Htah
42
18
V. High
78
36
Super
6
0
As with the 1.00/8.0/2.0 cutpoint, single-mode repairs that reduced the
5015 mode emissions by 80 or 90 percent would succeed in "successfully"
93
-------
repairing only 108 (18 percent of the 587 failing vehicles)
of those 108 vehicles is given in the following table.
The distribution
Flaafc nlafeglbuteion Failing 0 10/9.0/1 5
Passing after Reducing 5015 Mod* by 80 or 90%
NOx * 2.0
NOx > 2.0
Laboratory Sample
FTP HC/CO Emissions
Pass
1
0
Marginal
5
0
Highs
0
0
V.Hfeh
0
0
Super
0
0
Simulated Lane Rest
Pass
6
0
FTP HC/CO Emissions
Marginal
102
0
High
0
0
V. High
0
0
Super
0
0
Rather than attempting to reduce the emissions on the 5015 mode by a flat
percentage, the mechanic could target the typical emissions on the 5015 mode
of vehicles whose ASM composite emissions pass the 0.40/8.0/1.5 cutpoint.
Such a repair strategy would result in *successfully* repairing only 54 (16%)
of the originally failing vehicles. The distribution of those 54 passing
vehicles is given in the following table.
Ifloofr. nlafcglhufcion Falllny 0 10/8 0/1 S
Passing after Reducing 5015 Mod* to Nominal Score
NOx $2.0
NOx > 2.0
Laboratory Sample
FTP HC/CO Emissions
Psss
1
0
Marginal
2
0
Highs
0
0
V.Htah
0
0
Super
0
0
Simulated Lane Fleet
FTP HC/CO Emissions
Pass
6
0
Marginal
48
0
High
0
0
V. High
0
0
Super
0
0
If mechanics were able to repair those 587 vehicles so that the emissions
on the 2525 mode were reduced by 80 percent (while the emissions on the other
three modes remained unchanged), then only 138 of those 587 would be able to
pass the 0.40/8.0/1.5 cutpoint. The distribution of those 138 vehicles is
given in the following table.
94
-------
ylMfc Platributeion galling Q.AQ/a.O/l.S
Passing after Reducing 2525 Mod* by 80%
NOx * 2.0
NOx > 2.0
Laboratory Sample
FTP HC/CO Emissions
Paaa
1
0
Marginal
9
0
Highs
0
0
V. High
1
0
Super
0
0
Simulated Lane Fleet
FTP HC/CO Emissions
Paaa
6
0
Marginal
126
0
High
0
0
V. High
6
0
Super
0
0
Reducing the 2525 mode emissions by 90 percent would add 12 vehicles (6
passing HC/CO with NOx £ 2.0 and 6 marginal HC/CO emitters with NOx > 2.0) to
the 138 whose estimated ASM composite score would pass the cutpoint. Thus, a
repair strategy that targeted only the 2525 mode would be successful on only
about 26 percent of the originally failing vehicles.
Rather than attempting to reduce the emissions on the 2525 mode by a flat
percentage, the mechanic could target the typical emissions on the 2525 mode
of vehicles whose ASM composite emissions pass the 0.40/8.0/1.5 cutpoint.
Such a repair strategy would result in "successfully" repairing only 102 (17%)
of the originally failing vehicles. The distribution of those 102 passing
vehicles is given in the following table.
glMt Distribution gailinqr 0.40/8.0/1.S
Passing after Reducing 2525 Mode to Nominal Score
NOx i 2.0
NOx > 2.0
Laboratory Sample
FTP HC/CO Emission*
Pass
0
0
Marginal
4
0
•jt—f^^
nlQnS
0
0
V.Htah
1
0
Super
0
0
Simulated Lane Reet
FTP HC/CO Emissions
Pass
0
0
Marginal
96
0
High
0
0
V. High
6
0
Super
0
0
If mechanics were able to repair those 587 vehicles so that the emissions
on the 50 mph cruise mode were reduced by 80 percent (while the emissions on
the other three* modes remained unchanged), then 365 of those 587 would be able
to pass thev 0.40/8.0/1.5 cutpoint. The distribution of those 365 passing
vehicles is- given in the following table.
95
-------
Flaat ptafcribiifeian galling 0 40/8.0/1.5
Passing after Reducing 50 mph Cruise Mod* by 80%
NOx S 2.0
NOx > 2.0
Laboratory Sample
FTP HC/CO Emission*
Pass
6
0
Marginal
12
1
High*
4
0
V.Htah
2
0
Super
0
0
Simulated Lane Fla«t
FTP HC/CO Emission*
Paaa
72
0
Marginal
251
6
High
24
0
V. High
1 2
0
Super
0
0
Reducing the SO mph cruise emissions by 90 percent would add 12 vehicles (all
with NOx > 2.0; 6 of which with marginal HC/CO and 6 with high HC/CO) to the
365 whose estimated ASM composite score would pass the outpoint. Thus, a
repair strategy that targeted only the 50 mph cruise mode would result in
"successfully" repairing about 64 percent of the originally failing vehicles.
Rather than attempting to reduce the emissions on the 50 mph cruise mode
by a flat percentage, the mechanic could target the typical emissions on the
50 mph cruise mode. of vehicles whose ASM composite emissions pass the
0.40/8.0/1.5 cutpoint. Such a repair strategy would result in "successfully"
repairing only 275 (47%) of the originally failing vehicles. The distribution
of those 275 passing vehicles is given in the following table.
rimmti Plat rlbufc ion Failing 0.40/8. 0/1. S
Passing after Reducing 50 mph Cruise Mod* to Nominal Score
NOx S 24
NOx > 2.0
Laboratory Sample
FTP HC/CO EmiaaJoro
Pass
2
0
Marginal
8
0
Htobe
4
0
V.HIgh
2
0
Supar
0
0
Simulated Lana Fleet
Paaa
48
0
FTP HC/CO Emlsalona
Marginal
191
0
High
24
0
V. High
12
0
Super
0
0
If mechanics were able to repair those 587 vehicles so that the emissions
on the idle node) were reduced by 80 or 90 percent (the model yields the same
result for each} while the emissions on the other three modes remained
unchanged, tnearonly 42 of those 587 would be able to pass the 0.40/8.0/1.5
cutpoint. Thua>« a repair strategy that targeted only the idle mode would
result in 'successfully* repairing only about seven percent of the originally
failing vehicles. The distribution of those 42 passing vehicles is given in
the following table.
96
-------
Tl««» nlaferibufcion Failing Q.AQ/8.Q/1.S
Paaaing after Reducing Idle Mod* by 80 or 90%
NOx S 2.0
NOx > 2.0
Laboratory Sample
FTP HC/CO Emlaalona
Paaa
0
0
Marginal
1
0
Htaha
0
0
V. High
0
0
Supar
0
0
Slmulatad Lana Float
Paaa
0
0
FTP HC/CO Emlaalona
Marginal
42
0
High
0
0
V. High
0
0
Supar
0
0
Rather than attempting to reduce the emissions (only HC and CO) on the
idle mode by a flat percentage, the mechanic could target the typical
emissions on the idle mode of vehicles whose ASM composite emissions pass the
0.40/8.0/1.5 cutpoint. Such a repair strategy would have produced exactly the
same result (i.e., 42 passing vehicles) as would reducing the idle emissions
by a flat 80 or 90 percent.
From the preceding two examples (i.e., using outpoints of 1.00/8.0/2.0 and
0.40/8.0/1.5), the only potentially effective "single-mode ASM repairs" are
those repairs targeted at the 50 mph cruise mode (reducing emissions by 90%).
However, the model predicts that those repairs would not be successful on 27
to 36 percent of the originally failing vehicles. The distributions, of those
vehicles that are still failing the respective ASM cutpoint after repairs
targeted on the 50 mph cruise mode (reducing emissions by 90%), are given in
the following table.
Sfclll railing 0 .40/8.0/1.S
After Reducing SO mph Cruise Mod* by 90%
NOx £24
NOx>2A
Still Falling Cutpoint of 1.00/8.0/24
Slmulatad Lana Ftoet
FTP HC/CO Emlaalona
Paw
0
0
Marginal
6
6
Hton
12
0
V.Htah
54
18
Supar
6
0
Still Falling Cutpoint of 0.40/8,0/1 J
Slmulatad Lana Raat
FTP HC/CO Emlaalona
Paaa
0
0
Mtrglnal
60
12
HlQn
18
12
V. High
66
36
Super
6
0
From th*r- pr*c*ding table, we can see that the vehicles that the model
predicts will continu* to exceed the respective ASM outpoints, even after
single-mod* repairs targeted at the 50 mph cruise mode, are among the highest-
emitters in the simulated lane fleet.
97
-------
5 . 7 Purge Analyses
5.7.1 Introduction
In purge testing, the concern is not that too many malfunctioning vehicles
will pass a teat. Instead, the major concern is too many properly functioning
vehicles will fail a test that attempts to replace a test with real-world
driving behavior with a few steady-state modes. These steady-state modes may
not provide some vehicles with the opportunity to purge. Thus, the purpose of
this section was to compare the canister purge system false failure rates
(errors-of-commission or Ecs) for the ASM and IM240. The data indicate that
the IM240 is significantly less likely to falsely fail vehicles for purge than
the ASM.
Vehicle evaporative emissions contribute significantly to the VOC
inventory. Because vehicle fuel tanks and carburetors must vent to atmosphere
for proper vehicle operation, carbon canisters are added to collect
hydrocarbon molecules which would otherwise escape. Because the carbon
canister has a finite capacity, which if exceeded, allows hydrocarbons to
escape, the canister must be kept purged of stored hydrocarbon molecules. The
evaporative control system includes a purge system which draws stored
hydrocarbons into the engine where they are burned.
Most properly functioning canister purge systems do not purge constantly;
instead, most only purge when their ECM computer algorithms call for purge.
Oriveability or emission problems accompany purge that initiates during
unfavorable conditions, so purge algorithms are designed to take advantage of
opportune conditions. The purge algorithms are known to vary widely from
model to model. So, the main problem for an I/M test is to provide vehicle
operation that will coincide with the conditions necessary to induce the
system to activate canister purge. EPA has found that some vehicles only
purge during accelerations or decelerations, which is problematic for steady
state tests such as the ASM and could result in falsely failing vehicles with
purge systems that are properly functioning.
Also, iiMgi'ielili In I have timers that don't allow purge for several minutes
after the +*vyt*» is started or a specified operating temperature is reached,
so all els* being equal* the longer the test duration, the lower the
probability of purge false failures. This also makes the test order
important, since the test that was performed second is more likely to achieve
purge than the initial test. This is one reason the test procedure in our
study in Mesa required the test order to be reversed each time another car was
tested and why the engine was restarted just before each test.
98
-------
The IM240 purge flow was summed over the full IM240. The ASM purge flow
was measured and summed over the full four modes of the ASM including
transient segments of the ASM cycle, in contrast to the ASM exhaust emissions,
which were only measured during the four ASM steady-state modes. Because the
flow measuring equipment is the same for both transient and steady-state
tests, it was practical to measure purge flow on the ASM during the three
accelerations and the single deceleration needed to complete the four ASM
modes.
5.7.2 The Database
The database for the lane purge analysis was restricted to vehicles that
met all of the following criteria:
- as-received purge data were available for both the IM240 and the ASM,
the test order was known,
and the data passed the purge QC criteria (see Appendix C).
The resulting database consisted of 1170 vehicles. Of these, 577 received
the IM240 first and 593 received the ASM test first. The comparisons made for
this analysis included failure rate comparisons, and comparisons of vehicles
for which the ASM and ZM240 purge status did not agree, or "false failures".
In addition, these comparisons were made for data stratified by test order.
The standard used for these comparisons was 1.0 liter/test.
5.7.3 The Results
The results of the failure rate comparisons are as follows:
• Overall purge failure rates:
- The IM240 failed 7.43% (87 vehicles).
- The ASM failed 11.45% (134 vehicles).
• Initial test failure rates:
- Th* IM240.1st failed 6.93% (40 vehicles).
- TH* ASM.1st failed 11.13% (66 vehicles).
• Second test failure rates:
- The IM240.2nd failed 7.92% (47 vehicles).
- The ASM.2nd failed 11.79% (68 vehicles).
These higher failure rates for the ASM raise the question of whether the
ASM correctly identified non-purging vehicles that the IM240 missed, or
whether the ASM incorrectly failed vehicles. Passing the ZM240 purge test
requires either that purge actually occurs or that the measurement system
99
-------
falsely indicate that purge is occurring. Since the ASM and IM240 were run
with the same measurement system, and results were reported electronically
without human intervention, it is not conceivable that a measurement error
made some cars pass the IM240 and fail the ASM. Consequently, the ASM-
fail/IM240-pass cars must be considered improper fails by the ASM, and vice
versa, with a possibility that test order was a contributing factor in
specific cases despite the engine restart for both tests. However, since the
sample has essentially an equal number of each test order, test order should
not be a relevant factor overall.
The next set of statistics implies that both the ASM and the IM240 falsely
fail vehicles, but the ASM falsely fails more vehicles.
• Overall false failure rates (fails one test but not the other):
- 1.1% or 13 vehicles failed the IM240 but passed the ASM.
- 5.13% or 60 vehicles failed the ASM but passed the IM240.
• False failures on initial test:
- The IM240.1st falsely failed 1.21% (7 vehicles).
- The ASM.1st falsely failed 4.22% (25 vehicles).
• False failures on second test:
- The ZM240.2nd falsely failed 1.01% (6 vehicles).
- The ASM.2nd falsely failed 6.07% (35 vehicles).
Figure 5.7.1 graphically illustrates the comparison of false failure
rates.
100
-------
Overall
7.0% •
6.0% •
5.0% •
4.0% -
3.0% •
2.0% -
1.0% -
N-1170
5.1%
•
»
•
•
m
m
figure 5.7.1
Comparison of raise Failure Rate* for
IM240 purge va ASM purge
ASM.2nd
N-577
1.1%
ASM. 1st
N-593
4.2%
If
m
1
Overall - Falsa
Failure Rate
IM240.1st
N-577
1.2%
Initial - False
Failure Rate
6.1%
IM240.2nd
N-593
1.0%
Preconditioned
False Failure
Rate
DASM Purge - False Failure Rate H IM240 Purge - False Failure Rate
The most relevant comparison is the initial tests (ASM.1st 6 IM240.1st)
because they are more representative of the conditions and vehicle
preconditioning expected in official Z/M programs than the preconditioned
tests. The purge results for the initial tests were similar to the overall
results; the ASM.1st'a false failure rate was 3 percentage points higher than
for the IM240.1at.
with a 3 to 4% false failure rate, the ASM purge test could cause severe
problems to Z/M prograaa in the form of frustrated consumers and skeptical
mechanics.
As discusMd in the introduction, test order was expected to be important
because the. tejrfc that was performed second would b« more likely to achieve
purge than th« teat that waa performed first. Contrary to expectations,
however, the ASM.2nd exhibited a 0.63% increase in failure rate and a 1.85%
increase in false failuxea when compared to the ASM.1st. The the IM240.2nd
also produced a higher failure rate (+1.0%) than the ZM240.1st, but the
IM240.2nd's false failurea decreased 0.2% compared to the ZM240.1st.
Zn addition, the falae failure rate for the IM240.2nd is markedly better
than for the ASM.2nd when viewed as a percentage of failurea. Figure 5.7.2
101
-------
shows that the false failure rate dropped from 17.5% of the failing IM240.1st
vehicles to 12.8% of the failing IM240.2nd vehicles. In contrast, the false
failure rate increased to 51.5% of the ASM.2nd failing vehicles from 37.9% of
the ASM.1st failing vehicles. So although the false failure rate for both
tests is expected to decrease further if the engine restart is avoided before
performing the second-chance test, these data suggest that retesting is more
effective in reducing IM240 false failures than ASM false failures.
Figure 5.7.2
False Failure Rates as a Percentage of Failures
for IM240 purge vs ASM purge
Overall - False
Failure Rate
Initial - False
Failure Rate
Preconditioned
- False Failure
Rate
D ASM Purge-False Failure Rate
IM240 Purge-False Failure Rate
Overall, 74 vehicles failed both the ASM and IM240. The ASM falsely
failed 60 additional vehicles while the IM240 falsely failed only 13
additional vehicles. As shown in Figure 5.7.2, 44.8% of the 134 vehicles
failing the-ASM purge were false failures compared to 14.9% of the 87 vehicles
failing th*rXM240 purge. In addition, 37.9% (25 of 66) of the ASM.1st falsely
failed compered to 17.5% (7 of 40) IM240.1st false failures. Of the 68
vehicles that failed the ASM.2nd 35 had purged on the IM240, so 51.5% of the
ASM.2nd failures were false failures, whereas only 6 of the 47 vehicles that
failed the IM240.2nd had purged on the preceding ASM, so 12.8% of the
IM240.2nd failures were false failures.
These results indicate that the ASM error-of-commission rate will be
intolerably high with one third to one half of failing vehicles being false
failures. I/M programs could implement a second chance test immediately
following the first test without shutting off the engine to reduce false
102
-------
failures. However, it La speculative whether this will significantly reduce
the ASM false failure rate. Also, second-chance testing adds cost. Since the
false failure rate increases from the ASM.1st to the ASM.2nd and decreases for
the IM24.2nd compared to the IM240.1st (See Figure 5.7.2), the data indicates
that second-chance testing may not be as effective for the ASM as for the
IM240.
Second-chance testing costs will be lower for the IM240 because it fails
fewer cars initially, thus, requiring fewer retests and some vehicles just do
not purge during steady-state operation. For these vehicles, an alternate
cycle such as the IM240 would be required. Dynamometer costs would then
increase because inertia simulation is needed, but the more expensive IM240
exhaust measurement systems would not be needed.
In summary, the comparison of ASM purge and IM240 purge shows that the
IM240 is superior in correctly identifying vehicles with malfunctioning purge
systems. With false failure rates of 4 to 6% for the ASM (3 to 6 times higher
than IM240 false failure rates), an additional second-chance test will be
required. And since some vehicles simply do not purge on the ASM steady-state
modes, even with purge measured during the accelerations between modes, an
alternate cycle such as the IM240 may be required for retests. In conclusion,
the ASM purge test is substantially less effective than the IM240 purge test.
5.8 IM240 Improvement* and the roar-Mod* 1X240
The purpose of this section is to convey that refinements are possible
which would make the IM240's performance a "moving-target," and to further
reiterate why one sample should be used to develop the ASM-mode weighting
factors and an independent sample used to evaluate the ASM's effectiveness.
EPA's recommended IM240 outpoints of 0.80/15.0/2.0 + 0.50/12.0 represent a
compromise between failing high emitting vehicles and not failing clean
vehicles. As outpoints are tightened, the IDRs generally increase at the
expense of increasing the possibility of errors-of-commission. Increasing the
power of the te«t (i.e./ the ability to distinguish between malfunctioning and
properly functioning vehicles) serves the public good, in that the high
emitters not identified by the test are not repaired, so the cost of testing
such vehicles is not rewarded by air quality improvements that accrue from
identifying and repairing such vehicles. More tangible is that vehicle owner
satisfaction and acceptance of I/M programs increase with lower errors-of-
commission.
EPA is not content for all time with the absolute performance of the IM240
as now defined. Although in a relative sense, its performance is superior to
103
-------
any of the alternative Z/M teats, it can be improved. Consider for example the
IM240'a 15 g/mi CO outpoint. This is more than four times the FTP CO
standard, and unlike the FTP which includes a cold start9, the IM240 is to be
performed on fully warmed up vehicles. So the errors of omission (vehicles
which pass, but should not) are higher than if a more stringent CO standard
were used. EPA testing has shown that tighter outpoints will identify more
high emitters, but also fail some properly functioning vehicles. Although the
IM240 is considerably better than any alternative Z/M test, in this regard,
there is no question that its performance can be improved. The IM240's
performance can be improved in two areas:
Reduce the test-to-test variability so that outpoints can be tightened
without falsely failing clean vehicles.
- Use statistical techniques to improve the IM240's correlation with the FTP.
Such improvements will serve the public interest by increasing the air
quality yield per test-dollar, so alternative tests should be evaluated
against the state-of-the-art of IM240 testing rather than the ZM240
performance, as it existed, when the Z/M Rule was published. Proponents of
alternative Z/M tests may point out that if the IM240's performance, as it
stood in November 1992, was good enough to meet the performance standard, then
this performance standard should be the standard for alternative tests. While
such a policy may indeed "level-the-playing-field" for alternative tests and
is in fact what is allowed by EPA'3 Z/M Rule, it is difficult to argue that
this approach promotes the general welfare and should guide state and local
decision-makers concerned as much about clean air as about meeting minimum
requirements.
5.8.1 R«duc« T«*t-to-T«*t Variability
Test-to-test variability is the primary reason why the IM240's outpoints
are so much less stringent than the FTPs. The FTP controls a number of
variables that are widely known to affect a given vehicle's emissions. Some
variables that. ar« tightly controlled for FTPs were either more loosely
controlled oc:Dot controlled in EPA's IM240 lane tests. These include, among
others:
9 CO (and HC) emissions are considerably higher during warmup than during
fully warmed-up operation.
104
-------
- ambient temperature in the test cell
- humidity in the test cell
- engine temperature (FTP indirectly controls with preconditioning and
ambient conditions)
- catalyst temperature (FTP indirectly controls with preconditioning
and ambient conditions)
- vehicle operation prior to the emissions test (can affect emission
control system timers for purge, air switching, etc., and other
variables affecting emissions)
evaporative canister loading (FTP indirectly controls with ambient
conditions and vehicle operation during the 12 hours preceding the
FTP emissions test)
tire pressure
speed excursions from the nominal speed (±2 mph on FTP vs. driver
discretion for EPA's pilot IM240 testing to date)
exhaust system backpressure (NOx can be adversely affected by a
constant volume sampler if quality control is not adequate)
fuel composition
EPA has already made improvements that I/M programs will be required to
implement, but were not implemented during EPA's testing. For example, FTPs
are voided if speed excursions from the nominal speed exceed ±2 mph. In
contrast, much of EPA's data are from vehicles with speed excursions that
exceed ±2 mph. In a committee that included I/M contractors, state I/M
program officials, IM240 equipment manufacturers, and automobile
manufacturers, a consensus was reached on requiring this tighter speed
tolerance along with additional tighter controls that will reduce test-to-teat
variability10.
There are also variables that can not be controlled, such as ambient
temperature and canister loading, but can be compensated for to better
distinguish between malfunctioning and properly functioning vehicles. Given
enough data, computer algorithms can be developed that consider the more
important variables and apply adjustment factors to the official IM240 test
results.
Simplistic." approaches such as setting tire pressure or providing second-
chance tests for vehicles that are within 1.5 times the cutpoints seem costly
to implement because additional I/M lanes and personnel are needed, but are
judged to be cost effective since vehicles that should not fail but do, must
be retested after "repairs" anyway.
10 Draft High-Tech Test Procedures, Quality Control Requirements, and
Equipment Specifications, April 5, 1993.
105
-------
More sophisticated algorithms utilizing sensors to measure variables like
ambient temperature and catalyst temperature allow relationships to be
developed and used to compute scores. These are more efficient because they
can increase the power of the test without requiring additional I/M lanes and
personnel. Developing these techniques will require substantial data. When
the IM240 is implemented, much data will become available to allow development
of such algorithms.
5.8.2 Statistical Techniques to Improve the IM240's
Correlation With the FTP
Presently, the IM240 score constitutes the sum of the mass emissions
divided by the distance accumulated. Because almost every second of operation
is taken from various segments of the FTP, the two tests correlate better than
any existing alternative I/M test. But their correlation can be improved
using multiple regression. For example, the uncontrolled variables that
attend IM240 tests probably make it appropriate to de-emphasize the initial
operation of the IM240 in computing the score and emphasizing the later
operation. The later operation is somewhat preconditioned by the initial
operation. Also, the IM240 has a higher average speed than the FTP, so de-
emphasizing the high speed portions should produce a better correlation with
the FTP.
The data itself can be used to determine the more appropriate weighting
through the use of regression techniques. For example, EPA divided the IM240
into four modes as follows:
Mode 1: 0-60 seconds
Mode 2: 61-119 seconds
Mode 3: 120-174 seconds
Mode 4: 175-239 seconds
As for th* ASM, coefficients are developed by performing a multiple
regression wn*x«in the results from four modes are the independent variables
and the FTP results is the dependent variable, which allows the data to
determine th« appropriate mode weighting.
EPA tried this using the only substantial database with the information
needed (FTPs with IM240 4-mode results or second-by-second results), which
happens to be the vehicles on which this report focuses. So the coefficients
had to be developed on the same set of data to which they were applied. EPA
condemns this practice, as discussed in Section 5.5, but having no
106
-------
alternative, the results are presented only to provide an indication of how
the IM240's performance is enhanced through this approach.
Unfortunately, the legitimate performance increase gained by using
multiple regression to determine appropriate mode weighting can not be
isolated from the inappropriate application of these coefficients to the same
vehicles from which they were developed. So the performance presented in Table
5.8.2 gives an overly optimistic view, but also reiterates that ASM-advocates
who do not accept EPA's judgement that it is inappropriate to apply
coefficients to the vehicles from which they were developed, should then
compare the ASM performance with inter-linked coefficients to the IM240 also
utilizing inter-linked coefficients.
The multiple regression was performed on the first 91 vehicles in the
database only (this analysis was not repeated when additional data became
available). The results are presented in Tables 5.8.1 and 5.8.2. The
negative coefficients in Table 5.8.1 indicate that insufficient data is
available for developing logical coefficients, which will compensate, to some
degree for the inter-linked performance listed in Table 5.8.2.
Tabl« 5.8.1
Coefficients Developed froa
Multiple Regression of 4-Mode IM240 va. FTP
Mod*
Constant
1
2
3
4
R*
HC
0.03
-0.30
1.16
0.26
0.09
90.3%
CO
-0.28
-0.18
1.70
-0.01
-0.08
86.0%
NOx
-0.02
-0.07
0.37
0.02
0.56
69.6%
Table 5.V.2: illustrates how the 4-Mode test improves the tradeoff between
IDRs and Be rates at equivalent failure rates.
107
-------
Table 5.8.2
4-Mod* IM240 Performance Versus Normal ZM240
Fail
Rate
12%
13%
14%
15%
19%
20%
23%
ac
4 -Mode
88%
90%
91%
92%
91%
94%
95%
Regular
88%
90%
91%
91%
93%
93%
93%
IDRS
CO
4 -Mode
65%
66%
66%
66%
72%
72%
73%
Regular
63%
65%
66%
66%
68%
69%
69%
NOz
4 -Mode
72%
75%
78%
75%
83%
86%
88%
Regular
75%
76%
78%
78%
82%
82%
83%
ECS
4 -Mode
0.0%
0.0%
0.4%
0.4%
0.4%
0.4%
0.7%
Regular
0.4%
0.4%
0.7%
0.7%
1.4%
1.8%
3.9%
Notice the performance increase in that the IDRa increase and errors-of-
commission decrease.
In conclusion, developers of alternative I/M tests should not consider the
performance of the IM240 to be fixed, while better than any existing I/M
tests, IM240 improvements are possible and desirable. EPA's mission to
improve air quality and enhance the public welfare necessitates evaluating
alternative tests, not against the performance of the IM240 as it was in
November 1992 when the I/M Rule was published, but instead, against the state-
of-the-art.
108
-------
6 . Test Programs by Other Organizations
6. 1 Colorado Test Program
The Colorado Department of Health (CDH) completed an evaluation11
comparing the FTP and the IM240 to the following eight I/M test modes, in the
order the modes were performed:
- 35 mph road load
50 mph road load
- ASM 2545
- ASM 2525
- ASM 5015
- Idle test
- 2500 rpm
Their conclusions included the following:
"The loaded mode [IM240] tests (both [93] second and 240 second]
identify significantly more of the excessively emitting vehicles and more of the
excess emissions than do any of the steady-state tests. They also have fewer
errors of commission and less sensitivity to differences between FTP and short
test emission levels. With no other consideration, either the 95 second or the
240 second version of the [IM240] would be the clear choice for the most
accurate and effective identification of excessively emitting vehicles."
6.2 California Test Program
EPA has received a preliminary analysis12 from the California Air
Resources Board (CARS) comparing the ASM5015 and the ASM2525 to the IM240.
The CARB analysis looks favorably on the ASM modes and concludes that the ASM
tests are as effective as the IM240. However, there are significant concerns
with CARB's data. Thesa concerns include the following:
The CMS database is not representative of the newer fleet.
11
Ragazzi, et al.
12 Draft Memorandum from Jeff Long, Manager, Analysis Section to Mark
Carlock, Chief, Motor Vehicle Analysis Section, "Comparison of Excess
Emissions Identified by IM240, ASM5015 and ASM252S Tests," California Air
Resources Board, not dated, received April 15, 1993.
109
-------
- GARB'S testing is not representative of actual I/M testing.
- All of CARB's tests were preconditioned.
- CARB's ASM equations, when applied to EPA'3 data, demonstrate poor
performance.
EPA is preparing a separate document that will consider CARB's ASM test
program in more detail.
110
-------
7 . Teat Coat* Comparison
Supporters of the ASM have frequently suggested that it would be a more
cost-effective test than the IM240, given that the equipment cost is
significantly lower. As Table 5.12.1 shows, the equipment package for the ASM
series, with purge and pressure testing, does have a lower total cost than the
IM240, purge, and pressure equipment package.
Table 7.1
Equipment Costa for the ASM Series and IM240
IM240 ASM
Equipment Coat. Equipment Coafc
Pressure Rig $600 Pressure Rig $600
Purge Meter $500 Purge Meter $500
VDA $1,000 VDA $1,000
Dynamometer $25,000 Dynamometer $20,000
CVS & Analyzers $79,00013 BAR90 & NOx Bench $19,000
Total $106,100 $41,100
These figures reflect the most recent cost information that EPA has
received from industry. EPA has published previous estimates of the per
vehicle costs of ASM and IM240 testing in "X/M Costs, Benefits, and Impacts,"
in November, 1992. EPA found, and independent analyses confirmed, that
equipment costs, when spread over the useful life of the equipment, constitute
a relatively small portion of the per vehicle cost of an I/M test; labor and
•overhead costs are considerably higher. In analyzing the current average per
vehicle inspection cost in a centralized program of $8.50, EPA estimated that
equipment accounted for 21$, labor for 96C, 82$ went to defray construction
costs, the state oversight fee averaged $1.25, and the remaining $5.26 went to
cover various overhead costs (for a full discussion of EPA's cost estimation
assumptions and methodology the reader is referred to Sections 5.2 and 5.3 of
"I/M Costs, Benefits, and Impacts," contained in Appendix H). Current testing
stations have an average peak capacity of 25 vehicles per hour and enough
stations are constructed to avoid long lines on peak demand days. Given the
typical pattern of owners' choices about when to come for inspections, this
results in an average actual throughput of 12.5 vehicles per hour which
13 Letter from Kenneth W. Thomas, Marketing Manager, I/M Systems, Horiba
Instruments Incorporated, to Bill Pidgeon, U.S. Environmental Protection
Agency, April 7, 1973 and Quotation from Scott P. Corrunker, Sales Engineer,
Combined Fluid Products Company to Dan Sampson, U.S. Environmental Protection
Agency, January 27, 1993. These are attached as Appendices J and 1C.
Ill
-------
translates into 39,000 vehicles per year per lane, and costs are spread over a
multi-year period, five years in moat cases.
Throughput is the most critical variable in estimating costs since it
determines the size of the. inspection station network needed for a given area
and the number of vehicles over which costs for each lane are spread.
Inspection lanes usually have more than one position, with different parts of
the inspection performed at each one. Hence, throughput is governed not by
the time required to perform the total test sequence, but by the time required
at the longest position. Whether the test sequence consists of the IM240 with
purge and pressure testing or the ASM with purge and pressure testing, the
longest part of the sequence is the tailpipe emissions test.
The combined IM240 and purge test takes approximately three minutes (using
fast-pass and fast-fail) to perform on the average. Allowing an additional
minute to maneuver the vehicle onto the dynamometer and otherwise prepare the
vehicle for testing the total time at the longest position is estimated to be
four minutes. This translates into a peak lane capacity of 15 vehicles per
hour and an average actual throughput of 7.5 vehicles per hour. The ASM
consists of four modes lasting 40 seconds each with a few seconds in between
to change speed. This works out to approximately three minutes per test.
Allowing, again, an additional minute to maneuver the vehicle onto the
dynamometer and otherwise prepare it for testing, the total test time is about
four minutes, hence, the throughput rates for the ASM is the same as for the
IM240. Average throughput for both tests is 7.5 vehicles per hour. Assuming
that stations operate 60 hours per week, 52 weeks per year, and costs are
spread over a five year period, then equipment costs are spread over a total
-of 117,000 vehicles.
The optimum lane configuration for both tests is a three position lane
staffed by three inspectors. Consequently, as shown in "I/M Costs, Benefits,
and Impacts," staff, infrastructure and overhead costs are essentially the
same for both tests. The only difference is in the cost of equipment. Table
5.12.2 shows the estimated per vehicle costs for performing the ASM and the
IM240. Th» coats are derived using the same methodology and assumptions as in
Appendix R. Overhead costs for IM240 and ASM tests are estimated by factoring
the overhead for current centralized programs by the change in throughput.
Equipment, and construction costs are obtained by dividing those costs over
the total vehicle traffic in a five year period. Staff costs are obtained by
dividing inspectors' hourly wages (96.00) by the average number of vehicles
inspected in a hour. State oversight costs are estimated at $1.75 per vehicle
but could vary depending upon the intensity of the state oversight program;
they would not vary between the two test types.
112
-------
Despite the difference between the costs of the equipment packages
required for the two tests, the total cost per vehicle, factoring in all
necessary costs involved in a testing program, differs very little between the
two tests. In a high volume test program the per vehicle cost difference is
estimated at 74$; the per vehicle cost for the ASM is about 5 percent less
than for the IM240..
Table 7.2
Cost Component* and Coat per Vehicle for the ASM and XM240
IM24Q
$2.40
$1.75
$1.39
$1.71
$9.12
$16.37
Inspection Staff
State Oversight
Test -Equipment
Building Modification/Construction
Other Overhead
Total Cost Per Test
ASM
$2.40
$1.75
$0.65
$1.71
$9.12
$15.63
113
-------
8 . Evaluation of the Adequacy of the ASM for Enhanced I/M
• Programs
8 . 1 Introduction
The preceding chapters show that the four-mode ASM test is not equivalent
to the IM240 on a per-car basis. Even if ASM outpoints are selected so that
the same number of cars are failed, they will represent a smaller portion of
the fleet's excess emissions, and the cars will not be repaired as effectively
as if the IM240 were used for reinspection after repair. However, to some
extent this loss of emission reduction can be compensated for by improving
other I/M program features to make them more stringent than would otherwise be
required to meet the emission reduction performance standard in EPA's rule for
enhanced I/M programs. Among these other features are the inspection of
heavy-duty gasoline-fueled vehicles, the use of the ASM test for all 1981 and
newer vehicles rather than just the 1986 and newer vehicles which are assumed
to be tested with the IM240 in the model I/M program, a higher failure rate
for pre-1981 vehicles, purge testing for more model years than in the model
program, and more comprehensive tampering inspections.
Whether these improvements are enough to offset the loss of benefit
from the ASM is the decisive question that determines whether areas subject to
the enhanced I/M program requirement can rely on ASM testing instead of IM240
testing. Also of interest is whether it is possible to use the ASM and still
operate a biennial program. To answer these questions, EPA examined annual
and biennial scenarios in which the ASM cutpoints were made as stringent as
EPA believes is consistent with good engineering practice and the possible
.offsetting program improvements were made as large as EPA considers reasonably
possible. If this hypothetical best-possible ASM program cannot satisfy the
enhanced I/M performance standard, then no ASM program can.
Regarding best-possible ASM cutpoints, EPA haa assumed that the
failure rate associated with the most stringent IM240 cutpoints for which EPA
has provided emission reduction credits is the limit of good engineering
practice in aa-I/M program. These IM240 HC/CO/NOx cutpoints are 0.6/10.0/1.5,
compared to th% 0.8/20.0/2.0 used in the model enhanced Z/M program. The ASM
cutpoints that matched this failure rate in the full Mesa lane sample were
0.40/8.0/1.8. These ASM cutpoints can be expected to produce a higher error
of commission rate than the 0.6/20.0/2.0 IM240 cutpoints, but in the interest
of exploring the limits of ASM testing, EPA assumed that this did not make
them unacceptable. EPA calculated MOBILESa I/M credits for these ASM
cutpoints, using the same basic approach as originally used for the IM240
credits. We then used MOBILESa with these credits and appropriate assumptions
for the offsetting program improvements to determine the overall benefit of a
114
-------
beat-possible hypothetical ASM program. Further description of this process
follows.
8.2 MOBXLBSa Analysis
The I/M credits for the ASM teat procedure were determined using the
identification rate from the Arizona teat sample. The laboratory sample was
weighted as described in Section 5.2.5.8 to reverse the effect of the
recruitment bias. The fraction of total emissions identified by the ASM test
with beat-possible outpoints and the IM240 teat with ita atandard cutpoints
were determined for that sample*. Using the IM240 results for the Arizona
sample, the ASM identification rates were converted to a fraction of the IM240
results. These fractions were then used in the I/M credit model to adjust the
IM240 identification rates used in MOBILE 5 to represent the effect of the ASM
test.
For repair effects, based on current information, EPA can only give the
ASM test the same repair effect as the 2500/Idle test procedure for HC and CO.
For NOx, the ASM test was temporarily assumed to have the same repair effect
as the IM240 test procedure using a 2.0 NOx cutpoint, the nearest available to
the 1.8 ASM cutpoint. At this time, we made this temporary assumption for NOx
so that the ASM program can be analyzed for all three pollutants even though
the repair effectiveness problems found for HC and CO appear to be similar for
NOx. Unlike HC and CO, there is no set of alternative repair effectiveness
numbers available that could be used since steady-state tests have never been
used for NOx control in the past.
Using the ASM credit set described above, we proceeded to perform MOBILESa
runs for four separate I/M program scenarios: a no-I/M run, an enhanced I/M
performance standard run, and two ASM runs, one assuming an annual testing
program, and the other, a biennial program. All four scenarios assume
national default inputs for the local area parameter record - including
vehicle registration mix, ambient temperature, average VMT, fuel RVP, average
speed, etc. - and cover evaluation years ranging from 2000 to 2011. Depending
on the ozone classification, states must show in the 1993 SIP that the I/M
program selected meets the performance standard in these evaluation years.
Both ASM. runs were identical, with the exception of the above-noted
difference in teat frequency. The other program parameters assumed for the
* For convenience in calculations, MOBILESa I/M credits for a particular test
and cutpoints are developed starting with the total emissions identification
rate, rather than the excess emission identification rate used in earlier
sections, to more readily display the relative effectiveness of tests. The
difference does not affect the final result.
115
-------
ASM runa include a program atart year of 1983, a teat-only network, and ASM
testing of modal year 1981 and later light-duty vehicles and light-duty
trucks. The ASM runa also assumed evaporative system purge and pressure
testing, and visual inspection of the catalyst, inlet restrictor, gas cap, air
pump, EGR, tailpipe lead test, and PCV system on all 1971 and later model year
vehicles. Full purge benefits were given for ASM testing, since ASM purge
testing will fail virtually all cars that would fail the IM240 test. A pre-
1981 stringency of 40% was assumed, along with a 3% waiver rate and a 96%
program compliance rate.
Once these MOBILESa runs were complete, we compared the results for the
enhanced I/M performance standard run and the ASM runs with the no I/M case to
determine what percent reduction was required to meet the performance standard
and what reductions could be expected from the annual and biennial ASM
programs modeled. The results are shown in Table 8.1.
116
-------
Table 8.1
MOBXLXSa Emission ractora and Reductions from ASM Tasting
2000 No I/M
Enhanced Performance Standard
Maximum Annual ASM
Maximum Biennial ASM
2001 No I/M
Enhanced Performance Standard
Maximum Annual ASM
Maximum Biennial ASM
2003 No I/M
Enhanced Performance Standard
Maximum Annual ASM
Maximum Biennial ASM
2006 No I/M
Enhanced Performance Standard
Maximum Annual ASM
Maximum Biennial ASM
.2008 No I/M
Enhanced Performance Standard
Maximum Annual ASM
Maximum Biennial ASM
2011 Ho I/M
Enhanced Performance Standard
Maximum Annual. AM?
-*c
Maximum Biennirf ASJT
voc
g/ra Redux OK?
2.88
1.96 32.0%
2.00 30.5% NO
2.07 27.9% NO
2.66
1.68 36.6%
1.81 31.8% NO
1.87 29.4% NO
2.52
1.53 39.2%
1.71 32.3% NO
1.76 30.1% NO
2.47
1.47 40.3%
1.66 32.6% NO
1.72 30.4% NO
2.39
1.39 41.8%
1.60 33.1% NO
1.65 30.9% NO
CO
g/m Redux OK?
22.23
13.98 37.1%
15.08 32.1% NO
15.79 29.0% NO
NOx*
g/m Redux OK?
2.27
1.97 13.5%
1.93 15.0% YES
1.96 13.9% YES
2.10
1.77 15.8%
1.76 16.3% YES
1.78 15.0% NO
2.02
1.67 17.2%
1.67 17.0% NO
1.70 15.7% NO
1.97
1.62 17.8%
1.63 17.4% NO
1.66 16.1% NO
1.94
1.58 18.8%
1.60 17.6% NO
1.62 16.3% NO
* With temporary assumption for NOx repair benefits, as described in text.
By comparing tha ASM results to the performance standard, we conclude that
neither tha my»|m"r» annual nor tha i"*»tj"mn biennial ASM program would meet the
performance standard for HC or CO for any of tha milestone years. For NOx,
the biennial ASM program with tha temporary assumption for NOx repair benefits
meets the performance standard in 2000, but misses it for each successive
117
-------
milestone, while the annual ASM program meets the performance standard through
the 2003 milestone.
These NOx results include a caveat, however. The degree to which the ASM
NOx benefit in the table exceeds the performance standard is quite small. If
the percent NOx repair benefit for ASM testing is anything less than 90%
(i.e., 13.5%/15.0%) as good as for IM240 testing, the Maximum Annual ASM
program will not meet the NOx performance standard in 2000. The corresponding
"actual values" for the Maximum Biennial ASM program in 2000 and the Maximum
Annual ASM program in 2003 are 98% (13.5%/13.8%) and 97% (15.8%/16.2%),
respectively. While EPA for the present reserves judgment on exactly how much
NOx repair benefit is lost with ASM testing, (while we consider a test program
to further explore this question) it is clear from Section 5.3 that the loss
is certainly at least 10%. Thus, ASM testing cannot meet the performance
standard for any pollutant for any milestone date, and therefore is not an
acceptable test in any enhanced X/M program.
118
-------
9. Appendices Table of Contents
Appendix A Test Procedures A-l
Appendix B Emissions Data Listing for Vehicles Receiving Both Lane and
Laboratory Tests B-l
Appendix C QC Criteria for ASM/IM240 Database C-l
Appendix D IM240 Cutpoint Tables D-l
Appendix E ASM Cutpoint Tables E-l
Appendix F Scatter Plots and Regression Tables F-l
Appendix G ARCO, Sierra/ Environment Canada Data Analysis G-l
Appendix H Estimated Cost of High-Tech I/M Testing H-l
Appendix I ASM and IM240 Credits for State Implementation Plans With
MOBILES Runs 1-1
Appendix J Emissions Analyzer Price Information from Horiba J-l
Appendix K Centrifugal Blower Price Quotation from Combined Fluid
Products Company K-l
Appendix L Peat-Pas* and Fast-Fail L-l
119
-------
Appendix A: Test Procedures
-------
Test Procedure to Evaluate the Acceleration Simulation Mode and
the Emissions Measurement Capabilities of a BAR90 Certified
Analyzer With An Integrated Fuel Cell Type NO Analyzer
1.0 Objectives
The objective of this ASM test project was to collect data to compare the
effectiveness of a four-mode steady-state test procedure as an alternative I/M
test to the IM240. Emissions and canister purge flow data were collected
using the following vehicle operating modes:
Two Acceleration Simulation Modes (5015 and 2525)
A 50 mph steady state mode at road load
- An idle mode in Drive
- An idle mode in Neutral
These modes will subsequently be referred to as the ASM test. The same
data were collected for the IM240 test.
The lower cost of the emissions measurement equipment is the salient
feature that makes the ASM attractive to its proponents. Therefore, EPA made
ASM emissions measurements using a certified BAR90 analyzer (for HC, CO, and
C02) with an integrated NO analyzer of the fuel cell type for NO measurements.
For the IM240 a CVS-based emissions measurement system was used.
Canister purge flow measurements were made with the same 0-50 liter per
minute for both the ASM and IM240.
The testing was carried out in two locations, a single I/M lane in an
official Arizona I/M station and at a laboratory owned by Automotive Testing
Laboratories (ATL). Both were located in Mesa, Arizona.
2.0 Phoenix Lane Procedure
The following is a description of the I/M lane procedures.
This procedure was restricted to 1983 and newer light duty vehicles
with fuel injection, when available. Carbureted 1983 & newer
vehicles were tested when fuel injected vehicles were unavailable.
A-2
-------
Pre-1983 light duty vehicles were tested only when 1983 and newer
vehicles were unavailable.
- Each light duty vehicle received:
1. The ASM test that included the following modes in the sequence
listed:
- ASM5015 with purge,
- ASM2525 with purge,
50 mph at road load, with purge,
- idle test (automatic transmissions in drive),
idle test (automatic transmissions in neutral) for the first 50
cars. Car 51 and subsequent cars will not get the 5th mode.
These four or five modes will be referred to as the ASM series.
2. An IM240 with purge.
3. A pressure test.
4. An Arizona State I/M test.
2.1 Procedure Sequence
In general all odd numbered vehicles got the IM240 as the initial
test and all even numbered vehicles got the ASM as the initial test.
- Data collected included a number 1 or 2 in a field named "Test.Order"
to designate whether the ASM series procedure was run first or
second. Discrepancies between the Test.Order entry and even/odd
vehicle numbers are resolved by relying on the Test.Order entry, as
this was ATL's primary means to identify test order.
2.2 Measurement Equipment
For the ASM series, a certified BAR90 HC/CO/C02 exhaust emission
analyzer was used to measure HC, CO, and CO2, with an integrated NO
analyzer using a fuel cell sensor. ATL only acquired a NO
analyzer/BAR90 analyzer combination that provided second-by-second
data for HC, CO, C02, and NO. The data output for the ASM test went
A-3
-------
to 3-1/2 inch floppy discs that included run number, time (sec), mode
number, vehicle speed, purge flow, NO (ppm), HC (ppm), CO2 (%), CO
(%), actual torque, required torque, actual horsepower and required
horsepower.
- A 50 liter/min Sierra flow meter was used to measure total canister
purge flow. The flow meter system output was the cumulative second-
by-second data for total flow recorded on the 3-1/2" floppy discs
discussed above.
For the IM240, normal measurements with the CVS system continued at
the lane. The data collected included time (sec), bag number,
ambient measurements, NOx (grama/second), HC (g/sec), CO2 (g/sec), CO
(g/sec), and purge in standard liters.
2.3 Procedure Details
An electric Clayton dynamometer was used for both the IM240 and the
ASM series. The dynamometer horsepower settings for the ASMs were as
follows:
• 5015 HP = (ETW / 250)
• 2525 HP = (ETW / 300)
• 50 Mph HP - Road Load
The horsepower and inertia weight settings for the IM240 were as
normally performed. The minimum inertia weight setting (2,000 Ibs.)
was used for the ASMs.
- Manual transmission vehicles were tested in second gear for both the
ASM5015 and the ASM2525. The 50 mph road load mode used the top non-
overdrive gear, typically 4th gear on a 5-speed, 4th gear on a 4-
speed, and 3rd gear on a 3-speed. Drivers used a lower gears for
vehicles that were lugging.
- The engine was shut off prior to the IM240 and the ASM5015 (as will
normally be done by I/M programs to connect the purge meter),
regardless of which procedure was performed first, and restarted just
prior to initiating these procedures. The engine was not shut off
between ASM modes, and the vehicle was accelerated from the current
mode up to the next mode speed, without first returning to zero.
- The ASM emission sampling period and the canister purge flow
measurement period were as follows:
A-4
-------
1. Each ASM mode was initiated after the vehicle speed had achieved
the nominal speed (15, 25, or 50 mph, and 0 mph idle) ±2 mph. Once
up to speed, emissions sampling of one second average concentrations
continued for 40 seconds. Emission scores for HC, CO, CO2 and NO
were reported for each second. Emissions scores for the first 10
seconds of each mode were ignored to allow the dynamometer to
stabilize and to allow for transport time to the analyzers.
2. The purge flow reported was the second by second cumulative flow
over the entire ASM cycle, including transient accelerations. The
nominal acceleration rate was 3.3 mph/sec., with a minimum
acceleration rate of 1.8 mph/sec and a maximum of 4.3 mph/sec. The
table below lists the minimum, nominal, and maximum acceleration
times used to accelerate from one mode to another. For example, the
table shows that the time to accelerate from 25 mph to 50 mph should
be 7.6 seconds., but can take as long as 13.9 seconds., and as little
as 5.8 seconds. The zero to 60 mph time is provided to indicate how
the specified acceleration times relate to a commonly known reference
of vehicle performance. ATL used a video driver's aid with the
nominal acceleration rate.
Minimum
Nominal
Maximum
Acceleration
Rate
(mph/sec)
4.3
3.3
1.8
Time
0-15
mph
(sees)
3.5
4.5
8.3
to Accelerate from-to:
15-25
mph
(sees)
2.3
3.0
5.6
25-50
mph
(sees)
5.8
7.6
13.9
0-60
mph
(sees)
14.0
18.2
33.3
During the accelerations between modes, the dynamometer load setting
did not exceed road load. This was specified to enhance the
opportunity for canister purge during the ASM accelerations. The
combination of the ASM load and the base 2,000 Ib. inertia may load
some vehicles to heavily to allow purge to initiate.
3.0
Lab Recruitment
Light duty vehicles that received all of the lane tests (IM240, ASM
series, and Arizona I/M test), were recruited for testing at ATL's laboratory.
Cars were categorized as passing or failing using the IM240 cutpoints in the
table below:
A-5
-------
Phoenix Lane IM240 Cutpoints for Lab Procurement
Model
Years
1983+
HC
g/mile
>0.80
CO
g/mile
>15.0
NOx
g/mile
>2.0
The following table provides the laboratory recruitment goals for the
pass/fail categories listed as a percentage of the total number of cars
recruited to the lab for this task. The initial recruitment target was 100
vehicles.
Phoenix Lab Recruitment Goals Using Lane IM240 Categories
Model
Years
1986+
1983-85
HC/CO
Pass
15%
10%
HC/CO
Fail
15%
10%
NOx
Pass
15%
10%
NOx
Fail
15%
10%
4.0 Commercial Repair Recruitment
Owners of vehicles that failed the Arizona I/M test, and received and
IM240/ASM series, were offered $50 to return to the lane for after-repair
tests. These vehicle owners were only offered this incentive if they refused
to participate in the laboratory testing program or if their vehicles were not
needed for laboratory recruitment. Recruiting vehicles for laboratory tests
was a higher priority than for commercial repair participation.
The owners were informed that they must return with repair receipts
indicating repairs by a commercial establishment with itemized labor and parts
costs to qualify for the $50 incentive. ATL included either the original
receipts or copies in the vehicle test packets that ATL provided to EPA. In
addition, ATL provided summarized comments and data for these vehicles on
electronic disk.
Vehicles returning after commercial repairs followed the same procedures.
5.0
Lab Procedure
The lab procedure is summarized in Attachment 1, so this section will only
add explanations to the procedure listed in Attachment 1.
A-6
-------
5.1 Two Groups
The vehicles recruited to the lab were separated into two groups:
1. Those whose initial lane test was the IM240 and were repaired to IM240
targets. For the vehicles in this group, the IM240 always precedes the
ASM series (see Attachment 1).
2. Those whose initial lane test was the ASM series and were repaired to
ASM targets. There were not enough data to set ASM repair targets, so
IM240 targets were used for both groups. For the vehicles in this group,
however, the ASM series always preceded the IM240 (see Attachment 1).
5.2 Repair Targets
The repair targets were to achieve 0.80/15.0/2.0 on the IM240 for both the
ASM-targeted group and the IM240-targeted group. Initially, repair targets
were to be provided to ATL for the ASM targeted group to replace the IM240
targets. However, due to time and data constraints this proved impossible.
For the initial repair attempt, the mechanic was only aware of the lane
IM240 score for both vehicle groups (initial lane test: ASM or IM240). For
subsequent repair attempts, the mechanics were only be aware of lane and lab
IM240 scores. FTP scores were not provided to the mechanics for either group.
Repairs were limited to $1,000.
5.3 Laboratory Test Equipment
Due to time and financial constraints, EPA was unable to develop lab ASM
capability. The IM240 and FTP were measured with a CVS system. Modal or
second-by-second CVS capability was not available at the laboratory.
A-7
-------
Appendix A: Attachment 1
ASM/IM240 Lab Procedure
Revision Date: 10/21/92
Number tested =
Recruitment: 1983+ fuel injected only.
Repairs: Get IM240 Indolene to
.8/15/2.0. The mechanic should only be
aware of IM240 scores for the IM240
targetted repairs. $1,000 repair limit/car
- catalyst if necessary, aftmrkt preferred.
Develop explanations for any IM240
failures that pass FTP, while veh is still at
lab.
Tank Fuel
On-Road Warmup
Tank Fuel IM240
9.0 RVP Indolene As-Received
LA-4 Prep cycle @ SOT
No Diurnal
FTP Exhaust
No Hot Soak
IM240 Indolene (with purge if available)
Repair to get IM240 Indolene to
.8/15/2.0. The mechanic should only be
aware of IM240 scores - not FTP scores,
and only perform minimum repairs
necessary to achieve targets.
Report After-First-Repair Indolene
IM240 regardless of outcome. Mechanic
will only be aware of lane EM240 score
for first repair, not lab tank fuel score.
Continue repairs if necessary. Don't
perform FTP until .8/15/2.0 is achieved.
9.0 RVP Indolene After-Repair to
IM240 0.8/15/2.0
3 LA-4 Prep cycles @ 80°F for all
vehicles
Top off to 40% fill - dont drain.
FTP Exhaust
IM240 Indolene RM1 (w/purge if
available)
Stop repairs even if failing FTP.
Indolene Lane Tests For Vehicles
Whose Initial Lane Test Was IM240
Indolene Lane Tests Procedure for
Vehicles Whose Initial Lane Test Was
ASM Series
On-Road Warmup
Lane Indolene IM240
ASM Series
On-Road Warmup
ASM Series
Lane Indolene IM240
A-8
-------
Appendix B
Data Listings
-------
I. Cutpoint Table Analyses Laboratory Recruited Vehicles
Vehicle Information
Vahf Run* Test Date Order
3148 1672 TST17 921030 IM240.2nd
3149 1685 TST17 921102 IM240.1st
3150 1692 TST17 921030 IM240.2nd
3151 1696 TST17 921102 IM240.2nd
3152 1709 TST17 921103 IM240.1st
3154 1739 TST17 921103 IM240.1st
3155 1735 TST17 921103 IM240.1st
3156 1726 TST17 921104 IM240.2nd
3157 1747 TST17 921104 IM240.1st
3158 1753 TST17 921104 IM240.1st
3159 1752 TST17 921105 IM240.2nd
3160 1749 TST17 921105 IM240.1st
3161 1754 TST17 921105 IM240.2nd
3162 1777 TST17 921106 lM240.1St
3165 1810 TST17 921106 IM240.2nd
3169 1677 TST17 921109 IM240.1st
3170 1879 TST17 921109 IM240.1st
3171 1891 TST17 921118 IM240.1st
3172 1895 TST17 921120 IM240.1st
3173 1804 TST17 921111 IM240.2nd
3174 1688 TST17 921111 IM240.2nd
3175 1907 TST17 921120 IM240.1st
3178 1965 TST17 921118 IM240.1st
3179 1966 TST17 921220 IM240.2nd
3180 2005 TST17 921123 IM240.1st
3181 2015 TST17 921120 IM240.1St
3182 2019 TST17 921120 IM240.1st
3183 2024 TST17 921124 IM240.2nd
3184 2128 TST17 921125 IM240.2nd
3185 2130 TST17 921125 IM240.2nd
3186 2152 TST17 921125 IM240.2nd
3187 2131 TST17 921127 IM240.1st
3188 2160 TST17 921127 IM240.2nd
FTP Scores
HC CO MOx
0.11 1.1 0.97
0.2 2.3 0.2
2.43 82.9 0.59
0.34 3.5 5.81
0.18 3.6 1.01
0.31 6.2 1.04
3.25 46.7 0.26
0.31 6.7 1.07
1.7 14.3 2.14
0.15 2.6 0.85
1.11 74.2 0.31
0.29 3.0 1.26
0.28 5.1 1.7
0.35 3.7 1.25
1.96 13.2 2.5
1.04 15.0 0.96
0.42 7.2 1.16
0.15 3.2 0.52
0.16 3.3 0.73
0.37 6.7 0.82
0.74 16.3 1.88
0.4 13.1 0.46
0.2 1.6 0.82
2.9 77.6 2.06
0.96 9.8 1.22
0.2 3.4 0.48
1.47 26.2 1.12
3.13 66.3 0.7
0.3 3.0 0.48
0.43 7.4 1.27
0.19 2.3 0.17
0.26 2.3 0.66
4.49 17.8 0.2
XM240 SCORES
Composites
HC CO NOx
0.03 1.3 0.55
0.21 2.8 1.19
1.44 32.6 0.95
0.2 2.0 4.54
0.12 2.8 1.34
0.34 4.4 2.24
2.77 24.1 1.2
0.45 6.1 2.07
2.84 34.7 2.95
0.16 2.4 2.44
0.75 41.2 0.49
0.21 2.8 2.27
0.24 4.1 2.42
0.77 6.7 3.08
1.59 7.3 2.48
0.85 14.2 0.98
0.34 7.6 2.02
0.1 2.3 0.46
0.12 2.1 3.3
0.18 3.7 0.84
0.76 19.3 2.5
0.9 47.8 0.63
0.09 1.8 0.72
1.84 55.9 1.6
1.33 8.5 2.34
0.13 1.2 0.69
1.53 18.0 1.36
1.29 26.0 0.85
0.14 3.5 0.52
0.33 7.7 1.32
0.15 1.4 0.17
0.23 2.7 1.42
4.02 14.4 0.1
Bag 2 score
HC2 C02
0.04 1.1
0.16 2.2
0.79 14.7
0.1 2.7
0.09 2.6
0.23 3.4
2.06 17.2
0.45 6.2
2.26 15.7
0.07 1.6
0.83 49.2
0.18 2.6
0.23 4.6
0.35 2.9
1.5 6.9
0.79 14.7
0.24 6.6
0.07 2.8
0.09 1.8
0.13 2.9
0.62 16.6
1.04 59.7
0.08 1.6
1.71 54.3
1.09 5.9
0.1 1.3
1.44 17.5
1.01 20.6
0.15 2.9
0.31 7.7
0.19 1.7
0.18 2.5
3.21 14.2
ASM
Composite Scores
ASM HC ASM CO ASM NOx
0.13 3.8 0.51
0.19 3.3 0.77
1.40 72.5 0.56
0.47 3.3 2.09
0.10 4.0 0.35
0.31 8.2 1.70
0.43 5.8 0.50
0.19 3.8 1.04
1.59 11.3 2.56
0.11 2.9 0.90
0.57 38.8 1.01
0.15 3.4 1.31
0.16 5.8 2.04
0.11 3.1 1.05
0.40 4.7 1.14
1.68 17.3 0.74
0.16 3.9 1.32
0.12 2.9 1.07
0.21 3.6 0.61
0.11 2.9 0.91
0.31 9.8 2.03
0.20 12.9 0.84
0.20 3.3 0.64
2.44 89.0 2.10
0.16 3.6 0.68
0.22 3.3 0.77
1.07 22.1 2.16
1.92 34.8 1.59
0.34 4.4 1.09
0.74 11.8 2.86
0.18 3.9 0.31
0.12 3.5 0.87
3.78 15.1 0.31
B-2
-------
I. Cutpoint Table Analyses Laboratory Recruited Vehicles
Vehicle Information
Vehf Runt Test Date Order
3189 2165 TST17 921127 IM240.1st
3190 2161 TST17 921201 IM240.1st
3191 2164 TST17 921130 IM240.2nd
3192 1995 TST17 921130 IM240.1st
3193 2176 TST17 921130 IM240.2nd
3194 2202 TST17 921201 IM240.2nd
3195 2200 TST17 921202 IM240.2nd
3196 2230 TST17 921201 IM240.2nd
3197 2238 TST17 921201 IM240.2nd
3198 2198 TST17 921201 IM240.2nd
3199 2244 TST17 921203 IM240.2nd
3200 2245 TST17 921203 IM240.1st
3201 2237 TST17 921202 IM240.l3t
3202 2273 TST17 921203 IM240.1st
3203 2261 TST17 921203 IM240.1st
3204 2280 TST17 921203 IM240.2nd
3205 2302 TST17 921204 IM240.2nd
3206 2317 TST17 921207 IM240.1st
3207 2319 TST17 921207 IM240.1st
3208 2324 TST17 921207 IM240.2nd
3209 2326 TST17 921207 IM240.2nd
3210 2337 TST17 921207 IM240.1st
3211 2330 TST17 921207 IM240.2nd
3212 2352 TST17 921208 IM240.2nd
3213 2368 TST17 921208 IM240.2nd
3214 2369 TST17 921208 IM240.1st
3216 2379 TST17 921210 IM240.1st
3217 2376 TST17 921209 IM240.2nd
3218 2419 TST17 921210 IM240.1st
3219 2416 TST17 921210 IM240.2nd
3220 2424 TST17 921210 IM240.2nd
3221 2451 TST17 921211 IM240.2nd
3222 2435 TST17 921211 IM240.!st
FTP Scores
HC CO NOx
0.4 5.9 1.24
13.07 42.0 0.56
0.32 3.3 0.56
0.49 6.3 0.53
0.61 5.0 0.97
2.29 47.1 1.92
0.51 5.8 0.66
2.87 26.5 5.81
1.29 3.5 2.42
1.77 10.2 1.8
0.53 10.9 1.53
0.59 0.3 0.69
0.94 19.7 1.72
0.5 7.5 7.56
0.96 6.4 4.17
0.34 6.4 0.47
0.33 5.6 0.89
0.51 10.2 0.34
3.33 87.3 0.92
2.38 113.4 0.31
0.2 2.5 0.53
1.4 20.3 1.21
0.48 10.8 0.57
0.37 3.9 1.11
0.33 4.3 0.93
1.15 12.9 2.5
0.3 3.2 0.65
0.8 9.7 2.02
0.2 2.7 0.3
0.33 4.0 0.78
1.22 12.9 1.56
0.39 4.5 0.57
0.32 4.7 0.64
IM240 SCORES
Composites
BC CO NOx
0.17 3.2 1.41
7.06 24.8 0.79
0.17 3.5 0.45
0.5 8.7 0.76
0.86 6.6 1.33
1.42 20.2 1.69
0.21 3.5 0.9
2.73 13.9 5.1
0.99 8.3 2.66
1.51 8.5 2.04
0.3 9.6 1.15
0.29 1.6 2.49
1.15 8.8 1.82
0.23 3.6 7.88
0.74 5.9 4.37
0.16 4.1 0.45
0.17 4.1 0.9
0.28 5.4 0.58
3.22 77.3 0.97
1.87 74.4 0.41
0.11 2.1 0.6
1.04 13.0 2.98
1.42 93.1 0.53
0.15 1.5 5.15
0.54 19.6 1.17
2.01 23.4 2.96
0.96 14.8 1.04
0.53 6.5 2.22
0.1 1.2 0.87
0.23 4.0 0.81
1.05 13.3 1.78
0.35 4.0 0.8
0.15 3.0 1
Bag 2 score
HC2 CO2
0.15 3.1
5.77 23.1
0.18 3.6
0.41 9.0
0.72 5.6
1.39 20.4
0.2 3.7
2.54 13.7
0.88 8.7
1.31 8.2
0.25 9.1
0.27 1.5
0.52 6.0
0.2 3.2
0.71 6.3
0.15 3.4
0.16 4.2
0.26 6.2
3.19 79.3
1.83 71.9
0.12 2.6
0.9 " 13.2
1.94 129.3
0.15 1.7
0.59 24.7
1.83 21.4
0.12 0.9
0.54 7.4
0.08 1.0
0.2 3.1
0.95 14.1
0.27 3.5
0.09 2.1
ASM
Composite Scores
ASM HC ASM CO ASM NOx
0.12 4.1 1.33
4.27 12.7 0.80
0.22 4.4 0.62
0.42 6.6 0.92
0.41 3.4 2.79
1.04 43.2 2.16
0.19 3.3 1.27
1.49 15.9 5.88
0.86 7.7 2.97
1.07 7.4 2.50
0.37 4.1 1.74
0.13 2.9 1.58
0.19 4.5 1.08
0.17 3.1 6.51
0.41 3.6 4.61
0.15 3.3 0.57
0.28 4.6 1.57
0.29 8.2 0.57
2.19 70.8 1.04
1.59 73.7 0.73
0.17 4.3 0.67
0.55 7.2 3.65
0.63 64.9 0.58
0.26 4.7 4.04
0.16 3.0 0.55
0.85 6.6 2.03
0.47 7.1 0.55
0.31 4.5 2.06
0.33 3.3 0.72
0.38 4.0 0.89
1.05 6.9 1.99
0.36 3.6 0.60
0.77 3.2 0.76
B-3
-------
I. Cutpoint Table Analyses Laboratory Recruited Vehicles
Vehicle Information
Vehf Runf Test Date Order
3223 2440 TST17 921211 IM240.2nd
3224 2441 TST17 921215 IM240.1st
3225 2446 TST17 921214 IM240.2nd
3226 2447 TST17 921214 IM240.1st
3227 2449 TST17 921214 IM240.1st
3228 2450 TST17 921214 IM240.2nd
3229 2453 TST17 921214 IM240.1st
3230 2445 TST17 921215 IM240.1st
3231 2463 TST17 921216 IM240.1st
3232 2464 TST17 921216 IM240.2nd
3233 2469 TST17 921216 IM240.1st
3234 2470 TST17 921216 IM240.2nd
3235 2479 TST17 921217 IM240.1st
3236 2483 TST17 921217 IM240.1st
3237 2488 TST17 921217 IM240.2nd
3238 2489 TST17 921217 IM240.1st
3239 2490 TST17 921217 IM240.2nd
3240 2492 TST17 921217 IM240.2nd
3241 2496 TST17 921218 IM240.2nd
3242 2499 TST17 921218 IM240.1st
3243 2507 TST17 921218 IM240.1st
3244 2516 TST17 921218 IM240.2nd
3245 2529 TST17 921218 IM240.1st
3246 2563 TST17 921221 IM240.2nd
3247 2548 TST17 921221 IM240.2nd
3248 2830 TST17 930112 IM240.2nd
3249 2835 TST17 930112 IM240.1st
3250 2845 TST17 930113 IM240.1st
3251 2914 TST17 930114 IM240.2nd
3252 2945 TST17 930114 IM240.1st
3254 3080 TST17 930128 IM240.2nd
3255 3174 TST17 930129 IM240.2nd
3256 3208 TST17 930202 IM240.2nd
FTP Scores
HC CO MOx
0.53 4.4 0.93
1.8 21.4 3.23
0.57 5.7 0.77
0.31 3.7 0.99
0.42 7.6 1.25
0.44 19.9 0.8
0.3 4.4 0.43
0.41 4.0 0.12
0.04 3.7 0.2
0.18 3.2 0.18
0.24 2.0 0.28
0.25 2.9 0.94
0.38 2.4 0.34
0.73 8.6 1.84
0.33 3.0 0.28
0.35 2.3 0.35
0.24 1.5 0.72
0.27 2.7 1.14
0.3 5.5 0.83
0.39 5.8 1.91
0.67 8.5 2.18
0.22 3.1 0.47
0.56 4.7 1.63
0.33 8.6 1.29
0.84 11.4 1.99
0.39 3.3 1.51
0.2 3.8 2.25
1.55 5.1 1.06
1.31 16.9 4.26
1.03 12.5 1.34
1.87 35.9 1.16
0.18 1.3 0.23
0.23 2.5 0.26
IM240 SCORES
Composites
HC CO NO*
0.37 4.3 1.16
1.15 10.3 4.47
0.19 7.9 1.33
0.19 3.2 1.59
0.23 1.4 1.82
0.09 3.4 0.61
0.11 2.7 0.88
1.04 8.2 0.78
0.11 1.4 0.25
0.33 11.3 0.75
0.13 1.2 0.21
0.16 2.6 1.41
0.82 6.4 1.8
1.04 13.9 4.01
0.22 2.3 0.19
0.27 4.7 0.43
0.03 0.5 2.4
0.07 1.5 2.55
0.19 3.4 1.11
1.05 13.7 3.34
0.81 7.6 3.38
0.09 3.5 2.34
0.49 4.2 4.52
0.42 11.1 2.02
0.69 13.5 3.78
0.21 2.5 2.55
0.15 4.6 3.56
1.57 8.0 1.38
0.25 3.6 5.25
0.99 8.8 1.76
2.26 28.6 1.5
0.1 0.8 0.14
0.1 0.7 0.19
Bag 2 score
HC2 C02
0.35 4.1
0.93 9.5
0.18 6.3
0.18 3.1
0.15 1.2
0.09 2.7
0.07 2.9
0.41 6.8
0.09 1.5
0.44 15.4
0.09 0.5
0.17 2.5
0.5 5.0
1.06 14.6
0.14 1.3
0.23 5.3
0.02 0.1
0.06 1.2
0.11 1.8
0.71 9.2
0.55 6.9
0.09 3.9
0.48 4.4
0.52 12.0
0.6 13.1
0.15 2.1
0.15 4.9
1.27 5.9
0.25 4.2
0.85 8.5
2 31.6
0.12 0.9
0.12 0.8
ASM
Composite Scores
ASM HC ASM CO ASM NOx
0.48 4.1 1.03
2.36 18.9 3.28
0.76 8.3 1.07
0.30 4.3 1.44
0.37 3.2 0.80
0.29 3.9 0.62
0.29 3.0 0.76
0.74 4.5 0.34
0.19 3.2 0.38
0.34 12.8 0.84
0.26 2.9 0.34
0.27 3.4 1.19
0.70 4.1 0.93
0.56 5.5 3.03
0.31 3.4 0.35
0.39 3.7 0.40
0.28 3.0 1.07
0.20 3.0 2.55
0.16 3.0 0.87
0.16 3.7 0.98
0.40 5.5 2.58
0.16 3.8 2.39
0.19 3.6 1.99
0.50 5.4 2.08
0.53 6.5 3.09
0.15 3.1 2.05
0.12 4.0 1.97
0.93 2.9 1.09
0.59 5.8 3.46
0.96 8.4 1.69
0.34 6.3 0.99
0.16 3.1 0.38
0.25 3.2 0.42
B-4
-------
I. Cutpoint Table Analyses Laboratory Recruited Vehicles
Vehicle Information
Vehf Run* Teat Date Order
3257 3213 TST17 930202 IM240.1st
3259 3250 TST17 930209 IM240.2nd
3260 3438 TST17 930216 IM240.2nd
3261 3475 TST17 930216 IM240.l3t
3262 3480 TST17 930217 IM240.2nd
3264 3519 TST17 930218 IM240.1st
3265 3530 TST17 930223 IM240.2nd
FTP Scores
HC CO NOse
1.26 8.6 0.9
1.94 15.0 0.53
0.2 3.0 0.66
0.72 12.5 0.37
0.34 3.7 1.88
1.36 20.3 1.06
2.7 14.8 2.59
IM240 SCORES
Composites
HC CO HOx
1.92 10.6 1.55
1.5 9.3 0.88
0.11 3.4 0.85
0.91 19.1 0.69
0.19 3.7 2.26
2.16 20.1 1.65
2.49 9.9 3.32
Bag 2 score
HC2 CO2
1.48 9.4
1.15 8.3
0.12 3.7
0.79 19.3
0.17 3.4
1.44 16.0
2.22 9.1
ASM
Composite Scores
ASM HC ASM CO ASM NOx
0.27 4.6 1.03
0.26 4.6 1.22
0.23 3.0 1.46
0.48 11.8 0.76
0.16 3.3 1.56
0.54 5.6 1.55
1.81 5.5 2.84
B-5
-------
II. Contractor Repair Data
Vehicle Information
Vehft Run* Test Date Order
3150 1692 TST17 921028 IM240.2nd
3150 1924 TST2 921113 IM240.2nd
3151 1696 TST17 921028 IM240.2nd
3151 2145 TST27 921123 IM240.2nd
3154 1739 TST17 921030 IM240.!st
3154 1923 TST2 921113 IM240.1st
3155 1735 TST17 921030 IM240.1st
3155 1901 TST2 921106 IM240.1st
3156 1726 TST17 921029 IM240.2nd
3156 1926 TST2 921113 IM240.2nd
3157 1747 TST17 921030 IM240.1st
3157 2025 TST2 921118 IM240.1st
3159 1752 TST17 921030 IM240.2nd
3159 2032 TST2 921118 IM240.2nd
3160 1749 TST17 921030 IM240.1st
3160 1925 TST2 921113 IM240.l3t
3165 1810 TST17 921103 IM240.2nd
3165 2141 TST27 921123 IM240.2nd
3169 1677 TST17 921027 IM240.1st
3169 1927 TST2 921113 IM240.1st
3172 1895 TST18 921106 IM240.1st
3172 2335 TST2 921203 IM240.1st
3174 1688 TST17 921028 IM240.2nd
3174 2174 TST2 921124 IM240.2nd
FTP Scores
HC CO NOx
2.43 82.9 0.59
0.45 7.6 0.95
0.34 3.5 5.81
0.28 3.3 0.88
0.31 6.2 1.04
0.30 5.2 1.15
3.25 46.7 0.26
0.35 3.3 0.37
0.31 6.7 1.07
0.27 7.0 1.12
1.70 14.3 2.14
0.24 2.4 0.53
1.11 74.2 0.31
0.28 7.6 0.13
0.29 3.0 1.26
0.30 3.7 1.49
1.96 13.2 2.50
0.29 1.3 0.98
1.04 15.0 0.96
0.34 1.3 1.81
0.16 3.3 0.73
0.15 2.0 0.52
0.74 16.3 1.88
0.19 4.6 1.06
IM240 SCORES
Composites
HC CO NOx
1.44 32.6 0.95
0.19 2.8 1.40
0.20 2.0 4.52
0.14 7.4 0.87
0.34 4.4 2.24
0.32 4.7 1.59
2.77 24.1 1.20
0.30 4.5 0.75
0.45 6.1 2.07
0.16 4.5 1.16
2.84 34.7 2.95
0.04 1.3 0.28
0.74 40.3 0.48
0.12 4.6 0.14
0.21 2.8 2.26
0.16 1.4 1.87
1.58 7.3 2.48
0.08 0.4 0.96
0.85 14.1 0.97
0.27 0.5 1.43
0.13 2.1 3.30
0.04 0.8 0.44
0.76 19.3 2.50
0.14 6.6 1.37
Bag 2 score
HC2 C02
0.79 14.7
0.21 3.2
0.10 2.7
0.16 10.3
0.23 3.4
0.23 3.7
2.06 17.2
0.37 5.8
0.45 6.2
0.12 4.1
2.26 15.7
0.05 1.4
0.83 49.2
0.10 4.4
0.18 2.6
0.15 1.6
1.50 6.9
0.09 0.5
0.79 14.7
0.23 0.5
0.09 1.8
0.05 0.7
0.62 16.6
0.08 3.4
ASM
Composite Scores
ASM HC ASM CO ASM NOx
1.35 86.9 0.62
0.18 3.4 0.95
0.36 3.6 3.19
0.09 2.9 0.66
0.24 7.3 1.46
0.23 5.4 1.87
0.34 6.0 0.50
0.22 4.9 0.66
0.15 3.7 1.04
0.15 3.5 1.12
0.95 11.1 2.56
0.13 2.9 0.61
0.48 32.9 0.88
0.13 5.3 0.39
0.13 3.4 1.31
0.19 6.8 1.98
0.39 5.6 1.50
0.09 3.0 1.01
0.57 16.7 0.68
0.26 5.7 1.04
0.11 3.3 0.51
0.09 3.2 0.41
0.20 9.0 1.81
0.08 3.1 1.18
B-5
-------
II. Contractor Repair Data
Vehicle Information
Vehf Runt Test Date Order
3175 1907 TST17 921106 IM240.1st
3175 2364 TST2 921204 IM240.1st
3179 1966 TST17 921116 IM240.2nd
3179 2206 TST2 921125 IM240.2nd
3180 2005 TST17 921118 IM240.1st
3180 2433 TST27 921209 IM240.l3t
3183 2024 TST17 921118 IM240.2nd
3183 2288 TST2 921201 IM240.2nd
3188 2160 TST17 921124 IM240.2nd
3188 2382 TST27 921207 IM240.2nd
3190 2161 TST17 921124 IM240.1St
3190 2456 TST27 921210 IM240.1st
3196 2230 TST17 921127 IM240.2nd
3196 2511 TST27 921214 IM240.2nd
3197 2238 TST17 921127 IM240.2nd
3197 2432 TST27 921208 IM240.2nd
3198 2198 TST17 921125 IM240.2nd
3198 2431 TST27 921208 IM240.2nd
3200 2245 TST17 921128 IM240.1st
3200 2457 TST27 921210 IM240.1st
3201 2237 TST17 921127 IM240.1St
3201 2388 TST27 921207 IM240.1st
3202 2273 TST17 921201 IM240.1st
3202 2487 TST27 921211 IM240.1st
FTP Scores
HC CO NOz
0.40 13.1 0.46
0.42 7.9 0.71
2.90 77.6 2.06
0.22 4.1 1.18
0.96 9.8 1.22
0.69 8.8 0.74
3.13 66.3 0.70
0.25 3.2 1.21
4.49 17.8 0.20
0.63 5.1 0.54
13.07 42.0 0.56
0.19 0.6 0.82
2.87 26.5 5.81
0.34 1.2 0.83
1.29 3.5 2.42
0.14 1.7 0.22
1.77 10.2 1.80
0.18 1.9 0.05
0.59 0.3 0.69
0.62 1.6 0.42
0.94 19.7 1.72
0.47 4.1 1.11
0.50 7.5 7.56
0.42 7.1 1.25
IM240 SCORES
Composites
HC CO NO*
0.90 47.8 0.63
0.43 9.6 0.60
1.84 55.9 1.60
0.07 2.1 1.37
1.33 8.5 2.34
0.57 4.8 1.12
1.29 26.0 0.85
0.17 8.1 1.13
4.02 14.4 0.10
0.13 2.7 0.66
7.06 24.8 0.79
0.10 2.7 0.66
2.73 13.9 5.10
0.16 1.1 0.94
0.99 8.3 2.66
0.00 0.3 0.07
1.51 8.5 2.04
0.05 0.8 0.03
0.29 1.6 2.47
0.44 0.3 0.80
1.15 8.8 1.82
0.49 5.8 1.11
0.23 3.6 7.88
0.25 7.2 1.66
Bag 2 score
HC2 CO2
1.04 59.7
0.48 12.0
1.71 54.3
0.09 2.7
1.09 5.9
0.30 3.7
1.01 20.6
0.20 10.9
3.21 14.2
0.11 2.7
5.77 23.1
0.10 3.6
2.55 13.6
0.17 1.3
0.88 8.7
0.00 0.3
1.31 8.2
0.06 1.1
0.27 1.5
0.38 0.3
0.52 6.0
0.31 3.7
0.20 3.2
0.25 8.8
ASM
Composite Scores
ASM HC ASM CO ASM NOx
0.17 12.9 0.84
0.08 4.1 0.54
1.48 89.0 2.10
0.13 3.0 1.53
0.15 3.6 0.68
0.36 3.8 0.84
1.11 35.3 1.59
0.10 3.1 1.26
3.61 18.2 0.32
0.16 3.5 0.76
3.70 12.8 0.80
0.19 2.9 0.68
1.21 15.8 5.88
0.16 2.9 1.15
0.60 7.8 2.97
0.21 3.1 0.30
0.86 7.4 2.50
0.32 4.6 0.29
0.13 2.9 1.58
0.45 2.9 0.53
0.17 4.7 1.22
0.27 4.4 0.77
0.18 3.2 8.60
0.31 4.1 1.16
B-6
-------
II. Contractor Repair Data
Vehicle Information
Vehf Runt Test Date Order
3203 2261 TST17 921130 IM240.1st
3203 2569 TST2 921217 IM240.1st
3207 2319 TST17 921203 IM240.1st
3207 2459 TST27 921210 IM240.1st
3208 2324 TST17 921203 IM240.2nd
3208 2468 TST27 921211 IM240.2nd
3210 2337 TST17 921203 IM240.1st
3210 2643 TST2 921222 IM240.1st
3211 2330 TST17 921203 IM240.2nd
3211 2461 TST27 921210 IM240.2nd
3212 2352 TST17 921204 IM240.2nd
3212 2494 TST27 921212 IM240.2nd
3213 2368 TST17 921205 IM240.2nd
3213 2493 TST27 921212 IM240.2nd
3214 2369 TST17 921205 IM240.1st
3214 2518 TST27 921214 IM240.1st
3217 2376 TST17 921205 IM240.2nd
3217 2515 TST27 921214 IM240.2nd
3220 2424 TST17 921208 IM240.2nd
3220 2570 TST2 921217 IM240.2nd
3224 2441 TST17 921209 IM240.1st
3224 2680 TST2 921224 IM240.1st
3236 2483 TST17 921211 IM240.1st
3236 2608 TST2 921221 IM240.1st
FTP Scores
HC CO NOx
0.96 6.4 4.17
0.75 5.7 1.30
3.33 87.3 0.92
0.44 2.7 1.11
2.38 113.4 0.31
0.22 1.7 0.96
1.40 20.3 1.21
0.34 1.2 0.34
0.48 10.8 0.57
0.38 2.8 0.50
0.37 3.9 1.11
0.33 3.7 1.05
0.33 4.3 0.93
0.30 4.1 0.78
1.15 12.9 2.50
0.15 1.6 0.32
0.80 9.7 2.02
0.67 8.7 1.36
1.22 12.9 1.56
0.24 1.6 0.57
1.80 21.4 3.23
0.20 1.4 0.36
0.73 8.6 1.84
0.15 2.6 0.53
IM240 SCORES
Composites
HC CO NOx
0.74 5.9 4.37
0.87 11.1 1.32
3.20 76.9 0.97
0.29 2.0 1.35
1.86 74.3 0.41
0.02 0.9 0.61
1.03 12.9 2.97
0.12 0.5 0.12
1.42 93.1 0.53
0.04 1.4 0.83
0.15 1.5 5.15
0.14 2.9 0.83
0.54 19.6 1.17
0.13 1.1 1.03
2.00 23.2 2.94
0.11 1.2 0.39
0.53 6.5 2.22
0.79 18.9 1.54
1.05 13.3 1.78
0.03 0.5 0.44
1.15 10.3 4.47
0.04 1.0 0.06
1.04 13.9 4.01
0.08 1.7 0.66
Bag 2 score
HC2 C02
0.71 6.3
0.82 13.6
3.19 79.3
0.35 2.0
1.83 71.9
0.02 0.7
0.90 13.2
0.10 0.5
1.94 129.3
0.02 0.7
0.15 1.7
0.13 2.8
0.59 24.7
0.12 1.3
1.83 21.4
0.09 0.9
0.54 7.4
0.89 24.7
0.95 14.0
0.03 0.6
0.93 9.5
0.05 1.0
1.06 14.6
0.08 1.5
ASM
Composite Scores
ASM HC ASM CO ASM NOz
0.27 3.4 3.52
0.47 3.9 1.49
1.50 70.8 1.04
0.41 3.4 1.31
1.24 73.7 0.73
0.22 2.9 0.62
0.35 6.6 3.17
0.29 3.5 0.38
0.67 75.2 0.63
0.71 68.9 0.43
0.24 4.7 4.04
0.14 3.3 0.93
0.14 3.0 0.55
0.18 2.9 0.66
0.69 6.6 2.03
0.20 3.1 0.51
0.23 4.2 1.80
0.28 3.9 1.71
0.73 6.3 1.70
0.12 2.9 0.48
1.03 16.4 3.28
0.20 3.1 0.29
0.47 5.5 3.03
0.16 3.0 0.99
B-7
-------
II. Contractor Repair Data
Vehicle Information
Vehl Run* Test Date Order
3239 2490 TST17 921211 IM240.2nd
3239 2646 TST2 921222 IM240.2nd
3240 2492 TST17 921212 IM240.2nd
3240 2676 TST2 921224 IM240.2nd
3242 2499 TST17 921214 IM240.1st
3242 2663 TST2 921223 IM240.1st
3243 2507 TST17 921214 IM240.1st
3243 2670 TST2 921223 IM240.1st
3244 2516 TST17 921214 IM240.2nd
3244 2671 TST2 921223 IM240.2nd
3245 2529 TST17 921215 IM240.1st
3245 2679 TST2 921224 IM240.1st
3247 2548 TST17 921216 IM240.2nd
3247 2681 TST2 921224 IM240.2nd
3248 2830 TST17 930106 IM240.2nd
3248 3105 TST27 930121 IM240.2nd
3249 2835 TST17 930107 IM240.1st
3249 3056 TST27 930119 IM240.1st
3250 2845 TST17 930107 IM240.1st
3250 3183 TST27 930127 IM240.1st
3252 2945 TST17 930113 IM240.1st
3252 3192 TST27 930127 IM240.1st
3257 3213 TST17 930128 IM240.1st
3257 3637 TST27 930225 IM240.1st
FTP Scores
HC CO NOx
0.24 1.5 0.72
0.30 1.0 0.85
0.27 2.7 1.14
0.26 2.6 1.23
0.39 5.8 1.91
0.15 1.4 0.38
0.67 8.5 2.18
0.21 1.8 0.27
0.22 3.1 0.47
0.25 3.8 0.42
0.56 4.7 1.63
0.11 0.6 0.35
0.84 11.4 1.99
0.12 0.8 0.35
0.39 3.3 1.51
0.27 1.5 0.33
0.20 3.8 2.25
0.18 1.6 0.69
1.55 5.1 1.06
0.68 1.5 1.20
1.03 12.5 1.34
0.12 1.1 0.17
1.26 8.6 0.90
0.70 3.9 0.26
IM240 SCORES
Composites
HC CO NOx
0.03 0.5 2.40
0.07 1.4 0.96
0.07 1.5 2.52
0.15 5.4 1.18
1.05 13.7 3.34
0.15 6.7 0.51
0.81 7.6 3.38
0.24 4.0 0.42
0.09 3.5 2.34
0.11 5.3 0.38
0.49 4.2 4.52
0.04 0.4 0.31
0.69 13.5 3.78
0.04 0.7 0.25
0.21 . 2.5 2.53
0.03 1.3 0.20
0.15 4.5 3.51
0.16 4.0 0.89
1.52 7.8 1.35
0.25 0.2 1.37
0.99 8.8 1.76
0.13 1.5 0.27
1.92 10.6 1.55
0.48 2.5 0.11
Bag 2 score
HC2 CO2
0.02 0.1
0.02 0.4
0.06 1.2
0.17 6.6
0.71 9.2
0.18 9.0
0.55 6.9
0.20 4.1
0.09 3.9
0.11 5.7
0.48 4.4
0.03 0.4
0.60 13.1
0.03 0.6
0.15 2.1
0.03 1.2
0.15 4.9
0.19 4.9
1.26 5.9
0.21 0.2
0.85 8.4
0.12 1.9
1.48 9.4
0.55 2.7
ASM
Composite Scores
ASM HC ASM CO ASM NOx
0.20 3.0 1.07
0.28 2.9 0.84
0.16 3.0 2.55
0.17 3.1 1.30
0.15 3.7 0.98
0.18 3.2 0.40
0.33 5.5 2.58
0.32 3.7 0.44
0.13 3.8 2.39
0.29 3.7 0.52
0.13 3.6 1.99
0.15 3.0 0.42
0.47 7.2 3.66
0.21 3.1 0.36
0.13 3.1 2.05
0.11 2.9 0.50
0.10 4.0 1.97
0.12 3.3 0.96
0.63 2.9 1.09
0.30 2.9 1.06
0.52 8.4 1.69
0.20 4.0 0.52
0.25 4.6 1.03
0.29 3.8 0.31
B-8
-------
II. Contractor Repair Data
Vehicle Information
Vehf Runt Test Date Order
3259 3250 TST17 930201 IM240.2nd
3259 3518 TST2 930217 IM240.2nd
3261 3475 TST17 930212 IM240.1st
3261 3581 TST2 930222 IM240.1st
3264 3519 TST17 930217 IM240.1st
3264 3671 TST29 930226 IM240.1st
3265 3530 TST17 930217 IM240.2nd
3265 3704 TST27 930310 IM240.2nd
FTP Scores
HC CO NOx
1.94 15.0 0.53
0.23 3.5 0.74
0.72 12.5 0.37
0.60 8.1 0.79
1.36 20.3 1.06
0.49 4.9 1.05
2.70 14.8 2.59
0.10 0.7 0.11
IM240 SCORES
Composites
HC CO NOx
1.50 9.3 0.88
0.11 2.0 0.95
0.91 19.1 0.68
0.45 11.5 0.79
2.16 20.1 1.66
0.33 3.7 0.96
2.49 9.9 3.31
0.01 0.4 0.07
Bag 2 score
HC2 CO2
1.15 8.3
0.10 1.8
0.79 19.3
0.41 12.4
1.44 16.0
0.23 3.3
2.21 9.1
0.01 0.3
. ASM
Composite Scores
ASM HC ASM CO ASM NOx
0.23 4.6 1.22
0.22 3.4 0.96
0.34 11.8 0.76
0.18 3.4 0.63
0.49 5.6 1.55
0.13 3.3 1.20
1.22 5.5 2.84
0.10 2.9 0.26
B-9
-------
III. Commercial Repair Data
Vehicle Information
CR# Vehi Run# Test Date Order
1 11898 1898 TST17 921106 IM240.2nd
1 11898 1906 TST19 921106 IM240.2nd
1 11898 2012 TST20 921118 IM240.2nd
2 12636 2636 TST17 921222 IM240.2nd
2 12636 2662 TST19 921223 IM240.2nd
3 12644 2644 TST17 921222 IM240.2nd
3 12644 2720 TST19 921230 IM240.2nd
8 12771 2771 TST17 930104 IM240.1st
8 12771 2977 TST19 930114 IM240.1st
8 12771 3168 TST20 930126 IM240.1st
6 12794 2794 TST17 930105 IM240.2nd
6 12794 2975 TST19 930114 IM240.1st
6 12794 3137 TST20 930125 IM240.1st
10 12798 2798 TST17 930105 IM240.2nd
10 12798 3049 TST19 930119 IM240.1st
10 12798 3064 TST20 930119 IM240.2nd
4 12853 2853 TST17 930107 IM240.1st
4 12853 2861 TST19 930108 IM240.1st
5 12863 2863 TST17 930108 IM240.1st
5 12863 2901 TST19 930111 IM240.1st
7 12968 2968 TST17 930113 IM240.2nd
7 12968 2976 TST19 930114 IM240.2nd
9 12981 2981 TST17 930114 IM240.1st
9 12981 2988 TST19 930114 IM240.1st
Arizona I/M Test
Loaded Mode
HC CO
70 2.15
116 2.97
1 0
70 0.02
39 0.02
260 6.88
14 0
86 1.55
38 0.63
75 0.38
51 0.11
298 10
46 0.15
279 2.53
229 0.54
100 0.15
201 4.33
81 0.95
433 8.72
7 0
397 1.58
177 2.16
108 1.46
78 0.37
Idle Mode
HC CO
26 0.18
43 0.09
1 0
545 0.11
140 0.02
45 0.01
13 0
87 0.38
835 0.07
41 0.06
455 7.03
480 7.21
106 1.44
152 2.29
122 0.06
141 0.29
637 7.41
12 0
1540 10
10 0
466 1.79
427 0.83
27 0.03
43 0.14
IM240 Scores
Composite
HC CO NOz
1.00 55.1 0.48
0.59 33.3 0.63
0.09 5.7 0.42
0.36 1.6 2.45
0.27 1.7 1.78
2.69 140.9 0.11
1.21 79.6 0.20
1.51 12.2 2.86
1.38 4.1 2.59
1.01 4.4 3.01
2.43 80.2 0.50
2.09 72.5 0.39
1.55 55.8 0.47
3.64 64.6 1.41
2.36 20.1 2.08
2.14 35.8 1.13
2.08 52.3 0.28
0.90 27.9 0.36
5.86 164.3 0.72
0.25 2.8 1.76
6.00 37.0 1.19
5.69 29.7 1.22
1.46 15.0 3.71
1.20 7.9 3.88
Bag 2 Score
HC2 C02
1.28 66.60
0.71 36.40
0.11 7.80
0.30 2.10
0.21 2.20
2.75 144.80
1.36 92.10
1.32 9.50
1.16 4.30
0.91 4.30
2.68 99.80
2.19 86.50
1.69 68.00
3.54 65.60
2.32 23.50
1.72 34.60
2.02 56.60
0.91 29.50
5.58 170.30
0.14 2.80
5.10 35.30
4.85 28.40
1.31 12.50
1.12 8.10
ASM
Composite Scores
ASM HC ASM CO ASM NOx
0.68 49.8 0.56
0.23 8.2 0.60
0.17 3.4 0.37
0.41 3.1 2.17
0.16 2.9 1.71
1.66 112.0 0.35
0.51 56.0 0.33
0.46 5.7 2.19
0.33 3.5 1.63
0.16 3.4 1.51
1.41 51.9 0.68
1.27 61.0 0.49
0.40 24.4 0.57
1.62 46.8 1.18
1.15 15.9 1.59
0.28 6.9 0.89
1.29 52.5 0.63
0.54 20.4 0.55
3.65 160.2 0.62
0.17 3.2 0.97
2.45 23.3 1.02
2.14 16.3 0.97
0.78 9.1 2.20
0.62 4.1 2.75
B-10
-------
III. Commercial Repair Data
Vehicle Information
CR# Veh# Run* Teat Date Order
11 13084 3084 TST17 930120 IM240.2nd
11 13084 3104 TST19 930121 IM240.2nd
14 13124 3124 TST17 930122 IM240.2nd
14 13124 3181 TST19 930127 IM240.2nd
12 13125 3125 TST17 930122 IM240.1st
12 13125 3129 TST19 930122 IM240.!st
13 13146 3146 TST17 930126 IM240.2nd
13 13146 3156 TST19 930126 IM240.2nd
15 13202 3202 TST17 930128 IM240.2nd
15 13202 3231 TST19 930129 IM240.1st
21 13263 3263 TST17 930201 IM240.!st
21 13263 3379 TST19 930205 IM240.1st
21 13263 3561 TST20 930219 IM240.1st
16 13306 3306 TST17 930203 IM240.2nd
16 13306 3310 TST19 930203 IM240.2nd
22 13349 3349 TST17 930204 IM240.1st
22 13349 3381 TST19 930205 IM240.1st
22 13349 3453 TST20 930211 IM240.l8t
22 13349 3548 TST21 930218 IM240.1St
23 13375 3375 TST17 930205 IM240.1st
23 13375 3388 TST19 930208 IM240.2nd
27 13471 3471 TST17 930212 IM240.1st
27 13471 3757 TST19 930316 IM240.!st
Arizona I/M Test
Loaded Mode
HC CO
15 0.02
13 0
111 1.42
6 0.01
110 0.34
74 0.41
117 1.91
40 0.16
129 0.97
112 0.36
172 0.99
93 0.5
303 1.24
191 0.45
73 0.26
251 8.68
194 6.97
71 1.89
261 9.51
136 0.38
7 0
10 1.61
12 0.01
Idle Mode
HC CO
1517 .0.01
370 4.7
17 0
12 0
853 0.09
16 0
84 0.14
26 0
712 10
178 0.63
673 2.49
601 0.75
46 0.02
428 1.84
75 0.58
115 3.28
125 3.22
54 0.36
185 3.89
132 2.81
2 0
1 0
4 0
IM240 Scores
Composite
HC CO NOx
1.16 7.8 0.20
0.12 2.2 0.60
0.41 11.1 0.52
0.13 2.4 1.44
1.06 10.7 1.77
1.08 15.1 1.26
3.25 50.7 1.51
0.80 13.5 0.89
1.33 16.8 3.50
1.11 6.3 3.55
6.02 21.4 1.68
5.74 17.5 1.65
5.87 25.5 1.69
3.42 19.6 4.25
1.34 4.9 2.40
5.88 162.5 0.20
4.85 145.7 0.25
1.84 25.9 1.13
8.48 224.2 0.12
0.09 1.2 0.78
0.03 0.3 0.84
0.24 35.0 0.21
0.17 3.3 1.33
Bag 2 Score
HC2 C02
1.13 7.50
0.09 1.90
0.44 12.80
0.14 2.80
0.93 9.70
0.81 10.60
2.87 46.50
0.66 11.40
1.22 8.40
1.08 6.00
5.42 20.00
5.20 16.10
5.35 23.60
2.89 16.00
1.24 3.60
5.22 141.30
4.25 121.10
1.48 22.90
7.44 199.50
0.07 1.00
0.02 0.30
0.22 37.00
0.17 4.50
ASM
Composite Scores
ASM HC ASM CO ASM NOx
0.19 3.0 0.35
0.30 3.5 0.83
0.29 7.8 1.04
0.28 3.2 2.45
0.48 4.3 1.21
0.29 3.5 0.82
1.03 42.2 1.64
0.38 6.2 0.96
0.62 6.3 3.32
0.52 5.4 3.16
1.52 12.8 1.01
1.01 9.2 1.15
1.71 16.3 1.08
1.35 4.8 3.16
0.61 4.0 1.72
1.90 100.4 0.39
1.12 73.2 0.38
0.45 17.8 0.77
1.89 118.2 0.32
0.20 3.5 0.45
0.22 3.7 0.53
0.14 7.2 0.34
0.18 3.1 0.45
B-ll
-------
III. Commercial Repair Data
Vehicle Information
CRf Vehf Runf Test Date Order
25 13504 3504 TST17 930216 IM240.2nd
25 13504 3511 TST19 930216 IM240.1st
26 13616 3616 TST17 930224 IM240.2nd
26 13616 3680 TST19 930301 IM240.2nd
Arizona I/M Test
Loaded Mode
HC CO
49 0.01
23 0
189 0.19
25 0.04
Idle Mode
HC CO
150 3.02
5 0
349 10
11 0
IM240 Scores
Composite
HC CO NOx
0.41 6.7 5.73
0.13 0.2 5.04
2.77 44.8 0.37
0.22 3.9 1.15
Bag 2 Score
HC2 C02
0.33 4.30
0.13 0.20
2.58 37.30
0.18 3.80
ASM
Composite Scores
ASM HC ASM CO ASM NOx
0.31 19.3 4.44
0.19 2.9 3.54
0.50 18.7 0.41
0.38 3.9 1.16
TST17 - Initial Test
TST19 - After 1st Repair
TST20 - After 2nd Repair
TST21 - After 3rd Repair
B-12
-------
ASM/IM240 Phoenix Lane Data Request Form
Please Fax or Mail Your Data Requests to:
Attn: William M. Pidgeon
U. S. Environmental Protection Agency
2565 Plymouth Road
Ann Arbor, MI 48105-2425
Phone #: (313) 668-4416
Fax #: (313) 668-4497
Requestor:
Organization:
Phone #:
Fax #:
Mailing Address:
Specify Disk Format
(3.5 inch only)
IBM
Mac
High Density
Low Density
Specify File Format:
Text ASCII File
(tab separated)
CSV ASCII File
(comma separated)
SYLK File
Lotus File (.WKS)
Excel File (v2.1)*
*Note: Oldest available formats were chosen for maximum compatibility.
These formats should be compatible with newer versions.
B-13
-------
Appendix C:
QC Steps for ASM Analysis Database
-------
The Phoenix lane data used in these analyses were reported to EPA by the
testing contractor as total values (concentrations, mass, or flow) for the
entire test mode as well as in second-by-second form. The following automated
quality control (QC) checks were performed by EPA on the data. Tests that
were flagged by one (or more) of these QC checks were then manually verified.
Second by Second ASM Tolerance Checks:
• Speed Tolerance - ± 15% of nominal speed for Modes 1,2,3. Allowed
tolerance to be exceeded for less than 3 seconds in duration. Also
checked Idle for Modes 4,5
• Mode Length - Checked to ensure that each mode contained at least 20 and
not more than 30 "stable" seconds.
• Hp/Torque Tolerance - Compared required and actual horsepower (Hp) and
Torque and flagged differences > ±10% for at least 5 seconds.
All vehicles with test weights above 4000 pounds exceeded this tolerance
because of the capacity of the dynamometer. These cars were not removed
from these analyses. Smaller vehicles exceeding this tolerance were
removed.
• Calculated average concentrations and cumulative purge for all ASM
modes. Average concentrations were calculated as the average
concentration from second 10 to second 39 of each mode. The first 10
seconds of each mode were ignored to allow for the dynamometer
stabilization and exhaust transport time. Vehicles with less than 30
seconds per mode were noted and vehicles with less than 20 seconds per
mode were excluded. Purge values were calculated as the total purge in
liters over the entire ASM including transient accelerations.
Second by Second IM240 Tolerance Checks:
• Speed Tolerance - ± 4 mph at ± 1 sec of nominal speed. Allowed
tolerance to be exceeded for less than 3 seconds in duration. Also
speeds exceeding 70 mph, and less than 0 mph were flagged.
• Background Concentration Tolerances - Flagged background readings
outside the following ranges:
1.8 < HC < 10.0 (ppm)
-10.0 < CO < 30.0 (ppm)
- 0.5 < NOX < 1.25 (ppm)
0.0 < C02 < 0.15 (percent)
• Test Length - Checked to ensure that the full 240 seconds were present.
• Distance Tolerance - Flagged distance > ± 5% of nominal distance
Bag 1: 0.532 < dist 1 < 0.588
Bag 2: 1.393 < dist 2 < 1.469
C-2
-------
• Fuel Economy Tolerance - Flagged fuel economies < 10 mpg and >50 mpg
• Sample Continuity and Integrity - Ensured that the sampling was
continuous (i.e., sec(I) = I for I = 1 to 240) and that gram and
concentration values were non-zero (HC, CO and C02 cannot all be zero
for fuel economy calculations or dilution factors).
Non-zero concentrations were not mandatory for the Phoenix data because
second by second concentrations received were calculated, not measured.
The calculated concentrations were based on the reported grains per
second results. These vehicles were still flagged for low
concentrations but were not removed from the analyses for this reason.
• Comparison of composite and bag results calculated from the second by
second data with composite and bags results received from ATL.
Differences of > 10% were flagged.
Purge Flow Data QC
• Comparison of second-by-second purge flow to the reported cumulative
purge flow and pass/fail status reported by ATL. All significant
differences were flagged.
• Vehicles exhibiting a non-zero constant purge rate for more than 20
seconds and at various speeds were flagged. Purge data was rounded to
nearest 0.01 liter/second prior to processing.
Bag FTP Tolerance Checks:
• The ratios of corresponding emissions (HC, CO, and NOX) and fuel economy
for each of the three bags that were not within expected ranges were
flagged.
• The temperatures, barometric pressures, and distances that were not
within expected ranges were flagged.
Bag IM240 Tolerance Checks:
• Bag-1 emissions (HC, CO, and NOX) and fuel economy were compared to the
corresponding Bag-2 results (based on regression analyses previously
performed on the Indiana data). All significant differences were
flagged.
• The Bag-1 and Bag-2 fuel economies were also compared to the test
weight. All fuel economy values that were not within an expected range
(based on test weight) were flagged.
C-3
-------
• The Bag-1 and Bag-2 distances not with the following ranges were
flagged:
Bag 1: 0.545 5 dist 1 £ 0.586
Bag 2: 1.365 S dist 2 £ 1.435
Bag IM240/FTP Tolerance Checks:
• For the laboratory recruited vehicles, the composite IM240 emissions
(HC, CO, and NOX) and fuel economy were compared to the corresponding
FTP results (based on regression analyses previously performed on the
Indiana data). All significant differences were flagged.
Dynamometer Loading Tolerance Checks:
• The test weights and horsepower settings had to be within 10% for all
tests performed on each vehicle.
Excluded Data Summary
This section of Appendix C details the vehicles excluded from the various
databases.
Purge Analysis - 1725 of the 1758 lane tests contained the necessary
data to be included into this analysis. Of these 153 were removed because of
a malfunctioning purge meter, 184 tests were repeat tests for vehicles
previously tested and were removed, 5 cars had purge flow status fields which
indicated missing data, 95 additional vehicles had no indication of test order
and were removed, and 118 of the remaining vehicles exhibited non-zero
constant purge rates over varied vehicle speeds and were removed. The result
was a database of 1170 lane tested vehicles.
Outpoint Table Analysis - This analysis required laboratory FTP data.
Therefore, only lab recruited vehicles were considered for this analysis. Of
the 127 recruited vehicles, 17 did not receive initial ASM tests and one (veh#
3258) did not receive an as-received FTP. Of the remaining 109 vehicles one
vehicle (vehf 2177) was removed because the ambient FTP temperature exceeded
allowable tolerances, one vehicle (vehf 3253) was removed due to extremely
low HC emissions at the lane caused by a flame-out in the FID HC analyzer, and
vehf 3164 was removed due to unacceptable speed deviations on its initial ASM
test. The resulting database contained 106 lab recruited vehicles.
C-4
-------
Commercial Repair Analysis - Of the 27 vehicles recruited for this
program only 23 had completed after repair tests at the time of this analysis.
One vehicle, #13239 (CR# 24) was removed from the database due to unacceptable
speed deviations on its initial ASM test, leaving 22 vehicles available for
analysis. For the analysis of Section 5.6.2, 5 vehicles failed to pass the
Arizona state test on the subsequent retest and were removed. The resulting
database used for this analysis consisted of 17 vehicles which received
"successful" commercial repairs. For the commercial repair analysis of
Section 5.6.3, only vehicles initially failing ASM outpoints were included.
The result was 17 vehicles. Commercial repairs did not have to be successful
for this analysis and the two data sets contained slightly different cars.
Regression Coefficient Analysis - For this analysis all lab recruited
vehicles and commercial repair vehicles were removed from the analysis to
prevent the application of coefficients to data used to develop those
coefficients. Therefore, 1422 of the 1758 vehicles were considered for
inclusion into this analysis. Ten vehicles were removed because the composite
IM240 data was not available. The following vehicles were removed because
there was insufficient second by second data to calculate composite IM240
results:
Run # Reason for Removal
1027 Test has only 93 seconds
1855 Missing second by second
2231 Test has only 93 seconds
3066 Sampling Discontinuity
3077 Sampling Discontinuity
3079 Sampling Discontinuity
3081 Sampling Discontinuity
Of the 1405 remaining vehicles, 1192 passed all QC tolerances. Purge
tolerances were not considered for this analysis. The following table lists
the QC tolerances checks for which vehicles were removed from this analysis.
C-5
-------
Tolerance
Flagged
ASM Speed
Short ASM Mode
ASM Horsepower
IM240 Speed
IM240 Fuel Economy
IM240 Background
IM240 Sample
Number of
Vehicles
8
2
10
14
4
163
18
Note: 1405 minus the above vehicles does not equal 1192 because some
vehicles exceeded more than one tolerance
Six hundred and eight (608) of the 1192 tests remaining received the IM240
second and were chosen for this analysis.
C-6
-------
Appendix D
IM240 Outpoint Tables
-------
Appendix D: IM240 Cutpoint Tables
IM240
Failure Rate
12%
13%
13%
13%
13%
13%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
Cutpoints
Composite + Mode 2
1.20 / 20.0 / 2.5 + 0.75 / 16.0
1.00 / 20.0 / 2.4 + 0.62 / 16.0
1.00 / 20.0 / 2.5 + 0.62 / 16.0
1.20 / 18.0 / 2.5 + 0.75 / 14.4
1.20 / 15.0 / 2.5 + 0.75 / 12.0
1.20 / 20.0 / 2.4 + 0.75 / 16.0
0.80 / 20.0 / 2.5 + 0.50 / 16.0
1.00 / 12.0 / 2.4 + 0.62 / 9.6
1.20 / 12.0 / 2.4 + 0.75 / 9.6
1.00 / 12.0 / 2.5 + 0.62 / 9.6
1.20 / 12.0 / 2.5 + 0.75 / 9.6
1.00 / 18.0 / 2.4 + 0.62 / 14.4
1.00 / 15.0 / 2.4 + 0.62 / 12.0
1.00 / 18.0 / 2.5 + 0.62 / 14.4
1.00 / 15.0 / 2.5 + 0.62 / 12.0
1.00 / 20.0 / 2.3 + 0.62 / 16.0
1.20 / 18.0 / 2.3 + 0.75 / 14.4
1.20 / 15.0 / 2.3 + 0.75 / 12.0
1.20 / 18.0 / 2.4 + 0.75 / 14.4
1.20 / 15.0 / 2.4 + 0.75 / 12.0
1.20 / 20.0 / 2.3 + 0.75 / 16.0
0.60 / 20.0 / 2.4 + 0.37 / 16.0
0.80 / 20.0 / 2.4 + 0.50 / 16.0
0.80 / 18.0 / 2.4 + 0.50 / 14.4
0.80 / 15.0 / 2.4 + 0.50 / 12.0
0.80 / 12.0 / 2.4 + 0.50 / 9.6
0.60 / 20.0 / 2.5 + 0.37 / 16.0
0.60 / 18.0 / 2.5 + 0.37 / 14.4
0.80 / 18.0 / 2.5 + 0.50 / 14.4
0.60 / 15.0 / 2.5 + 0.37 / 12.0
0.80 / 15.0 / 2.5 + 0.50 / 12.0
0.60 / 12.0 / 2.5 + 0.37 / 9.6
0.80 / 12.0 / 2.5 + 0.50 / 9.6
1.00 / 12.0 / 2.3 + 0.62 / 9.6
1.20 / 12.0 / 2.3 + 0.75 / 9.6
Excess Emissions Identified
HC
343
351
346
347
347
347
362
357
357
356
356
353
353
352
352
351
348
348
348
348
347
365
365
365
365
365
364
364
364
364
364
364
364
357
357
CO
5286
5419
5342
5422
5422
5363
5627
5548
5548
5548
5548
5478
5478
5478
5478
5429
5432
5432
5422
5422
5373
5707
5704
5709
5709
5709
5707
5713
5709
5713
5709
5713
5709
5558
5558
NOx
253
262
257
258
258
258
263
262
262
262
262
262
262
262
262
266
263
263
258
258
263
268
268
268
268
268
268
268
268
268
268
268
268
266
266
Identification
HC
86.2%
88.2%
87.4%
87.2%
87.2%
87.0%
90.9%
89.6%
89.6%
89.3%
89.3%
88.7%
88.7%
88.4%
88.4%
88.2%
87.4%
87.4%
87.4%
87.4%
87.0%
91.6%
91.6%
91.6%
91.6%
91.6%
91.3%
91.3%
91.3%
91.3%
91.3%
91.3%
91.3%
89.6%
89.6%
CO
61.5%
63.1%
62.2%
63.1%
63.1%
62.4%
65.5%
64.6%
64.6%
64.6%
64.6%
63.8%
63.8%
63.8%
63.8%
63.2%
63.2%
63.2%
63.1%
63.1%
62.6%
66.5%
66.4%
66.5%
66.5%
66.5%
66.5%
66.5%
66.5%
66.5%
66.5%
66.5%
66.5%
64.7%
64.7%
Rates
NOx
74.0%
76.5%
74.9%
75.5%
75.5%
75.5%
76.8%
76.5%
76.5%
76.5%
76.5%
76.5%
76.5%
76.5%
76.5%
77.7%
76.7%
76.7%
75.5%
75.5%
76.7%
78.3%
78.3%
78.3%
78.3%
78.3%
78.3%
78.3%
78.3%
78.3%
78.3%
78.3%
78.3%
77.7%
77.7%
Fails
257
275
263
275
275
269
293
293
293
287
287
287
287
281
281
293
298
298
281
281
287
316
304
310
310
310
310
316
304
316
304
316
304
310
310
Errors of
Commission
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
12
12
12
0
0
12
6
0
0
0
0
6
6
0
6
0
6
0
12
12
EC Rate*
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.6%
0.6%
0.6%
0.0%
0.0%
0.6%
0.3%
0.0%
0.0%
0.0%
0.0%
0.3%
0.3%
0.0%
0.3%
0.0%
0.3%
0.0%
0.6%
0.6%
Discrepant
Failures
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Probable
EC Rate
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.6%
0.6%
0.6%
0.0%
0.0%
0.6%
0.3%
0.0%
0.0%
0.0%
0.0%
0.3%
0.3%
0.0%
0.3%
0.0%
0.3%
0.0%
0.6%
0.6%
D-2
see text in Section 5 for explanation
-------
Appendix D: IM240 Outpoint Tables
DVU40
Failure Rate
15%
15%
15%
15%
15%
15%
15%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
16%
Outpoints
Composite + Mode 2
1.00 / 18.0 / 2.3 + 0.62 / 14.4
1.00 / 15.0 / 2.3 + 0.62 / 12.0
1.00 / 20.0 / 2.2 + 0.62 / 16.0
1.20 / 20.0 / 2.1 + 0.75 / 16.0
1.20 / 18.0 / 2.2 + 0.75 / 14.4
1.20 / 15.0 / 2.2 + 0.75 / 12.0
1.20 / 20.0 / 2.2 + 0.75 / 16.0
0.80 / 20.0 / 2.2 + 0.50 / 16.0
0.60 / 20.0 / 2.3 + 0.37 / 16.0
0.80 / 20.0 / 2.3 -I- 0.50 / 16.0
0.60 / 18.0 / 2.3 + 0.37 / 14.4
0.80 / 18.0 / 2.3 + 0.50 / 14.4
0.60 / 15.0 / 2.3 + 0.37 / 12.0
0.80 / 15.0 / 2.3 + 0.50 / 12.0
0.60 / 12.0 / 2.3 + 0.37 / 9.6
0.80 / 12.0 / 2.3 + 0.50 / 9.6
0.60 / 18.0 / 2.4 + 0.37 / 14.4
0.60 / 15.0 / 2.4 + 0.37 / 12.0
0.60 / 12.0 / 2.4 + 0.37 / 9.6
1.00 / 12.0 / 2.1 + 0.62 / 9.6
1.20 / 12.0 / 2.1 + 0.75 / 9.6
1.00 / 12.0 / 2.2 + 0.62 / 9.6
1.20 / 12.0 / 2.2 + 0.75 / 9.6
1.00 / 10.0 / 2.4 + 0.62 / 8.0
1.20 / 10.0 / 2.4 + 0.75 / 8.0
1.00 / 10.0 / 2.5 + 0.62 / 8.0
1.20 / 10.0 / 2.5 + 0.75 / 8.0
1.00 / 18.0 / 2.0 + 0.62 / 14.4
1.00 / 15.0 / 2.0 + 0.62 / 12.0
1.00 / 18.0 / 2.1 + 0.62 / 14.4
1.00 / 15.0 / 2.1 + 0.62 / 12.0
1.00 / 20.0 / 1.9 + 0.62 / 16.0
1.00 / 20.0 / 2.0 + 0.62 / 16.0
1.00 / 20.0 / 2.1 + 0.62 / 16.0
1.00 / 18.0 / 2.2 + 0.62 / 14.4
Excess Emissions Identified
HC
353
353
351
349
348
348
347
365
365
365
365
365
365
365
365
365
365
365
365
359
359
357
357
357
357
356
356
356
356
356
356
354
354
354
353
CO
5489
5489
5431
5429
5434
5434
5375
5716
5718
5714
5723
5719
5723
5719
5723
5719
5713
5713
5713
5614
5614
5560
5560
5579
5579
5579
5579
5565
5565
5545
5545
5559
5505
5485
5491
NOx
266
266
273
276
269
269
269
279
272
272
272
272
272
272
272
272
268
268
268
279
279
273
273
264
264
264
264
279
279
279
279
282
279
279
273
Identification
HC
88.7%
88.7%
88.2%
87.6%
87.4%
87.4%
87.0%
91.6%
91.6%
91.6%
91.6%
91.6%
91.6%
91.6%
91.6%
91.6%
91.6%
91.6%
91.6%
90.2%
90.2%
89.6%
89.6%
89.6%
89.6%
89.3%
89.3%
89.2%
89.2%
89.2%
89.2%
88.8%
88.8%
88.8%
88.7%
CO
63.9%
63.9%
63.2%
63.2%
63.3%
63.3%
62.6%
66.6%
66.6%
66.5%
66.6%
66.6%
66.6%
66.6%
66.6%
66.6%
66.5%
66.5%
66.5%
65.4%
65.4%
64.7%
64.7%
65.0%
65.0%
65.0%
65.0%
64.8%
64.8%
64.6%
64.6%
64.7%
64.1%
63.9%
63.9%
Rates
NOx
77.7%
77.7%
79.7%
80.5%
78.7%
78.7%
78.7%
81.5%
79.5%
79.5%
79.5%
79.5%
79.5%
79.5%
79.5%
79.5%
78.3%
78.3%
78.3%
81.5%
81.5%
79.7%
79.7%
77.0%
77.0%
77.0%
77.0%
81.6%
81.6%
81.5%
81.5%
'82.4%
81.6%
81.5%
79.7%
Fails
304
304
310
316
316
316
304
340
334
322
340
328
340
328
340
328
322
322
322
340
340
328
328
340
340
334
334
340
340
334
334
340
328
322
322
Errors of
Commission
12
12
12
12
12
12
12
12
18
12
18
12
18
12
18
12
6
6
6
12
12
12
12
42
42
42
42
12
12
12
12
12
12
12
12
EC Rate*
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.9%
0.6%
0.9%
0.6%
0.9%
0.6%
0.9%
0.6%
0.3%
0.3%
0.3%
0.6%
0.6%
0.6%
0.6%
2.0%
2.0%
2.0%
2.0%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
Discrepant
Failures
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Probable
EC Rate
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.9%
0.6%
0.9%
0.6%
0.9%
0.6%
0.9%
0.6%
0.3%
0.3%
0.3%
0.6%
0.6%
0.6%
0.6%
2.0%
2.0%
2.0%
2.0%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
D-3
see text in Section 5 for explanation
-------
Appendix D: XM240 Outpoint Tables
IM240
Failure Rate
16%
16%
16%
16%
16%
16%
16%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
17%
18%
18%
18%
18%
Outpoints
Composite + Mode 2
1.00 / 15.0 / 2.2 + 0.62 /
1.20 / 18.0 / 2.0 + 0.75 /
1.20 / 15.0 / 2.0 + 0.75 /
1.20 / 18.0 / 2.1 + 0.75 /
1.20 / 15.0 / 2.1 + 0.75 /
1.20 / 20.0 / 1.9 + 0.75 /
1.20 / 20.0 / 2.0 + 0.75 /
0.80 / 20.0 / 2.0 + 0.50 /
0.80 / 20.0 / 2.1 + 0.50 /
0.80 / 18.0 / 2.1 + 0.50 /
0.80 / 15.0 / 2.1 + 0.50 /
0.80 / 12.0 / 2.1 + 0.50 /
0.60 / 20.0 / 2.2 + 0.37 /
0.60 / 18.0 / 2.2 + 0.37 /
0.80 / 18.0 / 2.2 + 0.50 /
0.60 / 15.0 / 2.2 + 0.37 /
0.80 / 15.0 / 2.2 + 0.50 /
0.60 / 12.0 / 2.2 + 0.37 /
0.80 / 12.0 / 2.2 + 0.50 /
0.80 / 10.0 / 2.4 + 0.50 /
0.80 / 10.0 / 2.5 + 0.50 /
1.00 / 12.0 / 1.9 + 0.62 /
1.20 / 12.0 / 1.9 + 0.75 /
1.00 / 12.0 / 2.0 + 0.62 /
1.20 / 12.0 / 2.0 + 0.75 /
1.00 / 10.0 / 2.3 + 0.62 /
1.20 / 10.0 / 2.3 + 0.75 /
1.00 / 18.0 / 1.9 + 0.62 /
1.00 / 15.0 / 1.9 + 0.62 /
1.20 / 18.0 / 1.9 + 0.75 /
1.20 / 15.0 / 1.9 + 0.75 /
0.40 / 20.0 / 2.4 + 0.25 /
0.40 / 18.0 / 2.4 + 0.25 /
0.40 / 15.0 / 2.4 + 0.25 /
0.40 / 12.0 / 2.4 + 0.25 /
12.0
14.4
12.0
14.4
12.0
16.0
16.0
16.0
16.0
14.4
12.0
9.6
16.0
14.4
14.4
12.0
12.0
9.6
9.6
8.0
8.0
9.6
9.6
9.6
9.6
8.0
8.0
14.4
12.0
14.4
12.0
16.0
14.4
12.0
9.6
Excess Emissions Identified
HC
353
351
351
351
351
349
349
367
367
367
367
367
365
365
365
365
365
365
365
365
364
359
359
359
359
357
357
356
356
351
351
371
371
371
371
CO
5491
5508
5508
5488
5488
5502
5449
5790
5770
5776
5776
5776
5720
5725
5721
5725
5721
5725
5721
5740
5740
5688
5688
5634
5634
5589
5589
5618
5618
5562
5562
5923
5923
5923
5923
NOx
273
276
276
276
276
279
276
286
285
285
285
285
279
279
279
279
279
279
279
270
270
282
282
279
279
268
268
282
282
279
279
276
276
276
276
Identification
HC
88.7%
88.0%
88.0%
88.0%
88.0%
87.6%
87.6%
92.2%
92.2%
92.2%
92.2%
92.2%
91.6%
91.6%
91.6%
91.6%
91.6%
91.6%
91.6%
91.6%
91.3%
90.2%
90.2%
90.2%
90.2%
89.6%
89.6%
89.3%
89.3%
88.0%
88.0%
93.0%
93.0%
93.0%
93.0%
CO
63.9%
64.1%
64.1%
63.9%
63.9%
64.1%
63.4%
67.4%
67.2%
67.2%
67.2%
67.2%
66.6%
66.7%
66.6%
66.7%
66.6%
66.7%
66.6%
66.8%
66.8%
66.2%
66.2%
65.6%
65.6%
65.1%
65.1%
65.4%
65.4%
64.8%
64.8%
69.0%
69.0%
69.0%
69.0%
Rates
NOx
79.7%
80.6%
80.6%
80.5%
80.5%
81.4%
80.6%
83.4%
83.3%
83.3%
83.3%
83.3%
81.5%
81.5%
81.5%
81.5%
81.5%
81.5%
81.5%
78.8%
78.8%
82.4%
82.4%
81.6%
81.6%
78.2%
78.2%
82.4%
82.4%
81.4%
81.4%
80.7%
.80.7%
80.7%
80.7%
Fails
322
334
334
328
328
334
322
358
352
358
358
358
352
358
346
358
346
358
346
358
352
358
358
346
346
358
358
352
352
346
346
382
382
382
382
Errors of
Commission
12
12
12
12
12
12
12
12
12
12
12
12
18
18
12
18
12
18
12
42
42
12
12
12
12
54
54
12
12
12
12
6
6
6
6
EC Rate*
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.9%
0.9%
0.6%
0.9%
0.6%
0.9%
0.6%
2.0%
2.0%
0.6%
0.6%
0.6%
0.6%
2.6%
2.6%
0.6%
0.6%
0.6%
0.6%
0.3%
0.3%
0.3%
0.3%
Discrepant
Failures
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Probable
EC Rate
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.9%
0.9%
0.6%
0.9%
0.6%
0.9%
0.6%
2.0%
2.0%
0.6%
0.6%
0.6%
0.6%
2.6%
2.6%
0.6%
0.6%
0.6%
0.6%
0.3%
0.3%
0.3%
0.3%
D-4
see text in Section 5 for explanation
-------
Appendix D: IM240 Outpoint Tables
IM240
Failure Rate
18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
I 18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
18%
19%
19%
19%
19%
19%
19%
Outpoints
Composite + Mode 2
0.40 / 20.0 / 2.5 -f 0.25 / 16.0
0.40 / 18.0 / 2.5 + 0.25 / 14.4
0.40 / 15.0 / 2.5 + 0.25 / 12.0
0.40 / 12.0 / 2.5 -I- 0.25 / 9.6
0.60 / 20.0 / 1.9 + 0.37 / 16.0
0.80 / 20.0 / 1.9 + 0.50 / 16.0
0.80 / 18.0 / 1.9 + 0.50 / 14.4
0.80 / 15.0 / 1.9 + 0.50 / 12.0
0.80 / 12.0 / 1.9 + 0.50 / 9.6
0.50 / 12.0 / 2.0 + 0.31 / 10.0
0.60 / 12.0 / 2.0 -I- 0.38 / 10.0
0.60 / 20.0 / 2.0 + 0.37 / 16.0
0.60 / 18.0 / 2.0 + 0.37 / 14.4
0.80 / 18.0 / 2.0 -1- 0.50 / 14.4
0.60 / 15.0 / 2.0 + 0.37 / 12.0
0.80 / 15.0 / 2.0 + 0.50 / 12.0
0.60 / 12.0 / 2.0 + 0.37 / 9.6
0.80 / 12.0 / 2.0 + 0.50 / 9.6
0.50 / 12.0 / 2.1 + 0.31 / 10.0
0.60 / 12.0 / 2.1 + 0.38 / 10.0
0.60 / 20.0 / 2.1 + 0.37 / 16.0
0.60 / 18.0 / 2.1 + 0.37 / 14.4
0.60 / 15.6 / 2.1 + 0.37 / 12.0
0.60 / 12.0 / 2.1 + 0.37 / 9.6
0.80 / 10.0 / 2.3 + 0.50 / 8.0
0.60 / 10.0 / 2.4 -I- 0.37 / 8.0
0.60 / 10.0 / 2.5 + 0.37 / 8.0
1.00 / 10.0 / 2.2 + 0.62 / 8.0
1.20 / 10.0 / 2.2 + 0.75 / 8.0
0.40 / 20.0 / 2.3 + 0.25 / 16.0
0.40 / 18.0 / 2.3 + 0.25 / 14.4
0.40 / 15.0 / 2.3 + 0.25 / 12.0
0.40 / 12.0 / 2.3 + 0.25 / 9.6
0.50 / 12.0 / 1.9 + 0.31 / 10.0
0.60 / 12.0 / 1.9 + 0.38 / 10.0
Excess Emissions Identified
HC
370
370
370
370
367
367
367
367
367
367
367
367
367
367
367
367
367
367
367
367
367
367
367
367
365
365
364
357
357
371
371
371
371
367
367
CO
5923
5923
5923
5923
5847
5844
5849
5849
5849
5799
5799
5794
5799
5796
5799
5796
5799
5796
5779
5779
5774
5779
5779
5779
5750
5744
5744
5591
5591
5933
5933
5933
5933
5853
5853
NOx
276
276
276
276
288
288
288
288
288
286
286
286
286
286
286
286
286
286
285
285
285
285
285
285
274
270
270
274
274
281
281
281
281
268
288
Identification Rates 1
HC
92.8%
92.8%
92.8%
92.8%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
91.6%
91.6%
91.3%
89.6%
89.6%
93.0%
93.0%
93 . 0,%
93.0%
92.2%
92.2%
CO
69.0%
69.0%
69.0%
69.0%
68.1%
68.0%
68.1%
68.1%
68.1%
67.5%
67.5%
67.5%
67.5%
67.5%
67.5%
67.5%
67.5%
67.5%
67.3%
67.3%
67.2%
67.3%
67.3%
67.3%
66.9%
66.9%
66.9%
65.1%
65.1%
69.1%
69.1%
69.1%
69.1%
68.1%
68.1%
NOx 1
80.7%
80.7%
80.7%
80.7%
84.2%
84.2%
84.2%
84.2%
84.2%
83.4%
83.4%
83.4%
83.4%
83.4%
83.4%
83.4%
83.4%
83.4%
83.3%
83.3%
83.3%
83.3%
83.3%
83.3%
80.0%
78.8%
78.8%
80.2%
80.2%
81.9%
81.9%
81.9%
81.9%
84.2%
84.2%
Fails
376
376
376
376
382
370
376
376
376
376
376
370
376
364
376
364
376
364
370
370
364
370
370
370
376
370
364
376
376
400
400
400
400
388
388
Errors of
Commission
6
6
6
6
18
12
12
12
12
18
18
18
18
12
18
12
18
12
18
18
18
18
18
18
54
48
48
54
54
18
18
18
18
18
18
EC Rate*
0.3%
0.3%
0.3%
0.3%
0.9%
0.6%
0.6%
0.6%
0.6%
0.9%
0.9%
0.9%
0.9%
0.6%
0.9%
0.6%
0.9%
0.6%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
2.6%
2.3%
2.3%
2.6%
2.6%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
Discrepant
Failures
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Probable
EC Rate
0.3%
0.3%
0.3%
0.3%
0.9%
0.6%
0.6%
0.6%
0.6%
0.9%
0.9%
0.9%
0.9%
0.6%
0.9%
0.6% |
0.9%
0.6%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
2.6%
2.3%
2.3%
2.6%
2.6%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
D-5
* see text in Section 5 for explanation
-------
Appendix D: IM240 Outpoint Tables
IM240
Failure Rate
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
Outpoints
Composite + Mode 2
0.50 / 12.0 / 1.8 + 0.31 / 10.0
0.60 / 12.0 / 1.8 + 0.38 / 10.0
0.60 / 18.0 / 1.9 + 0.37 / 14.4
0.60 / 15.0 / 1.9 + 0.37 / 12.0
0.60 / 12.0 / 1.9 + 0.37 / 9.6
0.80 / 10.0 / 2.2 + 0.50 / 8.0
0.60 / 10.0 / 2.3 + 0.37 / 8.0
1.00 / 10.0 / 1.9 + 0.62 / 8.0
1.20 / 10.0 / 1.9 + 0.75 / 8.0
1.00 / 10.0 / 2.0 + 0.62 / 8.0
1.20 / 10.0 / 2.0 + 0.75 / 8.0
1.00 / 10.0 / 2.1 + 0.62 / 8.0
1.20 / 10.0 / 2.1 + 0.75 / 8.0
0.40 / 12.0 / 2.1 + 0.25 / 10.0
0.40 / 12.0 / 2.0 + 0.25 / 10.0
0.40 / 20.0 / 2.2 + 0.25 / 16.0
0.40 / 18.0 / 2.2 + 0.25 / 14.4
0.40 / 15.0 / 2.2 + 0.25 / 12.0
0.40 / 12.0 / 2.2 + 0.25 / 9.6
0.40 / 20.0 / 2.1 + 0.25 / 16.0
0.40 / 18.0 / 2.1 + 0.25 / 14.4
0.40 / 15.0 / 2.1 + 0.25 / 12.0
0.40 / 12.0 / 2.1 + 0.25 / 9.6
0.40 / 20.0 / 2.0 + 0.25 / 16.0
0.40 / 18.0 / 2.0 + 0.25 / 14.4
0.40 / 15.0 / 2.0 -I- 0.25 / 12.0
0.40 / 12.0 / 2.0 + 0.25 / 9.6
0.40 / 10.0 / 2.4 + 0.25 / 8.0
0.40 / 10.0 / 2.5 + 0.25 / 8.0
0.80 / 10.0 / 1.9 + 0.50 / 8.0
0.50 / 11.0 / 2.0 + 0.31 / 9.0
0.60 / 11.0 / 2.0 + 0.38 / 9.0
0.50 / 10.0 / 2.0 + 0.31 / 8.0
0.60 / 10.0 / 2.0 + 0.38 / 8.0
0.60 / 10.0 / 2.0 + 0.37 / 8.0
Excess Emissions Identified
HC
367
367
367
367
367
365
365
359
359
359
359
359
359
371
371
371
371
371
371
371
371
371
371
371
371
371
371
371
370
367
367
367
367
367
367
CO
5853
5853
5853
5853
5853
5752
5754
5688
5688
5665
5665
5645
5645
5952
5952
5935
5935
5935
5935
5952
5952
5952
5952
5952
5952
5952
5952
5923
5923
5849
5830
5830
5830
5830
5830
NOx
288
288
288
288
288
281
274
282
282
281
281
281
281
287
287
287
287
287
287
287
287
287
287
287
287
287
287
276
276
288
287
287
287
287
287
Identification
HC
92.2%
92.2%
92.2%
92.2%
92.2%
91.6%
91.6%
90.2%
90.2%
90.2%
90.2%
90.2%
90.2%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
92.8%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
CO
68.1%
68.1%
68.1%
68.1%
68.1%
67.0%
67.0%
66.2%
66.2%
66.0%
66.0%
65.7%
65.7%
69.3%
69.3%
69.1%
69.1%
69.1%
69.1%
69.3%
69.3%
69.3%
69.3%
69.3%
69.3%
69.3%
69.3%
69.0%
69.0%
68.1%
67.9%
67.9%
67.9%
67.9%
67.9%
Rates
NOx
84.2%
84.2%
84.2%
84.2%
84.2%
82.0%
80.0%
82.4%
82.4%
82.1%
82.1%
82.0%
82.0%
83.9%
83.9%
83.9%
83.9%
83.9%
83.9%
83.9%
83.9%
83.9%
83.9%
83.9%
83.9%
83.9%
83.9%
80.7%
80.7%
84.2%
83.9%
83.9%
83.9%
83.9%
83.9%
Fails
388
388
388
388
388
394
388
400
400
394
394
388
388
424
424
418
418
418
418
424
424
424
424
424
424
424
424
424
418
418
424
424
424
424
424
Errors of
Commission
18
18
18
18
18
54
60
54
54
54
54
54
54
18
18
18
18
18
18
18
18
18
18
18
18
18
18
48
48
54
60
60
60
60
60
EC Rate*
0.9%
0.9%
0.9%
0.9%
0.9%
2.6%
2.9%
2.6%
2.6%
2.6%
2.6%
2.6%
2.6%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
2.3%
2.3%
2.6%
2.9%
2.9%
2.9%
2.9%
2.9%
Discrepant
Failures
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Probable
EC Rate
0.9%
0.9%
0.9%
0.9%
0.9%
2.6%
2.9%
2.6%
2.6%
2.6%
2.6%
2.6%
2.6%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
2.3%
2.3%
2.6%
2.9%
2.9%
2.9%
2.9%
2.9%
D-6
* see text in Section 5 for explanation
-------
Appendix D: IM240 Outpoint Tables
IM240
Failure Rate
20%
20%
20%
20%
20%
20%
20%
20%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
21%
22%
22%
22%
23%
23%
23%
23%
Cutpolnts
Composite + Mode 2
0.80 / 10.0 / 2.0 + 0.50 t
0.50 / 11.0 / 2.1 + 0.31 i
0.60 / 11.0 / 2.1 + 0.38 /
0.50 / 10.0 / 2.1 + 0.31 i
0.60 / 10.0 / 2.1 4 0.38 /
0.60 / 10.0 / 2.1 + 0.37 t
0.80 / 10.0 / 2.1 + 0.50 /
0.60 / 10.0 / 2.2 + 0.37 ,
0.40 / 12.0 / 1.9 + 0.25 t
0.40 / 12.0 / 1.8 + 0.25 t
0.40 / 20.0 / 1.9 •»• 0.25 ,
0.40 / 18.0 / 1.9 + 0.25 ,
0.40 / 15.0 / 1.9 + 0.25 ,
0.40 / 12.0 / 1.9 4 0.25
0.40 / 10.0 / 2.3 + 0.25 ,
0.50 / 12.0 / 1.7 + 0.31 t
0.60 / 12.0 / 1.7 + 0.38 ,
0.50 / 12.0 / 1.6 + 0.31 t
0.60 / 12.0 / 1.6 + 0.38 t
0.50 / 11.0 / 1.9 + 0.31 t
0.60 / 11.0 / 1.9 + 0.38 i
0.50 / 10.0 / 1.9 + 0.31 t
0.60 / 10.0 / 1.9 + 0.38 >
0.50 / 11.0 / 1.8 + 0.31 ,
0.60 / 11.0 / 1.8 + 0.38 i
0.50 / 10.0 / 1.8 + 0.31 i
0.60 / 10.0 / 1.8 -I- 0.38 t
0.60 / 10.0 / 1.9 4 0.37 /
0.50 / 9.0 / 2.1 + 0.31 >
0.60 / 9.0 / 2.1 + 0.38 >
0.40 / 10.0 / 2.2 + 0.25 /
0.50 / 9.0 / 1.9 + 0.31 /
0.60 / 9.0 / 1.9 4 0.38 >
0.50 / 9.0 / 1.8 + 0.31 i
0.60 / 9.0 / 1.8 + 0.38 ,
' 8.0
f 9.0
1 9.0
f 8.0
t 8.0
f 8.0
f 8.0
1 8.0
1 10.0
f 10.0
1 16.0
f 14.4
f 12.0
f 9.6
1 8.0
f 10.0
f 10.0
f 10.0
f 10.0
' 9.0
( 9.0
I 8.0
f 8.0
t 9.0
f 9.0
f 8.0
f 8.0
f 8.0
f 7.0
t 7.0
f 8.0
t 7.0
f 7.0
t 7.0
' 7.0
Excess Emissions Identified
HC
367
367
367
367
367
367
367
365
371
371
371
371
371
371
371
368
368
368
368
367
367
367
367
367
367
367
367
367
372
372
371
372
372
372
372
CO
5826
5810
5810
5810
5810
5810
5807
5756
5975
5975
5975
5975
5975
5975
5933
6029
6029
6029
6029
5853
5853
5853
5853
5853
5853
5853
5853
5853
6124
6124
5935
6167
6167
6167
6167
NOx
287
287
287
287
287
287
287
281
288
288
288
288
288
288
281
299
299
299
299
288
288
288
288
288
288
288
288
288
309
309
287
310
310
310
310
Identification
HC
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
91.6%
93.1%
93.1%
93.1%
93.1%
93.1%
93.1%
93.0%
92.3%
92.3%
92.3%
92.3%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
92.2%
93.5%
93.5%
93.0%
93.5%
93.5%
93.5%
93.5%
CO
67.8%
67.6%
67.6%
67.6%
67.6%
67.6%
67.6%
67.0%
69.6%
69.6%
69.6%
69.6%
69.6%
69.6%
69.1%
70.2%
70.2%
70.2%
70.2%
68.1%
68.1%
68.1%
68.1%
68.1%
68.1%
68.1%
68.1%
68.1%
71.3%
71.3%
69.1%
71.8%
71.8%
71.8%
71.8%
Rates
NOx
83.
83.
83.
83.
83.
83.
83.
82.
84.
84.
84.
84.
84.
84.
81.
87.
87.
87.
87.
84.
84.
84.
84.
84.
84.
84.
84.
84.
90.
90.
83.
90.
90.
90.
90.
9%
8%
8%
8%
8%
8%
8%
0%
2%
2%
2%
2%
2%
2%
9%
3%
3%
3%
3%
2%
2%
2%
2%
2%
2%
2%
2%
2%
3%
3%
9%
7%
7%
7%
7%
Falls
412
418
418
418
418
418
406
406
430
430
430
430
430
430
442
430
430
430
430
430
430
430
430
430
430
430
430
430
460
460
460
472
472
472
472
Errors of
Commission
54
60
60
60
60
60
54
60
18
18
18
18
18
18
60
18
18
18
18
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
EC Rate*
2.6%
2.9%
2.9%
2.9%
2.9%
2.9%
2.6%
2.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
2.9%
0.9%
0.9%
0.9%
0.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
Discrepant
Failures
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Probable
EC Rate
2.6%
2.9%
2.9%
2.9%
2.9%
2.9%
2.6%
2.9%
0.9%
0.9%
0.9%
0.9%
0.9%
0.9%
2.9%
0.9%
0.9%
0.9%
0.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
D-7
* see text in Section 5 for explanation
-------
Appendix D: XM240 Outpoint Tables
IM240
Failure Rate
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
23%
24%
24%
25%
25%
25%
25%
25%
25%
25%
25%
Cutpolnts
Composite + Mode 2
0.50 / 9.0 / 2.0 + 0.31 /
0.60 / 9.0 / 2.0 + 0.38 /
0.40 / 12.0 / 1.7 + 0.25 /
0.40 / 12.0 / 1.6 + 0.25 /
0.40 / 11.0 / 1.9 + 0.25 /
0.40 / 10.0 / 1.9 + 0.25 /
0.40 / 11.0 / 1.8 + 0.25 /
0.40 / 10.0 / 1.8 + 0.25 /
0.40 / 10.0 / 1.9 + 0.25 /
0.40 / 11.0 / 2.1 + 0.25 /
0.40 / 10.0 / 2.1 + 0.25 /
0.40 / 11.0 / 2.0 + 0.25 /
0.40 / 10.0 / 2.0 + 0.25 /
0.40 / 10.0 / 2.1 + 0.25 /
0.40 / 10.0 / 2.0 + 0.25 /
0.50 / 11.0 / 1.7 + 0.31 /
0.60 / 11.0 / 1.7 + 0.38 /
0.50 / 10.0 / 1.7 + 0.31 /
0.60 / 10.0 / 1.7 + 0.38 /
0.50 / 11.0 / 1.6 + 0.31 /
0.60 / 11.0 / 1.6 -»• 0.38 /
0.50 / 10.0 / 1.6 + 0.31 /
0.60 / 10.0 / 1.6 + 0.38 /
0.50 / 12.0 / 1.5 + 0.31 /
0.60 / 12.0 / 1.5 + 0.38 /
0.50 / 8.0 / 2.1 + 0.31 /
0.60 / 8.0 / 2.1 + 0.38 /
0.40 / 9.0 / 1.9 + 0.25 /
0.40 / 8.0 / 1.9 + 0.25 /
0.50 / 8.0 / 1.9 + 0.31 /
0.60 / 8.0 / 1.9 + 0.38 /
0.40 / 9.0 / 1.8 + 0.25 /
0.40 / 8.0 / 1.8 + 0.25 /
0.50 / 8.0 / 1.8 + 0.31 /
0.60 / 8.0 / 1.8 + 0.38 /
Excess Emissions Identified
7.0
7.0
10.0
10.0
9.0
8.0
9.0
8.0
8.0
9.0
8.0
9.0
8.0
8.0
8.0
9.0
9.0
8.0
8.0
9.0
9.0
8.0
8.0
10.0
10.0
6.0
6.0
7.0
6.0
6.0
6.0
7.0
6.0
6.0
6.0
1 HC
372
372
371
371
371
371
371
371
371
371
371
371
371
371
371
368
368
368
368
368
368
368
368
368
368
376
376
376
376
376
376
376
376
376
376
CO
6144
6144
6151
6151
5975
5975
5975
5975
5975
5952
5952
5952
5952
5952
5952
6029
6029
6029
6029
6029
6029
6029
6029
6041
6041
6246
6246
6289
6289
6289
6289
6289
6289
6289
6289
NOx
310
310
299
299
288
288
288
288
288
287
287
287
287
287
287
299
299
299
299
299
299
299
299
299
299
309
309
310
310
310
310
310
310
310
310
Identification
HC
93.5%
93.5%
93.2%
93.2%
93.1%
93.1%
93.1%
93.1%
93.1%
93.0%
93.0%
93.0%
93.0%
93.0%
93.0%
92.3%
92.3%
92.3%
92.3%
92.3%
92.3%
92.3%
92.3%
92.3%
92.3%
94.3%
94.3%
94.3%
94.3%
94.3%
94.3%
94.3%
94.3%
94.3%
94.3%
CO
71.5%
71.5%
71.6%
71.6%
69.6%
69.6%
69.6%
69.6%
69.6%
69.3%
69.3%
69.3%
69.3%
69.3%
69.3%
70.2%
70.2%
70.2%
70.2%
70.2%
70.2%
70.2%
70.2%
70.3%
70.3%
72.7%
72.7%
73.2%
73.2%
73.2%
73.2%
73.2%
73.2%
73.2%
73.2%
Rates
NOx
90.4%
90.4%
87.3%
87.3%
84.2%
84.2%
84.2%
84.2%
84.2%
83.9%
83.9%
83.9%
83.9%
83.9%
83.9%
87.3%
87.3%
87.3%
87.3%
87.3%
87.3%
87.3%
87.3%
87.3%
87.3%
90.3%
90.3%
90.7%
90.7%
90.7%
90.7%
90.7%
90.7%
90.7%
90.7%
Fails
466
466
472
472
472
472
472
472
472
466
466
466
466
466
466
472
472
472
472
472
472
472
472
472
472
502
502
514
514
514
514
514
514
514
514
Errors of
Commission
60
60
18
18
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
18
18
60
60
60
60
60
60
60
60
60
60
EC Rate*
2.9%
2.9%
0.9%
0.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
0.9%
0.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
Discrepant
Failures
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
48
48
0
0
0
0
0
0
0
0
0
0
Probable
EC Rate
2.9%
2.9%
0.9%
0.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
3.2%
3.2%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
D-8
* see text in Section 5 tor explanation
-------
Appendix D: IM240 Outpoint Tables
IM240
Failure Rate
25%
25%
25%
25%
25%
25%
25%
25%
25%
25%
25%
25%
25%
25%
25%
25%
25%
25%
25%
27%
27%
27%
27%
27%
27%
27%
27%
27%
27%
27%
27%
29%
29%
29%
29%
Outpoints
Composite •»• Mode 2
0.40 / 9.0 / 2.1 + 0.25 / 7.0
0.40 / 8.0 / 2.1 + 0.25 / 6.0
0.40 / 9.0 / 2.0 + 0.25 / 7.0
0.40 / 8.0 / 2.0 + 0.25 / 6.0
0.50 / 8.0 / 2.0 + 0.31 / 6.0
0.60 / 8.0 / 2.0 -I- 0.38 / 6.0
0.50 / 9.0 / 1.7 + 0.31 / 7.0
0.60 / 9.0 / 1.7 + 0.38 / 7.0
0.50 / 9.0 / 1.6 + 0.31 / 7.0
0.60 / 9.0 / 1.6 + 0.38 / 7.0
0.40 / 11.0 / 1.7 + 0.25 / 9.0
0.40 / 10.0 / 1.7 + 0.25 / 8.0
0.40 / 11.0 / 1.6 + 0.25 / 9.0
0.40 / 10.0 / 1.6 + 0.25 / 8.0
0.40 / 12.0 / 1.5 + 0.25 / 10.0
0.50 / 11.0 / 1.5 + 0.31 / 9.0
0.60 / 11.0 / 1.5 + 0.38 / 9.0
0.50 / 10.0 / 1.5 + 0.31 / 8.0
0.60 / 10.0 / 1.5 + 0.38 / 8.0
0.40 / 9.0 / 1.7 + 0.25 / 7.0
0.40 / 8.0 / 1.7 + 0.25 / 6.0
0.50 / 8.0 / 1.7 + 0.31 / 6.0
0.60 / 8.0 / 1.7 + 0.38 / 6.0
0.40 / 9.0 / 1.6 + 0.25 / 7.0
0.40 / 8.0 / 1.6 + 0.25 / 6.0
0.50 / 8.0 / 1.6 + 0.31 / 6.0
0.60 / 8.0 / 1.6 + 0.38 / 6.0
0.50 / 9.0 / 1.5 + 0.31 / 7.0
0.60 / 9.0 / 1.5 + 0.38 / 7.0
0.40 / 11.0 / 1.5 + 0.25 / 9.0
0.40 / 10.0 / 1.5 + 0.25 / 8.0
0.30 / 12.0 / 2.1 + 0.19 / 10.0
0.30 / 11.0 / 2.1 + 0.19 / 9.0
0.30 / 10.0 / 2.1 + 0.19 / 8.0
0.30 / 12.0 / 2.0 + 0.19 / 10.0
Excess Emissions Identified
HC
376
376
376
376
376
376
373
373
373
373
371
371
371
371
371
368
368
368
368
376
376
376
376
376
376
376
376
373
373
371
371
377
377
377
377
CO
6266
6266
6266
6266
6266
6266
6343
6343
6343
6343
6151
6151
6151
6151
6163
6041
6041
6041
6041
6465
6465
6465
6465
6465
6465
6465
6465
6355
6355
6163
6163
6228
6228
6228
6228
NOx
310
310
310
310
310
310
321
321
321
321
299
299
299
299
299
299
299
299
299
321
321
321
321
321
321
321
321
321
321
299
299
300
300
300
300
Identification
HC
94.3%
94.3%
94.3%
94.3%
94.3%
94.3%
93.6%
93.6%
93.6%
93.6%
93.2%
93.2%
93.2%
93.2%
93.2%
92.3%
92.3%
92.3%
92.3%
94.4%
94.4%
94.4%
94.4%
94.4%
94.4%
94.4%
94.4%
93.6%
93.6%
93.2%
93.2%
94.5%
94.5%
94.5%
94.5%
CO
73.0%
73.0%
73.0%
73.0%
73.0%
73.0%
73.9%
73.9%
73.9%
73.9%
71.6%
71.6%
71.6%
71.6%
71.8%
70.3%
70.3%
70.3%
70.3%
75.3%
75.3%
75.3%
75.3%
75.3%
75.3%
75.3%
75.3%
74.0%
74.0%
71.8%
71.8%
72.5%
72.5%
72.5%
72.5%
Rates
NOx
90.4%
90.4%
90.4%
90.4%
90.4%
90.4%
93.8%
93.8%
93.8%
93.8%
87.3%
87.3%
87.3%
87.3%
87.3%
87.3%
87.3%
87.3%
87.3%
93.8%
93.8%
93.8%
93.8%
93.8%
93.8%
93.8%
93.8%
93.8%
93.8%
87.3%
87.3%
87.5%
87.5%
87.5%
87.5%
Falls
508
508
508
508
508
508
514
514
514
514
514
514
514
514
514
514
514
514
514
556
556
556
556
556
556
556
556
556
556
556
556
598
598
598
598
Errors of
Commission
60
60
60
60
60
60
60
60
60
60
60
60
60
60
18
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
EC Rate*
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
0.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
Discrepant
Failures
0
0
0
0
0
0
0
0
0
6
0
0
0
0
48
48
48
42
42
0
0
0
0
0
0
0
0
42
42
48
42
0
0
0
0
Probable
EC Rate
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
3.2%
5.2%
5.2%
4.9%
4.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
4.9%
4.9%
5.2%
4.9%
2.9%
2.9%
2.9%
2.9%
D-9
* see text in Section 5 for explanation
-------
Appendix D: IM240 Outpoint Tables
IM240
Failure Rate
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
29%
31%
31%
31%
31%
31%
31%
31%
31%
31%
31%
31%
31%
31%
31%
33%
33%
33%
33%
33%
33%
33%
35%
35%
Cutpolnts
Composite + Mode 2
0.30 / 11.0 / 2.0 + 0.19 / 9.0
0.30 / 10.0 / 2.0 + 0.19 / 8.0
0.30 / 12.0 / 1.9 + 0.19 / 10.0
0.30 / 11.0 / 1.9 + 0.19 / 9.0
0.30 / 10.0 / 1.9 + 0.19 / 8.0
0.30 / 12.0 / 1.8 + 0.19 / 10.0
0.30 / 11.0 / 1.8 + 0.19 / 9.0
0.30 / 10.0 / 1.8 + 0.19 / 8.0
0.40 / 9.0 / 1.5 + 0.25 / 7.0
0.40 / 8.0 / 1.5 + 0.25 / 6.0
0.50 / 8.0 / 1.5 + 0.31 / 6.0
0.60 / 8.0 / 1.5 + 0.38 / 6.0
0.30 / 9.0 / 2.1 + 0.19 / 7.0
0.30 / 8.0 / 2.1 + 0.19 / 6.0
0.30 / 9.0 / 2.0 + 0.19 / 7.0
0.30 / 8.0 / 2.0 + 0.19 / 6.0
0.30 / 9.0 / 1.9 + 0.19 / 7.0
0.30 / 8.0 / 1.9 + 0.19 / 6.0
0.30 / 9.0 / 1.8 + 0.19 / 7.0
0.30 / 8.0 / 1.8 + 0.19 / 6.0
0.30 / 12.0 / 1.7 + 0.19 / 10.0
0.30 / 11.0 / 1.7 + 0.19 / 9.0
0.30 / 10.0 / 1.7 + 0.19 / 8.0
0.30 / 12.0 / 1.6 + 0.19 / 10.0
0.30 / 11.0 / 1.6 + 0.19 / 9.0
0.30 / 10.0 / 1.6 + 0.19 / 8.0
0.30 / 9.0 / 1.7 + 0.19 / 7.0
0.30 / 8.0 / 1.7 + 0.19 / 6.0
0.30 / 9.0 / 1.6 + 0.19 / 7.0
0.30 / 8.0 / 1.6 + 0.19 / 6.0
0.30 / 12.0 / 1.5 + 0.19 / 10.0
0.30 / 11.0 / 1.5 + 0.19 / 9.0
0.30 / 10.0 / 1.5 + 0.19 / 8.0
0.30 / 9.0 / 1.5 + 0.19 / 7.0
0.30 / 8.0 / 1.5 + 0.19 / 6.0
Excess Emissions Identified
HC
377
377
377
377
377
377
377
377
376
376
376
376
382
382
382
382
382
382
382
382
377
377
377
377
377
377
382
382
382
382
377
377
377
382
382
CO
6228
6228
6228
6228
6228
6228
6228
6228
6477
6477
6477
6477
6543
6543
6543
6543
6543
6543
6543
6543
6405
6405
6405
6405
6405
6405
6719
6719
6719
6719
6417
6417
6417
6731
6731
NOx
300
300
300
300
300
300
300
300
321
321
321
321
322
322
322
322
322
322
322
322
310
310
310
310
310
310
332
332
332
332
310
310
310
332
332
Identification
HC CO
94.5% 72.5%
94.5% 72.5%
94.5% 72.5%
94.5% 72.5%
94.5% 72.5%
94.5% 72.5%
94.5% 72.5%
94.5% 72.5%
94.4% 75.4%
94.4% 75.4%
94.4% 75.4%
94.4% 75.4%
95.8% 76.2%
95.8% 76.2%
95.8% 76.2%
95.8% 76.2%
95.8% 76.2%
95.8% 76.2%
95.8% 76.2%
95.8% 76.2%
94.6% 74.6%
94.6% 74.6%
94.6% 74.6%
94.6% 74.6%
94.6% 74.6%
94.6% 74.6%
95.9% 78.2%
95.9% 78.2%
95.9% 78.2%
95.9% 78.2%
94.6% 74.7%
94.6% 74.7%
94.6% 74.7%
95.9% 78.4%
95.9% 78.4%
Rates
NOx
87.5%
87.5%
87.5%
87.5%
87.5%
87.5%
87.5%
87.5%
93.8%
93.8%
93.8%
93.8%
94.0%
94.0%
94.0%
94.0%
94.0%
94.0%
94.0%
94.0%
90.6%
90.6%
90.6%
90.6%
90.6%
90.6%
97.1%
97.1%
97.1%
97.1%
90.6%
90.6%
90.6%
97.1%
97.1%
Falls
598
598
598
598
598
598
598
598
598
598
598
598
639
639
639
639
639
639
639
639
639
639
639
639
639
639
681
681
681
681
681
681
681
723
723
Errors of
Commission
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
EC Rate*
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
Discrepant
Failures
0
0
0
0
0
0
0
0
42
42
42
42
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
48
48
42
42
42
Probable
EC Rate
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
4.9%
4.9%
4.9%
4.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
2.9%
5.2%
5.2%
4.9%
4.9%
4.9%
D-10
* see text In Section 5 for explanation
-------
Appendix D: IM240 Outpoint Tables
IM240
Failure Rate
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
47%
49%
49%
49%
49%
49%
Outpoints
Composite + Mode 2
0.20 / 12.0 / 2.1 + 0.13 / 10.0
0.20 / 11.0 / 2.1 + 0.13 / 9.0
0.20 / 10.0 / 2.1 + 0.13 / 8.0
0.20 / 9.0 / 2.1 + 0.13 / 7.0
0.20 / 8.0 / 2.1 + 0.13 / 6.0
0.20 / 12.0 / 2.0 + 0.13 / 10.0
0.20 / 11.0 / 2.0 + 0.13 / 9.0
0.20 / 10.0 / 2.0 -I- 0.13 / 8.0
0.20 / 9.0 / 2.0 + 0.13 / 7.0
0.20 / 8.0 / 2.0 + 0.13 / 6.0
0.20 / 12.0 / 1.9 + 0.13 / 10.0
0.20 / 11.0 / 1.9 + 0.13 / 9.0
0.20 / 10.0 / 1.9 + 0.13 / 8.0
0.20 / 9.0 / 1.9 + 0.13 / 7.0
0.20 / 8.0 / 1.9 + 0.13 / 6.0
0.20 / 12.0 / 1.8 + 0.13 / 10.0
0.20 / 11.0 / 1.8 + 0.13 / 9.0
0.20 / 10.0 / 1.8 + 0.13 / 8.0
0.20 / 9.0 / 1.8 + 0.13 / 7.0
0.20 / 8.0 / 1.8 + 0.13 / 6.0
0.20 / 12.0 / 1.7 + 0.13 / 10.0
0.20 / 11.0 / 1.7 + 0.13 / 9.0
0.20 / 10.0 / 1.7 + 0.13 / 8.0
0.20 / 9.0 / 1.7 + 0.13 / 7.0
0.20 / 8.0 / 1.7 -I- 0.13 / 6.0
0.20 / 12.0 / 1.6 + 0.13 / 10.0
0.20 / 11.0 / 1.6 + 0.13 / 9.0
0.20 / 10.0 / 1.6 + 0.13 / 8.0
0.20 / 9.0 / 1.6 + 0.13 / 7.0
0.20 / 8.0 / 1.6 + 0.13 / 6.0
0.20 / 12.0 / 1.5 + 0.13 / 10.0
0.20 / 11.0 / 1.5 + 0.13 / 9.0
0.20 / 10.0 / 1.5 + 0.13 / 8.0
0.20 / 9.0 / 1.5 + 0.13 / 7.0
0.20 / 8.0 / 1.5 -I- 0.13 / 6.0
(Excess Emissions Identified
HC
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
390
CO
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7126
7138
7138
7138
7138
7138
NOx
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
332
Identification
HC
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
98.0%
CO
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.0%
83.1%
83.1%
83.1%
83.1%
83.1%
Rates
NOx
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
97.1%
Fails
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
975
1017
1017
1017
1017
1017
Errors of
Commission
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
227
EC Rate*
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
Discrepant
Failures
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
48
48
42
42
42
Probable
EC Rate
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
11.0%
13.3%
13.3%
13.0%
13.0%
13.0%
D-ll
* see text In Section 5 for explanation
-------
Appendix E
ASM Outpoint Tables
-------
Appendix E: ASM Outpoint Tables
ASM
Failure Rate
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
10%
11%
11%
11%
11%
11%
Cutpolnts
0.80 / 15.0 / 2.5
0.80 / 15.0 / 2.4
1.00 / 15.0 / 2.4
1.20 / 15.0 / 2.4
0.80 / 20.0 / 2.5
0.80 / 18.0 / 2.5
0.80 / 20.0 / 2.4
1.00 / 20.0 / 2.4
1.20 / 20.0 / 2.4
0.80 / 18.0 / 2.4
1.00 / 18.0 / 2.4
1.20 / 18.0 / 2.4
0.80 / 20.0 / 2.2
1.00 / 20.0 / 2.2
1.20 / 20.0 / 2.2
0.80 / 18.0 / 2.2
1.00 / 18.0 / 2.2
1.20 / 18.0 / 2.2
0.80 / 20.0 / 2.1
1.00 / 20.0 / 2.1
1.20 / 20.0 / 2.1
0.80 / 18.0 / 2.1
1.00 / 18.0 / 2.1
1.20 / 18.0 / 2.1
1.00 / 15.0 / 2.5
1.20 / 15.0 / 2.5
1.00 / 20.0 / 2.5
1.20 / 20.0 / 2.5
1.00 / 18.0 / 2.5
1.20 / 18.0 / 2.5
0.60 / 15.0 / 2.5
0.60 / 15.0 / 2.4
0.60 / 20.0 / 2.5
0.60 / 18.0 / 2.5
0.60 / 20.0 / 2.4
Excess Emissions Identified
HC
279
279
279
279
275
275
275
275
275
275
275
275
275
275
275
275
275
275
275
275
275
275
275
275
271
271
267
267
267
267
295
295
291
291
291
CO
4631
4631
4631
4631
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4562
4591
4591
4521
4521
4521
4521
4754
4754
4685
4685
4685
NOx
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
222
218
218
218
218
218
218
235
235
235
235
235
Identification Rates 1
HC CO
70.0% 53.9%
70.0% 53.9%
70.0% 53.9%
70.0% 53.9%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
69.1% 53.1%
68.0% 53.4%
68.0% 53.4%
67.1% 52.6%
67.1% 52.6%
67.1% 52.6%
67.1% 52.6%
74.1% 55.4%
74.1% 55.4%
73.1% 54.5%
73.1% 54.5%
73.1% 54.5%
NOx 1
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
63.6%
63.6%
63.6%
63.6%
63.6%
63.6%
68.7%
68.7%
68.7%
68.7%
68.7%
Falls
215
215
215
215
209
209
209
209
209
209
209
209
215
215
215
215
215
215
215
215
215
215
215
215
209
209
203
203
203
203
233
233
227
227
227
Errors of
Commission
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
EC Rate*
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
Discrepant
Failures
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
Probable
EC Rate
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
E-2
* see Section 5 for explanation
-------
Appendix E: ASM Outpoint Tables
ASM
Failure Rate
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
11%
12%
12%
12%
12%
12%
12%
12%
12%
12%
12%
12%
12%
12%
12%
12%
Outpoints
0.60 / 18.0 / 2.4
0.60 / 20.0 / 2.2
0.60 / 18.0 / 2.2
0.60 / 20.0 / 2.1
0.60 / 18.0 / 2.1
0.80 / 15.0 / 2.0
1.00 / 15.0 / 2.0
1.20 / 15.0 / 2.0
0.80 / 15.0 / 2.2
1.00 / 15.0 / 2.2
1.20 / 15.0 / 2.2
0.80 / 15.0 / 2.1
1.00 / 15.0 / 2.1
1.20 / 15.0 / 2.1
0.80 / 20.0 / 2.0
1.00 / 20.0 / 2.0
1.20 / 20.0 / 2.0
0.80 / 18.0 / 2.0
1.00 / 18.0 / 2.0
1.20 / 18.0 / 2.0
0.60 / 15.0 / 1.9
0.60 / 15.0 / 2.0
0.60 / 15.0 / 2.2
0.60 / 15.0 / 2.1
0.60 / 20.0 / 1.9
0.60 / 18.0 / 1.9
0.60 / 20.0 / 2.0
0.60 / 18.0 / 2.0
0.80 / 15.0 / 1.9
1.00 / 15.0 / 1.9
1.20 / 15.0 / 1.9
0.80 / 20.0 / 1.9
1.00 / 20.0 / 1.9
1.20 / 20.0 / 1.9
0.80 / 18.0 / 1.9
Excess Emissions Identified
HC
291
291
291
291
291
279
279
279
279
279
279
279
279
279
275
275
275
275
275
275
296
295
295
295
292
292
291
291
284
284
284
281
281
281
281
CO
4685
4685
4685
4685
4685
4631
4631
4631
4631
4631
4631
4631
4631
4631
4562
4562
4562
4562
4562
4562
4764
4754
4754
4754
4695
4695
4685
4685
4698
4698
4698
4629
4629
4629
4629
NOx
235
235
235
235
235
225
225
225
222
222
222
222
222
222
225
225
225
225
225
225
249
238
235
235
249
249
238
238
246
246
246
246
246
246
246
Identification
HC
73.1%
73.1%
73.1%
73.1%
73.1%
70.0%
70.0%
70.0%
70.0%
70.0%
70.0%
70.0%
70.0%
70.0%
69.1%
69.1%
69.1%
69.1%
69.1%
69.1%
74.3%
74.1%
74.1%
74.1%
73.4%
73.4%
73.1%
73.1%
71.4%
71.4%
71.4%
70.4%
70.4%
70.4%
70.4%
CO
54.5%
54.5%
54.5%
54.5%
54.5%
53.9%
53.9%
53.9%
53.9%
53.9%
53.9%
53.9%
53.9%
53.9%
53.1%
53.1%
53.1%
53.1%
53.1%
53.1%
55.5%
55.4%
55.4%
55.4%
54.7%
54.7%
54.5%
54.5%
54.7%
54.7%
54.7%
53.9%
53.9%
53.9%
53.9%
Rates
NOx
68.7%
68.7%
68.7%
68.7%
68.7%
65.9%
65.9%
65.9%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.9%
65.9%
65.9%
65.9%
65.9%
65.9%
72.8%
69.6%
68.7%
68.7%
72.8%
72.8%
69.6%
69.6%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
Falls
227
233
233
233
233
227
227
227
221
221
221
221
221
221
221
221
221
221
221
221
257
245
239
239
251
251
239
239
245
245
245
239
239
239
239
Errors of
Commission
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
EC Rate*
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
Discrepant
Failures
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
Probable
EC Rate
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
E-3
* see Section 5 for explanation
-------
Appendix E: ASM Outpoint Tables
ASM
Failure Rate
12%
12%
12%
12%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13%
13t
Outpoints
1.00 / 18.0 / 1.9
1.20 / 18.0 / 1.9
1.00 / 12.0 / 2.5
1.20 / 12.0 / 2.5
0.60 / 20.0 / 1.7
0.60 / 18.0 / 1.7
0.80 / 15.0 / 1.7
1.00 / 15.0 / 1.7
1.20 / 15.0 / 1.7
0.80 / 20.0 / 1.7
1.00 / 20.0 / 1.7
1.20 / 20.0 / 1.7
0.80 / 18.0 / 1.7
1.00 / 18.0 / 1.7
1.20 / 18.0 / 1.7
0.80 / 10.0 / 2.5
0.80 / 10.0 / 2.4
1.00 / 10.0 / 2.4
1.20 / 10.0 / 2.4
0.80 / 10.0 / 2.2
1.00 / 10.0 / 2.2
1.20 / 10.0 / 2.2
0.80 / 10.0 / 2.1
1.00 / 10.0 / 2.1
1.20 / 10.0 / 2.1
0.80 / 12.0 / 2.0
1.00 / 12.0 / 2.0
0.80 / 12.0 / 2.0
1.00 / 12.0 / 2.0
1.20 / 12.0 / 2.0
0.80 / 12.0 / 2.5
0.80 / 12.0 / 2.4
1.00 / 12.0 / 2.4 '
1.20 / 12.0 / 2.4
0.80 / 12.0 / 2.2
Excess Emissions Identified
HC
281
281
271
271
297
297
289
289
289
285
285
285
285
285
285
281
281
281
281
281
281
281
281
281
281
279
279
279
279
279
279
279
279
279
279
CO
4629
4629
4648
4648
4851
4851
4853
4853
4853
4784
4784
4784
4784
4784
4784
4743
4743
4743
4743
4743
4743
4743
4743
4743
4743
4689
4689
4689
4689
4689
4689
4689
4689
4689
4689
NOx
246
246
218
218
267
267
263
263
263
263
263
263
263
263
263
222
222
222
222
222
222
222
222
222
222
225
225
225
-225
225
222
222
222
222
222
Identification
HC
70.4%
70.4%
68.0%
68.0%
74.4%
74.4%
72.5%
72.5%
72.5%
71.5%
71.5%
71.5%
71.5%
71.5%
71.5%
70.5%
70.5%
70.5%
70.5%
70.5%
70.5%
70.5%
70.5%
70.5%
70.5%
70.0%
70.0%
70.0%
70.0%
70.0%
70.0%
70.0%
70.0%
70.0%
70.0%
CO
53.9%
53.9%
54.1%
54.1%
56.5%
56.5%
56.5%
56.5%
56.5%
55.7%
55.7%
55.7%
55.7%
55.7%
55.7%
55.2%
55.2%
55.2%
55.2%
55.2%
55.2%
55.2%
55.2%
55.2%
55.2%
54.6%
54.6%
54.6%
54.6%
54.6%
54.6%
54.6%
54.6%
54.6%
54.6%
Rates
NOx
71.8%
71.8%
63.6%
63.6%
77.9%
77.9%
76.8%
76.8%
76.8%
76.8%
76.8%
76.8%
76.8%
76.8%
76.8%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.0%
65.9%
65.9%
65.9%
65.9%
65.9%
65.0%
65.0%
65.0%
65.0%
65.0%
Falls
239
239
257
257
275
275
269
269
269
263
263
263
263
263
263
269
269
269
269
275
275
275
275
275
275
275
275
275
275
275
263
263
263
263
269
Errors of
Commission
0
0
42
42
0
0
0
0
0
0
0
0
0
0
0
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
EC Rate*
0.0%
0.0%
2.0%
2.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
, 2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
Discrepant
Failures
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
Probable
EC Rate
0.3%
0.3%
2.3%
2.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
E-4
see Section 5 for explanation
-------
Appendix E: ASM Outpoint Tables
ASM
Failure Rate
13%
13%
13%
13%
13%
13%
13%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
Cutpolnts
1.00 / 12.0 / 2.2
1.20 / 12.0 / 2.2
0.80 / 12.0 / 2.1
1.00 / 12.0 / 2.1
1.20 / 12.0 / 2.1
1.00 / 10.0 / 2.5
1.20 / 10.0 / 2.5
0.60 / 15.0 / 1.7
0.60 / 11.0 / 2.0
0.60 / 10.0 / 2.0
0.60 / 9.0 / 2.0
0.60 / 10.0 / 2.0
0.60 / 10.0 / 2.5
0.60 / 10.0 / 2.4
0.60 / 10.0 / 2.2
0.60 / 10.0 / 2.1
0.60 / 12.0 / 2.0
0.60 / 12.0 / 2.0
0.60 / 12.0 / 2.5
0.60 / 12.0 / 2.4
0.60 / 12.0 / 2.2
0.60 / 12.0 / 2.1
0.80 / 11.0 / 1.8
1.00 / 11.0 / 1.8
0.80 / 10.0 / 1.8
1.00 / 10.0 / 1.8
0.80 / 9.0 / 1.8
1.00 / 9.0 / 1.8
0.80 / 10.0 / 1.9
1.00 / 10.0 / 1.9
1.20 / 10.0 / 1.9
0.80 / 12.0 / 1.8
1.00 / 12.0 / 1.8
0.80 / 12.0 / 1.9
1.00 / 12.0 / 1.9
Excess Emissions Identified
HC
279
279
279
279
279
273
273
300
297
297
297
297
297
297
297
297
295
295
295
295
295
295
286
286
286
286
286
286
286
286
286
284
284
284
284
CO
4689
4689
4689
4689
4689
4702
4702
4920
4866
4866
4866
4866
4866
4866
4866
4866
4812
4812
4812
4812
4812
4812
4810
4810
4810
4810
4810
4810
4810
4810
4810
4756
4756
4756
4756
NOx
222
222
222
222
222
218
218
267
238
238
238
238
235
235
235
235
238
238
235
235
235
235
246
246
246
246
246
246
246
246
246
246
246
246
246
Identification
HC
70.0%
70.0%
70.0%
70.0%
70.0%
68.5%
68.5%
75.4%
74.5%
74.5%
74.5%
74.5%
74.5%
74.5%
74.5%
74.5%
74.1%
74.1%
74.1%
74.1%
74.1%
74.1%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.4%
71.4%
71.4%
71.4%
CO
54.6%
54.6%
54.6%
54.6%
54.6%
54.7%
54.7%
57.3%
56.7%
56.7%
56.7%
56.7%
56.7%
56.7%
56.7%
56.7%
56.0%
56.0%
56.0%
56.0%
56.0%
56.0%
56.0%
56.0%
56.0%
56.0%
56.0%
56.0%
56.0%
56.0%
56.0%
55.4%
55.4%
55.4%
55.4%
Rates
NOx
65.0%
65.0%
65.0%
65.0%
65.0%
63.6%
63.6%
77.9%
69.6%
69.6%
69.6%
69.6%
68.7%
68.7%
68.7%
68.7%
69.6%
69.6%
68.7%
68.7%
68.7%
68.7%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
71.8%
Falls
269
269
269
269
269
263
263
281
299
299
299
299
287
287
293
293
293
293
281
281
287
287
299
299
299
299
299
299
299
299
299
293
293
293
293
Errors of
Commission
42
42
42
42
42
42
42
0
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
EC Rate*
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
0.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
Discrepant
Failures
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
Probable
EC Rate
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
0.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
E-5
see Section 5 for explanation
-------
Appendix E: ASM Outpoint Tables
ASM
Failure Rate
14%
14%
14%
14%
14%
14%
14%
14%
14%
14%
15%
15%
15%
15%
15%
15%
15%
15%
15%
16%
16%
16%
16%
16%
18%
18%
19%
19%
19%
19%
19%
19%
19%
19%
19%
(Excess Emissions Identified
Outpoints
1.20 / 12.0 / 1.9
0.80 / 11.0 / 2.0
1.00 / 11.0 / 2.0
0.80 / 10.0 / 2.0
1.00 / 10.0 / 2.0
0.80 / 9.0 / 2.0
1.00 / 9.0 / 2.0
0.80 / 10.0 / 2.0
1.00 / 10.0 / 2.0
1.20 / 10.0 / 2.0
0.60 / 11.0 / 1.8
0.60 / 10.0 / 1.8
0.60 / 9.0 / 1.8
0.60 / 10.0 / 1.9
0.60 / 12.0 / 1.8
0.60 / 12.0 / 1.9
0.80 / 12.0 / 1.7
1.00 / 12.0 / 1.7
1.20 / 12.0 / 1.7
0.60 / 10.0 / 1.7
0.60 / 12.0 / 1.7
0.80 / 10.0 / 1.7
1.00 / 10.0 / 1.7
1.20 / 10.0 / 1.7
0.80 / 8.0 / 2.0
1.00 / 8.0 / 2.0
0.40 / 20.0 / 2.0
0.40 / 18.0 / 2.0
0.40 / 15.0 / 2.0
0.40 / 20.0 / 2.5
0.40 / 18.0 / 2.5
0.40 / 15.0 / 2.5
0.40 / 20.0 / 2.4
0.40 / 18.0 / 2.4
0.40 / 15.0 / 2.4
1 HC
284
281
281
281
281
281
281
281
281
281
298
298
298
298
296
296
289
289
289
302
300
291
291
291
298
298
316
316
316
316
316
316
316
316
316
CO
4756
4743
4743
4743
4743
4743
4743
4743
4743
4743
4876
4876
4876
4876
4822
4822
4911
4911
4911
5032
4978
4965
4965
4965
5251
5251
5100
5100
5100
5100
5100
5100
5100
5100
5100
NOx
246
225
225
225
225
225
225
225
225
225
249
249
249
249
249
249
263
263
263
267
267
263
263
263
233
233
241
241
241
237
237
237
237
237
237
Identification
HC
71.4%
70.5%
70.5%
70.5%
70.5%
70.5%
70.5%
70.5%
70.5%
70.5%
74.8%
74.8%
74.8%
74.8%
74.3%
74.3%
72.5%
72.5%
72.5%
75.8%
75.4%
72.9%
72.9%
72.9%
74.7%
74.7%
79.4%
79.4%
79.4%
79.4%
79.4%
79.4%
79.4%
79.4%
79.4%
CO
55.4%
55.2%
55.2%
55.2%
55.2%
55.2%
55.2%
55.2%
55.2%
55.2%
56.8%
56.8%
56.8%
56.8%
56.1%
56.1%
57.2%
57.2%
57.2%
58.6%
58.0%
57.8%
57.8%
57.8%
61.1%
61.1%
59.4%
59.4%
59.4%
59.4%
59.4%
59.4%
59.4%
59.4%
59.4%
Rates
NOx
71.8%
65.9%
65.9%
65.9%
65.9%
65.9%
65.9%
65.9%
65.9%
65.9%
72.8%
72.8%
72.8%
72.8%
72.8%
72.8%
76.8%
76.8%
76.8%
77.9%
77.9%
76.8%
76.8%
76.8%
68.0%
68.0%
70.3%
70.3%
70.3%
69.4%
69.4%
69.4%
69.4%
69.4%
69.4%
Falls
293
281
281
281
281
281
281
281
281
281
311
311
311
311
305
305
317
317
317
335
329
323
323
323
377
377
400
400
400
388
388
388
388
388
388
Errors of
Commission
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
6
6
6
6
6
6
6
6
6
EC Rate*
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
Discrepant
Failures
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
Probable
EC Rate
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
E-6
* see Section 5 for explanation
-------
Appendix E: ASM Outpoint Tables
ASM
Failure Rate
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
19%
20%
20%
20%
20%
21%
21%
21%
21%
21%
21%
21%
21%
21%
22%
Cutpolnts
0.40 / 20.0 / 2.2
0.40 / 18.0 / 2.2
0.40 / 15.0 / 2.2
0.40 / 20.0 / 2.1
0.40 / 18.0 / 2.1
0.40 / 15.0 / 2.1
0.60 / 8.0 / 2.0
0.60 / 11.0 / 1.5
0.60 / 10.0 / 1.5
0.60 / 9.0 / 1.5
0.60 / 12.0 / 1.5
0.80 / 11.0 / 1.5
1.00 / 11.0 / 1.5
0.80 / 10.0 / 1.5
1.00 / 10.0 / 1.5
0.80 / 9.0 / 1.5
1.00 / 9.0 / 1.5
0.80 / 12.0 / 1.5
1.00 / 12.0 / 1.5
0.80 / 8.0/1.8
1.00 / 8.0/1.8
0.40 / 20.0 / 1.9
0.40 / 18.0 / 1.9
0.40 / 15.0 / 1.9
0.60 / 8.0/1.8
0.40 / 20.0 / 1.7
0.40 / 18.0 / 1.7
0.40 / 15.0 / 1.7
0.40 / 10.0 / 2.5
0.40 / 10.0 / 2.4
0.40 / 12.0 / 2.5
0.40 / 12.0 / 2.4
0.40 / 12.0 / 2.2
0.40 / 12.0 / 2.1
0.40 / 11.0 / 2.0
Excess Emissions Identified
HC
316
316
316
316
316
316
314
313
313
313
311
306
306
306
306
306
306
304
304
303
303
317
317
317
315
321
321
321
318
318
316
316
316
316
318
CO
5100
5100
5100
5100
5100
5100
5375
5282
5282
5282
5228
5272
5272
5272
5272
5272
5272
5218
5218
5318
5318
5110
5110
5110
5385
5266
5266
5266
5212
5212
5158
5158
5158
5158
5212
NOx
237
237
237
237
237
237
245
274
274
274
274
274
274
274
274
274
274
274
274
253
253
252
252
252
257
269
269
269
237
237
237
237
237
237
241
Identification
HC CO
79.4% 59.4%
79.4% 59.4%
79.4% 59.4%
79.4% 59.4%
79.4% 59.4%
79.4% 59.4%
78.7% 62.6%
78.5% 61.5%
78.5% 61.5%
78.5% 61.5%
78.0% 60.9%
76.8% 61.4%
76.8% 61.4%
76.8% 61.4%
76.8% 61.4%
76.8% 61.4%
76.8% 61.4%
76.3% 60.8%
76.3% 60.8%
76.0% 61.9%
76.0% 61.9%
79.6% 59.5%
79.6% 59.5%
79.6% 59.5%
78.9% 62.7%
80.7% 61.3%
80.7% 61.3%
80.7% 61.3%
79.8% 60.7%
79.8% 60.7%
79.4% 60.1%
79.4% 60.1%
79.4% 60.1%
79.4% 60.1%
79.8% 60.7%
Rates
NOx
69.4%
69.4%
69.4%
69.4%
69.4%
69.4%
71.7%
80.1%
80.1%
80.1%
80.1%
80.0%
80.0%
80.0%
80.0%
80.0%
80.0%
80.0%
80.0%
73.9%
73.9%
73.5%
73.5%
73.5%
75.0%
78.6%
78.6%
78.6%
69.4%
69.4%
69.4%
69.4%
69.4%
69.4%
70.3%
Falls
394
394
394
394
394
394
394
400
400
400
394
394
394
394
394
394
394
388
388
394
394
412
412
412
406
436
436
436
442
442
436
436
442
442
454
Errors of
Commission
6
6
6
6
6
6
42
42
42
42
42
42
42
42
42
42
42
42
42
42
42
6
6
6
42
6
6
6
48
48
48
48
48
48
48
EC Rate*
0.3%
0.3%
0.3%
0.3%
0.3%
0.3%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
2.0%
0.3%
0.3%
0.3%
2.0%
0.3%
0.3%
0.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
Discrepant
Failures
6
6
6
6
6
6
6
48
48
48
48
48
48
48
48
48
48
48
48
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
Probable
EC Rate
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
2.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
2.3%
2.3%
0.6%
0.6%
0.6%
2.3%
0.6%
0.6%
0.6%
2.6%
2.6%
2.6%
2.6%
2.6%
2.6%
2.6%
E-7
* see Section 5 for explanation
-------
Appendix E: ASM Outpoint Tables
ASM
Failure Rate
22%
22%
22%
22%
22%
22%
22%
22%
22%
23%
23%
23%
23%
23%
23%
23%
23%
24%
24%
24%
24%
24%
24%
24%
24%
24%
24%
24%
24%
24%
24%
25%
26%
26%
26%
Outpoints
0.40 / 10.0 / 2.0
0.40 / 9.0 / 2.0
0.40 / 10.0 / 2.0
0.40 / 10.0 / 2.2
0.40 / 10.0 / 2.1
0.40 / 12.0 / 1.8
0.40 / 12.0 / 1.9
0.40 / 12.0 / 2.0
0.40 / 12.0 / 2.0
0.60 / 8.0 / 1.5
0.40 / 12.0 / 1.7
0.40 / 11.0 / 1.8
0.40 / 10.0 / 1.8
0.40 / 9.0 / 1.8
0.40 / 10.0 / 1.9
0.80 / 8.0 / 1.5
1.00 / 8.0 / 1.5
0.60 / 11.0 / 1.4
0.60 / 10.0 / 1.4
0.60 / 9.0 / 1.4
0.60 / 12.0 / 1.4
0.40 / 8.0 / 2.0
0.40 / 10.0 / 1.7
0.80 / 11.0 / 1.
1.00 / 11.0 / 1.
0.80 / 10.0 / 1.
1.00 / 10.0 / 1.
0.80 / 9.0 / 1.
1.00 / 9.0 / 1.
0.80 / 12.0 / 1.
1.00 / 12.0 / 1.
0.40 / 8.0 / 1.8
0.60 / 11.0 / 1.3
0.60 / 10.0 / 1.3
0.60 / 9.0 / 1.3
Excess Emissions Identified
HC
318
318
318
318
318
317
317
316
316
324
321
319
319
319
319
317
317
327
327
327
325
324
323
320
320
320
320
320
320
318
318
325
327
327
327
CO
5212
5212
5212
5212
5212
5168
5168
5158
5158
5659
5323
5222
5222
5222
5222
5649
5649
5694
5694
5694
5640
5572
5377
5684
5684
5684
5684
5684
5684
5630
5630
5582
5706
5706
5706
NOx
241
241
241
237
237
252
252
241
241
274
269
252
252
252
252
274
274
306
306
306
306
246
269
306
306
306
306
306
306
306
306
257
306
306
306
Identification
HC
79.8%
79.8%
79.8%
79.8%
79.8%
79.6%
79.6%
79.4%
79.4%
81.2%
80.7%
80.1%
80.1%
80.1%
80.1%
79.5%
79.5%
82.1%
82.1%
82.1%
81.6%
81.4%
81.1%
80.4%
80.4%
80.4%
80.4%
80.4%
80.4%
79.9%
79.9%
81.6%
82.1%
82.1%
82.1%
CO
60.7%
60.7%
60.7%
60.7%
60.7%
60.2%
60.2%
60.1%
60.1%
65.9%
62.0%
60.8%
60.8%
60.8%
60.8%
65.8%
65.8%
66.3%
66.3%
66.3%
65.7%
64.9%
62.6%
66.2%
66.2%
66.2%
66.2%
66.2%
66.2%
65.5%
65.5%
65.0%
66.4%
66.4%
66.4%
Rates
NOx
70.3%
70.3%
70.3%
69.4%
69.4%
73.5%
73.5%
70.3%
70.3%
80.1%
78.6%
73.5%
73.5%
73.5%
73.5%
80.0%
80.0%
89.5%
89.5%
89.5%
89.5%
71.8%
78.6%
89.4%
89.4%
89.4%
89.4%
89.4%
89.4%
89.4%
89.4%
75.1%
89.5%
89.5%
89.5%
Fails
454
454
454
448
448
460
460
448
448
484
484
466
466
466
466
478
478
502
502
502
496
502
490
496
496
496
496
496
496
490
490
514
544
544
544
Errors of
Commission
48
48
48
48
48
48
48
48
48
42
48
48
48
48
48
42
42
84
84
84
84
48
48
84
84
84
84
84
84
84
84
48
84
84
84
EC Rate*
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.3%
2.0%
2.3%
2.3%
2.3%
2.3%
2.3%
2.0%
2.0%
4.0%
4.0%
4.0%
4.0%
2.3%
2.3%
4.0%
4.0%
4.0%
4.0%
4.0%
4.0%
4.0%
4.0%
2.3%
4.0%
4.0%
4.0%
Discrepant
Failures
6
6
6
6
6
6
6
6
6
48
6
6
6
6
6
48
48
48
48
48
48
6
6
48
48
48
48
48
48
48
48
6
90
90
90
Probable
EC Rate
2.6%
2.6%
2.6%
2.6%
2.6%
2.6%
2.6%
2.6%
2.6%
4.3%
2.6%
2.6%
2.6%
2.6%
2.6%
4.3%
4.3%
6.4%
6.4%
6.4%
6.4%
2.6%
2.6%
6.4%
6.4%
6.4%
6.4%
6.4%
6.4%
6.4%
6.4%
2.6%
8.4%
8.4%
8.4%
E-8
* see Section 5 for explanation
-------
Appendix E: ASM Outpoint Tables
ASM
Failure Rate
26%
26%
26%
26%
26%
26%
26%
26%
26%
26%
26%
26%
26%
27%
27%
27%
27%
27%
27%
27%
27%
28%
28%
28%
28%
28%
28%
28%
28%
28%
28%
28%
28%
28%
29%
Cutpolnts
0.60 / 12.0 / 1.3
0.40 / 11.0 / 1.5
0.40 / 10.0 / 1.5
0.40 / 9.0 / 1.5
0.40 / 12.0 / 1.5
0.80 / 11.0 / 1.3
1.00 / 11.0 / 1.3
0.80 / 10.0 / 1.3
1.00 / 10.0 / 1.3
0.80 / 9.0 / 1.3
1.00 / 9.0 / 1.3
0.80 / 12.0 / 1.3
1.00 / 12.0 / 1.3
0.30 / 12.0 / 1.8
0.30 / 11.0 / 1.8
0.30 / 10.0 / 1.8
0.30 / 9.0 / 1.8
0.30 / 12.0 / 2.0
0.30 / 11.0 / 2.0
0.30 / 10.0 / 2.0
0.30 / 9.0 / 2.0
0.60 / 8.0 / 1.4
0.80 / 8.0 / 1.4
1.00 / 8.0 / 1.4
0.60 / 12.0 / 1.2
0.40 / 8.0 / 1.5
0.80 / 11.0 / 1.2
1.00 / 11.0 / 1.2
0.80 / 10.0 / 1.2
1.00 / 10.0 / 1.2
0.80 / 9.0 / 1.2
1.00 / 9.0 / 1.2
0.80 / 12.0 / 1.2
1.00 / 12.0 / 1.2
0.30 / 8.0/2.0
Excess Emissions Identified
HC
325
324
324
324
323
320
320
320
320
320
320
318
318
345
345
345
345
344
344
344
344
338
331
331
329
329
324
324
324
324
324
324
323
323
351
CO
5652
5472
5472
5472
5418
5696
5696
5696
5696
5696
5696
5642
5642
5594
5594
5594
5594
5584
5584
5584
5584
6071
6061
6061
5753
5755
5797
5797
5797
5797
5797
5797
5743
5743
5943
NOx
306
274
274
274
274
306
306
306
306
306
306
306
306
262
262
262
262
251
251
251
251
306
306
306
308
274
308
308
308
308
308
308
308
308
256
Identification Rates
HC
81.6%
81.4%
81.4%
81.4%
80.9%
80.4%
80.4%
80.4%
80.4%
80.4%
80.4%
79.9%
79.9%
86.6%
86.6%
86.6%
86.6%
86.4%
86.4%
86.4%
86.4%
84.8%
83.1%
83.1%
82.7%
82.5%
81.4%
81.4%
81.4%
81.4%
81.4%
81.4%
80.9%
80.9%
88.0%
CO
65.8%
63.7%
63.7%
63.7%
63.1%
66.3%
66.3%
66.3%
66.3%
66.3%
66.3%
65.7%
65.7%
65.1%
65.1%
65.1%
65.1%
65.0%
65.0%
65.0%
65.0%
70.7%
70.6%
70.6%
67.0%
67.0%
67.5%
67.5%
67.5%
67.5%
67.5%
67.5%
66.9%
66.9%
69.2%
NOx
89.5%
80.1%
80.1%
80.1%
80.1%
89.4%
89.4%
89.4%
89.4%
89.4%
89.4%
89.4%
89.4%
76.7%
76.7%
76.7%
76.7%
73.4%
73.4%
73.4%
73.4%
89.5%
89.4%
89.4%
89.9%
80.1%
89.8%
89.8%
89.8%
89.8%
89.8%
89.8%
89.8%
89.8%
74.9%
Fails
538
544
544
544
538
538
538
538
538
538
538
532
532
568
568
568
568
556
556
556
556
586
580
580
586
586
586
586
586
586
586
586
580
580
604
Errors of
Commission
84
48
48
48
48
84
84
84
84
84
84
84
84
90
90
90
90
90
90
90
90
84
84
84
84
48
84
84
84
84
84
84
84
84
90
EC Rate*
4.0%
2.3%
2.3%
2.3%
2.3%
4.0%
4.0%
4.0%
4.0%
4.0%
4.0%
4.0%
4.0%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.0%
4.0%
4.0%
4.0%
2.3%
4.0%
4.0%
4.0%
4.0%
4.0%
4.0%
4.0%
4.0%
4.3%
Discrepant
Failures
90
48
48
48
48
90
90
90
90
90
90
90
90
0
0
0
0
0
0
0
0
48
48
48
132
48
132
132
132
132
132
132
132
132
0
Probable
EC Rate
8.4%
4.6%
4.6%
4.6%
4.6%
8.4%
8.4%
8.4%
8.4%
8.4%
8.4%
8.4%
8.4%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
6.4%
6.4%
6.4%
10.4%
4.6%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
4.3%
E-9
see Section 5 for explanation
-------
Appendix E: ASM Outpoint Tables
ASM
Failure Rate
29%
29%
29%
30%
30%
30%
30%
31%
31%
31%
31%
31%
31%
31%
31%
32%
32%
33%
33%
33%
33%
33%
33%
33%
35%
35%
36%
36%
36%
36%
36%
36%
36%
38%
38%
Cutpolnts
0.60 / 11.0 / 1.2
0.60 / 10.0 / 1.2
0.60 / 9.0 / 1.2
0.30 / 8.0 / 1.8
0.60 / 8.0 / 1.3
0.80 / 8.0 / 1.3
1.00 / 8.0 / 1.3
0.30 / 12.0 / 1.5
0.30 / 11.0 / 1.5
0.30 / 10.0 / 1.5
0.30 / 9.0 / 1.5
0.40 / 11.0 / 1.4
0.40 / 10.0 / 1.4
0.40 / 9.0 / 1.4
0.40 / 12.0 / 1.4
0.80 / 8.0 / 1.2
1.00 / 8.0 / 1.2
0.30 / 8.0 / 1.5
0.40 / 8.0 / 1.4
0.60 / 8.0 / 1.2
0.40 / 11.0 / 1.3
0.40 / 10.0 / 1.3
0.40 / 9.0 / 1.3
0.40 / 12.0 / 1.3
0.40 / 8.0 / 1.3
0.40 / 12.0 / 1.2
0.30 / 12.0 / 1.4
0.30 / 11.0 / 1.4
0.30 / 10.0 / 1.4
0.30 / 9.0 / 1.4
0.40 / 11.0 / 1.2
0.40 / 10.0 / 1.2
0.40 / 9.0 / 1.2
0.30 / 8.0 / 1.4
0.30 / 12.0 / 1.3
Excess Emissions Identified
HC
331
331
331
351
338
331
331
351
351
351
351
339
339
339
337
335
335
355
343
342
339
339
339
337
343
341
356
356
356
356
343
343
343
360
356
CO
5807
5807
5807
5953
6083
6073
6073
5813
5813
5813
5813
5885
5885
5885
5830
6174
6174
6096
6167
6184
5897
5897
5897
5843
6180
5943
6167
6167
6167
6167
5997
5997
5997
6450
6179
NOx
308
308
308
268
306
306
306
283
283
283
283
306
306
306
306
308
308
283
306
308
306
306
306
306
306
308
306
306
306
306
308
308
308
306
306
Identification
HC CO
83.1% 67.6%
83.1% 67.6%
83.1% 67.6%
88.2% 69.3%
84.8% 70.8%
83.1% 70.7%
83.1% 70.7%
88.0% 67.7%
88.0% 67.7%
88.0% 67.7%
88.0% 67.7%
85.0% 68.5%
85.0% 68.5%
85.0% 68.5%
84.5% 67.9%
84.1% 71.9%
84.1% 71.9%
89.0% 71.0%
86.1% 71.8%
85.8% 72.0%
85.0% 68.7%
85.0% 68.7%
85.0% 68.7%
84.5% 68.0%
86.1% 71.9%
85.6% 69.2%
89.3% 71.8%
89.3% 71.8%
89.3% 71.8%
89.3% 71.8%
86.1% 69.8%
86.1% 69.8%
86.1% 69.8%
90.3% 75.1%
89.3% 71.9%
Rates
NOx
89.9%
89.9%
89.9%
78.2%
89.5%
89.4%
89.4%
82.7%
82.7%
82.7%
82.7%
89.5%
89.5%
89.5%
89.5%
89.8%
89.8%
82.7%
89.5%
89.9%
89.5%
89.5%
89.5%
89.5%
89.5%
89.9%
89.5%
89.5%
89.5%
89.5%
89.9%
89.9%
89.9%
89.5%
89.5%
Falls
592
592
592
616
628
622
622
640
640
640
640
646
646
646
640
670
670
682
688
676
688
688
688
682
729
729
735
735
735
735
735
735
735
777
777
Errors of
Commission
84
84
84
90
84
84
84
90
90
90
90
90
90
90
90
84
84
90
90
84
90
90
90
90
90
90
132
132
132
132
90
90
90
132
132
EC Rate*
4.0%
4.0%
4.0%
4.3%
4.0%
4.0%
4.0%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
4.0%
4.0%
4.3%
4.3%
4.0%
4.3%
4.3%
4.3%
4.3%
4.3%
4.3%
6.4%
6.4%
6.4%
6.4%
4.3%
4.3%
4.3%
6.4%
6.4%
Discrepant
Failures
132
132
132
0
90
90
90
42
42
42
42
48
48
48
48
132
132
42
48
132
90
90
90
90
90
132
42
42
42
42
132
132
132
42
84
Probable
EC Rate
10.4%
10.4%
10.4%
4.3%
8.4%
8.4%
8.4%
6.4%
6.4%
6.4%
6.4%
6.6%
6.6%
6.6%
6.6%
10.4%
10.4%
6.4%
6.6%
10.4%
8.7%
8.7%
8.7%
8.7%
8.7%
10.7%
8.4%
8.4%
8.4%
8.4%
10.7%
10.7%
10.7%
8.4%
10.4%
E-10
* see Section 5 for explanation
-------
Appendix E: ASM Outpoint: Tables
ASM
Failure Rate
38%
38%
38%
38%
40%
40%
40%
40%
40%
40%
40%
40%
40%
40%
40%
40%
40%
40%
40%
40%
40%
42%
42%
42%
42%
43%
43%
43%
43%
45%
46%
46%
46%
46%
48%
Outpoints
0.30 / 11.0 / 1.3
0.30 / 10.0 / 1.3
0.30 / 9.0 / 1.3
0.40 / 8.0 / 1.2
0.60 / 11.0 / 1.0
0.80 / 11.0 / 1.0.
1.00 / 11.0 / 1.0
0.60 / 10.0 / 1.0
0.80 / 10.0 / 1.0
1.00 / 10.0 / 1.0
0.60 / 9.0 / 1.0
0.80 / 9.0 / 1.0
1.00 / 9.0 / 1.0
0.60 / 12.0 / 1.0
0.80 / 12.0 / 1.0
1.00 / 12.0 / 1.0
0.30 / 12.0 / 1.2
0.30 / 11.0 / 1.2
0.30 / 10.0 / 1.2
0.30/ 9.0/1.2
0.30 / 8.0 / 1.3
0.60 / 8.0 / 1.0
0.80 / 8.0 / 1.0
1.00 / 8.0 / 1.0
0.30 / 8.0 / 1.2
0.40 / 11.0 / 1.0
0.40 / 10.0 / 1.0
0.40 / 9.0 / 1.0
0.40 / 12.0 / 1.0
0.40 / 8.0 / 1.0
0.30 / 12.0 / 1.0
0.30 / 11.0 / 1.0
0.30 / 10.0 / 1.0
0.30 / 9.0 / 1.0
0.30 / 8.0 / 1.0
Excess Emissions Identified
HC
356
356
356
347
364
364
364
364
364
364
364
364
364
362
362
362
360
360
360
360
360
368
368
368
364
364
364
364
362
368
381
381
381
381
385
CO
6179
6179
6179
6280
6429
6429
6429
6429
6429
6429
6429
6429
6429
6375
6375
6375
6280
6280
6280
6280
6462
6712
6712
6712
6562
6485
6485
6485
6431
6768
6767
6767
6767
6767
7050
NOx
306
306
306
308
325
325
325
325
325
325
325
325
325
325
325
325
308
308
308
308
306
325
325
325
308
325
325
325
325
325
325
325
325
325
325
Identification
HC CO
89.3% 71.9%
89.3% 71.9%
89.3% 71.9%
87.1% 73.1%
91.3% 74.9%
91.3% 74.9%
91.3% 74.9%
91.3% 74.9%
91.3% 74.9%
91.3% 74.9%
91.3% 74.9%
91.3% 74.9%
91.3% 74.9%
90.9% 74.2%
90.9% 74.2%
90.9% 74.2%
90.3% 73.1%
90.3% 73.1%
90.3% 73.1%
90.3% 73.1%
90.3% 75.2%
92.4% 78.1%
92.4% 78.1%
92.4% 78.1%
91.4% 76.4%
91.3% 75.5%
91.3% 75.5%
91.3% 75.5%
90.9% 74.9%
92.4% 78.8%
95.6% 78.8%
95.6% 78.8%
95.6% 78.8%
95.6% 78.8%
96.6% 82.1%
Rates
NOx
89.5%
89.5%
89.5%
89.9%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
89.9%
89.9%
89.9%
89.9%
89.5%
95.0%
95.0%
95.0%
89.9%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
95.0%
Fails
777
777
777
777
837
837
837
837
837
837
837
837
837
831
831
831
825
825
825
825
819
879
879
879
867
897
897
897
891
939
945
945
945
945
987
Errors of
Commission
132
132
132
90
174
174
174
174
174
174
174
174
174
174
174
174
132
132
132
132
132
174
174
174
132
180
180
180
180
180
180
180
180
180
180
EC Rate*
6.4%
6.4%
6.4%
4.3%
8.4%
8.4%
8.4%
8.4%
8.4%
. 8.4%
8.4%
8.4%
8.4%
8.4%
8.4%
8.4%
6.4%
6.4%
6.4%
6.4%
6.4%
8.4%
8.4%
8.4%
6.4%
8.7%
8.7%
8.7%
8.7%
8.7%
8.7%
8.7%
8.7%
8.7%
8.7%
Discrepant
Failures
84
84
84
132
221
221
221
221
221
221
221
221
221
221
221
221
126
126
126
126
84
180
180
180
126
138
138
138
138
138
132
132
132
132
132
Probable
EC Rate
10.4%
10.4%
10.4%
10.7%
19.1%
19.1%
19.1%
19.1%
19.1%
19.1%
19.1%
19.1%
19.1%
19.1%
19.1%
19.1%
12.4%
12.4%
12.4%
12.4%
10.4%
17.1%
17.1%
17.1%
12.4%
15.3%
15.3%
15.3%
15.3%
15.3%
15.0%
15.0%
15.0%
15.0%
15.0%
E-ll
* see Section 5 for explanation
-------
Appendix F
Scatter Plots and Regression Tables
-------
Regression Tables
All Vehicles
Dependant Variable is: HC FTP
RX2 - 81.9%
s - 0.6266 with 106 - 2 - 104 DOF
Standard Error: 0.62 g/mi
Source Sum of Square*
Regression 184.815
Residual 40.839
Variable
Constant
HC IM240
Coefficient
-0.118
1.318
*.«. of Coeff
0.078
0.061
Dependant Variable is: HC FTP
R*2 - 73.4%
3 - 0.7602 with 106 - 2 - 104 DOF
Standard Irror: 0.76 g/ai
Source STUB of Square*
Regression 165.550
Residual 60.103
Variable
Constant
HC ASM
Coefficient
-0.056
2.264
«.e. of Coeff
0.094
0.134
Dependent Variable is: CO FTP
RA2 = 54.2%
s - 13.47 with 106 - 2 - 104 DOF
Standard Error: 13.4 g/mi
Source Sum of Square*
Regression 22318.900
Residual 18857.200
Variable Coefficient
Constant 2.625
CO IM240 0.929
*.e. of Coeff
1.609
0.084
Dependent Variable is: CO FTP
RA2 - 67.9%
s - 11.27 with 106 -2-104 DOF
Standard Error: 11.2 g/mi
Source Sum of Squares
Regression 27959.200
Residual 13216.900
Variable
Constant
CO ASM
Coefficient
3.358
0.970
*.e. of Coeff
1.274
0.065
Dependent Variable is: NOx FTP
RA2 - 69.7%
3 - 0.6570 with 106 - 2 - 104 DOF
Standard Error: 0.6S g/mi
Source Sum of Square*
Regression 103.202
Residual 44.889
Variable Coefficient
Constant -0.046
NOx IM240 0.724
. of Coeff
0.104
0.047
Dependent Variable is: NOx FTP
RA2 - 71.4%
a - 0.6386 with 106 - 2 - 104 DOF
Standard Error: 0.64 g/mi
Source Sum of Square*
Regression 105.685
Residual 42.406
Variable
Constant
NOx ASM
Coefficient
-0.002
0.831
. of Coeff
0.098
0.052
-------
Figure F-l
HC Scatterplots
All Vehicles
HC IM240 vs FTP
2.00 4.00 6.00
HC XM240 (g/mi)
H
8.00
HC ASM VS FTP
.00
1.00 2.00 3.00
ASM Predicted HC
4.00
-------
CO Scatterplots
All Vehicles
120 T
CO XM240 va FTP
20 40 60
CO XM240 (g/mi)
80
100
120 T
CO ASM vs FTP
20 40 60 80
ASM Predicted CO
100
-------
Figure F-3
NOx Scatterplots
All Vehicles
NOx IM240 vs FTP
HOx IM240 (g/mi)
NOx ASM vs FTP
468
ASM Predicted NOx
10
-------
Regression Tables
Vehicle 3211 Removed
Dependent Variable is: HC FTP
RA2 =• 82.6%
3 - 0.6169 with 105 - 2 - 103 DOF
Standard Error: 0.61 g/mi
Source Sum of Square*
Regression 186.255
Residual 39.194
Variable
Constant
HC IM240
Coefficient
-0.112
1.326
. of Coeff
0.077
0.060
Dependent Variable is: HC FTP
RA2 • 73.8%
s - 0.7578 with 105 - 2 - 105 DOF
Standard Error: 0.75 g/mi
Source Sum of Square*
Regression 166.299
Residual 59.149
Variable
Constant
HC ASM
Coefficient
-0.050
2.271
*.e. of Coeff
0.094
0.133
Dependent Variable is: CO FTP
R*2 - 74.6%
3 - 10.08 with 105 - 2 - 103 DOF
Standard Error: 10.0 g/mi
Source
Regression
Residual
Variable
Constant
CO IM240
Sum of Squares
30708.400
10462.600
Coefficient
-0.164
1.269
a.e. of Coeff
1.243
0.073
Dependent Variable is: CO FTP
RA2 - 80.2%
s - 8.892 with 105 - 2 - 103 DOF
Standard Error: 8.9 g/mi
Source STUB of Square*
Regression 33027.000
Residual 8144.000
Variable
Constant
CO ASM
Coefficient
2.394
1.140
s.e. of Coeff
1.012
0.056
Dependent Variable is: NOx FTP
RA2 - 69.6%
3 - 0.6598 with 105 - 2 - 103 DOF
Standard Error: 0.66 g/mi
Source Sum of Square*
Regression 102.819
Residual 44.834
Variable
Constant
NOx IM240
Coefficient
-0.051
0.725
•.e. of Coeff
0.106
0.047
Dependent Variable is: NOx FTP
RA2 - 71.4%
s - 0.6386 with 105 - 2 - 103 DOF
Standard Error: 0.64 g/mi
Source Sum of Square*
Regression 105.685
Residual 42.406
Variable
Constant
NOx ASM
Coefficient
-0.002
0.831
«.e. of Coeff
0.098'
0.052
-------
Figure F-4
HC Scatterplots
Vehicle 3211 Removed
HC IM240 va FTP
2.00 4.00 6.00
HC IM240 (g/mi)
8.00
HC ASM vs FTP
,00
1.00 2.00 3.00
ASM Predicted HC
4.00
-------
Figure F-5
CO Scatterplots
Vehicle 3211 Removed
CO IM240 vs FTP
120
20 40 60
CO IM240 (g/mi)
80
120 T
CO ASM vs FTP
20 40 60 80
ASM Predicted CO
100
-------
F— 6
NOx Scatterplots
Vehicle 3211 Removed
NOx IM240 vs FTP
NOx IM240
-------
Tattlo Tf-
Regression Tables
Vehicles Near Standards
Dependent Variable is: HC FTP
RA2 = 63.0%
3 - 0.1953 with 43-2-41 DOF
Standard Error: 0.19 g/mi
Source Soa of Squares
Regression 2.662
Residual 1.563
Variable
Constant
HC IM240
Coefficient
0.271
0.531
. of Coeff
0.048
0.064
Dependent Variable is: HC FTP
RA2 - 17.6%
3 - 0.2915 with 43-2-41 DOF
Standard Zrror: 0.28 g/mi.
Source Sum of Squares
Regression 0.742
Residual 3.483
Variable
Constant
HC ASM
Coefficient
0.342 '
0.818
. of Coeff
0.092
0.277
Dependent Variable is: CO FTP
RA2 - 24.8%
s - 4.360 with 43-2-41 DOF
Standard Zrror: 4.3 g/mi
Source Sum of Square*
Regression 257.268
Residual 779.566
Variable
Constant
CO IM240
Coefficient
4.308
0.490
i.e. of Coeff
1.243
0.133
Dependent Variable is: CO FTP
RA2 - 13.3%
3 - 4.683 with 43-2-41 DOF
Standard Zrror: 4.6 g/mi
Source Sum of Squares
Regression 137.732
Residual 899.102
Variable
Constant
CO ASM
Coefficient
4.622
0.684
. of Coeff
1.586
0.273
Dependent Variable is: NOx FTP
RA2 a 45.6%
s - 0.3349 with 43-2-41 DOF
Standard Error: 0.34 g/mi
Source Sum of Squares
Regression 3.848
Rasldual 4.599
Variable Coefficient
Constant 0.670
NOX IM240 0.276
s.e. of Coeff
0.103
0.047
Dependent Variable is: NOx FTP
RA2 - 26.0%
s - 0.3904 with 43-2-41 DOF
Standard Zrror: 0.39 g/mi
Source Sum of Squares
Regression 2.200
Residual 6.248
Variable Coefficient
Constant 0.792
NOx ASM 0.266
. of Coeff
0.121
0.070
-------
Figure F-7
HC Scatterplots
Vehicles Near Standards
HC IM240 vs FTP
0.00
0.00 0.50 1.00 1.50
HC IM240 (g/mi)
2.00
2.50
1.40
1.20
^ 1.00
I
t 0.80
u
0.60
0.40
0.20
0.00
HC ASM vs FTP
-t-
0.00 0.20 0.40 0.60
ASM Predicted HC
—I
0.80
-------
Figure P-fl
CO Scatterplots
Vehicles Near Standards
CO IM240 vs FTP
25 T
10 15
CO IM240 (g/mi)
20
25
25 T
20
u
8
10
5 -
CO ASM vs FTP
10
ASM Predicted CO
15
20
-------
Figure F-9
NOx Scatterplots
Vehicles Near Standards
NOx IM240 vs FTP
^
i
2.5 T
2 -
1.5
1 -
0.5 -
H 1 1 1-
1234
NOz XM240 (g/mi)
2.5 T
2
Oi
H
0.5 -
NOx ASM vs FTP
B-. '
2 3
ASM Predicted NOz
-------
Appendix G:
ARCO, Sierra, Environment Canada Data Analysis
-------
1.0
The objective of this report is to respond to the pilot ASM test
programs performed by Sierra Research, Inc., ARCO Products Company, and
Environment Canada. Sierra and ARCO both previously published papers
praising the capabilities of the ASM, and both concluded that some form of the
ASM could replace the IM240 as an enhanced I/M test.
EPA has concluded that the ARCO and Sierra reports are incorrect in
claiming the ASM as equal to the IM240. Based on a comparison with a similar
database of IM240 vehicles, the ASM is inferior to the IM240 at identifying
excess emissions without committing false failures. Moreover, a series of
regressions were run for both the ASM and the IM240 versus the FTP. The
scatterplots for these regressions, contained in the Appendix to this report,
show significant variability for the ASM at predicting FTP values, compared to
the IM240.
A contractor for EPA is currently testing a number of vehicles at a
state I/M lane in Mesa, Arizona on both the IM240, and a 4 mode steady state
test, which includes two ASMs, the ASM2525 and the ASM5015. A sample of
vehicles is being recruited to the contractor's lab for further FTP testing.
The data from that program will give EPA a chance to determine, with greater
confidence, if some form of the ASM is as effective as the IM240.
This report focuses on a small dataset of vehicles, therefore the
conclusions made in this report are subject to change when more data is
available to EPA. However, from the data that has been presented to EPA to
date on the ASMs, the IM240 remains the only enhanced I/M test.
2 . 0 Database Deaeriptien
There are 31 vehicles in the ASM database EPA used for this analysis.
The data were gathered from programs performed by three different
organizations: Environment Canada^, Sierra Research^, and ARCO Products^.
Ballantyne, Vera F. Draft. Steady State Teatincy Report: and Data.
Environment Canada, August 28, 1992.
Austin, Thomas C., Sherwood, Larry, Development of Tmproved Loaded—Mode
Test Procedures fov Inspection and Maintenance Programs. Sierra Research,
Inc. and California Bureau of Automotive Repair, SAE Paper No. 891120,
Government/Industry Meeting and Exposition, May 2-4, 1989.
G-2
-------
EPA started performing ASM teats in Mesa Arizona on September 10, 1992. These
data will be the topic of a separate analysis.
A number of vehicles in the ASM database were tested with and without
implanted defects, so 51 test configurations were used for this analysis. All
the vehicles tested by the three different organizations received the ASM5015
and the FTP, but ARCO did not perform the ASM2525. This left 39 test
configurations receiving multiple-mode ASM tests and FTPs.
2.1 ASM Vehicles Removed from Database
There were originally 55 vehicles tested in the three programs, resulting
in 117 test configurations, broken down as follows: Environment Canada (32
vehicles or 36 configurations); Sierra Research (18 vehicles or 51
configurations), and ARCO Products (5 vehicles or 30 configurations).
Vehicles were removed from the database for reasons which are discussed below.
First, all pre-1983 vehicles were removed to focus on newer
technology vehicles. So 3 Canadian vehicles and 5 Sierra vehicles were
removed, leaving 29 Canadian vehicles with 33 configurations and 13 Sierra
vehicles also with 33 configurations.
Next all pre-1988 Canadian vehicles were removed. Canadian vehicle
standards were not lowered to 0.41/3.4/1.0 until the 1988 model year, so the
prior model years could not be used. So 13 Canadian vehicles were removed,
leaving 16 Canadian vehicles with 20 configurations.
Next, all ARCO vehicles that were not certified to the 50-state standards
of 0.41/3.4/1.0 were removed. Three ARCO vehicles were certified to
California-only standards, so they were removed, leaving 2 ARCO vehicles with
12 configurations.
Finally, all Sierra configurations that received hot-start FTPs instead
of cold-start FTPs were removed. Because the normal cold-start FTP is more
variable than hot-start FTPs, short test comparisons should be made using
cold-start FTPs. Also, vehicles are certified using cold-start FTPs, so the
results are more relevant. So 14 Sierra configurations were removed, leaving
13 Sierra vehicles with 19 configurations.
Boekhaus Kenneth L., et al. Evaluation of Enhanced Inspection Techniques on
State-of-the-Art Automobiles. ARCO Products Company Report, May 8,1992.
G-3
-------
2 .2
Selection of IM240 Vehicles Used in Database
In order to compare the ASM to the IM240, the analysis should be performed
on a set of vehicles that have received both tests. However, none of the ASM
vehicles received the IM240, therefore 39 vehicles were randomly selected
from the Indiana laboratory IM240 database. These vehicles were chosen from
those used in the IM240 cutpoint table analysis in EPA's I/M Coata. Benefits.
and Impacts Analysisf which included 274 vehicles with both IM240 and FTP
results. In order to make the IM240 database similar to the ASM database, the
following process was used.
First, the ASM vehicles were categorized by emission levels according to
the following table:
Table 1
Number of Vehicles in Database per Emittant Category.
HC/CO
Category
Normal
Normal
High
Very High
Very High
NOz
Category
Normal
High
Normal
Normal
High
HC Range*
OSHC<0.82
OSHC<0.82
0.82£HC<1.64
1.64£HC<10.0
1.64SHC<10.0
CO Range*
OSCCK10.2
OSCCK10.2
10.2SCCK13.6
13.6SCCX150
13.6SCCK150
NOx Range
0£NOx<2
2£NOx<4
OSNOx<2
OSNOx<2
2£NOx<4
# in
Dataset
29
1
2
4
3
* These are the same categories as those used in the I/M Technical Support
Document
Second, the Lab IM240 database was broken down into these same categories.
All vehicles were 1983+ model years, and only vehicles that received the lab
IM240 after the FTP were kept in the database. This kept the IM240 database
as similar as possible to the ASM database. From the remaining vehicles, a
random sample was chosen from each category so that both databases had the
same number of vehicles in each category.
By selecting the same number of vehicles from each emittant range, it
prevents one test from getting an unfair advantage in achieving identification
rates. For example, if the IM240 database included considerably higher FTP
scores, it would have identified much more excess emissions, thus making its
Identification Rates (IDRs) higher.
G-4
-------
3 . 0 Calculating ASM Ma**
Sierra indicated (SAE Paper No. 891120) that calculated ASM mass emissions
correlate better to the FTP than concentration measurements, so their method
of converting ASM NOx concentration measurements to "mass" emissions was
applied to this ASM database for HC and CO, as well as NOx. This was done by
multiplying the emission concentrations (ppm for HC and NOx, and % for CO) by
the vehicles' Inertia Weights (IW), yielding the following units: kiloton-ppm
for HC (IW * ppm/103), ton-% for CO (IW * %), and megaton-ppm for NOx (IW *
ppm/10^). These are the values EPA used for the regressions in this report.
3 .1 EPA Equations Versus Sierra Equations
In their test program, Sierra measured the ASM emissions on both a
concentration basis and mass basis. This allowed them to regress
Concentration * Inertia Weight (IW) versus mass emissions for the same test,
and develop equations that convert [Concentration * IW] to Mass. As expected,
these mass calculations correlated very well with the measured mass emissions.
Sierra's next step was to regress the measured steady state mass emissions
against the FTP emissions and report r^s for these regressions. They did not
actually use the calculated mass emissions to predict FTP scores. This is
where EPA's analysis of the ASMs was slightly different. EPA regressed the
[Concentration * IW] values against the FTP emissions for each vehicle. This
was done because EPA did not have measured mass emissions from all three test
programs compiled in this report. However, the major benefit of the ASMs,
according to Sierra and ARCO, is the ability to use the less expensive BAR90
type analyzers when measuring the exhaust concentrations. Since this is a
claimed benefit of the ASMs, the readings from these less expensive analyzers
should be used when comparing the ASM to the IM240.
4 . 0 Mu3.fci.ple T.int»ar Rt»qrreaalona tor fchg ASM
Using data from all three previously mentioned programs, EPA calculated
the IW * Concentration for each emittant. Then a multiple linear regression
was performed, using the calculated (IW*Concentration) ASM2525 and ASM5015
scores as two separate variables vs measured FTP emissions. Equations were
developed from these regressions that predict an FTP score from a combination
of the ASM2525 and ASM5015 concentrations * IW scores:
G-5
-------
Table 2.
Eonafciona Developed to Predict FTP from ASM Modes.
Predicted FTP = [IW (A*ASM2525 + B*ASM5015) + C]
Emittant
HC (ppm)
CO (%)
NOx (ppm)
A
-3.96xlO"7
2.64xlO-3
1.13xlO~7
B
4.60xlO~7
5.10xlO~5
1.30xlO~7
C
0.523520
4.222840
0.515531
r2
49.2%
43.5%
71.4%
4 .1
Simple Linear Regressions
Aside from those already mentioned, regressions were also run for each
individual ASM mode vs the FTP, and for the IM240 vs the FTP. Since the IM240
is a transient test, like the FTP, it correlates much better to the FTP than
the ASM modes.
4.1.1
Coefficient of Determination (r2)
The r2 may be interpreted as the proportion of the total FTP variability
that was predicted by the short test. For example, if the r2 equalled 100%,
the short test would have perfectly predicted the FTP scores for these cars.
If the r2 for these vehicles was zero, the short test would not have any
linear relationship to the FTP.
The r2 data, listed in table below, show that the IM240 is considerably
better than the ASM tests in predicting FTP HC, CO, and NOx scores. For HC
and CO, less than half of the FTP variation is explained by the ASM scores.
Table 3.
Statistical Comparison of the FTP Versus I/M Teata
r2
HC
IM240
95%
5015
36%
2525
20%
CO
ZM240
92%
5015
45%
2525
44%
NOx
IM240
84%
5015
62%
2525
70%
4.2
Scatterplots
For an I/M test, more important than the r2 is the ability to identify
high proportions of dirty cars without falsely failing vehicles. The IM240
also has a significant advantage at identifying more of the dirty cars while
failing less of the clean cars. The scatterplots in the appendix show this
clearly. When viewing the plots, consider the following chart for a
reference.
G-6
-------
Scattsrolot Used^_ to Define Terms
5
4 -•
3-.
o. 2 +
H
bi
Excess
Emissions
Omitted
Excess
Emissions
Identified
Clean Cars
both tests)
1
Errors of Commission
Short Test Score
The short test outpoint under consideration is the vertical line, and
the FTP standard is the horizontal line. The intersection of these lines
splits the chart into quadrants. The goal of the short test is to maximize
the number of FTP failing vehicles into the upper right quadrant, while
minimizing the false failures in the lower right quadrant.
The more vehicles that appear in the upper left quadrant, the less
effective the test becomes, because these are all dirty cars that are not
identified by the short test. From this perspective, the advantages of the
IM240 is clear. Every IM240 chart shows that an x-axis value (cutpoint) can
be selected that clearly places the vast majority of dirty cars in the upper-
right quadrant, without errors of commission. The ASM tests do not display
this trait nearly as well. Only the 2-mode ASM tests and the IM240
scatterplots have the horizontal and vertical lines on them, so the reader can
examine different cutpoint scenarios.
G-7
-------
4.2.1 Scatterplot Statistics
Each of the regression scatterplots contains the following information:
• The best-fit regression line showing predicted FTP for a continuum of
short test scores, developed from a regression of the actual data.
• 'Boundary Lines' at + 2 and - 2 standard error from the predicted
value.
• A horizontal dotted line at the FTP standard.
• A box containing descriptive statistics.
On each Regression plot, a box in the upper-left corner provides the
following statistics: 1) The equation of the line used to predict FTP values
from the short test's score. 2) r2, discussed above. 3) The standard error*
, a statistic that describes the variability of the FTP score predicted from
the selected short test. The next section discusses standard error in more
detail.
4.3 Standard Error as a Measure of Variability
The weakness of the ASM tests regarding r2 and the low proportion of
cars that can be identified as dirty while simultaneously avoiding false
failures, is related to test variability. The standard error is an objective
measurement of test variability. The following shows that the ASM tests are
significantly more variable than the IM240, using the standard error as an
objective measure of variability.
4.3.1 Assumptions Made for Using Standard Error
The following assumptions were made in order to use standard error as it
is used in this report:
• Linear relationship between the FTP and the short tests.
• Normally distributed data.
• Homoscedastic distribution (i.e., the standard deviation of FTP values
is constant for all short test values).
What is referred to in this report is formally termed standard error of
estimate, but for convenience purposes, will simply be called standard
error.
G-8
-------
The standard error is similar to standard deviation because a bandwidth
of ±1 std. error includes =-68% of the data and ±2 std. error includes =95% of
the data.
4.3.2 Example Using Standard Error
Consider a 3000 Ib. vehicle that emits 1500 ppm NOx on both the ASM5015
and ASM2525. Plugging these numbers into the equation for predicting FTP
values (Table 2) yields 1.61 g/mi. However, because the standard error for
ASM NOx (see Table 4) is 0.36 g/mile, roughly 5% of the FTP scores predicted
by the ASM result will be greater than 2.33 g/mile (1.61 + 2*0.36) or less
than 0.89 g/mile (1.61 - 2*0.36). Since half of these will err on the low
side, it is probable that =2.5% of the vehicles identified as failures by an
ASM cutpoint of 1.61 g/mi would be false failures.
4.3.3 Effect of Standard Error on "Safe FTP Predictions"
In order to be confident the false failure rate would be less than 2.5% the
selected cutpoint should predict an FTP value of 2 standard errors greater than
the FTP standard. This ensures that the low values (FTP prediction - 2 std.
error) are still failing the FTP.
For example,
FTP NOx standard is 1.0 g/mi
The ASM NOx std. error is 0.36 g/mi
FTPgtandard + 2 std- errors - 1.72 g/mi
So, the selected ASM cutpoint should predict an FTP of no less than 1.72
g/mi. Applying the same logic to the IM240, whose standard error is 0.28 g/mi, a
predicted FTP score of 1.56 g/mi (1.0 + 2*0.28) will also yield an error of
commission rate less than 2.5%. But because the "safe" predicted FTP score is
more stringent, the excess emissions identified will be higher. The standard
errors and predicted FTP levels that are expected to limit false failures to
approximately 2.5% are compared in Table 4 below.
G-9
-------
Table 4. Comparison of ASM and IM240 Standard errors And Thgjr Effect on
Predicted FTP Stringency at a 2.5% Falae Failure
1 std. error (g/mi)
Predicted FTP Level
@ =2.5% EC (g/ mi)
HC
ZM240
0.24
0.89
ASM
0.60
1.61
CO
IM240
3.8
11.0
ASM
4.8
13.0
NOx
ZM240
0.28
1.56
ASM
0.36
1.72
5.0 CutpQi.nt Tables
Another way to assess the effectiveness of I/M tests is to evaluate the
following factors, which were discussed in detail in Section 4.2.1 of EPA's
T/M Costa. Benefit.3. and Impaetia Analysis; excess emission identification
rates, failure rates, error-of-commission rate, the failure rate among
vehicles that pass FTP standards, and the failure rate for so-called "normal
emitters," which may fail an FTP standard (normal emitters are defined as
vehicles whose FTP HC < 0.82 g/mi and FTP CO < 10.2 g/mi), but are clean
enough to make the cost effectiveness of repairs an issue. These factors are
highly interactive, for example, high IDRs can be achieved with stringent
cutpoints, but this will adversely affect failure rates.
Cutpoint tables for the ASM tests and the IM240 in the appendix allow
these factors to be compared. The cutpoints for the tables were chosen using
an iterative process. The goal was to select cutpoints that would give
reasonable identification rates while limiting errors of commission. The goal
was to keep the EC rate at 0% for both procedures.
For both cutpoint tables, four different cutpoints were selected for each
of the three emittants, resulting in 64 different cutpoint combinations. For
the IM240, the "Two Ways to Pass Criteria" was used, as described in Section
4.2.3.2 of EPA's I/M Coafca. Benefits, and Impaeta Analysis. This is a method
of combining the composite HC and CO scores with the bag 2 HC and CO scores in
order to minimize Errors of Commission on vehicles with cold start problems,
while maintaining high Identification Rates.
5 .1 Selecting ASM Cutpointa
Scatterplots were done plotting Measured FTP vs. Calculated FTP from the
ASM scores. From these scatterplots, EPA determined a range of cutpoints to
use for the cutpoint tables. For example, looking at Chart x, FTP CO vs ASM
G-10
-------
Prediction CO, it can be determined that an ASM Prediction between 6 and 10
grams/mile would identify most dirty vehicles (those above the standard of 3.4
g/mi) without failing the clean vehicles. It is also obvious that a cutpoint
of 5 would falsely fail at least one vehicle while achieving no added benefit.
Consequently, the chosen CO cutpoints for the ASM range from 6 to 20. The
range of cutpoints for HC and NOx were chosen the same way.
5.1.1 Using Standard error to Predict Reasonable Cutpoints
Although the cutpoint tables do include a wide range of cutpoints, there is
still a concern that the errors of commission are not representative of what they
might be in a real world scenario. For this reason, the cutpoints shaded at the
end of each table were selected using the standard error.
The ASM cutpoints used are identical to the "safe FTP predictions" in
Table 4. This is because the values are obtained from calculations using both
ASM scores. Each mode has a "sliding scale" of cutpoints, dependent on the other
mode results. In other words, no single ASM5015 or ASM2525 value can be used for
a cutpoint since a vehicle might be clean on one mode and very dirty on the
other. The cutpoints for the IM240, on the other hand, are direct IM240 scores,
in grams per mile. The "safe cutpoints" for the IM240 were determined by
calculating the IM240 score that would predict the "safe FTP Level =2.5% EC"
(see Table 4), using the equations on each respective IM240 scatterplot.
For example, the IM240 HC FTP Level » 2.5% EC is 0.89 g/mi. The regression
equation on the IM240 vs FTP scatterplot is:
FTPpred> = 1.429 * IM240 + 0.04;
Since we want to predict an FTP of no less than 0.89 g/mi, setting FTPprec:.
equal to 0.89 yields an IM240 score of 0.60 g/mi. This was done to calculate
each "safe cutpoint" for the IM240.
5.2 Limitations of Cutpoint Tables
It is important to recognize several limitations in these tables. Most
important is that the database is very small and does not represent the in use
fleet. Additionally, the vehicles were preconditioned by the FTP before the
ASM test and before the IM240 tests, so the correlation between these short
tests will be much better than can be expected for vehicles tested in an I/M
lane, because of all the uncontrolled variables associated with I/M lane tests
like temperature, fuel RVP, distance driven prior to the test, catalyst
temperature, etc. Because all of these variables were controlled for the
G-ll
-------
vehicles in the ASM and IM240 databases, the outpoints can be very stringent
while still avoiding false failures. For example, the IM240 table shows that
outpoints of 0.4/6/1.0 yield IDRs of 97%, 93%, and 100%, for HC, CO and NOx,
respectively, without errors of commission. If cutpoints this stringent were
used for random vehicles tested in I/M lanes, the error of commission rate
would be unacceptably high. Similarly, because of the introduced
malfunctions, the failure rates are not representative of the in-use fleet
failure rates for an acceptable I/M program. So, while it is valid to use
these cutpoint tables to compare the ASM to the IM240, it is not valid to
assume that the rates are representative of those that will be realized in a
real I/M program. The ASM and IM240 testing that EPA is currently sponsoring
in Mesa will provide the actual in-use fleet rates.
5.3 IM240 Identifies Much More Excess Emissions
Using the cutpoint tables to compare the two procedures, the IM240 did
considerably better than the Two-Mode ASM at each tests' optimal cutpoints* .
The IM240 identified 97% of excess HC, 93% of excess CO, and 100% of excess
NOx at cutpoints of 0.4/6/1.0 (HC/CO/NOx). The Two-Mode ASM identified 87%,
80%, and 75% of HC, CO, and NOx, respectively at cutpoints of 0.6/6/1.50.
As discussed in the Variability section, using the standard error of
estimate to choose cutpoints that should prevent exceeding an error of
commission rate of 2.5% can help in assessing the performance of I/M tests.
The shaded cutpoints at the end of each test's cutpoint table suggest that the
IM240's performance is significantly better than the two-mode ASMs. Using the
"safe" cutpoints, the IM240 identifies 92%, 84%, and 71% excess of the excess
HC, CO, and NOx, respectively - the Two-Mode ASM only identifies 75%, 63%, and
64%.
6 . o
The ASM tests were considerably more variable than the IM240 under
controlled laboratory conditions, as evidenced by subjective analyses of the
scatter plots and objective measurements using the standard error statistic.
Testing at real-world I/M lanes will add considerably more variability to both
tests, because conditions known to affect emissions such as temperature,
humidity, and vehicle operating conditions prior to the test. These
uncontrolled variables are expected to add proportionally more variability to
'Optimal Cutpoints', as used here, is the lowest cutpoints the test could go
to and still have zero errors of commission.
G-12
-------
a steady state test like the ASM, but data are not available to evaluate the
validity of the hypothesis.
On the other hand, the increased variability associated with actual I/M
testing will be somewhat offset for the ASM by adding two additional modes; a
50 mph steady mode at road-load horsepower, and an idle mode. This four-mode
ASM procedure is now being performed by EPA in a Mesa Arizona I/M lane.
The result of these offsetting effects on variability will determine the
viability of the ASM as a lower cost substitute for the IM240. A final
conclusion should be postponed until enough Mesa data can gathered for a valid
evaluation.
G-13
-------
Appendix A
Cutpoint Tables
-------
"Two ffaya to Paaa" Lab XM240a
39 1983+
_
«<<"— 1 «• -*« !•«*
ZM240 Cut-Point*
C_B + o
0.4/15/2
0.8/15/2
1.0/15/2
1.2/15/2
0.4/10/2
0.8/10/2
1.0/10/2
1.2/10/2
0.4/ 6/2
0.8/ 6/2
l.O/ 6/2
1.2/ 6/2
0.4/ 5/2
0.8/ 5/2
l.O/ 5/2
1.2/ 5/2
0.4/15/1
0.8/15/1
1.0/15/1
1.2/15/1
0.4/10/1
0.8/10/1
1.0/10/1
1.2/10/1
0.4/ 6/1
0.8/ 6/1
l.O/ 6/1
1.2/ 6/1
0.4/ 5/1
0.8/ 5/1
l.O/ 5/1
1.2/ 5/1
0.4/15/1
0.8/15/1
1.0/15/1
1.2/15/1
bad Va
hiolai
i
Identification
Rat*
Number
of
Failure
ao_2 BC CO IIOz r«ilur«« tim±m
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.00
.50
.50
.50
.50
.50
.50
.50
.50
.50
.50
.50
.50
.50
.50
.50
.50
.25
.25
.25
.25
+ 0.3/12
+ 0.5/12
+ 0.6/12
+ 0.8/12
•f 0.3/ 8
+ 0.5/ 8
+ 0.6/ 8
+ 0.8/ 8
+ 0.3/4.5
+ 0.5/4.5
-t- 0.6/4.5
+ 0.8/4.5
+ 0.3/ 4
+ 0.5/ 4
+ 0.6/ 4
•f 0.8/ 4
+ 0.3/12
+ 0.5/12
+ 0.6/12
+ 0.8/12
+ 0.3/ 8
+ 0.5/ 8
+ 0.6/ 8
+ 0.8/ 8
+ 0.3/4.5
+ 0.5/4.5
+ 0.6/4.5
+ 0.8/4.5
+ 0.3/ 4
+ 0.5/ 4
+ 0.6/ 4
+ 0.8/ 4
+ 0.3/12
+.0.5/12
+ 0.6/12
+ 0.8/12
96%
91%
91%
91%
96%
91%
91%
91%
97%
94%
94%
94%
97%
94%
94%
94%
96%
91%
91%
91%
96%
91%
91%
91%
97%
94%
94%
94%
97%
94%
94%
94%
96%
91%
91%
91%
86%
80%
80%
80%
86%
82%
82%
82%
91%
89%
89%
89%
94%
92%
92%
92%
86%
80%
80%
80%
86%
82%
82%
82%
91%
89%
89%
89%
94%
92%
92%
92%
87%
81%
61%
81%
93%
71%
71%
71%
93%
71%
71%
71%
97%
75%
75%
75%
97%
75%
75%
75%
93%
71%
71%
71%
93%
71%
71%
71%
97%
73%
75%
75%
97%
75%
75%
75%
95%
73%
73%
73%
22
17
16
16
22
18
17
17
26
23
22
22
30
27
26
26
22
17
16
16
22
18
17
17
26
23
22
22
30
27
26
26
23
18
17
17
48%
37%
35%
35%
48%
39%
37%
37%
57%
50%
48%
48%
65%
59%
57%
57%
48%
37%
35%
35%
48%
39%
37%
37%
57%
50%
48%
48%
65%
59%
57%
57%
50%
39%
37% .
37%
Number
of
Sail
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
0
0
0
0
Ke
Rate
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
2%
2%
2%
2%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
2%
2%
2%
2%
0%
0%
0%
0%
Failure Rate Failure Rate
for FTP for Normal
Packing emitting
Vehiolea
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
10%
10%
10%
10%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
10%
10%
10%
10%
0%
0%
0%
0%
y— W4-.1— -
16%
3%
3%
3%
16%
6%
6%
6%
29%
23%
23%
23%
42%
35%
35%
35%
16%
3%
3%
3%
16%
6%
6%
6%
29%
23%
23%
23%
42%
35%
35%
35%
19%
6%
6%
6%
A-l
G-/4
-------
'Two Nay* to Pa**" Lab XM240*
39 1983+ Randomly Selected V*hiol**
Identification
ZM240 Cut-Point* Rate*
f»naMrl 0.8/12
+ 0.3/ 8
+ 0.5/ 8
+ 0.6/ 8
•t- 0.8/ 8
+ 0.3/4.5
+ 0.5/4.5
+ 0.6/4.5
•t- 0.8/4.5
+ 0.3/ 4
+ 0.5/ 4
+ 0.6/ 4
+ 0.8/ 4
+ 0.5/ 8
fiC
96%
91%
91%
91%
97%
94%
94%
94%
97%
94%
94%
94%
97%
95%
95%
95%
97%
95%
95%
95%
97%
95%
95%
95%
97%
95%
95%
95%
92%
Cfi
87%
83%
83%
83%
91%
89%
89%
89%
94%
92%
92%
92%
91%
90%
90%
90%
91%
90%
90%
90%
93%
92%
92%
92%
9C%
94%
94%
94%
84%
BQs
95%
73%
73%
73%
97%
75%
75%
75%
97%
75%
75%
75%
100%
93%
93%
93%
100%
93%
93%
93%
100%
93%
93%
93%
100%
93%
93%
93%
71%
of
r.ilm^a
23
19
18
18
26
23
22
22
30
27
26
26
27
25
24
24
27
25
24
24
29
27
26
26
32
30
29
29
19
Failure
Bat*.
50%
41%
39%
39%
57%
50%
48%
48%
65%
59%
57%
57%
59%
54%
52%
52%
59%
54%
52%
52%
63%
59%
57%
57%
70%
65%
63%
63%
41%
of
tsLm
0
0
0
0
0
0
0
0
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
0
Xo
Rate
0%
0%
0%
0%
0%
0%
0%
0%
2%
2%
2%
2%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
2%
2%
2%
2%
0%
Failure Rate
for FTP
Pa**iag
Ythtnlti
0%
0%
0%
. °*
0%
0%
0%
0%
10%
10%
10%
10%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
10%
10%
10%
10%
0%
Failure Rate
for Normal
Emitting
Yahinlif
19%
10%
10%
10%
29%
23%
23%
23%
42%
35%
35%
35%
32%
29%
29%
29%
32%
29%
29%
29%
39%
35%
35%
35%
48%
45%
45%
45%
10%
Idaotifioation Rate** la bold indicate* that th* outpoint for
that •aittaot ha* oau«*d th* Error (•) of Co*mi»ioi
R-2
-------
Two - Mod* ASM T««t — A8M501S «ad A8M2525
39 Vahlolaa
Mnlti-ASM Mod
Cut-Points
HC/CO/NOx
0.6/20/2.00
0.8/20/2.00
1.0/20/2.00
1.5/20/2.00
0.6/15/2.00
0.8/15/2.00
1.0/15/2.00
1.5/15/2.00
0.6/ 8/2.00
0.8/ 8/2.00
l.O/ 8/2.00
1.5/ 8/2.00
0.6/ 6/2.00
0.6/ 6/2.00
l.O/ 6/2.00
1.5/ 6/2.00
0.6/20/1.50
0.8/20/1.50
1.0/20/1.50
1.5/20/1.50
0.6/15/1.50
0.8/15/1.50
1.0/15/1.50
1.5/15/1.50
0.6/ 8/1.50
0.8/ 8/1.50
l.O/ 8/1.50
1.5/ 8/1.50
0.6/ 6/1.50
0.6/ 6/1.50
l.O/ 6/1.50
1.5/ 6/1.50
0.6/20/1.25
0.8/20/1.25
1.0/20/1.25
1.5/20/1.25
« Identification
Rata
ac
88%
75%
74%
74%
88%
75%
75%
75%
88%
76%
76%
76%
88%
87%
87%
87%
88%
75%
74%
74%
88%
75%
75%
75%
88%
76%
76%
76%
88%
87%
87%
87%
88%
75%
74%
74%
C2
77%
62%
60%
60%
77%
62%
62%
62%
81%
67%
67%
67%
82%
80%
80%
80%
78%
63%
61%
61%
78%
63%
63%
63%
82%
67%
67%
67%
83%
80%
80%
80%
78%
63%
61%
61%
B2X
68%
49%
43%
43%
68%
49%
49%
49%
68%
49%
49%
49%
68%
59%
59%
59%
83%
64%
58%
58%
83%
64%
64%
64%
83%
64%
64%
64%
83%
75%
75%
75%
83%
64%
38%
38%
Hnabar
of
Failnraa
13
7
6
6
13
7
7
7
14
8
8
8
15
11
11
11
14
8
7
7
14
8
8
8
15
9
9
9
16
12
12
12
14
9
8
8
Failura
Brta.
33%
18%
15%
15%
33%
18%
18%
18%
36%
21%
21%
21%
38%
28%
28%
28%
36%
21%
18%
18%
36%
21%
21%
21%
38%
23%
23%
23%
41%
31%
31%
31%
36%
23%
21%
21%
Numbar
of
Xali
i
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
1
1
1.
Failure Rata
fox FTP
Xe Paaaing
B*fc*
3%
0%
0%
0%
3%
0%
0%
0%
3%
0%
0%
0%
3%
0%
0%
0%
3%
0%
0%
0%
3%
0%
0%
0%
3%
0%
0%
0%
3%
0%
0%
0%
3%
3%
3%
3%
Vahiolaa
9%
0%
0%
0%
9%
0%
0%
0%
9%
0%
0%
0%
9%
0%
0%
0%
9%
0%
0%
0%
9%
0%
0%
0%
9%
0%
0%
0%
9%
0%
0%
0%
9%
9%
9%
9% '
Failure Rata
foe Normal
Emitting
Yfhi"1*"
17%
3%
0%
0%
17%
3%
3%
3%
20%
7%
7%
7%
23%
10%
10%
10%
20%
7%
3%
3%
20%
7%
7%
7%
23%
10%
10%
10%
27%
13%
13%
13%
20%
10%
7%
7%
A-3
G-/6
-------
Two - Mod* ASM
— ASMS01S and ASM2S25
39 V«hiol«»
ttalti-ASM Mod
Cut-Point*
HC/CO/NOx
0.6/15/1.25
0.8/15/1.25
1.0/15/1.25
1.5/15/1.25
0.6/ 8/1.25
0.8/ 8/1.25
l.O/ 8/1.25
1.5/ 8/1.25
0.6/ 6/1.25
0.6/ 6/1.25
l.O/ 6/1.25
1.5/ 6/1.25
0.6/20/1.00
0.8/20/1.00
1.0/20/1.00
1.5/20/1.00
0.6/15/1.00
0.8/15/1.00
1.0/15/1.00
1.5/15/1.00
0.6/ 8/1.00
0.8/ 8/1.00
l.O/ 8/1.00
1.5/ 8/1.00
0.6/ 6/1.00
0.6/ 6/1.00
l.O/ 6/1.00
1.5/ 6/1.00
1.6/13/1.72
!• Identification
BC
88%
75%
75%
75%
88%
76%
76%
76%
88%
87%
87%
87%
94%
82%
80%
80%
94%
82%
82%
82%
93%
82%
82%
82%
95%
94%
94%
94%
75%
Rat*
£Q
78%
63%
63%
63%
82%
67%
67%
67%
83%
80%
80%
80%
85%
72%
70%
70%
85%
72%
72%
72%
90%
76%
76%
76%
91%
90%
90%
90%
63%
BOX
83%
64%
64%
64%
83%
64%
64%
64*%
83%
75%
75%
75%
96%
86%
80%
80%
96%
86%
86%
86%
96%
86%
86%
86%
96%
96%
96%
96%
64%
Number Number
of
Tmllnx
14
9
9
9
15
10
10
10
16
13
13
13
18
14
13
13
18
14
14
14
19
15
15
15
20
18
18
18
8
failure
•ff Rate
36%
23%
23%
23%
38%
26%
26%
26%
41%
33%
33%
33%
46%
36%
33%
33%
46%
36%
36%
36%
49%
38%
38%
38%
51%
46%
46%
46%
21%
of
la'.
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
0
2
2
2
2
2
2
2
2
0
Ke
Rate
3%
3%
3%
3%
3%
3%
3%
3%
3%
3%
3%
3%
5%
5%
5%
5%
5%
5%
5%
0%
5%
5%
5%
5%
5%
5%
5%
5%
0%
Failure R*t«
for FTP
Paaaing
VehialM
9%
9%
9%
9%
9%
9%
9%
9%
9%
9%
9%
9%
.18%
18%
18%
18%
18%
18%
18%
0%
18%
18%
18%
18%
18%
18%
18%
18%
0%
Failure Rat*
for Normal
Emitting
YihH nl tt
20%
10%
10%
10%
23%
13%
13%
13%
27%
17%
17%
17%
30%
23%
20%
20%
30%
23%
23%
23%
33%
27%
27%
27%
37%
30%
30%
30%
7%
Identification Rat*« IB bold indioat* that th« outpoint for
that wait t ant (•) hM oaq«ad th* lrror(«) of Commi»ion.
A-4
-------
Appendix B
Sc»tt«rplots
G-/S
-------
Figure B-l
HC Bnissions IM240 vs FTP
(39 1983 & Newer Vehicles)
With 95% Confidence Bands
5 —
4.5
4 —
3.5 -
1
0 -:
Y = 1.429(IM240) + 0.04
r"2 = 94.7%
Std. Error = 0.24 q/mi
0.5
1.5
IM240 HC (g/mi)
2.5
FTP Values
Predicted Values + Std. Error
- Std. Error
-------
Figure B-2
CO Emissions IM240 vs FTP
(39 1983 & Newer Vehicles)
With 95% Confidence Bands
80 -
70 -
Y = I. 11 (11-1240) - 0. 60
r~2 = 91.8%
Std. Error =3.84 q/mi
60 -
(^
K)
0
|
8
&
50 -
40 "
30 -r-
20 -
10
20
30
IM240 CO (g/mi)
40
50
60
FTP Values
Predicted Values
+ Std. Error
- Std. Error
-------
Figure B-3
NOx emissions IM240 vs FTP
(39 1983 t Hewer Vehicles)
With 95% Confidence Bands
4 —
3.5
3 —
2.5-
2 —
1.5 —
0.5 —
Y = 0.944(IM240) + 0.04
r~2 = 84.3%
Std. Error = 0.28 g/mi
0.5
1.5 2 2.5
IM240 NOx (g/mi)
3.5
FTP Values
Predicted Values
+ Std. Error
- Std. Error
-------
3.00 -f
K)
KJ
~ 2.50 -r
o>
o
2.00 -r
1.50 —
Figure B-4
HC emissions ASM 2-Mode vs FTP
(39 1983 & Newer Vehicles)
4.00 —
Y = IW*[-3.96 E-07 (ASM2525) -1-4.60 E-07 (ASM5015) ] -t-0.52
r"2 = 49.2%
Std Error = 0.60 g/mi
1.00 -r
0.50 -|-
0.00 —
0.00
0.50
1.00
1.50 2.00 2.50
ASM HC Prediction
3.00
3.50
4.00
-------
25.00 —
OJ
8
20.00 —
15.00 -T
10. 00
Figure B-5
CO Emissions ASM 2-Mode vs FTP
(39 1983 & Newer Vehicles)
35.00 —
30.00 —
Y = IW*[2.64 E-03(ASM2525) + 5.10 E-05(ASM5015)] + 4.22
r'2 =43.5%
Std Error = 4.8 g/mi
5.00
0.00
0.00
**
I!
5.00
10.00 15.00
ASM CO Prediction
20.00
25.00
-------
3.00 —
2.50
Figure B-6
NOx Emissions ASM 2-Mode vs FTP
(39 1983 & Newer Vehicles)
Y = IW*[1.13 E-07(ASM2525) + 1.30 E-07(ASM5015)] + 0.52
r"2 = 71.4%
Std. Error = 0.36 g/mi
2.00
1.50
i.oo -•-
0.50 --
0.00
0.00
0.50
1.00 1.50 2.00
A.SM NOx Prediction
2.50
3.00
-------
Figure B-7
HC Emissions ASM5015 vs FTP
(69 1983 & Newer Vehicles - Including CA Certified Vehicles)
With 95% Confidence Interval
5.00 -
4 . 50 -
4.00 -
y = 0.0003x + 0.51
r"2 = 36.4% .
Std. Error =0.57 g/mi
N
VA
3.50 -
3.00 —
1000
2000
3000
4000 5000 6000
A.SM5015 HC (IW*ppm / 10A
7000
8000
9000 10000
FTP Values
Predicted Values + Std. Error
- Std. Error
-------
Figure B-8
CO Emissions ASM5015 vs FTP
(69 1983 & Newer Vehicles - Including CA Certified Vehicles)
With 95% Confidence Interval
80.00 -
70.00 —
I
60.00 -
y = 0.0015x -M . 4c
r-2 = 44.9%
Std. Error = 5.04 g/mi
N)
50.00 -
i
8«o.ooT
h 30.00
0.00 -?
0
5000
10000
A.SM5015 CO (IW*%)
15000
20000
FTP Values
Predicted Values
• + Std. Error
- Std. Error
-------
Figure B-9
NOx Emissions ASM5015 vs FTP
(67 1983 & Hewer Vehicles - Including CA Certified Vehicles)
With 95% Confidence Interval
4.00 -
3.50 -
y = 0.253J: + 0. 50
r"2 = 62.1%
Std. Error = 0.46 q/ni:
3.00 -
2.50 -
O>
2.00 —
1.50 -,,.,»*•'*
1 .00 —
0.50
0.. 00
0.00
1.00 2.00
3.00 4.00 5.00 6.00
A.SM5015 NOx (IW*ppm / 10*6)
7.00
.00 9.00 10.00
FTP Values
Predicted Values
Std. Error
- Std. Error
-------
Figure B-10
HC Emissions ASM2525 vs FTP
(39 1983 & Hewer Vehicles)
With 95% Confidence Interval
5.00 -
4.50 -
4.00 -
3.50 -
y = 0.0003* + 0. 52
r"2 = 20.2%
Std. Error =0.75 g/mi
500
1000
1500
2000 2500 3000
ASM2525 HC (IW*ppm / 10A3)
3500
4500 5000
FTP Values
Predicted Values
+ Std. Error
- Std. Error
-------
Figure B-ll
CO Emissions ASM2525 vs FTP
(39 1983 t Newer Vehicles)
With 95% Confidence Interval
80.00 -
70.00 -
60.00 —
y = 0.0027i: 44.20
r"2 = 43.5%
Std. Error = 4.85 g/mi
NJ
\0
•g 50.00 -
3 |
n 40.00 --
30.00 j
20.00 -
0.00
1000
2000
3000 4000
ASM2525 CO (IW*%)
5000
6000
7000
FTP Values
Predicted Values
+ Std. Error
- Std. Error
-------
Figure B-12
NOx Emissions ASM2525 vs FTP
(39 1963 & Newer Vehicles)
With 95% Confidence Interval
u>
o
4.50 -
4.00 -
y = 0.250:: + 0. 5-2
r"2 = 70.3%
Std. Error =0.37 g/rai
1.00 2.00 3.00
—I— _
4.00 5.00 6.00 7.00
A.SM2525 HOx (IH*ppm / 10^6)
8.00 9.00 10.00
FTP Values
Predicted Values + Std. Error
- Std. Error
-------
Appendix H:
Estimated Cost of High-Tech I/M Testing
-------
5.2.1 General Methodology
EPA's estimates of the costs of high-tech test procedures are driven
by a number of assumptions. Costs in conventional centralized and
decentralized test-and-repair programs were derived using current inspection
costs in I/M programs as they are reported to EPA as the starting point. For
decentralized test-only networks costs are modelled in a manner similar to
centralized programs, since all current test-only programs are centralized,
however, costs are estimated using a range of test volumes and a higher level
of state oversight is assumed since the network is composed of independent
operators and may have a higher number of test sites than in centralized
programs.
Another key assumption is that adding the new tests will increase
inspection costs in programs that are now efficiently designed and operated.
In programs that are not now well designed, current costs are likely to be
higher than necessary and the cost increase less if efficiency improvements
are made simultaneously. In order to perform the high-tech tests new
equipment will have to be acquired and additional inspector time will be
required for some test procedures. The amount of the cost increase will be
determined to a large degree by the costs of acquiring new equipment and the
impact of the longer test on throughput in a high volume operation. Average
test volume in decentralized programs is low enough to easily absorb the
additional test time involved (although at a cost in labor time). Equipment
costs are analyzed in terms of the additional cost to equip each inspection
site (i.e., each inspection lane in centralized inspection networks, and each
licensed inspection station in decentralized networks).
By focusing on the inspection lane or station as the basic unit of
analysis, the resulting cost estimates are equally applicable in large
programs, with many subject vehicles and inspection sites, or small programs,
with few subject vehicles and inspection sites. Previous EPA analyses of
costs in I/M programs have found that the major determinants of inspection
costs are test volume and the level of sophistication of the inspection
equipment. Costs of operating programs were not found to be measurably
affected by the size of the program (for further information the reader may
refer to EPA's report entitled, "I/M Network Type: Effects on Emission
Reductions, Cost, and Convenience"). Figures on inspection volumes at
inspection stations and lanes are available from I/M program operating data.
This information enables the equipment cost per vehicle and the additional
staff cost per vehicle to be calculated for each test procedure.
H-2
-------
The equipment cost figures presented in this paper are based on the
costs of the equipment EPA believes is best suited for high-tech testing.
They are current prices quoted by manufacturers, and do not reflect what the
per unit prices might be if this equipment were purchased in volume. Staff
costs are based on prevailing wage rates for inspectors in both types of
programs as reported in conversations with state I/M program personnel.
Construction costs in centralized programs are based on estimates supplied by
centralized contractors. Other site costs and management overhead in
centralized programs are back calculated from current inspection costs. For
decentralized networks, it is assumed that longer test times could be absorbed
with no increase in sites. The current average volume in decentralized
stations is 1,025 vehicles per year (between 3 and 4 vehicles per day,
depending upon the number of days per year the station is open).
Consequently, increasing the length of the test, to the degree that the new
procedures would, is not expected to impact the number of inspections that can
be performed.
5.2.2 Equipment Needs and Coats
A pressure metering system, composed of a cylinder of nitrogen gas
with a regulator, and hoses connecting the tank to a pressure meter, and to
the vehicle's evaporative system is needed to perform evaporative system
pressure testing. Hardware to interface the metering system with a
computerized analyzer is also needed and is included in the cost estimate.
Purge testing can be performed by adding a flow sensor with a computer
interface, a dynamometer, and a Video Driver's Aid. With the further addition
of a Constant Volume Sampler (CVS) and a flame ionization detector (FID) for
HC analysis, two nondispersive infrared (NDIR) analyzers for CO and carbon
monoxide (CO2), and a chemiluminescent (CI) analyzer for NOX, transient
testing can be performed.
The analyzers used for the transient test are laboratory grade
equipment. They are designed to higher accuracy and repeatability
specifications than the NDIR analyzers used to perform the current I/M tests.
Table 5-4 shows the estimated cost of equipment for conducting high-tech
tests. This quality of technology is essential for accurate instantaneous
measurements of low concentration mass emission levels.
H-3
-------
Table 5-4
Equipment Coats for New Tests
Teat Equipment Price
Pressure Metering System $600
Purge Flow Sensor $500
Dynamometer $45,000
Video Drivers Aid $3,000
Transient CVS & Analyzers $95,000
TOTAL $144,100
The figures in Table 5-4 do not include the costs of expendable
materials. Nitrogen gas is used up in performing the pressure test.
Additionally, the FID burns hydrogen fuel. Calibration gases are needed for
each of the analyzers used in the transient test. Because the analyzers used
in the transient test are designed to more stringent specifications than the
analyzers currently used in the field, bi-blends, gaseous mixtures composed of
one interest gas in a diluent (usually nitrogen) are used to calibrate them.
Multi-blend gases, such as are typically used to calibrate current I/M
equipment, are not suitable. Current estimates for expendables are shown in
Table 5-5. The replacement intervals are estimated based on the usage rates
observed in the EPA Indiana pilot program and typical inspection volumes as
presented later in this section. Calculations of per vehicle equipment costs
presented throughout this report include per vehicle costs of these
expendables as well.
Table 5-5
Expendables for New Tests
Replacement Interval
Test Material Cost Centralized Decentralized
Pressure N2 Gas $30 250 tests 250 tests
Transient H2 Fuel $60 2 months 1000 tests
HC Cal Gas $60 2 months 1000 tests
CO Cal Gas $60 2 months 1000 tests
C02 Cal Gas $60 2 months 1000 teats
Staff costs have been found to vary between centralized and
decentralized programs, as does the effect on the number of sites in the
network infrastructure. Therefore, the following sections are devoted to
separate cost analyses for each network type.
H-4
-------
5.2.3 Coat to Upgrade Centralized Networks
5.2.3.1 Basic Assumptions
The starting point in this analysis is the current average per
vehicle inspection cost in centralized programs. A figure of $8.50 was used
based upon data from operating programs. This figure includes the cost of one
or more retests and network oversight costs. The key variables to consider in
estimating the costs in centralized networks are throughput, equipment, and
staff coats. Data on these variables were obtained by contacting program
managers in a number of these programs, and by surveying program contracts and
Requests for Proposal.
Throughput refers to the number of vehicles per hour that can be
tested in a lane. The higher the throughput rate, the greater the number of
vehicles over which costs are spread, and the lower the per vehicle cost. EPA
contacted program managers and consulted the contracts in a number of
centralized programs to determine peak period throughput rates in the
different systems. Rates were as reported in Table 5-6.
Table 5-6
Peak Period Throughput Rates in Centralized I/M Programs
Program Vehicles Tested per Hour
Arizona 20
Connecticut 25-30
Illinois 25
Maryland 25-35
Wisconsin 25-30
On the basis of this information, 25 vehicles per hour was assumed to
represent the typical peak period throughput rate or design capacity in
centralized I/M programs. During off-peak hours and days, throughput is lower
since there is not a constant stream of arriving vehicles. Conversations with
individuals in the centralized inspection service industry indicate that
inspectors start at minimum wage or slightly higher, that by the end of the
first year they earn $5.50 to $6 per hour, and that they generally stay with
the job for one to three years. Thus, $€ per hour was used to estimate the
average inspector's hourly wage.
Estimates of the costs of adding pressure testing, purge testing, and
transient tailpipe testing were derived by taking the current costs for the
new equipment to perform the new tests, dividing it by the number of
inspections expected to be performed in the lane over a five year period and
H-5
-------
adding it to the current $8.50 per vehicle cost, with a further adjustment for
the impact of test time on throughput, and thus on the number of sites and
site costs. The same is done to estimate additional personnel costs
associated with adding the new tests. When independent programs were surveyed
to determine the length of a typical contract, it was discovered that
Illinois, Florida, and Minnesota all have five year contracts, Arizona has a
seven year contract, and the program in the State of Washington is operating
under a three year contract, resulting in an average contract length of five
years among the five programs surveyed. Five years was therefore chosen as
the typical contract length.
The number of inspections expected to be performed over the five year
contract period was derived by calculating the total number of hours of lane
operation, estimating the average number of vehicles per lane and multiplying
the two. A lane is assumed to operate for 60 hours a week (lane operation
times were found to vary from 54 to 64 hours per week), 52 weeks a year for
five years for a total of 15,600 hours. Lanes are assumed to have a peak
throughput capacity of 25 vehicles per hour. Modern centralized inspection
networks are designed so that they can accommodate peak demand periods with
all lanes operating at this throughput rate. Networks are usually designed so
that average throughput is 50-65% of peak capacity or 13-15 vehicles per hour.
When operating for 15,600 hours over the life of a contract, a centralized
inspection lane is estimated to perform a total of 195,000 inspections, or
about 39,000 per year.
5.2.3.2 The Effect of Changing Throughput
The addition of evaporative system pressure testing to a centralized
program would result in a slight decrease in the throughput capacity. The
addition of purge and transient testing, along with pressure testing, would
result in a further decrease.
Assuming the same test frequency (i.e., annual or biennial) the
reduced throughput rate means that the number of lanes needed to test a given
number of vehicles would increase accordingly, as would the size of the
network infrastructure needed to support the test program. The result is an
increase in the cost per vehicle. Actual consumer cost depends on the test
frequency; EPA would encourage states to adopt biennial programs to reduce the
costs and imposition of the program. Less frequent testing only slightly
reduces the emission reduction benefits while cutting test costs almost in
half.
H-6
-------
One way to estimate the coat would be to simulate an actual network
of stations and lanes in a given city. One could attempt to assess land
costs, building costs, staff and equipment costs, costs for all necessary
support systems, and other cost factors. However, this approach would be very
time consuming and would rely on information which is proprietary to the
private contractors that operate the programs and is, therefore, unavailable.
Instead, the cost of the increased number of lanes and stations is derived by
analyzing current costs and subtracting out equipment, direct personnel,
construction, and state agency oversight costs. The remainder is adjusted by
the change in throughput in the new system. Then, new estimates of equipment,
personnel, construction, and oversight costs are added back in to obtain the
estimated total cost.
As discussed previously, the typical high volume station can test 25
vehicles per hour, performing (in most cases) a test consisting of 30 seconds
of high speed preconditioning or testing, followed by 30 seconds of idle
testing. In addition, a short time is spent getting the vehicle into position
and preparing it for testing. This leads to a two to three minute test time
on average, depending upon what short test is performed. EPA recently issued
alternative test procedures for steady-state tests that reduce various
problems associated with those tests, especially false failures, but at a cost
of longer average per test time.
Current costs were estimated by contacting operating program
personnel, equipment vendors and contractors. The most sophisticated
equipment installation (i.e., the equipment for loaded steady-state testing)
was used to estimate current equipment costs.
The cost to acquire and install a single curve dynamometer and an
analyzer in existing networks is about $40,000 or 21$ per vehicle using the
basic test volume assumptions. As indicated previously, a staff person is
assumed to earn $6.00 per hour. When this figure is multiplied by 15,600
total contract hours and divided by 195,000 vehicles, direct staff costs are
estimated at 48C per vehicle. Existing centralized networks typically have
two staff per lane. Thus, total staff costs work out to 96C per vehicle.
Total average construction costs are estimated at $800,000 for a five lane
station, yielding an average per vehicle cost of 82$. In this analysis a
figure of $1.25 is used to estimate the amount of the state retainer. This
reflects EPA's best estimate of the per vehicle expense for a good state
quality assurance program in a centralized network. Equipment, staff,
construction, and state costs add up to $3.24 per vehicle. Subtracting this
amount from the current average of $8.50 leaves $5.26 in infrastructure costs
and other overhead expenses including employee benefits and employer taxes as
H-7
-------
shown in Table 5-7. This amount is then factored by the change in the
throughput rate and the equipment, oversight, and staff costs for the new
tests are then added.
Table 5-7
Current Program Costs
Total Cost Less
Increments Per Vehicle Cost Increments
Current $8.50
Equipment $0.21 $8.29
Staff $0.96 $7.33
Construction $0.82 $6.51
State Retainer $1.25 $5.26
5.2.3.2 Costs of Mew Tests
Most centralized programs use a two position test queue; emission
test are done in one position while emission control devices are checked in
the other, along with other functions such as fee collection. In this type of
system the throughput rate is determined by the length of time required to
perform the longest step in the sequence, not by length of the entire test
sequence. The new tests would likely be performed in a three position test
queue, with one position dedicated to fee collection and other administrative
functions, one to performing the pressure test, and the third to performing
the transient and purge tests. The transient/purge test is a longer test
procedure than the ones currently used in most I/M programs and is the longest
single procedure in the whole inspection process. Thus, it is the determining
factor in lane throughput and will therefore influence the number of test
sites required.
The transient test takes a maximum of four minutes to perform. An
additional minute is assumed to prepare the vehicle for testing, for a maximum
total of five minutes. The pressure test would take approximately two
minutes, and could be shortened through such potential strategies as
computerized monitoring of the rate of pressure drop. EPA is in the process
of looking at potential fast-pass and fast-fail strategies, and'preliminary
results suggest that roughly 33% of the vehicles tested could be fast passed
or failed based upon analysis of data gathered during the first 93 seconds of
the IM240 (i.e., Bag 1) using separate fast-pass and fast-fail outpoints.
Hence, EPA estimates that the average total test time could be shortened to at
least four minutes per vehicle. This translates into a throughput capacity of
15 vehicles per hour. To accommodate peak demand periods and maintain short
wait times, a design throughput rate of half of capacity is assumed, for a
H-8
-------
typical throughput rate of 7.5 vehicles per hour. Assuming the same number of
hours of lane operation as previously, the total number of tests per lane in a
transient lane is estimated to be 117,000 over the five year contract period.
State quality assurance program costs would increase given the
complexity and diversity of the test system; an estimate of an additional 50$
is used here but the amount could vary depending upon the intensity of the
oversight function the state chooses. Staff costs per vehicle are calculated
using the same assumptions for wages and hours of operation as shown in Table
5-7; however, the cost is spread over 117,000 tests over the life of the
contract rather than 195,000. The result is staff costs of 80* per staff per
vehicle. Three staff per lane are assumed to perform the tests. The
additional tasks performed by inspectors in conducting the new tests - i.e.,
disconnecting vapor lines and connecting them to analytical equipment for the
evaporative tests and driving the vehicle through the transient driving cycle
- do not require that inspectors have higher levels of skill than they do
presently. Rather, these tasks can be performed by comparably skilled
individuals trained to these specific tasks. Total staff costs work out to
$2.40 per vehicle. Equipment costs for each test procedure are derived by
taking the equipment costs from Table 5-4 and calculating the costs of five
years worth of expendables using the figures in Table 5-5 and dividing by
117,000. Construction costs for a five lane station are assumed to rise to
$1,000,000. This is due to the fact that slightly longer lanes may be needed
in order to accommodate test equipment and facilitate faster throughput.
Dividing this figure by 117,000 vehicles per lane yields a per vehicle cost of
$1.71. The resulting costs estimates are shown in Table 5-8. Table 5-8 shows
the result of factoring the figure of $5.26, from Table 5-7, by the change in
the throughput rate and adding in the equipment, staff, construction and state
costs associated with the new test procedures. The figure of $5.26 is
multiplied by 12.5/7.5, i.e., the ratio of the design throughput rate in the
current program to the design throughput rate in a program conducting pressure
purge and transient testing.
H-9
-------
Table 5-8
Coats to Add Proposed Tests to Centralized Programs
Running Total
Increments Per Vehicle Cost Cost per Vehicle
Adjust for Throughput $5.26 * 12.5/7.5 $9.12
Staff $2.40 $11.52
Construction $1.71 $13.23
Oversight $1.75 $14.98
Pressure Test $0.13 $15.11
Purge Test $0.41 $15.52
Transient Test $0.87 $16.40
Thus, the cost of adding the new tests to centralized networks is
found to be about double the current average cost. The cost of centralized
test systems has been dropping in the past few years as a result of
competitive pressures and efficiency improvements. These factors may drive
down the costs of the new tests as well, especially as they relate to
equipment costs. Given that conservative assumptions were made regarding
equipment costs of $144,000 per lane, and low throughput rates, the cost
estimate presented here can be fairly viewed as a worst case assumption. As
discussed earlier, the important issue is the quality of the test, not the
frequency, so doing these tests on a biennial basis would offset the increased
per test cost.
5.2.4 Cost to Upgrade Decentralized Programs
5.2.4.1 Basic Assumptions
The methodology used to estimate costs in decentralized programs is
similar to that described above for centralized programs. Equipment and labor
costs are key variables as they were in determining costs for centralized
programs. However, estimates of costs for decentralized programs presented
here do not include estimates of land costs and overhead. While inspections
in decentralized programs are generally conducted in pre-existing facilities
rather than newly built ones, there are nonetheless a variety of overhead
expenses as well as opportunity costs associated with making space available
for inspections in a facility that provides a number of other services as
well. Data on these costs are not available and they cannot be deduced from
reported inspection fees since, in most programs, fees are capped by law and,
hence, do not reflect the actual cost of providing an inspection.
H-10
-------
Total test volume rather than throughput and test time are the
critical factors affecting cost in decentralized programs. Licensed
inspection stations at present only perform, on the average, about 1,025
inspections per year, as shown in Table 5-9 (note that this number is a
station-weighted average). Test volumes among stations in a single program
can vary widely as shown in Section 7.0. It should also be noted that all
decentralized programs in enhanced I/M areas, except for California, Virginia,
and Colorado (which tests vehicles five years old and newer biennially, and
vehicles older than five years annually) are annual programs. In this
analysis the effect on per vehicle costs of switching from an annual
inspection frequency to biennial, as well the effect of varying inspection
volume, will be examined.
Table 5-9
Inspection Volumes in Licensed Inspection Stations
Program
California
Colorado
Dallas/Ft. Worth
El Paso
Georgia
Houston
Louisiana
Massachusetts
Nevada
New Hampshire
New York
Pennsylvania
Rhode Island
Virginia
Weighted Average
Vehicles per Year
6,180,093
1,655,897
1,948,333
278,540
1,118,448
1,482,349
145,175
3,700,000
523,098
137,137
4,605,158
3,202,450
650,000
481,305
Vehicles per Station
799
1,104
1,624
1,161
1,729
1,348
1,037
1,321
1,260
564
1,071
834
684
1,301
1,025
Annual tests of 1,025 vehicles per station is equivalent to between
three and four inspections per day depending upon the number of days per week
the facility is open and inspections are available. This is far below the 75
inspections per day projected in a multi-position high volume lane with three
inspectors conducting high-tech tests, and significantly below the 16
inspections per day that could be done in a single position inspection bay
with only one inspector (the derivation of this figure is detailed below).
Two conclusions can be drawn from this. The first is that the additional time
requirements of the new tests will not force a reduction in the total number
of inspections that most stations can perform. The second is that, because
costs are spread over a smaller number of vehicles than in the case of high-
H-ll
-------
volume, centralized stations, the cost per vehicle for the new tests will be
larger in this type of inspection network.
The higher costs for high-tech testing equipment have prompted
questions of whether all current inspection stations would choose to stay in
the inspection business with the implementation of an enhanced program, and
how high a drop-out rate programs would experience if some did not. EPA knows
of no data or reasonable assumptions by which a station drop-out rate could be
reliably estimated. In this analysis inspection costs for high-tech testing
are estimated for three scenarios: one where all stations remain in the
inspection business, one where 50% of the stations drop out, and one where
enough stations drop out such that those that remain are operating at maximum
possible volume assuming that each has one inspection bay which has not been
improved for high throughput and one inspector performing all parts of the
inspection. In all three scenarios a biennial inspection frequency is
assumed.
The current average test fee for vehicle inspection in decentralized
programs is about $17.70 (again, the derivation of this figure can be found in
EPA1s technical information document, "I/M Network Type: Effects on Emission
Reductions, Cost, and Convenience"). Note that this figure may substantially
underestimate actual costs since most states limit the inspection fee that a
station may charge. In many cases, the actual fee is likely to be below cost;
stations presumably obtain sufficient revenue to stay in business by providing
other services, which may include repair. It should also be noted that the
intensity of the inspection and the sophistication and cost of the analyzer
vary significantly among programs. Average inspection costs and revenues by
program, taking these factors into account, are estimated in Section 7.4.1.
The costs for adding high-tech tests are derived by estimating the
per vehicle costs of the key components: labor; equipment, including
expendables; and support, i.e., service contracts and annual updates. Per
vehicle costs are estimated by deriving total costs for each component and
dividing by the number of vehicle inspections expected to be performed in a
year, again, taking into account variations in inspection volumes and changes
in frequency. Equipment costs are spread over the useful life of the
equipment. While a piece of equipment's useful life can vary considerably in
actual practice, a five year equipment life is assumed.
While large businesses, such as dealerships, may be able to afford to
purchase current analyzer equipment outright, the smaller gas stations and
garages typically have to finance these purchases (although in some cases they
may lease equipment). The higher cost of the equipment needed to perform
H-12
-------
purge and transient testing ($144,000 for the dynamometer, CVS, analyzers,
etc., as opposed to $12,000 to $15,000 for the most sophisticated of the
current NDIR-based analyzers) makes it even more likely that these purchases
will have to be financed for most inspection stations. Equipment costs are
amortized over five years at 12% interest in the analysis in this report.
Program personnel in decentralized programs were contacted to
determine inspector wage rates. In many cases, inspectors are professional
mechanics earning about $25 per hour. However, most states do not require
inspectors to be mechanics, and inspections may be performed by less skilled
individuals who typically earn $6 or $7 per hour. The prevalence of different
wage rates among inspectors is unknown. Therefore, EPA assumed an average
wage of $15 per hour for this analysis. An overhead rate of 40% is assumed,
for a total labor cost of $21 an hour.
5.2.4.3 Cost Components and Scenarios
The full test, including data entry on the computer, preparing the
vehicle for the different steps in the test procedure and conducting them, is
estimated to take 30 minutes with only one inspector performing all tasks in a
repair bay that is not configured specifically for inspection throughput.
With labor costs at $21 per hour, as described above, this works out to $11.50
per vehicle. Equipment costs are taken from Table 5-4 and are amortized over
a five year period at 12 percent annual interest (changing the assumed
interest rate does not significantly affect the total per vehicle cost). This
brings the total cost for the equipment package over the five year period to
$192,325. These costs are divided by five years worth of inspections. The
costs of expendables from Table 5-5 are added in according to the usage rates
assumed for decentralized programs. Two other expenses typically encountered
in decentralized programs are service contracts and software updates. Based
on information from states, service contracts are estimated at $200 per month
and annual software updates are assumed to cost $1,500.
Per vehicle costs are estimated for three scenarios, biennial testing
is assumed in all three. In the first, all stations remain in the inspection
program. In the second, 50 percent of the stations drop out of the program,
and in third there are only the minimum number of stations in the program to
enable each to inspect at full volume with one inspector performing all parts
of the inspection and a service station bay that has not been improved for
high throughput.
In the first scenario, the switch to biennial would mean that annual
volume is cut in half, or 513 vehicles per year. In the second scenario the
H-13
-------
50 percent reduction in the number of stations brings the annual inspection
volume back to 1,025. In the fourth scenario, it is assumed that each station
inspects at maximum capacity, i.e., one vehicle every thirty minutes, and that
an inspector is available 50 hours per week. This results in an annual volume
of 5,200 vehicles.
Table 5-10
Costs to Conduct High-Tech Testing in Decentralized Programs
Scenario
No Drop-out
50% Drop-out
72% Drop-out
(Maximum volume)
Annual Volume
513
1,025
5,200
Cost per Vehicle
$106
$58
$32
Note that while reducing inspection frequency to biennial cuts
motorists' costs in centralized programs, in decentralized programs such cost
reductions are only achieved by reducing opportunities for stations to
participate. In the scenario in which 50 percent of the stations drop out and
testing is biennial, annual station volume is the same as if testing were
annual and no stations dropped out. Hence, the estimated per vehicle cost in
a biennial program with a 50 percent station drop-out rate is the same as
would be derived for an annual program with no stations dropping out.
Reducing inspection frequency to biennial, while maintaining the same number
of stations, has the effect of almost doubling the per vehicle cost since
operating costs are spread over half as many vehicles. Note also that the per
vehicle cost far exceeds the per vehicle cost in centralized programs except
in the scenario where 72 percent of the stations drop-out.
5.3
Costs of Four-Mode, Purge and Pressure Testing
It has been proposed that a series of simpler, loaded mode and other
steady-state tests would provide equivalent emission reductions to the IM240
at a lower cost. The emission reduction potential of this approach is
currently being evaluated at EPA's test lane in Phoenix, Arizona. The
information needed to do a cost analysis can be approximated at this time
based upon the test process.
The test procedure being evaluated is a series of emission tests
referred to as the four-mode test: A 40 second 5015 mode (15 mph at a load
equivalent to ETW / 250), a 40 second 2525 mode (25 mph at load equivalent to
ETW / 300), a 40 second mode at 50 mph and normal road load, and a 40 second
idle mode. EPA anticipates a 30-60 second preconditioning mode would be
H-14
-------
needed to insure proper warm-up and canister purge down. Allowing also for
necessary time to transition between test modes (5-10 seconds), the four-mode
test would require a total of approximately four minutes. As with the IM240-
based test scenario, purge testing is assumed to occur simultaneously with the
tailpipe test and pressure testing would be done separately. It should be
noted, however, that some vehicles may not purge during this test and may
require a short transient retest to activate purge.
5.3.1 Equipment and Expendables
The equipment used for the four-mode test is simpler than for the
IM240 test. The dynamometer may not need inertia weights, and a raw gas
analyzer, like the ones used in the current I/M tests, is upgraded with a NOx
analyzer and an anemometer, to enable mass concentration calculations, for
this test. The equipment for the purge and pressure test are the same as
described previously. The estimated costs are shown in Table 5-11.
Table 5-11
Equipment and Costs for the ASM Test
Pressure System $600
Flow Sensor $500
Dynamometer $20,000
Anemometer $2,000
BAR90 w/NOx Analyzer $16,900
Total $40,000
Expendables for this test are nitrogen gas for the pressure test and
calibration gases for the analyzer. The cost of nitrogen gas is the same as
in the previous analysis on IM240 costs (the pressure test procedure is the
same regardless of the type of tailpipe test used). Current calibration gases
are multi-blends consisting of propane, CO, and C02. A cost of $45 per bottle
is used here. In this analysis, it is assumed that multi-blend gases that
include NO will be available at the same cost. Alternatively, one could
assume that two bottles of calibration gas, one current standard multi-blend
and a bottle of NO will be needed, however, the additional cost per test is
insignificant (less than 5C, even in a low volume situation).
5.3.2 Centralized Programs
The total test time per vehicle would be about 11 minutes, including
administrative processing in an efficiently run testing lane. In a multi-
position lane the throughput would be governed by test time at the longest
position, which would be four minutes. This translates into a peak throughput
H-15
-------
rate of 15 vehicles per hour and, using the standard design criteria for
centralized programs described earlier, an average throughput of 7.5 vehicles
per hour. Using the lane operation assumptions detailed earlier, this
translates into 23,400 vehicles per lane per year and 117,000 vehicles over an
assumed five year contract period. Three staff per lane would be needed to
perform the entire test sequence including inputting vehicle identification
information, conducting the tests and presenting and explaining the results to
the motorist.
The per vehicle cost of the four-mode test in centralized programs is
estimated by the same methodology as was used to estimate IM240 costs.
Current costs for test equipment, staff, state oversight, and construction are
subtracted from the current average per vehicle cost, this amount is factored
by the change in throughput, and estimated costs for equipment, staff,
construction, and state oversight in a four-mode test program are added to
obtain an estimated total cost.
Table 5-12
Coats to Add Proposed Testa to Centralized Programs
Running Total
Increments Per Vehicle Cost Cost per Vehicle
Adjust for Throughput $5.26 * 12.5/7.5 $9.12
Staff $2.40 $11.52
Construction $1.71 $13.23
Oversight $1.75 $14.98
Pressure Test $0.13 $15.11
Purge Test $0.18 $15.29
Four-mode Test $0.35 $15.64
5.3.3 Decentralized Programs
The same methodology used to estimate costs of IM240 testing is used
here. Most assumptions are unchanged. Total test time is thirty minutes,
equipment is amortized over a five year period. Two parameters are changed in
this analysis: equipment costs total $40,000 instead of $144,100, and state
costs include a cost for state mass emission testing.
Table 5-13
Costs to Conduct Four-Mode Testing In Decentralized Programs
Scenario Annual Volume Cost per Vehicle
No Drop-out 513 $51
50% Drop-out 1,025 $31
72% Drop-out 5,200 $25
H-16
-------
Appendix I:
ASM and IM240 Credits for State Implementation
Plans With MOBILES Runs
-------
May 6, 1993
ASM, Purge, Pressure (annual) 0.4/8.0/1.8
MOBILES* (26-Mar-93)
0
-Ml 14 Horning:
+ Purge Check emission benefits assume the use of a dyn<
the IM240 transient test procedure driving cycle.
OI/M program selected:
iter and
Start year (January 1):
Pre-1981 MYR stringency rate:
First model year covered:
Last model year covered:
Waiver rate (pre-1981):
Waiver rate (1981 and newer):
Compliance Rate:
Inspection type:
Inspection frequency
Vehicle types covered:
1981 t later MYR test type:
Outpoints, HC: 0.400 CO
OFunctional Check Program Description:
1983
40%
1968
2020
3.%
3.%
96.%
Test Only
Annual
LDGV - Yes
LDGT1 - Yes
LDGT2 - Yes
KDCV - Yes
Loaded / Idle
8.000 NOz: 1
800
OCheck Start Model Yrs Vehicle Classes Covered Inspection Comp
(Janl) Covered LDGV LDGT1 LDCT2 HDCV Type Freq Rate
Purge 1983 1971-2020 Yes Yes Yes Yes Test Only Annual 96.0%
ATP 1983 1971-2020 Yes Yes Yes Yes Test Only Annual 96.0%
OAir pump system disablements : Yes Catalyst removals : Yes
Fuel inlet restrictor disablements : Yes Tailpipe lead deposit test : Yes
ECR disablement: Yes Evaporative system disablements: Yes
PCV system disablements: Yes Missing gas caps: Yes
0 Minimum Temp: 72. (F» Maximum
Period 1 RVP : 10.5 Period 2 RVP : 8.7
OVOC HC emission factors include evaporative HC emission factors.
0
Tamo
Period 2 Start Yr
: 92.
: 1992
OCal. Year: 2000 I/M Program: Yes Ambient Temp: 87.5 / 87.5 / 87.5 (F) Region: Low
Anti-tarn. Program: Yes Operating Mode: 20.6 / 27.3 / 20.6 Altitude: 500
Reformulated Gas: No
0 Veh. Type: LDGV LDCT1 LDGT2 LDCT UDCV LDDV LDDT HDDV
•f
Veh. Speeds:
VtCT Mix:
19. 6
0
OComposite Emission
VOC HC:
Exhaust HC :
Bvaporat HC :
Refuel L HC:
Riming L HC:
Rsting L HC:
Exhaust CO :
Exhaust HOX:
1
1
0
0
0
0
13
1
.616
Factors
.73
.04
.15
.19
.29
.06
.31
.09
19. £
0.191
(Cm/Mile)
1.83
1.17
0.19
0.25
0.16
0.06
15.05
1.25
T9TS T5TS I9TS
0
2.
1
0.
0.
0.
0.
19.
1,
.086
.45
.66
.25
.25
.24
.06
.43
.68
2
1
0
0
0
0
16
1
.02
.32
.21
.25
.18
.06
.41
.38
0
6
2
1
0
0
0
56
4
.031
.02
.81
.95
.41
.74
.10
.19
.67
0
0.
O.
1.
1.
.002
.65
.65
.60
37
IfTS
0
0
0,
1.
1.
.001
.87
.87
.76
.51
19. 6
0
2
2
11
10
.068
.23
.23
.58
.69
(F)
. Ft.
MC
All Veh
T5T3
0.
5.
1.
3.
0.
24.
0.
.006
53
89
21
43
78
77
2.000
1.260
0.233
0.196
0.249
0.062
15.424
1.932
1-1
-------
May 6, 1993
OCal.
OEmiaaioa factora are aa of Jan. lat of the indicated calendar year.
~- ' Year: 2001 I/M Program: Yea Ambient Temp: 87.5 / 87.5 / 87.5 (F) Region:
Anti-tarn. Program: Yea Operating Mode: 20.6 / 27.3 / 20.6
Reformulated Gaa: Ho
LDCV LDGT1 LDGT2 LOOT HDCV LDDV LDDT
Low
Altitude: 500. Ft.
0 Veh. Type:
HDDV
MC
All Vah
Veh. Speeda:
VMT Mix:
19. <5
0.613
OCompoaite Emiaaion Factora
VOC HC:
Exhauat HC :
Evaporat HC :
Rafael L HC:
Runing L HC:
Rating L HC:
Exhauat CO :
Exhauat NOX:
OEmiaaion factora
OCal. Year: 2003
1.68
1.03
0.14
0.19
0.26
0.06
13.25
1.07
are aa of
I/M
Anti-tarn.
19.6
0.192
(Cm/Mile]
1.78
1.15
0.17
0.25
0.15
0.06
14.82
1.21
Jan . lat
Program :
Program:
Reformulated Gaa:
0 Vah. Type:
Veh. Speeda:
VMT Mix:
LDGV
19. i
0.606
OCompoaite Emiaaion Factor*
VOC HC:
Exhauat HC :
Evaporat HC :
Refuel L HC:
Runing L HC :
Rating L HC:
Exhauat CO :
Exhauat NOX:
OCal. Year: 2006
1.58
1.01
0.12
0.19
0.22
0.05
13.17
1.03
I/M
Anti-tarn .
LDGT1
19.6
0.194
(Cm/Mile)
1.69
1.13
0.14
0.25
0.13
0.05
14.52
1.16
Program:
Program:
Reformulated Caa:
0 Veh. Type:
Veh. Speed*:
VMT Mix:
LDGV
I9.fi
0.599
OCompoaite Emiaaion Factora
VOC HC:
Exhauat HC:
Evaporat HC :
Refuel L HC:
Runing L HC:
Rating L HC:
Exhauat CO :
Exhauat NOX:
1.49
O.98
0.10
0.19
0.18
0.04
13.08
0.99
LDGT1
19. 6
0.197
(Cm/Mile)
1.61
1.09
0.11
0.25
0.12
0.04
14.28
1.13
14. t!
0.086
1
2.38
1.64
0.22
0.25
0.22
0.06
19.27
1.64
of the
Yea
Yea
No
LDGT2
19. £
0.087
2.19
1.59
0.14
0.25
0.16
O.O5
19.01
1.57
of the
Yea
Yea
No
LDCT2
T9TS~
0.087
2.08
1.53
0.11
0.25
0.15
0.04
18.62
1.53
1.96
1.30
0.19
0.25
0.17
0.06
16.20
1.34
indicated
Ambient
Operating
LOOT
1.85
1.27
0.14
0.25
0.14
0.05
15.90
1.29
19.6
0.031
5.45
2.54
1.74
0.41
0.66
0.10
48.47
4.52
calendar year
Temp: 87.5 /
Mode: 20 . 6 /
HDCV
19. 6
0.031
4.55
2.13
1.41
0.41
0.52
0.08
36.06
4.36
indicated calendar year.
Ambient Temp: 87.5 /
Operating
LOOT
1.75
1.22
0.11
0.25
0.13
0.04
15.61
1.25
Mode: 20.6 /
HDGV
19. t
0.031
3.91
1.90
1.13
0.40
0.41
0.07
28.81
4.11
19. 6
0.001
0.61
0.61
1.55
1.30
87.5 /
27.3 /
LDDV
19.4
0.002
0.56
0.56
1.48
1.18
87.5 /
27.3 /
LDDV
19.6
0.002
0.51
0.51
157? T9~75
0.001 0.069
0.80
0.80
1.69
1.41
87.5 (F)
2.20
2.20
11.49
9.94
Ragion : Low
20.6 Altitude: 500.
LDDT
T57S —
0.002
0.74
0.74
1.63
1.31
HDDV
19.4
0.072
2.14
2.14
11.36
8.67
0.006
5.53
1.89
3.21
0.43
24.78
0.77
Ft.
MC
19. i
0.006
5.53
1.89
3.21
0.43
24.78
0.77
1
1
0
0
0
0
15
1
AH
1
1
0
0
0
0
14
1
.932
.238
.214
.195
.228
.057
.084
.865
Vah
.811
.203
.176
.195
.189
.048
.553
.757
87.5 (F) Region: Low
20.6 Altitude: 500.
LDDT
T9T? —
0.002
0.71
0.71
HDDV
19.6
0.076
2.11
2.11
Ft.
MC
19.4
0.006
5.53
1.89
3.21
All
1.
Vah
.705
1.169
0.148
0.194
0.156
1.42
1.10
1.59
1.25
11.24
7.74
0.43
24.78
0.77
0.038
14.173
1.673
X-2
-------
May 6, 1993
0Emission factora are aa of Jan. 1st of the indicated calendar year.
OCal. Year: 2008 I/M Program: Yea Ambient Temp: 87.5 / 87.5 / 87.5 (F) Region: Low
Anti-tarn. Program: Yea Operating Mod*: 20.6 / 27.3 / 20.6 Altitude: 500. Ft.
Reformulated Caa: No
0 Vah. Type: LDGV LDCT1 LDCT2 LOOT HDGV LDDV LDDT RDDV MC
All Veh
Veh . Speeds :
VMT Mix:
19.6
0.594
OConpoaite Emiaaion Factora
VOC HC:
Exhauat HC :
Evaporat HC :
Refuel L HC:
Runing L HC:
Rating L HC:
Exhauat CO :
Exhauat NOX:
OEmiaaion factora
OCal. Year: 2O11
1.44
0.97
0.10
0.19
0.16
0.03
13.08
0.97
are aa of
I/M
Anti-tarn .
19.6
0.199
(On/Mile)
1.58
1.08
0.11
0.25
0.11
0.04
14.28
1.09
Jan . lat
Program:
Program:
Reformulated Gaa :
0 Veh. Type:
Veh. Speeds:
VMT Mix:
LDGV
T3T3
0.588
OComipoaite Emiaaion Factora
VOC HC:
Exhauat HC :
Evaporat HC:
Refuel L HC:
Runing L HC:
Rating 1 HC:
Exhauat CO :
Exhaust NOX:
1.39
0.96
0.09
0.19
0.14
0.02
13.07
0.96
LDGT1
T5T3
0.201
(On/Mile)
1.54
1.07
0.09
0.25
0.10
0.03
14.31
1.08
19. 6
0.088
2.05
1.52
0.11
0.25
0.14
0.04
18.73
1.49
1.73
1.22
0.11
0.25
0.12
0.04
15.64
1.22
o£ the indicated
Yaa Ambient
Yea
No
LDGT2
19.3
0.088
2.00
1.52
0.09
0.25
0.12
0.03
18.88
1.48
Operating
LDCT
1.68
1.21
0.09
0.25
0.11
0.03
15.70
1.20
19.3
0.032
3.67
1.86
0.99
0.40
0.35
0.06
26.62
3.95
calendar year
Temp: 87 . 5 /
Mode: 20 . 6 /
HDGV
T37S
0.032
2.86
1.58
0.64
0.40
0.20
0.04
18.32
3.83
1973—
0.002
0.50
0.50
1.41
1.08
'87.5 /
27.3 /
LDDV
1573
0.002
0.51
0.51
1.42
1.07
19.3
0.003
0.69
0.69
1.57
1.22
TTTS—
0
2
2
11
7
87.5 (F) Region
20.6 Altitude
LDDT
19.6
0.003
0.69
0.69
1.58
1.22
.078
.11
.11
.21
.29
: Low
: 500.
HDDV
19"
0.
2.
2.
11.
6.
.080
.10
,10
18
86
19.6—
0.005
5.53
1.89
3.21
0.43
24.78
0.77
Ft.
MC
1973 —
0.005
5.53
1.89
3.21
0.43
24.78
0.77
1.663
1.161
0.136
0.194
0.139
0.033
14.108
1.629
All Veh
1.597
1.147
0.114
0.194
0.117
0.025
13.850
1.598
- 3
-------
May 6, 1993
1 Maximum ASM, Purge, Pressure (biennial) 0.4/8.0/1.8
MOBILES. (26-Mar-93)
0
-Mil4 Naming:
+ Purge Check emission benefits assume the use of a dynamometer and
the IM240 transient test procedure driving cycle.
OI/M program selected:
Start year (January 1}:
Pre-1991 MYR stringency rate:
First model year covered:
Last model year covered:
Maiver rate (pre-1981):
Waiver rate (1981 and newer):
Compliance Rate:
Inspection type:
Inspection frequency
Vehicle type* covered:
1983
40%
1968
2020
3.%
3.%
96.%
Test Only
Biennial
LDGV - Yes
LDCT1 - Yes
LDGT2 - Tes
HDCV - Yes
Loaded / Idle
8.000 HOx: 1.800
1981 I later MYR test type:
Outpoints, HC: 0.400 CO:
OFunctional Check. Program Description:
OCheck Start Model Yrs Vehicle Classes Covered
(Janl) Covered LDCV LDGT1 LDCT2 HDCV
Inspection
Type Freq
Press 1983 1971-202O Yes Yes
Purge 1983 1971-202O Yes Yes
ATP 1983 1971-2020 Yes Yes
OAir pump system disablements :
Fuel inlet restrictor disablements:
EGR disablement;
PCV system disablements :
0
Period 1 RVP:
OVOC HC emission factors include evaporative
0
Yes Yes Test Only Biennial
Yes Yes Test Only Biennial
Yes Yes Test Only Biennial
Yes Catalyst removals:
Yes Tailpipe lead deposit test:
Yes Evaporative system disablements :
Yes Missing gas caps:
Minimum Temp: 72. (F)
10.5 Period 2 RVP: 8.7 P<
HC emission factors.
Camp
Rate
96.0%
96.0%
96.0%
Yes
Yes
Yes
Yes
ciod 2
Temp:
Start Yr:
92.
1992
OEmission factors are as of Jan. 1st nf ^h* indimt^d r^l^n^r y^f
OCal. Year: 2000 I/M Program: Yes Ambient Temp: 87.5
Anti-tarn. Program: Yes Operating Mode: 20.6
Reformulated Gas: No
0 Veh. Type: LDGV LDCT1 LDGT2 LOOT HDCV
/ 87.5 / 87.5 (F) Region: Low
/ 27.3 / 20.6 Altitude: 500.
Veh. Speeds:
VMX Mix:
19. 6
0.616
OConposite Emission Factors
VOC HC: 1.80
Exhaust HCi
Evaporat HC:
Refuel L HC:
Runing L HC :
Rsting L HC:
Exhaust CO :
Exhaust HOX:
1.08
0.16
0.19
0.30
0.06
13.94
1.12
T9~S~
0.191
(Cm/Mile)
1.91
1.23
0.20
0.25
0.18
0.06
15.90
1.27
19. li
0.086
2.58
1.75
0.25
0.25
0.26
0.06
20.63
1.72
l9.g
0.031
2.12
1.39
0.21
0.25
0.20
0.06
17.37
1.41
6.23
2.99
1.97
0.41
0.76
0.10
59.30
4.71
LDDV
19..~
0.002
0.65
0.65
1.60
1.37
LDDT
19. 6
0.001
0.87
0.87
1.76
1.51
HDDV
19. 6
0.068
2.23
2.23
11.58
10.69
Ft.
MC
0.006
5.53
1.89
3.21
0.43
24.78
0.77
All Veh
2.073
1.308
0.241
0.196
0.266
0.062
16.172
1.958
X-4
-------
May 6, 1993
OEmission factors axe as of Jan. 1st of the indicated calendar year.
OCal. Year: 2001 I/M Program: Yea Ambient Temp: 87.5 / 87.5 / 87.5 (F) Region: Low
Anti-tarn. Program: Yes Operating Mode: 20.6 / 27.3 / 20.6 Altitude: 500. Ft.
Reformulated Gas: No
0 Veh. Type: LDGV LDGT1 LDGT2 LDGT HDGV LDDV LDDT HDDV MC
All Veh
•Veh. Speeds:
VMT Mix:
19.4
0.613
OCoaposite Emission Factors
VOC HC: 1.74
Exhaust HC:
Evaporat HC :
Refuel L HC:
Ruaing t HC:
Rsting.L HC:
Exhaust CO :
Exhaust NOX:
OEmission factors
OCal. Year: 2003
0 Veh . Type :
Veh. Speeds:
VMT Mix:
1.07
0.15
0.19
0.28
0.06
13.85
1.09
19.4
0.192
19.4
0.086
(Cm/Mile)
1.86 2.49
1.20
0.18
0.25
0.17
0.06
15.62
1.24
are aa of Jan. 1st
I/M Program:
Anti-tarn. Program:
Reformulated Gas:
LDGV
19.4
0.606
OCoaposite Emission Factors
VOC HC: 1.64
Exhaust HC:
Evaporat HC :
Rafael L HC:
Riming L HC:
Rating L HC:
Exhaust CO:
Exhaust NOX:
OEmission factors
OCal. Year: 2006
0 Veh. Type:
Veh . Speeds :
VMT Mix:
1.04
0.13
0.19
0.24
0.05
13.75
1.06
LDOT1
19.4
0.194
(Cm/Mile)
1.76
1.17
0.15
0.25
0.15
0.05
15.22
1.19
are as of Jan. 1st
I/M Program:
Anti-tarn . Program :
Reformulated Gas:
LDGV
19.4
0.599
OCoaposite Emission Factors
VOC HC: 1.54
Exhaust HC :
Evaporat HC:
Refuel L HC:
Riming L HCi
Rating L HCi
Exhaust CO:
Exhaust NOX:
1.01
0.11
0.19
0.19
0.04
13.63
1.02
LDOT1
19.4
O.197
(On/Mile)
1.67
1.13
0.12
0.25
0.13
0.04
14.92
1.15
1.72
0.23
0.25
0.24
0.06
20.38
1.68
of the
Yes
Yes
No
LDCT2
19.4
0.087
2.28
1.65
0.15
0.25
0.19
0.05
19.94
1.60
2.05
1.36
0.19
0.25
0.19
0.06
17.09
1.37
indicated <
Ambient
Oparating
LDGT
1.92
1.32
0.15
0.25
0.16
0.05
16.68
1.32
19.4
0.031
5.65
2.70
1.76
0.41
0.68
0.10
51.17
4.56
calendar year
Temp: 87.5 /
Mode: 20.6 /
HDGV
fa. 4
0.031
4.73
2.26
1.43
0.41
0.54
0.08
38.08
4.40
of the indicated calendar year
Yes Ambient Temp: 87.5 /
Yes Operating Mode: 20.6 /
No
LDCT2
19.4
O.087
2.16
1.57
0.12
0.25
0.17
0.04
19.42
1.57
LDGT
1.82
1.26
0.12
0.25
0.14
0.04
16.29
1.28
HDCV
19.4
0.031
4.08
2.02
1.16
0.40
0.43
0.07
30.45
4.15
0.001
0.61
0.61
1.55
1.30
87.5 /
27.3 /
LDDV
19.4
0.002
0.56
0.56
1.48
1.18
87.5 /
27.3 /
LDDV
19.4
0.002
0.51
0.51
1.42
1.10
T5TS 1573
0.001 0.069
0.80
0.80
1.69
1.41
2.20
2.20
11.49
9.94
87.5 (F) Region: Low
20.6 Altitude: 500.
LDDT HDDV
iTTS 157?
0.002 0.072
0.74
0.74
1.63
1.31
2.14
2.14
11.36
8.67
87.5 (F) Region: Low
20.6 Altitude: 500.
LDDT
19.4
0.002
0.71
0.71
1.59
1.25
HDDV
19.4
0.076
2.11
2.11
11.24
7.74
0.006
5.53
1.89
3.21
0.43
24.78
0.77
Ft.
MC
19.4
0.006
5.53
1.89
3.21
0.43
24.78
0.77
Ft.
MC
19.4
0.006
5.53
1.89
3.21
0.43
24.78
0.77
2
1
0
0
0
0
15
1
All
1.
1.
0,
0,
0.
0.
15.
1.
All
1.
1.
0.
0.
0.
0.
14.
1.
.001
.283
.221
.195
.245
.057
.786
.891
Veh
.874
.243
.183
,195
,206
.048
.189
783
Veh
761
203
155
194
171
038
749
698
X-5
-------
May 6, 1993
0 Emission factors
OCal. Year: 2O08
are as of
I/M
Anti-tarn.
Jan . 1st
Program :
Program:
Reformulated Can:
0 Veh. Type:
•f
Veh . Speeds :
VMI Mix:
LDCV
T77S
0.594
OComposita Emission Factors
VOC HC:
Exhaust HC :
Evaporat HC :
Refuel L HC:
Riming L HC:
Rating L BC:
Exhaust CO:
Exhaust NOX:
OEmission factors
OCal. Year: 2011
1.49
1.00
0.10
0.19
0.17
0.03
13.63
1.00
ace as of
I/M
Anti-tarn.
LDGT1
T97S
0.199
(Cm/Mile)
1.64
1.12
0.11
0.25
0.13
0.04
14.88
1.12
Jan . 1st
Program :
Program:
Reformulated Gas:
0 Veh . Type :
Veh. Speeds:
VMI Mix:
LDCV
19.6*
0.588
OComposite Emission Factors
VOC HC:
Exhaust HC :
Evaporat HC :
Refuel L HC:
Runing L HC :
Rating L HC:
Exhaust CO :
Exhaust NOX:
1.44
0.99
0.09
0.19
0.15
0.02
13.62
0.99
LDGT1
T373 —
0.201
(On/Mile)
1.60
1.11
0.10
0.25
0.12
0.03
14.92
1.11
of the indicated calendar year.
Yes
Yes
No
LDGT2
19.6
0.088
2.12
1.56
0.11
0.25
0.16
0.04
19.47
1.53
Ambient
Operating
LDCT
1.79
1.25
0.11
0.25
0.14
0.04
16.29
1.25
of the indicated <
Yes Ambient
Yes
No
LDCT2
19. d
0.088
2.07
1.56
0.10
0.25
0.14
0.03
19.62
1.52
Operating
LDCT
1.74
1.24
0.10
0.25
0.12
0.03
16.35
1.23
Temp: 87 . 5 /
Mode: 20.6 /
HDGV
19. 6
0.032
3.84
1.98
1.02
0.40
0.37
0.06
28.14
3.99
:alendar year .
Temp: 87.5 /
Mode: 20.6 /
HDCV
ITTS —
0.032
3.02
1.68
0.66
0.40
0.23
0.04
19.39
3.87
87.5 / 87
27.3 / 20
LDDV
19.6
0.002
0.50
0.50
1.41
1.08
87.5 / 87.
27.3 / 20.
LDDV
19.6
0.002
0.51
0.51
1.42
1.07
.5 (F) Region
.6 Altitude
: Low
: 500.
LDDT HDDV
T97? T9~76
0.003 0
0.69 2
0.69 2
1.57 11.
1.22 7
.5 (F) Region:
.6 Altitude:
.078
.11
.11
.21
.29
: Low
: 500.
LDDT HDDV
I5~3 — IT
0.003 0.
0.69 2.
0.69 2.
1.58 11.
1.22 6.
rs^
.080
.10
10
18
86
Ft.
MC
19.6
0.005
5.53
1.89
3.21
0.43
24.78
0.77
Ft.
MC
19.6
0.005
5.53
1.89
3.21
0.43
24.78
0.77
All
1
1
0
0
0
0
14
1
All
1.
1.
0.
0.
0.
0.
14.
1.
Veh
.717
.194
.143
.194
.153
.033
.669
.655
Veh
.650
.179
.122
194
131
025
395
623
T-6
-------
April 22, 1993
1 Enhanced Performance Standard
MOBILES* (26-M&T-93)
OI/M program fl selected:
OStart year (Jan 1): 1983
Pre-1981 stringency: 20%
First MYR covered: 1968
Last ten. covered: 2020
Raiver (pre-1981): 3.%
Waiver (1981+): 3.%
Compliance Rate: 96.%
Inspection type:
Teat Only
Inspection frequency: Annual
I/M program tl vehicle type*
LDGV - Yes
LDCT1 - Ye.
LDCT2 - Tea
HDGV - No
1981 S later MYR teat type:
2500 rpm / Idle
Cutpointa, BC: 220.000
Cutpointa, CO: 1.200
Cutpointa, NOx: 999.OOO
I/M program #2 selected:
Start year (Jan 1): 1983
Pre-1981 stringency: 20%
First MYR covered: 1986
Laat MYR covered: 2020
Naiver (pre-1981): 3.%
Haiver (1981+): 3.%
Compliance Rate: 96.%
Inspection type:
Test Only
Inspection frequency: Annual
I/M program t2 vehicle types
LDGV - Yes
LDCT1 - Yes
LDCT2 - Yea
HDGV - Yea
1981 t later MYR test type:
IM240 test
Outpoints, HC: 0.800
Cutpoints, CO: 20.000
Outpoints, HOz: 2.000
OFunctional Check Program Deacription:
OCheck Start Model Yra Vehicle Classes Covered Inspection
(Janl) Covered LDGV LDCT1 LDCT2 HDGV Type Freq
Preaa 1983 1983-2020 Yes
Purge 1983 1986-2020 Yes
ATP 1983 1984-2020 Yes
OJLLr pump system dXsablamenta:
Fuel inlet reatrictor disablementa: Yea
ECR disablement: No
PCV system disablementa: Ho
Yea Yea No Teat Only Annual
Yea Yes No Test Only Annual
Yea Yea No Teat Only Annual
No Catalyst removals:
Tailpipe lead deposit test:
Evaporative system disablements:
Missing gaa caps:
0 Minimum Temp: 72. (F)
Period 1 RVP: 10.5 Period 2 RVP: 8.7 S
OVOC HC emission factora include evaporative HC emission factors.
0
Camp
Rate
96.0%
96.0%
96.0%
Yea
No
No
No
Maximum Tamp: 92. (F)
ariod 2 Start Yr: 1992
OEmisaion factora axe as of Jan.1st of the indicated calendar year.
OCal. Year: 2000 I/M Program: Yes Ambient Temp: 87.5 / 87.5 / 87.5 (F) Region: Low
Anti-tarn. Program: Yea
Reformulated Gas: No
LDGV LDOT1 LDGT2
Operating Mode: 20.6 / 27.3 / 20.6 Altitude: 500. Ft.
0 Veh. Type:
Veh. Speeds:
VMT Mix: 0.616 0.191
OCooposite Emission Factors (Cm/Mile)
LOOT
T5T3 T5TTS
VOC HC:
Exhaust HC:
Evaporat HC:
Refuel L HC:
Baning L HC:
Rating L HC:
Exhaust CO:
Exhaust NOX:
1.63
0.88
0.18
0.19
0.32
0.06
11.89
1.14
.82
1.06
0.22
0.25
0.22
0.06
14.49
1.27
0.086
2.37
1.48
0.27
0.25
0.31
0.06
18.72
1.71
.99
.19
.24
0.25
0.25
0.06
15.81
1.41
HDGV
19. 6
0.031
6.94
3.05
2.32
0.41
1.05
0.10
63.66
4.71
LDDV
19. i
0.002
0.65
0.65
1.60
1.37
LDDT
19. 6"
0.001
0.87
0.87
1.76
1.51
HDDV
0.068
2.23
2.23
11.58
10.69
0.43
24.78
0.77
All Veh
1.955
1.132
0.267
0.196
0.298
0.062
14.612
1.968
1-7
-------
April 22, 1993
OEmission factors are as of Jan. 1st of the indicated calendar year.
OCal. Year: 2001 I/M Program: Yea Ambient Temp: 87.5 / 87.5 / 87.5 (F) Region: Low
Anti-tarn. Program: Yea Operating Mode: 20.6
Reformulated Oaa : No
0 Voh. Type: LDGV LDGT1 LDGT2 LDCT HDCV
Veh. Speeds: TITS I57S T37S T57?
VMT Mix: 0.613 0.192 0.086 0.031
OCoopoaite Emission Factors (Cm/Mile)
VOC HC: 1.54 1.71 2.23 1.87 6.38
Exhauat HC: 0.85 1.01 1.41 1.13 2.74
Evaporat HC: 0.16 0.20 0.24 0.21 2.14
Refuel L HC: 0.19 0.25 0.25 0.25 0.41
Runing L HC: 0.29 0.20 0.28 0.22 1.00
Rating L HC: 0.06 0.06 0.06 0.06 0.10
Exhaust CO: 11.59 13.90 18.06 15.19 54.65
Exhaust NOX: 1.10 1.22 1.65 1.35 4.56
OEmission factors are as of Jan. 1st of the indicated calendar yej
OCal. Year: 2003 I/M Program: Yes Ambient Temp: 87.5
Anti-tarn. Program: Yes Operating Mode: 20.6
Reformulated Gas : No
0 Veh. Type: LDGV LDCT1 LDCT 2 LDCT HDCV
•f
Veh. Speeds: T57S I57S T57S T3T8
VMT Mix: 0.606 0.194 0.087 0.031
OCocnpoaite Emission Factors (Cm/Mile)
VOC HC: 1.40 1.55 1.96 1.68 5.52
Exhauat HC: 0.80 0.93 1.28 1.04 2.26
Evaporat HC: 0.13 0.16 0.16 0.16 1.86
Refuel L HC: 0.19 0.25 0.25 0.25 0.41
Runing L HC: 0.24 0.16 0.22 0.18 0.91
Rating L BC: 0.05 0.05 0.05 0.05 0.08
Exhaust CO: 11.17 13.08 16.91 14.26 40.04
Exhaust NOX: 1.04 1.17 1.57 1.29 4.39
OEraiaaion f actora are as of Jan . 1st of the indicated calendar ye4
OCal. Year: 2006 I/M Program: Yes Ambient Temp: 87.5
Anti-tarn. Program: Yea Operating Mode: 20.6
Reformulated Gas: No
0 Veh. Type: LDCV LDGT1 LDCT 2 LDCT HDCV
Veh. Speeds: T373 T5TS 1375 1378
VMT Mix: 0.599 0.197 0.087 0.031
OConposite Emission Factors (Cm/Mile)
VOC BC: 1.26 1.40 1.73 1.50 4.93
Exhaust BC: 0.74 0.85 1.14 0.94 1.99
Evaporat BC: 0.11 0.12 0.13 0.12 1.63
Refuel L BC: 0.19 0.25 0.25 0.25 0.40
Runing L HC: 0.19 0.13 0.18 0.15 0.84
Rating L HC: O.O4 0.04 0.04 0.04 0.07
Exhauat CO: 1O.83 12.40 15.76 13.43 31.53
Exhaust NOX: O.99 1.11 1.51 1.23 4.14
/ 27.3 /
LDDV
19.6
0.001
0.61
0.61
1.55
1.30
ir .
/ 87.5 /
/ 27.3 /
LDDV
T97S
0.002
0.56
0.56
1.48
1.18
ir .
/ 87.5 /
/ 27.3 /
LDDV
19. <>
0.002
0.51
0.51
1.42
1.10
20.6 Altitude: 500.
LDDT
I57S —
0.001
0.80
0.80
1.69
1.41
HDDV
19.6
0.069
2.20
2.20
11.49
9.94
Ft.
MC
19,6
0.006
5.53
1.89
3.21
0.43
24.78
0.77
All Veh
1.852
1.086
0.245
0.195
0.270
0.057
13.978
1.890
87.5 (F) Region: Low
20.6 Altitude: 500.
LDDT
19. 6
0.002
0.74
0.74
1.63
1.31
HDDV
19.6
0.072
2.14
2.14
11.36
8.67
Ft.
MC
19.6
0.006
5.53
1.89
3.21
0.43
24.78
0.77
All Veh
1 . 682
1.014
0.203
0.195
0.223
0.048
13.005
1.767
87.5 (F) Region: Low
20.6 Altitude: 500.
LDDT
19. 6
0.002
0.71
0.71
1.59
1.25
HDDV
19. tf
0.076
2.11
2.11
11.24
7.74
Ft.
MC
19.6
0.006
5.53
1.89
3.21
0.43
24.78
0.77
All Veh
1.531
0.949
0.171
0.194
0.179
0.038
12.293
1.668
1-8
-------
April 22, 1993
OEmisaion factors or* as of Jan. 1st of the indicated calandax year.
OCal. Year: 2008 I/M Program
Anti— tarn. Program
Reformulated Gas
0 Veh. Type: LDOV LDCT1
Veh. Speeds:
VMT Mix:
19.6
0.594
OComposite Emission
VOC HC:
Exhaust HC:
Evaporat HC :
Refuel L RC:
Runing L HC :
Rating L HC:
Exhaust CO:
Exhaust NOX:
1
0
0
0
0
0
10
0
19
0
.6
.199
: Yes Ambient Temp: 87.5
: Yes Operating Mode: 20.6
: No
LDCT2 LOOT HDGV
T5TS
0.088
19.6
0.032
/ 87.5 / 87.5 (F) Region: Low
/ 27.3 / 20.6 Altitude: 500. Ft.
LDDV LDDT HDDV MC
TTT3
0.002
19.6
0.003
19
0
.078
All Veh
19.6
0.005
Factors (On/Mile)
.20
.73
.10
.19
.16
.03
.72
.97
1
0
0
0
0
0.
12
1
.34
.83
.11
.25
.12
.04
.30
.08
1
1
0
0.
0
0,
15.
1.
.65
.11
.11
.25
.15
.04
.68
.48
1.
0.
0.
0.
0.
0.
13.
1.
44
91
11
25
13
04
33
20
4
1
1
0
0
0
28
3
.71
.94
.51
.40
.79
.06
.91
.98
0.
0.
1.
1.
.50
.50
.41
.08
0.69
0.69
1.57
1.22
2
2
11
7
.11
.11
.21
.29
5.
1.
3.
0.
24.
0.
.53
.89
.21
.43
.78
,77
1.472
0.931
0.157
0.194
0.157
0.033
12.122
1.621
0 Emission factors are as of Jan. 1st of the indicated calendar year.
OCal. Year: 2011 I/M Program: Yea Ambient Temp: 87.5 / 87.5 / 87.5 (F) Region: Low
Anti-tarn. Program: Yes Operating Mode: 20.6 / 27.3 / 20.6 Altitude: 500. Ft.
Reformulated Gas: No
0 Veh. Type: LDGV LDGT1 LDGT2 LOOT KDCV LDDV LDDT HDDV MC
All Veh
Veh . Speeds :
VMT Mix:
ITT"
0.588
OComposite Emission Factors
VOC HC:
Exhaust HC:
Evaporat HC:
Refuel L HC:
Runiag L HC :
Rating L HC:
Exhaust CO:
Exhaust NOX:
1.14
0.71
0.09
0.19
0.14
0.02
10.62
0.94
19. 6
0.201
(On/Nile)
1.27
0.80
0.09
0.25
O.10
0.03
12.15
1.06
ia.t
0.088
1.54
1.05
0.09
0.25
0.12
0.03
15.43
1.43
TS7S
1.35
0.88
0.09
0.25
0.11
0.03
13.15
1.17
0
3
1
1
0
0
0
19
3
.032
.95
.61
.21
.40
.69
.04
.07
.85
19. 61
0.002
0.51
0.51
1.42
1.07
19.6
0.003
0.69
0.69
1.58
1.22
19.6
0.080
2.10
2.10
11.18
6.86
19. i
0.005
5.53
1.89
3.21
0.43
24.78
0.77
1.389
0.902
0.135
0.194
0.132
0.025
11.694
1.576
I- 9
-------
April 22, 1993
1 NO I/M (DEFAULT LAP)
MOBILESa (26-Mar-93)
Period 1 RVP: 10.5 Period 2 RVP: 8.7 Period 2 Start Yr: 1992
OVOC BC emission factora include evaporative HC emission factors.
OEmiaaion factors are aa of Jan. lat of the indicated calendar year
OCal. Year: 2000 I/M Program: No Ambient Temp: 87.5 /
Anti-tarn. Program: No Operating Mode: 20.6 /
Reformulated Gas: Ho
0 Veh. Type: LDGV IJX2T1 LDCT 2 LDCT HDGV
Veh. Speeda: 19T5 TST? 19T3 TJ7S
VMT Mix: 0.616 0.191 0.086 0.031
((Composite Emission Factora (Cm/Mile)
VOC HC: 2.56 2.93 3.90 3.23 7.14
Exhaust HC: 1.46 1.80 2.53 2.03 3.24
Evaporat BC: 0.28 0.34 0.40 0.36 2.32
Refuel L BC: 0.19 0.25 0.25 0.25 0.41
Runing L BC: 0.57 0.48 0.66 0.54 1.05
Rating L BC: 0.06 0.06 0.06 0.06 0.10
Exhaust CO: 20.09 23.94 32.84 26.70 66.26
Exhaust NOX: . 1.43 1.67 2.27 1.86 4.32
OCmisaion factors are as of Jan. lat of the indicated calendar year
OCal.. Year: 2001 I/M Program: No Ambient Temp: 87.5 /
Anti-tan. Program: Ho Operating Mode: 20.6 /
Reformulated Gas: No
0 Veh. Type: LDGV LDCT1 LDCT2 LDCT HDGV
Veh. Speeda: T97S T37S T9~7S T37S
VMT Mix: 0.613 0.192 0.086 0.031
OCompoaite Emission Factors (On/Mile)
VOC BC: 2.49 2.85 3.82 3.15 6.60
Exhaust BC: 1.44 1.75 2.50 1.98 2.95
Evaporat BC: 0.27 0.32 0.37 0.34 2.15
Refuel L BC: 0.19 0.25 0.25 0.25 0.41
Runing L BC: O.S4 0.47 0.65 0.52 1.00
Rating L BC: 0.06 0.06 0.06 0.06 0.10
Exhaust CO: 19.92 23.43 32.36 26.20 57.38
Exhaust NOX: 1.40 1.64 2.23 1.82 4.68
0 Emission factors are as of Jan . 1st of the indicated calendar year .
OCal. Year: 2003 I/M Program: Ho Ambient Temp: 87.5 /
Anti-tan. Program: Ho Operating Mode: 20.6 /
Reformulated Gas : No
0 Veh. Type: LDCV LDOT1 LDCT2 LDCT HDGV
Veh. Speeds: T57? T57? T57S T57S
VMT Mix: 0.606 0.194 0.087 0.031
OComposite Emission Factors (Gm/Mile)
VOC BC: 2.38 2.72 3.65 3.OO 5.76
Exhaust BC: 1.41 1.69 2.43 1.92 2.50
Evaporat BC: 0.24 0.29 0.30 0.29 1.87
Refuel L BC: 0.19 O.2S 0.25 0.25 0.41
Runing L BC: 0.49 0.45 0.61 0.50 0.91
Rating L BC: 0.05 0.05 O.OS 0.05 0.08
Exhaust CO: 19.73 22.72 31.27 25.35 43.09
Exhaust HOX: 1.36 1.60 2.18 1.78 4.54
87.5 /
27.3 /
LDDV
19. £
0.002
0.65
0.65
1.60
1.37
87.5 /
27.3 /
LDDV
19.6*
0.001
0.61
0.61
1.55
1.30
87.5 /
27.3 /
LDDV
19. £
0.002
0.56
0.56
1.48
1.18
87.5 (F) Region: Low
20.6 Altitude: 500.
LDDT HDDV
T57S T5TS
0.001 0.068
0.87 2.23
0.87 2.23
1.76 11.58
1.51 10.69
87.5 (F) Region: Low
20.6 Altitude: 500.
LDDT HDDV
T57S 1573
0.001 0.069
0.80 2.20
0.80 2.20
1.69 11.49
1.41 9.94
87.5 (F) Region: Low
20.6 Altitude: 500.
LDDT HDDV
T57S T575
0.002 0.072
0.74 2.14
0.74 2.14
1.63 11.36
1.31 8.67
l« 1
Ft.
MC
1576
0.006
5.53
1.89
3.21
0.43
24.78
0.77
Ft.
MC
19. <>
0.006
5.53
1.89
3.21
0.43
24.78
0.77
Ft.
MC
1975 —
0.006
5.53
1.89
3.21
0.43
24.78
0.77
All Vah
2.877
1.728
0.363
0.196
0.529
0.062
22.753
2.274
All Veh
2.796
1.692
0.344
0.195
0.507
0.057
22.228
2.207
All Veh
2.655
1.639
0.307
0.195
0.467
0.048
21.410
2.098
1-10
-------
April 22, 1993
OCal. Year: 2006
0 Veh. Type:
Veh. Speed*:
VMT Mix:
I/M Program
Anti— tarn. Program
Reformulated Gaa
LDCV LDCT1
UTS
0.599
OCompoaite Emiaaion Factora
VOC HC:
Eshauat HC:
Evaporat HC:
Refuel L HC:
Runing L HC:
Rating L HC:
Exhauat CO :
Exhauat NOX:
OCal. Yeax: 2008
2.25
1.37
0.22
0.19
0.44
0.04
19.55
1.32
I/M
Anti-tarn.
19.6
0.197
(On/Mile
2.58
1.62
0.26
0.25
0.42
0.04
22.20
1.56
Program :
Program :
Reformulated Gaa:
0 Veh. Type:
Veh. Speed*:
VMT Mix:
LDCV
1573 —
0.594
OCompoaita Emiaaion Factora
VOC HC:
Exhauat HC:
Evaporat HC :
Refuel L HC:
Runing L HC:
Rating L HC:
Exhauat CO:
Exhauat NOX:
OCal. Tear: 2011
2.20
1.36
0.21
0.19
0.41
0.03
19.54
1.30
I/M
Anti-tarn.
LDCT1
T573
0.199
(Cm/Mile)
2.54
1.60
0.25
0.25
0.41
0.04
22.17
1.53
Program :
Program :
Reformulated Caa :
0 Veh . Type :
Veh. Speed* :
VMT Mix:
LDCV
19.5
0.588
OCompoait* Emiaaion Factora
VOC HC:
Exhauat HC:
Evaporat HC :
Refuel L HC:
Bun Ing L HC:
Rating L HC:
Exhauat CO:
Exhauat NOX:
2.14
1.35
0.20
0.19
0.38
0.02
19.53
1.29
LDOT1
19. 5
0.201
(Cm/Mile)
2.48
1.60
0.23
0.25
0.38
0.03
22.25
1.53
No
No
No
LDCT2
indicated calendar yej
Ambient Temp: 87.5
Operating Mode : 20.6
LDGT HDGV
1378
0.087
3.47
2.31
0.28
0.25
0.59
0.04
3O.39
2.16
No
No
No
LDGT2
19. 6
0.088
3.42
2.30
0.27
0.25
0.57
0.04
30.53
2.15
No
No
No
LDCT2
19.
-------
Appendix J:
Emissions Analyzer Price Information from Horiba
-------
HORIBA
HORIBA INSTRUMENTS INCORPORATED
3901 Varsity Drive. Ann Arbor, Michigan 48108
Telephone: 1(800)3 HORIBA. In Mich. 1(800) 624-0899 or (313) 973-2171
Faxi (313) 973-7868
April 7, 1993
Environmental Protection Agency
2565 Plymouth Road
Ann Arbor, Ml 48105
Attn: Mr. Bill Pidgeon
Re: IM 240 Analyzer Information
Dear Mr. Pidgeon:
This letter is a follow-up to prior discussions we've had regarding the list price of IM 240
Analyzer Systems. We would like to thank you for the opportunity of discussing our equipment and
market with you and your staff.
We would like to make a clarification in reference to the IM 240 pricing. Horiba is actively
working with six of the seven centralized contractors. Four of these contractors currently have IM 240
analyzers installed. The current list price for the Analyzer/CVS System is $75,515. It should be noted
that this price does not include a blower or external sample line. As you can understand, this is a
"single unit price" and does not reflect discounting for quantity orders. For long-term pricing
considerations, it should be recognized that we also anticipate price reductions following improvements
in manufacturing efficiencies.
Horiba's analytical system can be supplied with other options, such as; a driver's aid, purge
ssure equipment, data collection and processing capabilities.
IM 240 Analyzer System
HC - FID
CO - NDIR
C02 - NDIR
NOx - Chemiluminescent
CVS - 500-700 CFM
Total: $75,515
and pressure
-------
EPA Page 2
Mr. Bill Pidgeon
We feel that our forte' is in the analytical and sample handling portion of the testing lane. For
this reason, we are providing you with analytical system pricing only. Most of our customers have
sourced or built the other components themselves.
If you should have any additional comments or questions, please feel free to contact me at 1-
800-3HORIBA.
Sincerely,/
Kenneth W. Thomas
Marketing Manager,
IM Systems
KWT/pm
cc: Neat Harvey
Andy Marko
kt041 .Itr
-------
Appendix K:
Centrifugal Blower Price Quotation
from Combined Fluid Products Company
-------
COMBINED
FLUID
PRODUCTS
COMPANY
TO: Environmental Protection Agency
2565 Plymouth Road
Ann Arbor/ Michigan 48105
QUOTATION
ISSUED FROM
Q 805 Oakwood Rd., Lake Zurich, IL 60047
Phone (708) 540-0054 FAX (708) 540-0513
Q 125 N. Executive Dr., Brookfield, Wl 53005
Phone {414) 258-7770 FAX (414) 821 -1492
Q P.O. Box 216, 24 S. Green St., Brownsbuig, IN 46112
Phone (317) 852-3961 FAX (317) 852-2337
& 5025 Venture Dr., Ann Arbor, Ml 48108
Phone (313) 930-2024 FAX (313) 747-7040
Please Note:
A/R Oept. is located in Lake Zurich, Illinois
ATTENTION:
Mr. Dan Sampson
EFFECTIVE DATE: January 27, 1993
REFERENCE:
EXPIRATION DATE: February 27, 1993 QUOTATION NO.: AA408
In Compliance With Your Request. We Are Pleased to Quote You As Follows:
QUANTITY
DESCRIPTION
PRICE
1-10
Paxton Centrifugal Blower Model RM-87, including: 10 HP
electric motor running at 3,600 RPM on 230/460 volt
vacuum, three-phase, 60 Hz,. TEPC motor.
- Inlet Filter/Silencer
$3,320.00
Per Unit Net
11-100
$2,656.00
Per Unit Net
DELIVERY Six to eight weeks F.O.B. Santa Monica* California
TERMS: Net 30 Days. Subject To Credit Approval
(Delivery Subject to Prior Sale)
Scott P. Corrunker, Sales Engineer
This quotation subject to standard terms and condition* of sale aa stated on reverae side.
Please use the Acceptance term to place your order.
-------
Appendix L:
Average IM240 Test Time Utilizing
Preliminary Fast-Pass and Fast-Fail Algorithms
-------
Average IM240 Test Time Utilizing
Preliminary Fast-Pass and Fast-Fail Algorithms
The objective of this analysis was to estimate the average IM240 test time
using algorithms that allow vehicles with very low emissions to fast-pass and
vehicles with very high emissions to fast-fail. This reduces the average time
required for the IM240, allowing higher throughput, which reduces the number
of inspection lanes required. The reduced number of lanes lowers equipment
and personnel costs, having the potential to significantly improve the cost
effectiveness of the I/M program.
This analysis describes the fast-pass and fast-fail algorithms used to
estimate the average IM240 test time. The results are preliminary,
representing what could be achieved in time to comply with the court ordered
deadline for this rulemaking. Developing these algorithms requires using
second-by-second data for HC, CO, NOx, and purge, which is very time
consuming, given the huge amount of data per vehicle.
The ideal fast-pass/fast-fail algorithm consists of two continuous
functions. One function represents emission levels at each second of the
IM240 that reliably predict a passing result while the other function
represents emission levels that reliably predict a failing result. Because
this requires evaluating the results at each second of the test for each of
the vehicles, we determined that this could not be achieved under the time
constraint. Instead, we evaluated nine segments (modes) of the IM240, which
significantly reduces the burden, but gives a less than optimal result.
So, additional fast-pass and fast-fail algorithms will be evaluated in the
future, and additional vehicles will be available for those analyses, so these
results should be regarded as preliminary. For example, very low emitters or
extremely high emitters can be fast-passed or fast-failed early in the IM240
cycle, while vehicles near the certification emission levels will require more
time to accurately predict a passing or failing result. The emission
reduction benefits, obtained from repairing vehicles whose emission levels are
slightly dirtier than their certification standards, are not very cost
effective. Similarly, it also may not be cost effective to run the full IM240
as required to accurately distinguish marginal emitters that pass the full
IM240 from marginal emitters that fail. This can be evaluated by comparing
IDRs, failure rates, and error of commission rates for each second of the
IM240 to determine the best tradeoff.
Another consideration is the IM240 reversed. The IM240 was designed as a
two-mode test. The second mode includes the maximum speed of 56.7 mph. The
L-2
-------
lM240-reversed starts with this high speed mode, then is followed by the low
speed mode. This may further reduce the average test time required to
distinguish malfunctioning cars from properly functioning cars. It should be
especially helpful in rapidly determining whether the purge system is
performing adequately.
The algorithm used in this analysis was comparatively crude due to time
and data handling constraints. Several discrete modes of the IM240 were
selected for determining passing and failing emission levels. These modes
were selected to avoid ending the test during an acceleration or deceleration
and to provide a reasonable duration for each of the nine modes. The average
IM240 test time was calculated as the average of the selected mode times
weighted by the number of vehicles passing or failing at each mode. A more
detailed description of the data and methodology used as well as the results
are included in the following sections.
The database used for this analysis conformed to the model I/M program, so
it was limited to 1986 and newer vehicles with second-by-second IM240 results
- 494 vehicles. These vehicles were tested between June 4, 1992 and August 4,
1992. Data were only used if the composite results calculated from the
second-by-second data had passed EPA's quality control measures. Due to the
volume of second-by-second data and the time constraints involved, the second-
by-second data were not QC'd separately.
The following nine modes were selected for pass/fail determinations:
Modes For Evaluating Fast-Pass And Fast-Fail
Mode
(#)
1
2
3
4
5
6
7
8
9
IM240Mode
(sees.)
0-34
0-60
0-74
0-93
0-113
0-154
0-173
0-206
0-239
IM240 Speed
@ End of Mode
(mph)
22.6
30.4
29.8
0.0
27.2
26.0
47.2
51.6
0.0
To determine the passing and failing emission levels for each mode, the
sample was divided into passing and failing vehicles. The pass/fail
determination was made based on the "two ways to pass" criteria with 0.8 g/mi
L-3
-------
HC, 15.0 g/mi CO and 2.0 g/mi NOx as composite IM240 outpoints and, 0.5 g/mi
HC and 12.0 g/mi CO bag 2 outpoints. One liter of purge volume was used as
the outpoint for purge flow. These criteria are illustrated below.
Pass/Fail Decisions Based On Two-Ways-To-Pass-Criteria
Decision
Fail
Fail
Fail
Fail
Pass
Pass
Pass
Pass
Pass
Pass
Pass
IM240
HC
g/mi
> 0.8
£0.8
£0.8
£0.8
£0.8
> 0.8
£0.8
> 0.8
£0.8
£0.8
£0.8
IM240
CO
g/mi
£ 15.0
>15.0
£ 15.0
£15.0
£ 15.0
>15.0
£ 15.0
£ 15.0
£ 15.0
>15.0
£ 15.0
Bag 2
HC
g/mi
> 0.5
£0.5
£0.5
£0.5
£0.5
£0.5
> 0.5
£0.5
> 0.5
£0.5
£0.5
Bag 2
CO
g/mi
£ 12.0
>12.0
£ 12.0
£12.0
£ 12.0
£ 12.0
>12.0
£ 12.0
£ 12.0
£ 12.0
>12.0
IM240
NOx
g/mi
£2.0
£2.0
> 2.0
£2.0
£2.0
£2.0
£2.0
£2.0
£2.0
£2.0
£2.0
Purge
liters
£1.0
£1.0
£1.0
>1.0
£1.0
£1.0
£1.0
£1.0
£1.0
£1.0
£1.0
Comments
Must fail HC on both
Composite & Bag 2 to
Must fail CO on both
Composite & Bag 2 to
Only 1 way to Pass:
Composite NOx £ 2.0
pass.
fail
fail
to
The minimum emission levels and maximum purge volume for failing vehicles
at each mode were used as fast-pass outpoints. Conversely, the maximum
emission levels for passing vehicles at each mode were used as fast-fail
outpoints. Vehicles were not fast-failed based on purge results since many
vehicles purge late in the IM240 cycle. As mentioned, the IM240-reversed may
help rapidly determine if the purge system is functioning adequately.
The modal cutpoint levels, the number of vehicles fast-passing or fast-
failing at each mode and the average IM240 test time as a result of the
application of this fast-pass/fast-fail algorithm are displayed in the
following table.
L-4
-------
Number
Fast-pass Outpoints of
Time Purge
2
3
4
5
6
7
8
9
0-34 <0.479/1.02/0.99 16
0-60 <0.487/0.89/0.99 2
>0.3
0-74 <0.429/0.929/0.90 1
>0.3
0-93 <0.377/0.921/0.84 0
>0.4
0-113 <0.460/0.932/0.89 3
>0.5
0-154 <0.567/1.088/0.96 3
>0.6
0-173 <0.697/3.52/1.33 65
>0.7
0-206 <0.916/14.99/1.77 210
>0.8
0-239 <0.805/15.05/2.05 45
Weighted Sum with
Fast-pass Only 102410
Average IM240 Test
Time with Fast-pass
Only = 207 sec
Number Time *
Number of Number
of Vehicles of
Vehicle Fast- Vehicles
Fast- Fast-fail Outpoints s Fast- passing with Fast
passing >HC/CO/NOx failing and Fast- Result
failing
>3.405/56.72/7.30 15 31 1054
>1.891/47.30/4.63 22
>1.648/38.09/3.58
>1.536/41.09/3.19
>1.518/36.78/3.02
>1.296/30.34/2.57 11
>1.120/25.22/2.65 11
>0.915/18.06/2.33 35
>0.805/15.05/2.05 33
24
8
9
9
14
76
245
78
1440
592
837
1017
2156
13148
50470
18642
Weighted
Sum 89356
Average
IM240
Test Time
180 sec
These results indicate that the test time for the IM240 can be reduced by
25% when fast-pass/fast-fail criteria are applied and a reduction of over half
a minute occurs when only fast-pass criteria are applied. Using only fast-
pass criteria allows for the collection of diagnostic data so that failing
cars may be repaired more effectively.
Because Hammond cars with second-by-second data were typically shut off
for 10 minutes, catalyst cool down could have caused high emissions during the
early parts of the test and adversely affected fast-pass and fast-fail.
Similarly, vehicles that drive a short distance to an I/M station may not be
fully warmed up when they start the test. Therefore, additional analyses were
L-5
-------
performed without integrating over the first part of the IM240. In effect,
utilizing the first segment of the IM240 as preconditioning. Three different
integration starting points were used. Since the accelerations contribute the
most toward catalyst light-off, these starting points follow the first three
accelerations of the IM240 cycle. The integrations begin after 17, 35 and 47
seconds of the test. The results of these analyses are displayed here.
ode*
1
2
3
4
5
6
7
8
9
Time
> (sec)
17-34
17-60
17-74
17-93
17-113
17-154
17-173
17-206
17-239
Fast-pass Outpoints
Purge
<0.525/0.95/1.33
<0.504/0.54/1.10
>0.3
<0.465/0.90/0.96
>0.3
<0.400/0.90/0.88
>0.4
<0.486/0.9 1/0.93
>0.5
<0.593/1.09/1.00
>0.6
<0.641/3.08/1.38
>0.7
<0.826/15.33/1.82
>0.8
SO.805/15.05/2.05
£1.0
Weighted Sum with
Fast-pass Only
Average IM240
Test Time with
Fast-pass Only
Number
of
Vehicles
Fast-
passing
11
1
4
0
5
3
56
217
48
103230
209 sec
Fast-fail Outpoints
>HC/CO/NOx
>2.643/76.94/10.33
> 1.892/53.86/5.11
> 1.615/41.40/3.77
> 1.498/45.64/3.27
> 1.484/40.16/3.08
>1.265/32.27/2.66
> 1.080/26.48/2.71
>0.936/18.44/2.32
>0.805/15.05/2.05
Number
of
Vehicles
Fast-
failing
19
11
11
10
7
10
10
37
34
Number
of
Vehicles
Fast-
passing
and Fast-
failing
30
12
15
10
12
13
66
254
82
Weighted
Sum
Average
IM240
Test Tune
Time*
Number of
Vehicles
with Fast
Result
1020
720
1110
930
1356
2002
11418
52324
19598
90478
183 sec
L-6
-------
Fast-pass Cutpoints
Tune Purge
1 N/A N/A
2 35-60 <0.493/0.79/0.90
>0.3
3 35-74 <0.403/0.73/0.79
>0.3
4 35-93 <0.340/0.69/0.75
>0.4
5 35-113 <0.454/0.91/0.82
>0.5
6 35-154 <0.585/1.10/0.93
>0.6
7 35-173 <0.575/2.85/1.37
>0.7
8 35-206 <0.795/15.17/1.84
>0.8
9 35-239 £0.805/15.05/2.05
21.0
Weighted Sum with
Fast-pass Only
Average IM240
Test Time with
Fast-pass Only
Number
of
Vehicles
Number Time *
Number of Number of
of Vehicles Vehicles
Vehicles Fast- with Fast
Fast- Fast-fail Cutpoints Fast- passing Result
passing >HC/CO/NOx failing and Fast-
failing
N/A N/A N/A N/A N/A
19 >1.983/41.71/3.71 41 60 3600
4
2
5
2
48
221
44
02452
07 sec
>1.499/31.32/3.08
>1.450/55.71/3.09
>1.406/47.21/3.07
>1. 299/35.99/2.59
> 1.06 1/28.83/2.81
>0.966/19.48/2.37
>0.805/15.05/2.05
8
5
3
7
7
35
43
12
7
8
9
55
256
87
Weighted
Sum
Average
IM240
Test Time
888
651
904
1386
9515
52736
20793
90473
183 sec
L-7
-------
Fast-pass Outpoints
Time Purge
1
2
3
4
5
6
7
8
9
N/A
47-60
47-74
47-93
47-113
47-154
47-173
47-206
47-239
N/A
<0.458/0.40/1.05
>0.3
<0.375/0.46/0.83
>0.3
<0.3 10/0.52/0.76
>0.4
<0.434/0.94/0.85
>0.5
<0.594/1. 14/0.96
>0.6
<0.550/2.88/1.43
>0.7
<0.751/14.82/1.91
>0.8
50.805/15.05/2.05
Number Number
of of
Vehicles Vehicles
Fast- Fast-fail Outpoints Fast-
passing >HC/CO/NOx failing
N/A N/A N/A
6 >2.089/37.20/3.67 41
9 > 1.282/33.21/2.99 14
5 > 1.737/67.22/3.10 0
15 > 1.619/54.68/3.08 2
4 >1.355/39.55/2.55 8
48 > 1.095/30.98/2.82 4
220 > 1.004/20.39/2.42 35
38 >0.805/15.05/2.05 45
Number
of
Vehicles
Fast-
passing
and Fast-
failing
N/A
47
Time *
Number of
Vehicles
with Fast
Result
N/A
2820
23
5
17
12
52
255
83
1702
465
1921
1848
89%
52530
19837
Weighted Sum with
Fast-pass Only 102119
Average IM240
Test Time with
Fast-pass Only 207 sec
Weighted
Sum 90119
Average
IM240
Test Time 182 sec
These results indicate, that for the data used in this analysis,
preconditioning has little effect on the average test time of the fast-
pass/fast-fail algorithm used. In spite of this, these estimates are
considered conservative for several reasons. First, older cars are excluded
from the analysis. Since most grossly emitting vehicles are older vehicles,
the inclusion of these cars would be expected to increase the number of fast-
failing vehicles and reduce the test time further. However, this reduction
may be offset by a reduction in the percentage of vehicles fast-passing. More
important than the vehicle sample is the algorithm used. If a continuous
function were used, actual test times could be used to calculate the average.
This should lead to significant time savings compared to using the last second
of a particular mode as the required test time for all vehicles that pass or
L-8
-------
fail during that mode. It is unlikely that all the vehicles failing or
passing a particular mode would have required the full mode to determine their
outcome. Therefore, average test times for vehicles passing the IM240 at
second 60 would be significantly less than 60 seconds. Likewise, this would
be true for each mode. On-going analyses are being performed to investigate
this and other alternatives such as the IM240-reversed. Finally, EPA will
continue to develop alternative algorithms which are also expected to reduce
the average test time for the IM240.
L-9
------- |