Report No. EPA 460/3-88-010
Analytical Support for
Emission Factors Development
and Air Quality Assessment
Work Assignment No. 0-01:
Analysis of California
I/M Review Committee Data
Task 5 Report
Analysis of inspection Inconsistencies
Between Different I/M Test Sites
prepared for:
U.S. Environmental Protection Agency
September 30, 1988
prepared by:
Sierra Research. Inc.
1521 I Street
Sacramento, California 95814
(916) 444-6666
-------
Report No. EPA 460/3-88-010
ANALYTICAL SUPPORT FOR
EMISSION FACTORS DEVELOPMENT
AND AIR QUALITY ASSESSMENT
EPA Contract No. 68-03-3474
Work Assignment No. 0-01:
Analysis of California
I/M Review Committee Data
Task 5 Report:
Analysis of Inspection Inconsistencies
Between Different I/M Test Sites
prepared for:
U.S. Environmental Protection Agency
September 30, 1988
prepared by:
Thomas C. Austin
Thomas R. Carlson
Kathryn A. Gianolini
Sierra Research, Inc.
1521 I Street
Sacramento, CA 95814
(916) 444-6666
-------
ANALYTICAL SUPPORT FOR
EMISSION FACTORS DEVELOPMENT
AND AIR QUALITY ASSESSMENT
Task 5 Report:
Analysis of Inspection Inconsistencies
Between Different I/M Test Sites
Table of Contents
page
1. Summary 1
2. Introduction and Methodology 6
3. Results 15
Emissions Measurement Variability 15
Underhood Inspection Variability 28
Emission Reductions 31
Effect of 207(b) Standards 37
-------
1. SUMMARY
During a recent evaluation of the California vehicle inspection and
maintenance program, vehicles were tested at several different
locations. Failure rates at a "screening facility" and at the
California Air Resources Board (ARB) laboratory were much higher than
those recorded at official Smog Check stations. The original analysis
of the data suggested that "false passes" at Smog Check stations were
due to incorrect underhood inspections of vehicles which failed only
for underhood reasons when tested by ARB. Further analysis of the
data has now been completed which confirms that inaccurate underhood
inspection results at Smog Check stations were the principal reason
for the difference in results between inspection facilities. However,
the analysis also indicates that inconsistencies in tailpipe emission
measurements were a contributing factor.
33% percent of the vehicles showed inconsistent tailpipe emission test
results (i.e., fail vs. pass) when comparing the screening facility to
the first Smog Check station the vehicles were taken to. However,
about one-third of those vehicles had inconsistent tailpipe emission
results when comparing the screening facility to the ARB laboratory
tests. In general, the vehicles that showed inconsistent tailpipe
results were primarily idle mode failures. In addition, there was a
-1-
-------
Figure 1
Screening Facility Tailpipe Failure Rate
For Vehicles That Passed
at the First Smog Check Station
Pre-75
75 - 79 1980 and Later
Model Year Group
Overall
strong relationship between tailpipe test inconsistency and model year
group. As illustrated in Figure 1, inconsistent results were much
more frequent for 1980 and later models.
One reason why Smog Check stations often reported lower tailpipe
emission levels for vehicles that failed the tailpipe test at the ARE
laboratory or the Screening Facility is that Smog Check mechanics will
sometimes perform repeated "initial" tests on a vehicle in an attempt
to make it pass the standards. The additional preconditioning that
occurs between tests may be contributing to a reduction in the number
of improper or pattern failures that occurred at Smog Check stations.
This hypothesis is supported by the fact that for 1975 and later
-2-
-------
models there was no significant reduction in HC or CO emissions for
vehicles that passed at the first Smog Check stations they were taken
to and then subsequently failed at a second Smog Check station.
Figure 2 provides another illustration of the inconsistency in
tailpipe emission measurements between the facilities. Under the 2500
RPM carbon monoxide test, 37% of the 1980 and later model vehicles
failed the test when tested at the screening facility. At the ARE
lab, the failure rate dropped to 6%; at the first Smog Check station,
none of the vehicles failed. As shown in the figure, the results at
Smog Check station #1 indicate a large increase in the number of
vehicles in the lowest emission level range (0-0.2%). A significant
80
-a 60
C 40
o>
o
s.
20
Figure 2
Distribution of 2500 RPM CO Emission Levels
For Vehicles that Passed at Smog Check #1
California Standard
I
bn H (B^ H^ H^
___JI^^j^n^_jiEa_JL^_^_jEa_
Screening Facility
ARB Laboratory
Smog Check #1
1980+ 3-way Catalyst Vehicles
Tailpipe Emissions (C0%)
-3-
-------
number of vehicles that failed the test at the screening facility or
at the ARE laboratory were well under the standard when tested at the
Smog Check station. Preconditioning differences are the expected
reason for the variability in failure rate.
As shown in Figure 3, most of the vehicles that passed at the first
Smog Check station were "underhood-only" failures. Based on tests at
the screening facility, about 80% of the pre-1980 vehicles that passed
at the ARB lab or the first Smog Check station were underhood-only
failures, this dropped to about 50% for the 1980 and later model
vehicles, where pattern failures are more of a problem.
100
Figure 3
Percent of Vehicles Missed at Smog Check #1
That Were Underhood-Only Failures
Pre-1975 1975-1979
Model Year Group
1980+
Screening Facility
ARB Lab Breezeway
-4-
-------
Although there was variability between model-year groups, repair of
those vehicles that passed the first Smog Check after failing at the
screening facility would have contributed to additional reductions of
hydrocarbon and oxides of nitrogen emissions. Vehicles that passed at
the first Smog Check station and then failed at the second station
experienced FTP emission reductions of 32.6% HC and 11.8% NOx when
repaired. Carbon monoxide emissions were essentially unaffected.
However, the hydrocarbon emission benefits from the repair of these
vehicles were almost exclusively from pre-1975 model vehicles. For
1975-1979 models, there were no significant benefit for any pollutant.
For 1980 and later models, significant reductions in NOx emissions
were recorded, but there were increases in HC and CO emissions. This
is the effect that would be expected from the correction of EGR
tampering that was missed during the visual inspection at the first
Smog Check station.
-5-
-------
2. INTRODUCTION AND METHODOLOGY
Under a contract with the U.S. Environmental Protection Agency (EPA)
for "Analytical Support for Emission Factors Development and Air
Quality Assessment," Sierra Research, Inc. (Sierra) performs a variety
of Work Assignments for the Emission Control Technology Division
(ECTD) of EPA's Motor Vehicle Emissions Laboratory in Ann Arbor,
Michigan. Work Assignment 0-01 directed Sierra to analyze California
I/M data for the ECTD Technical Support Staff (TSS). Task number 5 of
that Work Assignment required further evaluation of inspection
inconsistencies that were observed during the course of the recently
completed California I/M Evaluation Program.
The direction provided by TSS was as follows:
The analysis of inspection inconsistencies on pages 41-47 of
the Technical Appendix (to Sierra's previous report on the
California I/M Evaluation Program) for the most part does not
separate tailpipe inspection and underhood inspection as
causes of failure. The analysis is sufficient to suggest that
many or most "false passes" at the first smog check were due
to incorrect underhood inspections of vehicles which failed
only for underhood reasons at the ARE screening facility. The
contractor shall perform an analysis which looks solely at the
variation in tailpipe results between the four potential test
sites for each vehicle (screening facility, ARB lab/breezeway,
and the two smog check stations). Idle and 2500 rpm modes
shall be distinguished, and the analysis shall consider both
California and Federal (207(b)) outpoints. The as-received
FTP emission performance of the more- and less-variable
vehicle groups should be given, as well as the emission
reductions eventually delivered by repairs. Model year and
technology effects should be explored.
-6-
-------
Background
Under the California I/M Evaluation Program, "undercover" vehicles
obtained from the general population were given both I/M and FTP tests
at ARB's El Monte laboratory prior to being sent to a randomly
selected Smog Check station. All of the undercover vehicles initially
failed an inspection given by ARB technicians at the "screening
facilities" that were established on a temporary basis in various
locations throughout the South Coast Air Basin. Based on the
screening test, these vehicles were expected to fail a properly
conducted inspection at a Smog Check station.
Vehicles which either failed or received pre-inspection maintenance at
the first Smog Check station to which they were taken were referred to
as "F sample" vehicles. The emission reductions achieved through
repair of these vehicles were subsequently used to estimate the
benefits of the current Smog Check program.
Vehicles which passed at the first Smog Check station were taken to a
second, randomly selected Smog Check station. If pre-inspection
maintenance or a failure and repair occurred at the second station,
then the vehicles were placed in what ARB refers to as the "G sample"
category. Vehicles which passed at both Smog Check stations were
referred to as "E sample" vehicles.
Differences Between Vehicles that Passed and Failed
Table 1 is a copy of Table 5-1 from Sierra's earlier study for the
California I/M Review Committee ("Evaluation of the California Smog
-7-
-------
Check Program - Technical Appendix," April, 1987). The table shows
how the baseline Federal Test Procedure emissions compare for the
undercover vehicles, based on whether they passed or failed at the
Smog Check stations. As the table shows, vehicles which failed at the
first Smog Check station generally had significantly higher emissions
than those which passed. Similarly; vehicles which failed at the
second Smog Check station generally had higher emissions than those
which passed for a second time.
Table 1
Baseline FTP Emissions of Undercover Vehicles
That Failed at First Smog Check
vs. Vehicles That Passed
HC
CO
NOx
Sample
Size
Sample
Name
F
G
E
F
G
E
F
G
E
F
G
E
.L/ri J.BSL.
Smog
Check#l
Fail
Pass
Pass
Fail
Pass
Pass
Fail
Pass
Pass
Fail
Pass
Pass
Total
ix.eauj.i_i>
Smog
Check#2
NA
Fail
Pass
NA
Fail
Pass
NA
Fail
Pass
NA
Fail
Pass
Sample
Pre-'75s
10.49
7.16
4.98
70.73
53.44
62.26
3.31
3.87
3.25
139
13
53
205
Model Year Group
1975-1979
4.25
2.02
1.52
49.27
31.80
18.84
2.75
2.58
2.86
195
16
116
327
Post- '79
1.89
0.99
0.74
32.04
12.59
11.37
1.26
1.66
1.42
160
11
92
263
All
Vehicles
5.24
3.41
1.95
49.73
33.55
25.02
2.43
2.75
2.43
494
40
261
795
-8-
-------
Considering all model years together, hydrocarbon (HC), carbon
monoxide (CO), and oxides of nitrogen emissions (NOx) from all of the
vehicles that failed at the first Smog Check station were 5.24, 49.73,
and 2.43 grams/mile, respectively.
The grams/mile emissions from vehicles which passed the first time and
failed the second were 3.41 HC, 33.55 CO, and 2.75 NOx. NOx emissions
for this group were 13% higher, but HC and CO emissions were lower by
35% and 33%, respectively.
Vehicles which passed both Smog Checks had average emissions of 1.95
HC, 25.02 CO, and 2.43 NOx. This is 63% lower for HC and 50% lower
for CO than emissions from vehicles which failed at the first station.
NOx emissions were the same.
The sample size information at the bottom of Table 1 indicates that
494 of a total of 795 undercover cars failed at the first Smog Check
station. In other words, only 62% of the vehicles that were expected
to fail actually did fail. However, the cars with the highest
emissions, excluding NOx, were the ones that failed.
Table 2 (Table 5-2 from the earlier report) indicates why some cars
passed and others failed at Smog Check stations. As the table shows,
there were substantial differences in tailpipe failure rates between
vehicles that failed at the first station and vehicles that passed.
-9-
-------
Table 2
Tailpipe Failure Rate of Undercover Vehicles
That Failed at First Smog Check vs. Vehicles That Passed
Test Sample
Location Name
Smog Check
Station #1
ARB Lab
(Breezeway)
Screening
Facility
F
G
E
F
G
E
F
G
E
Smog Smog
Chk#l Chk#2
Fail
Pass
Pass
Fail
Pass
Pass
Fail
Pass
Pass
NA
Fail
Pass
NA
Fail
Pass
NA
Fail
Pass
Pre-'75s
63.8%
0.0%
0.0%
59.7%
15.4%
15.4%
64.7%
15.4%
20.8%
a.dj. j-p j-jje
lodel Year Gr
1975-1979
70.6%
0.0%
0.0%
60.0%
25.0%
12.1%
62.1%
37.5%
19.8%
r ctj.j_u.j. t: Raul
Post-'79
75.6%
0.0%
0.0%
56.9%
18.2%
21.7%
76.3%
45.5%
41.3%
All
Vehicles
70.3%
0.0%
0.0%
58.9%
20.0%
16.2%
67.4%
32.5%
27.6%
The group of vehicles that was passed at the first Smog Check station
had significantly lower tailpipe emissions failure rates when tested
by ARB. The Test Analyzer Systems used at Smog Check stations are
known to be reliable and accurate, and observations by ARB employees
proved that the test results reported are for the same vehicle.
"Test-to-test variability" appears to be the reason why some vehicles
failed the tailpipe standards when tested by ARB but not when tested
by the Smog Check station. Further evidence that test-to-test
variability is the problem can be seen from the difference between
failure rates at the ARB laboratory and at the screening facility.
ARB expects, and experience with repeated tests on a number of
vehicles indicates, that "preconditioning" differences are one factor
-10-
-------
affecting this test-to-test variability problem. Vehicles which are
not thoroughly warmed-up tend to have higher failure rates.
Table 3 (Table 5-3 from the earlier report) indicates another reason
why some vehicles failed when tested by ARE but not when tested at a
Smog Check station. This table indicates that most of the vehicles
contained visual or functional defects. According to inspection
results at the ARB lab, 84.6% of all vehicles that passed at both Smog
Check stations contained at least one visual or functional defect.
Due to the poor quality of the inspections performed at the Smog Check
stations, these defects were missed. When visual defects are missed
and the tailpipe emission levels meet the I/M standards, the vehicle
will obviously pass the test.
Table 3
Underhood Failure Rate of Undercover Vehicles
That Correctly Failed at First Smog Check
vs.
Vehicles That Passed
Test Sample
Location Name
Smog Check
Station #1
ARB Lab
(Breezeway)
Screening
Facility
F
G
E
F
G
E
F
G
E
J./11 O.tSSU.J-L.5
Smog Smog
Chk#l Chk#2
Fail
Pass
Pass
Fail
Pass
Pass
Fail
Pass
Pass
NA
Fail
Pass
NA
Fail
Pass
NA
Fail
Pass
Vff
Pre-'75s
55.8%
0.0%
0.0%
88.5%
84.6%
94.2%
84.2%
92.3%
88.7%
uiiucj.uuuu r
>del Year Gro
1975-1979
49.5%
0.0%
0.0%
88.2%
87.5%
93.1%
84.1%
81.3%
92.2%
cm.ui.ts .tt.ci.ce
UP
Post- '79
35.6%
0.0%
0.0%
57.5%
90.9%
68.5%
43.8%
81.8%
63.0%
All
Vehicles
46.7%
0.0%
0.0%
78.3%
87.5%
84.6%
71.1%
85.0%
81.2%
-11-
-------
Supplemental Analyses Performed
The Database - To provide the additional information requested by TSS,
Sierra utilized Test Analyzer System (TAS) data from the California
I/M Evaluation vehicles for which ARB had coded data from all of the
test locations (783 out of 795 vehicles). Although TSS wanted an
analysis of data at the screening facility, the ARB laboratory
"breezeway" and Smog Check stations #1 and #2, Sierra discovered that
very little data from Smog Check station #2 was in the database (to
date, ARB has coded a limited amount of data from Smog Check station
#2). Therefore, the analysis was restricted to the screening
facility, the ARB lab/breezeway and Smog Check station #1.
For the analysis of FTP emissions, the 783-vehicle sample was reduced
to 704 when only vehicles with both baseline and after-repair FTP test
data were included.
Analyses Conducted TAS data analysis involved disaggregation of the
sample into different technology categories and different failure
modes. Comparisons of the data subsets could then be used to
determine the relationship between test results at the different
facilities and the characteristics of the vehicles. In addition,
distributions of tailpipe emissions for vehicles with inconsistent
test results were constructed. Examples of changes in emission
measurements for individual vehicles were also prepared.
FTP data were used to determine the difference between the emission
reduction potential of vehicles that failed at the first Smog Check
-12-
-------
station and that of vehicles that did not fail until the second
station.
In the analysis conducted for the draft version of this report, a
group of 65 vehicles referred to as the "X-sample" were excluded from
&
the analysis. These vehicles were originally deleted from the sample
because they passed both the tailpipe and underhood portions of the
I/M test at the ARB laboratory, even though they failed at the
"screening facility". EPA speculated that these vehicles may have
been "pattern failure" vehicles that only passed at the ARB laboratory
because of differences in preconditioning. Depending on the relative
performance of the ARB lab in avoiding pattern failures, keeping these
vehicles in the sample could significantly affect the consistency
between test locations, especially for the 2500 rpm test mode.
Further analysis by Sierra indicated that 17 of the 65 vehicles were
(incorrect) underhood-only failures at the screening facility. These
* As might be expected, the FTP emission characteristics of the 43
1980 and later model X-sample vehicles were different from the other
vehicles that failed at the screening facility:
Sample
X-Sample
Others
HC
0.68
1.46
- gLcuua/ U10.0.C -
CO
9.57
24.30
NOx
0.79
1.34
In addition, the failure characteristics of the X-Sample vehicles were
significantly different. While 14.4% of the other vehicles were 2500
rpm-only failures, the 2500-only failure rate for the X-Sample was
53.5%.
-13-
-------
seventeen vehicles did not fail the tailpipe test, so they were not
pattern failures. All the remaining 48 vehicles in the X-sample
Passed the tailpipe test at Smog Check station 1. 43 of the 48 were
1980 or later model vehicles.
Given the observed pattern of failures for the X-sample, Sierra did
not believe it was appropriate to include the 48 tailpipe failure
vehicles in the sample with equal weighting. In fact, an argument
could be made that they should not be included at all because the way
vehicles were preconditioned at the screening facility does not
represent the Smog Check station environment. .(Routine analysis of
Smog Check station data indicates that it is common for mechanics to
run repeated "initial" tests on failing vehicles until they pass.) In
this analysis, however, the 1980 and later model X-sample vehicles
with tailpipe failures were included in the sample with a one-third
weighting factor to account for the fact that they failed at one-third
of the test sites. This increased the total sample of 1980 and later
models from 243 to 286.
All of the analyses conducted are summarized in the following section
of the report. Although not "F" sample cars by definition, the 48 "X"
sample cars were included with the "F" cars (fail Smog Check #1) in
the tables that follow (except Table 10). This is consistent with
treatment of X cars as vehicles which fail at a 1/3 detection rate.
-14-
-------
3. RESULTS
Emission Measurement Variability Table 4 provides more detail on the
tailpipe failure patterns for undercover vehicles than was presented
in the Technical Appendix to the California I/M Evaluation Program
study. The data presented in the table make it possible to
distinguish between idle and 2500 rpm-only failures. Several data
points in the table are preceded by an asterisk (*), indicating a
coding error. The failure rate should have been 0% for all vehicles
that were reported as passing the test at Smog Check station #1.
However, our latest analysis of the data included a routine under
which measured tailpipe emission levels were compared to the
standards. In a few cases, it is apparent that ARE had coded vehicles
as having passed at Smog Check station #1 when they actually failed.
As can be seen in Table 4, there were a significant number of 1980 and
later model vehicles (the only models subject to a 2500 rpm test) that
failed the 2500 rpm tailpipe test at the screening facility and at ARB
laboratory, but later passed the test at Smog Check station #1.
Inconsistencies in the tailpipe emission measurements at the various
inspection sites are more clearly illustrated in Figures 4, 5, 6, and
7. These four figures show the distribution of tailpipe emission
-15-
-------
Table 4
Modal Tailpipe Failure Rates of
Undercover Vehicles that Failed at First Smog Check
vs. Vehicles that were Incorrectly Passed,
By Model Year Range
Sample
Group
All
Vehicles
Test
Location
Smog Check #1
Sample
Size
1414
123
744
t
I/M Test Results
Smog #1 Smos #2
F
P
P
F
P
Tailpipe
Failure Rates (%)
Idle
62.2
*7.3
*0.4
Idle
and/or
2500
69.4
*7.3
*0.8
ARE Lab
(Breezeway)
1522
126
750
F
P
P
F
P
51.8
16.7
13.6
4.1
4.8
2.4
56.0
21.4
16.0
Screening
Facility
1514
126
753
F
P
P
F
P
61.2
28.6
20.3
7.1
4.8
3.6
68.4
33.3
23.9
pre-1975
models
Smog Check #1
393
39
156
F
P
P
F
P
65.6
0.0
0.0
0.0
0.0
0.0
65.6
0.0
0.0
ARB Lab
(Breezeway)
420
39
153
F
P
P
F
P
57.1
15.4
11.8
0.0
0.0
0.0
57,
15.
11
Screening
Facility
417
39
153
F
P
P
F
P
64.0
15.4
17.7
0.0
0.0
0.0
64.0
15.4
17.7
Sample size is the "weighted" sample: ("X" cars) + 3 x (non-"X" cars).
-16-
-------
Table 4 (continued)
Modal Tailpipe Failure Rates of
Undercover Vehicles that Failed at First Smog Check
vs. Vehicles that were Incorrectly Passed,
By Model Year Range
Sample
Group
1975 to
1979
models
Test
Location
Smog Check #1
Sample
Size
531
54
330
t
I/M Test Results
Smog #1 Smog #2
F
P
P
F
P
Tailpipe
Failure Rates (%)
2500
Only
0.0
0.0
0.0
Idle
and/or
ARB Lab
(Breezeway)
583
54
336
F
P
P
F
P
58.2
22.2
11.6
0.0
0.0
0.0
58.
22.
11.
Screening
Facility
583
54
336
F
P
P
F
P
62.9
38.9
17.0
0.0
0.0
0.0
62.9
38.9
17.0
1980 and
later
models
Smog Check #1
490
30
258
F
P
P
F
P
50.8
0.0
*1.2
20.8
0.0
*1.2
71.6
0.0
*2.3
ARB Lab
(Breezeway)
519
33
261
F
P
P
F
P
40.5
9.1
17.2
12.
18.
6.9
52.6
27.3
24.1
Screening
Facility
514
33
264
F
P
P
F
P
57.0
27.3
26.1
21.0
18.2
10.2
78.0
45.5
36.4
f Sample size is the "weighted" sample:
* Data coding errors.
("X" cars) + 3 x (non-"X" cars)
-17-
-------
measurements for HC and CO at each test condition (idle and 2500 rpm)
for 1980 and later model year vehicles equipped with 3-way catalysts.
Figure 4 indicates that there were many more vehicles that failed the
idle hydrocarbon standard at the screening facility than at either the
ARE laboratory or the first Smog Check station. As the figure shows,
there is a significant difference in the percent of vehicles with HC
emissions measured in the 0-50 ppm range at the screening facility and
at the other two locations. It should also be noted that two vehicles
passed at the Smog Check station and failed at the other two sites
Figure 4
Distribution of Idle HC Emission Levels
For Vehicles that Passed at Smog Check #1
IUU
RTI
"5
6
f^ 60
"o
0)
(_)
c5
Q.
20
'-
a
1
;
^
|
|
|
\
\
j
:
|
^P
7
K
--
lib Lk, h-, - - -
^ XX"* KX"' J.X-' JkX^ .Jk'' \.X^.\.'' \' ..V K' ^\X V "*
Hj Screening Facility
^^ AFB Laboratory
IS: Smog Check #1
Tailpipe Emissions (ppm HC)
1980+ 3-way Catalyst Vehicles
-18-
-------
because the Smog Check station improperly recorded the vehicle as
being equipped with an oxidation catalyst. As indicated in Table 5,
this changed the idle HC standard from 100 ppm to 150 ppm.
Table 5
Tailpipe Emission Standards
for the California Smog Check Program
Idle HC Idle CO 2500 HC 2500 CO
Category/Description (ppm) (%) (ppm) (%)
01 '55-'65, >4 CYL
02 '66-'70, >4 CYL, w/ AIR
03 '66-'70, >4 CYL, w/o AIR
04 '71-'74, >4 CYL, w/ AIR
05 '71-'74, >4 CYL, w/o AIR
06 '55-'67, 4 CYL
07 '68-'71, 4 CYL, w/ AIR
08 '68-'71, 4 CYL, w/o AIR
09 '72-'74, 4 CYL, w/ AIR
10 '72-'74, 4 CYL, w/o AIR
11 '75-'79, NO CAT
12 '75-'79, OX CAT, w/o AIR
13 '75-'79, OX CAT, w/ AIR
14 '75-'79, 3WY CAT
15 '80+ , NO CAT
16 '80+ , OX CAT, w/o AIR
17 '80+ , OX CAT, w/ AIR
18 '80+ , 3WY CAT
800
400
500
300
400
1200
450
700
350
350
200
250
150
100
150
150
150
100
8.0
4.5
6.5
3.5
6.5
7.5
5.5
7.0
5.0
6.5
3.5
4.5
1.5
1.5
2.5 220
2.5 220
1.2 220
1.2 220
1.2
1.2
1.2
1.2
-19-
-------
Figure 5 indicates that there were fewer failures at the 2500 rpm test
point. Only 4% of the vehicles that passed at Smog Check station #1
failed this test at the screening facility and none of the vehicles
failed the 2500 rpm HG test when tested at the ARE laboratory.
Figure 5
Distribution of 2500 RPM HC Emission Levels
For Vehicles that Passed at Smog Check #1
80
CO
3
8.
40
20
CaOfornia Standard
Screening FaciFrty
ARE Laboratory
Smog Check #1
Tailpipe Emissions (ppm HC)
1980+ 3-way Catalyst Vehicles
-20-
-------
Figure 6 indicates that idle-mode CO failures occurred more frequently
at the screening facility and the ARE lab than at the first Smog Check
station. However, Figure 7 indicates that 2500 rpm CO failures were
much more of a problem. Thirty-seven percent of the vehicles failed
the 2500 rpm CO standards when tested at the screening facility. This
failure rate dropped to 6% at the ARE lab and to 0% at the first Smog
Check station.
100
+S
"o
60
03
o
fc 40
Q-
20
Figure 6
Distribution of Idle CO Emission Levels
For Vehicles that Passed at Smog Check #1
man
Standard
Screening Facility
ARB Laboratory
Smog Check #1
CP
Tailpipe Emissions (C0%)
1980+ 3-way Catalyst Vehicles
-21-
-------
As shown in Figure 7, the results at Smog Check station #1 indicate a
large increase in the number of vehicles in the lowest emission level
range (0-0.2%). A significant number of vehicles that failed the test
at the screening facility or the ARE laboratory were well under the
standard when tested at the Smog Check station. Preconditioning
differences are the expected reason for the variability in failure
rate.
Figure 7
Distribution of 2500 RPM CO Emission Levels
For Vehicles that Passed at Smog Check #1
80
-= 60
c 40
d>
o
I
20
California Standard
i
I
B 1
Screening Facility
ARB Laboratory
Smog Check #1
Or-'
1980+ 3-way Catalyst Vehicles
'
Tailpipe Emissions (C0%)
-22-
-------
Figure 8 indicates why some of the vehicles that passed when they were
tested at the first Smog Check station failed when they were taken to
a second station. For 1980 and later models, 36.4% of the vehicles
failing at the second station had underhood defects that were missed
at the first station and detected at the second station. 63.6% of the
vehicles failed at the second station because of higher tailpipe
emissions.
Figure 8
Reasons Why Vehicles Failed 2nd Smog Check
After Passing First Smog Check
(1980 and Later Models)
Inconsistent Emissions Measurement63.6%
Underhood Defects Identified36.4%
-23-
-------
Figure 9 illustrates how variable the tailpipe emission measurements
&
of seven different vehicles were from site-to-site. The results
shown in the figure are for 1980 -and later model vehicles that passed
at the first Smog Check station and failed at the second one. Note
that in some cases the passing results at the first Smog Check station
are almost identical to the results at the ARE lab or the screening
facility. In other cases the failing results at the second station
are more like those at the ARB lab or the screening facility. Note
also that the variability between the screening facility and the ARB
Figure 9
Examples of Tailpipe Measurement Inconsistency
Between Inspection Sites
(19SO and Later Models)
700
Screening
Facility
ARB Lab
Smog Check *1
;!::| Smog Check #2
Car 83 Car 151 Car 166 Car 180 Car 320 Car 338 Car 361
(Hi CO) (Hi CO) (LoHC)
-------
Table 6 (continued)
Modal Tailpipe Failure Rates of
Undercover Vehicles that Failed at First Smog Check
vs. Vehicles that were Incorrectly Passed
By Technology Group (1980 and Later Models Only)
Tailpipe
Failure Rates (%)
Sample
Group
GARB/
OXD/NoAIR
Test
Location
Smog Check #1
ARB Lab
(Breezeway)
Screening
Facility
Sample
Size
19
6
24
19
6
24
19
6
24
I/M Test
Smog #1
F
P
P
F
P
P
F
P
P
Results
Smog #2
F
P
_
F
P
-
F
P
Idle
31.6
0.0
0.0
0.0
0.0
0.0
15.8
0.0
0.0
2500
Only
15.8
0.0
0.0
31.6
0.0
0.0
36.8
0.0
0.0
Idle
ana/or
2500
47.4
0.0
0.0
31.6
0.0
0.0
52.6
0.0
0.0
GARB
OXD/AIR
Smog Check #1
ARB Lab
(Breezeway)
Screening
Facility
104
0
75
107
0
78
101
0
81
F
P
P
F
P
P
F
P
P
-
F
P
-
F
P
-
F
P
63.5
-
0.0
47.7
-
15.4
58.4
-
18.5
23.1
-
0.0
16.8
-
7.7
23.8
-
11.1
86.5
-
0.0
64.5
-
23.1
82.2
-
29.6
CARS/
3WAY/OXD
Smog Check #1
ARB Lab
(Breezeway)
Screening
Facility
43
3
6
49
3
6
49
3
6
F
P
P
F
P
P
F
P
P
-
F
P
-
F
P
F
P
48.8
0.0
0.0
42.9
0.0
0.0
55.1
0.0
50.0
7.0
0.0
0.0
6.1
0.0
0.0
26.5
100.0
0.0
55.8
0.0
0.0
49.0
0.0
0.0
81.6
100.0
50.0
t Sample size is the "weighted" sample:
* Data coding errors.
("X" cars) + 3 x (non-"X" cars).
-26-
-------
Table 6 (continued)
Modal Tailpipe Failure Rates of
Undercover Vehicles that Failed at First Smog Check
vs. Vehicles that were Incorrectly Passed,
By Technology Group (1980 and Later Models Only)
Sample
Group
GARB/
3CL
Test
Location
Smog Check #1
ARE Lab
(Breezeway)
Screening
Facility
SampleT
Size
24
0
6
27
0
6
27
0
6
I/M Test
Smog #1
F
P
P
F
P
P
F
P
P
Results
Smog #2
F
P
.
F
P
F
P
Tailpipe
Failure Rates (%)
Idle
2.5
0.0
.5.6
10. 0
'4.1
2500
Only
0.0
0.0
0.0
0.0
3.7
Idle
and/ or
2500
62.5
0.0
55.6
50.0
77.8
00.0
0.0 100.0
GARB/
3CL/OXD
Smog Check #1
ARE Lab
(Breezeway)
Screening
Facility
221
15
87
237
18
87
238
18
87
F
P
P
F
P
P
F
P
P
F
P
-
F
P
F
P
40.7
0.0
0.0
35.4
16.7
10.3
50.0
33.3
17.2
27.2
0.0
*3.5
11.4
33.3
13.8
24.0
16.7
20.7
67.9
0.0
*3.5
46.8
50.0
24.1
74.0
50.0
37.9
FI/
/
3CL
Smog Check #1
ARE Lab
(Breezeway)
Screening
Facility
67
6
54
68
6
54
68
6
54
F
P
P
F
P
P
F
P
P
F
P
-
F
P
F
P
76.1
0.0
*5.6
48.5
0.0
38.9
86.8
50.0
55.6
4.5
0.0
0.0
4.4
0.0
0.0
0.0
0.0
0.0
80.6
0.0
*5.6
52.9
0.0
38.9
86.8
50.0
55.6
f Sample size is the "weighted" sample:
* Data coding errors.
("X" cars) + 3 x (non-"X" cars).
-27-
-------
100
80
O)
I 6°
u_
Figure 10
Percent of "Should Fail" Vehicles
Failing at Smog Check #1
(1980+ Tech Groups w/Sample Size >30)
o
©
CL
40
20
60
50
CARB/OC/AIR
CARB/3CL/OC
Technology Group
FI/3CL
Underhood Inspection Variability - It is apparent from the preceding
subsection that significant differences in tailpipe emission
measurements occurred at the three different inspection sites.
However, it is also clear from the data that most of the vehicles that
passed at Smog Check station #1, after failing at the screening
facility or the ARE laboratory, must have been underhood-only
failures. The percentage of the vehicles that were underhood-only
failures has now been precisely computed and is shown in Figure 11.
-28-
-------
Figure 11
100
8 80
o
w
CO
40
0
O
| 20
Percent of Vehicles Missed at Smog Check #1
That Were Underhood-Only Failures
81-880.3 80.6
76.0
52.1
55.4
Screening Facility
ARB Lab Breezeway
Pre-1975 1975-1979
Model Year Group
1980+
Based on tests at the screening facility, 52.1% of the 1980 and later
model vehicles that passed the test at Smog Check station #1 were
underhood-only failures. Based on tests at the ARB laboratory, almost
the same portion (55.4%) of these vehicles were underhood-only
failures.
-29-
-------
Figure 12 indicates that the Smog Check stations did catch some of the
vehicles that contained only underhood defects. Of those vehicles
that failed at the screening facility, over 50% of the pre-1980 models
contained only underhood defects. For 1980 and later models, the
underhood-only failures were just about 35% of the failing vehicles.
As the figure shows, less than half as many underhood-only failures
occurred at the first Smog Check station.
Figure 12
Percent of I/M Test Failures
Caused by Underhood Defects Only
Pre-1975 1975-1979
Model Year Group
Smog Check #1
Screening Facility
1980+
-30-
-------
Emission Reductions Table 7 shows the emission reductions that were
recorded for vehicles repaired at Smog Check stations. (Note that no
data are available for vehicles which passed at both Smog Check
stations because they were never repaired.) The data contained in
Table 7 have been translated into grams per mile changes, and both the
percent reductions and mass emission changes are illustrated for each
model-year range in Figures 13, 14, and 15. As the figures show,
model-year range seems to have a significant effect on the emissions
reductions achieved for vehicles that do not fail until the second
Smog Check station. On both a percent reduction and a mass emissions
Table 7
Baseline FTP Emissions and Average Emission Reductions of
Undercover Vehicles that Failed at First Smog Check
vs. Vehicles that were Incorrectly Passed
By Model Year Range
Sample Sample' I/M Test
Size Results
Baseline FTP Emissions
(g/mi)
HC CO NOx
Emissions Reductions
(in %)
HC CO NOx
All
Vehicles
Pre-1975
1975-1979
1980+
1532
129
573
422
33
108
583
57
213
527
39
252
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
5.09
3.43
1.91
10.39
8.06
4.92
4.23
2.40
1.66
1.80
1.04
0.82
48.41
34.89
27.54
69.99
58.52
71.88
49.02
36.40
21.78
30.44
12.68
13.40
2.37
2.70
2.17
3.30
3.78
3.06
2.74
2.84
2.66
1.21
1.58
1.37
32.5
29.9
"
36.8
53.5
"
25.9
3.8
"
30.2
-37.2
18.6
-5.8
"
9.2
10.8
20.5
-1.4
32.4
-89.1
8.5
12.0
9.4
18.2
9.8
3.5
3.3
21.9
t Sample size is the "weighted" sample: ("X" cars) + 3 x (non-"X" cars)
-31-
-------
basis, pre-1975 models that fail at the second Smog Check station
achieve just as great an emissions reduction as those that fail the
first time. However, Figure 14 indicates that 1975-1979 models that
pass the first Smog Check receive no significant benefit from repairs
performed at the second Smog Check station. Figure 15 indicates that
1980 and later models exhibit yet another trend. They obtain
significant NOx emission reductions when they are repaired after
failing at the second Smog Check station, but HC and CO emissions are
higher after repair.
The significant emission reductions achieved with pre-1975 vehicles
that do not fail until the second Smog Check might be expected to
result from the correction of underhood-only defects such as air
injection tampering and EGR disconnects. A greater percentage of EGR
problems could cause the simultaneous HC and CO increases and NOx
reductions observed after repair of the 1980 and later models. Review
of the individual vehicle data indicates that such defect
identification trends did, in fact, occur. Half of the 1980 and later
models which had underhood-only defects identified at the second Smog
Check station had EGR tampering corrected. The other half had minor
problems (such as disconnected heat stoves) identified and corrected
that would not be expected to contribute to significant emission
reductions. However, the sample sizes are just too small to draw any
firm conclusions about model-year range differences.
-32-
-------
Figure 13
Change in Emissions Due to Repairs
Pre-'75 Vehicles Failing 1st Smog Check vs.
Vehicles Passing 1st Check, Failing 2nd
Increased Emissions
Faffing 1st Check
Pass 1st, Fail 2nd
Increased Emissions
Failing 1st Check
Pass 1st, Fail 2nd
-33-
-------
Figure 14
Change in Emissions Due to Repairs
T5-'79 Vehicles Failing 1st Smog Check vs
Vehicles Passing 1st Check, Failing 2nd
ou
tn 40
c
0
CO
CO
'£ 9O
c *«
UJ
,c
CD
rn O
CO
O
k* C7O\
C V^U'
CD
O
&
Q_ f4T)\
V^^i
/Rn\
~
Increased Emissions
i
A
1
: ^^ ^^
^1 , -10
-26
Reduced Emissions
HHC
W$$ /~*if\
mm co
::: NOX
Failing 1st Check
Pass 1st, Fail 2nd
. .
O
-V,
O
O
5
_^
-------
Figure 15
Change in Emissions Due to Repairs
1980+ Vehicles Failing 1st Smog Check vs
Vehicles Passing 1st Check, Failing 2nd
60
40
(0
(0
"£ 20
LU
o
(0
£ (20)
s.
(40)
(60)
-Increased Emissions
i 28
T
Reduced Emissions
-19
Failing 1st Check
Pass 1st Fail 2nd
HC
CO
::: NOX
O
O
o
.E (2)
CD
D)
O
(3)
(4)
(5)
Increased Emissions
-0765"
'1-0.05
-0.32
ReducecTErnissions
Failing 1st Check
Pass 1st, Fail 2nd
CO
NOx
-35-
-------
Similar problems with small sample size frustrate analysis of
technology-specific differences in achieved emission reductions when
comparing vehicles that failed the first Smog Check with those that
passed the first Smog Check and failed the second. The results for
seven technology categories are shown in Table 8.
Table 8
Baseline FTP Emissions and Average Emission Reductions of
Undercover Vehicles that Failed at First Smog Check
vs. Vehicles that Incorrectly Passed
By Technology Group (1980 and Later Model Only)
Tech. Sample' I/M Test
Group Size Results
Baseline FTP Emissions
(g/mi)
HC
CO
NOx
Emissions Reductions
(in %)
1C CO NOx
No
Catalyst
GARB/
OXD/NoAIR
GARB/
OXD/AIR
GARB/
3WAY/OXD
GARB/
3CL
GARB/
3CL/OXD
FI/
3CL
12
0
3
19
9
24
110
0
72
48
3
3
29
0
6
240
21
87
69
6
57
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
2.43
-
2.23
0.54
0.63
0.49
2.16
-
0.99
1.82
1.95
1.02
1.75
-
0.33
1.53
1.25
0.77
2.38
0.43
0.80
32.29
-
20.18
7.82
6.08
4.83
35.12
21.06
32.19
31.75
20.18
28.60
-
5.04
25.23
14.91
13.42
46.59
5.25
7.48
1.76
2.00
1.50
0.97
1.09
1.33
-
1.63
1.26
1.29
0.57
0.96
0.71
1.13
1.84
1.25
1.17
1.71
1.40
12.3
--
2.9
10.9
28.4
6.7
34.6
37.9
32.1
-58.6
43.2
-86.9
--
14.0
--
11.2
-8.6
21.5
21.6
28.3
47.1
34.1
-143.5
47.1
-43.7
--
0.0
--
17.6
8.4
-13.3
--
3.0
6.0
12.7
--
11.5
23.0
--
-1.4
35.5
f Sample size is the "weighted" sample: ("X" cars) + 3 x (non-"X" cars)
-36-
-------
Effect of 207Cb) standards - Table 9 shows how the federal 207(b)
standards (220 ppm HC, 1.2% CO) failure rates for 1980 and later model
year vehicles at the various test sites compare to the failure rates
for California standards. As expected, fewer vehicles would have
failed if the less stringent 207(b) standards had been used.
Table 9
Comparison of California and 207(b) Failure Rates
for Failed Vehicles Taken to Smog Check Stations
Site
Sample' I/M Test
Size
Results
Failure Rate
California Standards 207(b) Standards
Idle Idle
Idle 2500 and/or Idle 2500 and/or
Only Only 2500 Only Only 2500
Smog
Check #1
ARE
Lab
Screening
Facility
490
30
258
519
33
261
514
33
264
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
Fail/N.A.
Pass/Fail
Pass/Pass
50.8
0.0
*1.2
40.5
9.1
17.2
57.0
27.3
26.1
i
20.8
0.0
*1.2
12.1
18.2
6.9
21.0
18.2
10.2
71.6
0.0
*2.3
52.6
27.3
24.1
78.0
45.5
36.4
i-
41.0
0.0
0.0
35.3
0.0
6.9
41.1
18.2
14.8
22.0
0.0
1.2
12.7
18.2
6.9
24.3
18.2
12.5
63.1
0.0
1.2
48.0
18.2
13.8
65.4
36.4
27.3
f Sample size is the "weighted" sample: ("X" cars) + 3 x (non-"X" cars)
* Data coding errors.
-37-
-------
Table 10 shows what the effect of the 207(b) standards would have been
on the average emissions of failing vehicles and the emission
reductions achieved as a result of repair. (Vehicles that failed
based on underhood inspection defects are included in both samples,
regardless of their emission levels.) As the table shows, the average
emissions of the vehicles failing the 207(b) standards are somewhat
higher but the percent reduction resulting from repair is almost
identical to the vehicles failing the California standards. Note,
however, that the total emission reduction achieved with the
California standards is 8.7% higher for HC, 1.9% higher for CO, and
51.7% higher for NOx. This is due to the fact that more vehicles fail
under the California standards (159 vs. 131).
Before I/M
After Repair
Reduction
Number of
Failing Vehicles
Increased Mass
Emissions
Table 10
Effect of Emission Standards
on I/M Emission Reductions
(1980 and Later Models)
California Standards
HC CO NOx
1.
1.
31
89
29
.5%
32.
21,
34
-- 1 '
,2
.2
.1%
59 --
1,
1
3
.25
.20
.7%
2.
1.
31
11
44
.9%
207(b) Standards
HC CO NOx
37.1 1.32
24.0 1.28
35.4% 3.3%
131
8.7%
1.9% 51.7%
* X sample cars did not receive I/M repair since they did not fail at
Smog Check #1. This table contains only "F" car reductions.
-38-
-------
Finally, Table 11 indicates that the more stringent California idle
outpoints contributed to the inconsistencies that were observed. For
vehicles that failed at the screening facility, a greater percentage
passed at the ARE breezeway and/or the first Smog Check station under
the California standards. The results presented in Table 11 are
consistent with the distribution of the idle hydrocarbon emissions
shown earlier in Figure 4. That figure showed that a significant
number of vehicles had emissions just slightly above 100 ppm HC at the
screening facility.
Table 11
Percentage of Inconsistent
Tailpipe Test Results
for 1980 and Later Models
Type of Inconsistency
S/B S/B/FT
California Stds.
Idle 21.9 31.6
2500 16.0 25.4
Idle and/or 2500 46.5 49.5
Fed. 207(bl Stds.
Idle 13.3 20.3
2500 16.4 26.1
Idle and/or 2500 27.0 39.5
Note: S = Screening facility.
B = ARB breezeway.
F - Field inspection (Smog Check #1).
*Vehicles with inconsistencies between screening facility and
ARB breezeway, irrespective of Smog Check #1 results
^Vehicles with inconsistencies between any two test locations
-39-
------- |