EPA-AA-IMS/81-13
Recommendations Regarding the
Selection of Idle Emission Cutpoints
for Inspection and Maintenance Programs
Requiring Only Carbon Monoxide Emission Reductions
by
Susan Vintilla
Inspection/Maintenance Staff
May, 1981
Notice
This Report does not necessarily represent final EPA decisions or positions.
It is intended to present technical analysis of the issue using data which are
currently available. The purpose in the release of such reports is to
facilitate the exchange of technical information and to inform the public of
technical developments which may form the basis for a final EPA decision,
position or regulatory action.
Inspection and Maintenance
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
-------
Preface
This report is a supplement to EPA-AA-IMS/81-1, "Recommendations Regarding the
Selection of Idle Emission Inspection Cutpoints for Inspection and Maintenance
Programs"[1]*, which was developed as a part of EPA's Model Program guidance.
The original report described methods for selecting HC and CO cutpoints given
an idle emission data base, and recommended specific HC and CO cutpoints by
model year groupings for desired program failure rates. The recommended
cutpoints in the original report were intended for use in I/M programs in
areas which require only HC, or both HC and CO, emission reductions in order
to attain the National Ambient Air Quality Standard for ozone, or ozone and
CO. This report Is intended to aid I/M programs which only require CO
reductions in the selection of cutpoints, prediction of failure rates, and
estimation of I/M emission benefits. Its recommendations are therefore
applicable to I/M in areas that .have -already attained the ozone standard or
that are projected to attain the ozone standard by 1982 without I/M.
^Numbers in brackets refer to a list of references at the end of the report.
-------
1.0 INTRODUCTION
1.1 Previous EPA Outpoint Selection Guidance
The recommendations presented in Reference 1 for pre-1981 vehicles were
derived froa a data base of vehicles which had undergone testing of idle HC
and CO emissions. Failure rates for given cutpoints were predicted for 5
different technology types corresponding to model year groups (pre-1968,
1968-1971, 1972-1974, 1975-1979 and 1980) by a computer model which accounted
for the increase in idle HC and CO scores with mileage. Failure rates for
1981 and future vehicles were predicted differently, due to their more
sophisticated emission control technology. A high proportion of these
vehicles will have computerized control of the air-fuel mixture, which is
expected to make them less sensitive to cutpoint variations.
Failure rates for each model year grouping in the previous report were
presented as the total percent of tested vehicles which fail the idle test for
HC and/or CO, given certain sets of HC and CO cutpoints. This report will
recommend CO cutpoints for desired failure rates for CO idle tests only. (As
explained below, EPA has determined that HC cutpoints are not crucial for
deriving a large CO-only emission benefit.) Naturally, some of these vehicles
would also fail HC idle tests but it is not the purpose of this report to
detail those failures. Failure rate predictions in this report will only
apply to the first functioning year of an I/M program; as explained in
Reference 1, failure rates for the following years can be estimated by -data
collected from the actual operation of the program.
1.2 Major Points Discussed In This Report
The following questions will be answered:
1) What, if any, should be the cutpoint locus; i.e., should there be an HC
cutpoint in addition to a CO cutpoint? If so, what multiple of the CO
cutpoint should it be?
2) How can states predict failure rates from specific CO cutpoints?
3) How can the emission benefits be calculated?
-------
2.0 SHOULD AN HC OUTPOINT BE USED?
2.1 Pre-1981 Vehicles
EPA's investigation of CO benefits from HC-only failures of pre-1981 vehicles
has shown that they typically amount to a small percent of total idle test
failures, and that they receive essentially no CO emission benefit from
repair. The lack of a CO benefit is due to the fact that these vehicles' FTP
CO emission levels at inspection are already low (as their ability to pass the
I/M CO outpoint would suggest) and the fact that no CO-oriented repairs need
to be performed for these vehicles to pass reinspection. Table 1 shows the
as-received FTP CO emissions and the change in CO emissions for vehicles from
repair of HC-only failures in EPA's Portland Study Sample.
Table 1
FTP CO Emission Change Resulting
From Maintenance On HC-Only Failures From
-v Elements II and III and Test Group 7
of the Portland I/M Fleet
As-received
Percent of idle test FTP CO Percent change in FTP CO
Vehicle failures which are emissions emissions from HC-only
model year HC-only failures (g/mi) failures after repair
1972-1974 6.3 27.98 - 2.2
1975-1977 7.4 16.70 +22.2
As can be seen, there was actually a small increase (3.7 g/mi) in CO emissions
after repair for the 1975-1977 model ..year group.
2.2 1981 and Future Vehicles
Emission control system failures of the 1981 and future fleet can be divided
into two categories: 1) failures of the closed-loop computer control system,
which result in rich operation and high HC and CO emissions 2) other failures
which are also common to older vehicles, caused by tampering, malmaintenance,
misfueling, and ignition-related problems. As explained in Reference 1, it is
the failures of the first category which are expected to be the main cause of
1981 and later vehicles emitting significantly above the standard. The
emission factors for this failure category, as predicted in MOBILE2 and. shown
in Table 2, demonstrate that both HC and CO emissions will be exceptionally
high [4]. A CO-only cutpoint is expected to identify essentially all of these
vehicles [3]. Failures of the second category will be similar to those of
pre-1981 vehicles, and so the same arguments for the suitability of a CO-only
cutpoint are applicable.
-------
5
Table 2
. ' Emission Rates for 1981 and Future Vehicles Which
Experience Failure of the Closed-Loop Control System
New Vehicle
Pollutant Emission Rate (g/mi)
HC 3.74
CO 107.36
2.3 Undesirable Consequences of Emphasizing HC Idle Failures
EPA has found that there are certain characteristics associated with HC
failures on the idle test which areas that need only CO reductions from the
I/M program would logically wish to avoid if possible. Cutpoint sets for HC
and CO which emphasize HC, for example, are accompanied by higher rates of
errors of commission (vehicles which fail the idle test but actually meet
federal emission standards). Also, idle failures for HC often necessitate
cftstly ignition repairs, which could needlessly drive up the average repair
cost in the I/M program with little resulting CO emission reduction. Yet
another issue related to the choice of idle inspection type is that of fuel
economy benefit from repair of failed vehicles. EPA has found that there is a
net fuel economy benefit of 0.8% for pre-1981 vehicles repaired by mechanics
who have been educated in conventional training programs. However, this fuel
economy benefit rises to 4% for these same vehicles when mechanics only adjust
carburetors to a specific idle CO level [2]. This would indicate that an I/M
program where repairs for pre-1981 vehicles are limited to carburetor adjust-
ments would yield the maximum CO and fuel economy benefit, at low cost. By
using only a CO outpoint, an I/M program can effectively limit repairs to
carburetor adjustments [3]. (The fuel economy benefit for 1981 and later
vehicles is not expected to be influenced by mechanic training [2].)
In summary, a program which consists of a CO-only idle inspection will deliver
full CO emission reduction benefits, and may actually deliver a greater fuel
economy benefit for pre-1981 vehicles since repairs can be limited to
carburetor adjustments. Also, a CO-only program operating at a particular
failure rate will deliver a greater CO benefit than an HC/CO program operating
at Chat same failure rate [3].
Additional discussion of the ramifications of using only a CO cutpoint can be
found in Reference 3.
-------
3.0 HOW CAN STATES PREDICT FAILURE RATES FROM SPECIFIC CO OUTPOINTS?
As discussed in Reference 1, EPA recommends using data from a pilot program or
a mandatory inspection/voluntary testing program in the intended geographic
location, for the most accurate estimation of failure rates. For situations
where this is not possible, EPA developed a model to predict failure rates for
various cutpoints, based on data collected from EPA's I/M Demonstration Vans,
EPA's Emission Factor Program and New Jersey's I/M program [1], Tables 3 and
4 present predicted minimum, maximum and mean failure rates for CO-only
cutpoints, based on this model. Table 3 is 'for programs beginning 1/1/82;
Table 4 for those starting a year later. The minimum is the failure rate for
the most recent and therefore the youngest and cleanest model year in the
group; the maximum is the failure rate for the oldest and therefore the
dirtiest model year; and the mean is the average for all the model years in
the group, taking the national average, registraton distribution of model years
within model year groups at the projected time into account.
Based on these tables, EPA has estimated the CO cutpoints that would be needed
at the beginning of an I/M program for a variety of failure rates. Table' 5
presents these cutpoints for I/M programs beginning 1/1/82 and 1/1/83. As
dxscussed in Reference 1, failure rates for pre-1981 vehicles should remain
approximately constant during the first year of a program if vehicle
inspections fall on vehicle birthdays. However, EPA's model predicts that
vehicle model year group failure rates at the second 'inspection will decrease,
for example, 4-14% for a given set of cutpoints designed to achieve an initial
failure rate of 35% [1]. Also, the data base used for prediction of 1980
model year failure rates was limited, and so these predictions may be
inaccurate [1]. Therefore, states will probably need to adjust cutpoints for
pre-1981 vehicles yearly if they wish to maintain a constant failure rate.
Also, it may be necessary for a state to revise cutpoints within the first few
months of the start of a program, to correct for differences caused by local
influences. • ' '
Failure rates for 1981 and later vehicles will be less sensitive to cutpoint
variations, due to the nature of "the typical failure of their computer-
controlled emission control systems. As mentioned above, data from these
vehicles indicate that these failures will cause a gross increase in CO
emissions, which will be easily identified by any reasonable CO cutpoint.
-------
Table 3
Initial Model Year Group Failure Rates
Predicted by EPA CO Outpoint Model
At I/M Program Start Date 1/1/82
Model Year Group Failure Rates (%)
CO Idle
CutpoinC (%)
Pre-1968
Mini/ Max2/ Mean3/
1968-71
Min Max Mean
1972-74
Min Max Mean
1975-79
Min Max Mean
1980 4/ 1981+ 5/
Mean
1.0
1.2
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
10.0
Registration
Fraction
64.3
60.9
55.3
49.6
43.9
39.1
36.3
30.7
25.0
20.5
15.8
11.1
1.8
64.6
61.5
55.3
51.8
45.1
39.8
36.9
31.8
26.2
22.2
17.8
13.5
2.5
.038
64.4
61.2
55.3
.50.7
44.4
39.6
36.6
31.4
25.7
21.4
16.8
12.4
2.2
58.4
53.4
48.9
43.0
37.2
32.7
27.2
22.1
19.6
14.9
11.0
58.4
54.1
50.0
44.4
37.8
33.9
28.6
24.3
20.4
15.9
13.3
.110
58.4
53.9
49.0
43.9
37.8
33.0
27.8
23.2
19.7
15.3
11.9
74.4
69.2
62.6
55.7
51.0
45.4
40.8
36.2
31.6
26.3
22.1
18.6
15.5
11.9
9.5
77.4
71.6
65.1
58.5
53.6
48.3
43.6
47.7
35.0
29.2
25.0
21.0
17.4
14.7
11.5
.184
75.7
70.5
63.8
57.0
51.9
46.8
42.0
47.3
32.8
27.8
23.6
19.7
16.2
13.1
10.4
46.3
39.6
34.9
30.8
27.2
. 23.7
20.6
17.3
14.5
12.6
10.4
8.7
7.0
5.7
4.1
65.2
47.5
42.3
37.9
33.5
30.2
26.7
23.7
20.9
18.3
15.4
13.3
11.6
9.9
8.4
.449
53.5
44.0
38.7
34.5
30.6
27.2
23.9
20.9
18.0
15.4
13.0
11.2
9.5
7.9
6.4
33.8
1.0-3.0
30.0
25.6
--21. 7
18.7
14.9
12.7
10.5
8.6
6.6
5.2
3.7
2.7
1.9
1.0
.095 .124
(1) Minimum model year failure rate within model year groups.
(2) Maximum model year failure rate within model year groups.
(3) Average failure rate in model year groups weighted by projected model year registration fractions.
(4) EPA predictions for 1980 vehicles are baaed on limited data and therefore may be less accurate
than those for pre-1980 vehicles.
(5) EPA recommends use of a 207(b) short test standard of 1.2% CO.
EPA expects that the failure rate from these cutpoints will never exceed 1-3%.
-------
Table 4
Initial Model Year Group Failyre Rates
Predicted by EPA CO Cutpoin't Model
At I/M Program Start Date 1/1/83
Model Year Group Failure Rates (%)
Outpoint (%)
CO Idle
Pre-1968
Mini/ Max2/ Mean3/
1968-71
Min Max Mean
1972-74
Min Max Mean
1975-79
Min Max Mean
1980 47 1981+ 5_/
Mean
1.0
1.2
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5 •'
6.0
6.5
7.0
7.5
8.0
8.5
10.0
Registration
Fraction
64.3
60.9
55.3
49.6
43.9
39.6
36.3
31.4
25.4
21.1
16.2
11.7
2.1
64.6
61.5
55.3
51.8
45.1
39.8
36.9
31.8
26.2
22.2
17.8
13.5
2.5
.027
64.5
61.3
55.3
51.2
44.7
39.7
36.7
31.7
26.0
21.8
17.2
12.4
2.4
58.4
54.1
48.9
44.4
37.8
32.7
27.8
23.1
19.6
15.3
11.9
58.4
54.1
50.0
45.1
38.7
33.9
28.6
24.5
20.4
17.0
13.3
.084
58.4
54.1
49.2
44.5
38.2
33.2
28.1
23.8
19.9
15.7
12.4
75.7
71.1
64.2
57.2
51.4
47.3
41.9
38.1
32.4
28.2
24.1
19.9
16.2
13.2
10.4
79.1
72.9
67.0
60.2
54.9
49.5
44.5
40.1
36.1
30.0
25.7
21.5
18.6
15.1
11.9
.152
77.1
71.7
65.2
58.5
53.0
48.2
43.1
38.9
34.2
29.0
24.8
20.7
17.2
14.2
11.1
48.1
42.3
37.2
33.3
29.1
25.6
• 22.2
19.1
16.5
13.9
11.9
10.1
8.4
6.6
5.5
67.1
50.6
44.1
38.6
34.5
31.7
28.0
25.0
21.7
19.1
16.5.
14.2
12.6
10.5
9.2
.447
56.9
45.8
40.2
35.7
31.8
28.5
25.1
22.1
19.1
16.5
14.0
12.1
10.4
8.7
7.2
35.8
1.0-3.0
32.0
27.4
24.6
20.9
18.2
14.5
12.6
10.5
8.7
7.1
5.7
4.2
3.0
2.2
.073 .217
(1) Minimum model year.failure rate within model year groups.
(2) Maximum model year failure rate within model year groups.
(3) Average failure rate in model year groups weighted by projected model year registration fractions,
(4) EPA predictions for 1980 vehicles are based on limited data and therefore may be less accurate
than those for pre-1980 vehicles.
(5) EPA recommends use of a 207(b) short test standard of 1.2% CO.
EPA expects that the failure rate from these cutpoints will never exceed 1-3%.
-------
Table 5
EPA Recommended I/M CO Outpoints
Outpoints predicted to give constant failure rates among pre-1981 vehicles for I/M starting
1/1/82 and 1/1/83 in initial year inspections by EPA CO cutpoint model.
I/M
Start
Date
1/1/82
1/1/83
Nominal
Failure
Rate (%)
20
25
30
35
20
25
30
35
Pre-1968
CO (%)
7.5
7.0
6.5
6.0
7.5
7.0
6.5
6.0
Recommended
1968-71
CO (Z)
7.0
6.5
6.0
5.5
7.0
6.5
6.0
5.5
CO Outpoints
1972-74
CO (%)
6.5
6.0
5.5
5.0
6.5
6.0
5.5
5.0
for Model
1975-79
CO (%)
4.5
4.0
3.0
2.5
5.0
4.0
3.5
2.5
Year Groups
1980 I/
CO (%)
3.0
2.0
1.5
1.0
3.0
2.5
1.5
1.0
198H- 2/
CO (%)
1.2
- 1.2
1.2
1.2
1.2
1.2
1.2
1.2
If EPA1s predictions for 1980 vehicles are based on limited data and therefore may be less accurate
than those for pre-1980 vehicles. .
21 All I/M programs should use the 207(b) cutpoint for 1981 and later vehicles for all program years,
that the failure rate from this cutpoint will never exceed 1-3%.
EPA expects
-------
10
4.0 HOW CAN THE EMISSION BENEFITS BE CALCULATED?
The MOBILE2 emission factors model can be used, with modifications. Orice a
state has decided on the failure rate desired, it must choose the appropriate
CO outpoints for each pre-1981 model year group based on data from the local
vehicle sample, or based on the recommendations in this report. These CO
cutpoints should then be matched to appropriate hypothetical HC cutpoints from
EPA's locus pairs as described in Table 1 of Reference 1. These HC/CO
cutpoint pairs should then be applied to the local data sample to find a total
failure rate for pre-1981 vehicles; or they should be used to look up a total
failure rate among pre-1981 vehicles in Tables 3 or 4 of Reference 1, by
weighting the failure rate for each pre-1981 model year group by its
registration fraction. This failure rate should then be used in MOBILE2 as
the stringency for the I/M program for pre-1981 vehicles. For 1981 and later
vehicles, the effect of an I/M program is calculated by inputting an
identification (ID) rate instead of a stringency rate. The ID rate describes
the fraction of gross HC and CO emitters (caused by closed-loop system
failure, as described in Reference 4) which will be identified by the I/M
test. If the basic idle-in-neutral test will be used for an area's I/M
program, then a 50% ID rate should be inputted into MOBILE2. (This ID rate is
th"e default value when the input flag IMFLG=1, therefore, only the pre-1981
stringency value needs to be entered in this case.) .If the I/M program will
use a two—speed idle or loaded mode test and will pass and fail on both modes
of the test, then an ID rate of 70% should be entered for 1981 and later
vehicles.
Presented below is a sample calculation for a hypothetical area which plans a
CO-only I/M program and which will use the idle-in-neutral test. The area
wants a failure rate of 25% for the first year of its I/M program, which will
begin on January 1, 1983. Based on Table 5 of this report, the following CO
cutpoints should be used for the first year of the program for pre-1981
vehicles:
Vehicle model year CO idle cutpoint, %
pre-1968 7.0
1968-1971 6.5
1972-1974 6.0
1975-1979 4.0
1980 2.5
These CO cutpoints are then matched to their hypothetical -EC cutpoints from
Table 1 of Reference 1, to yield the following pairs:
Vehicle model year HC (ppm) C0(%)
pre-1968 700 7.0
1968-1971 650 6.5
1972-1974 600 6.0
1975-1979 400 '• 4.0
" 1980 275 2.5
-------
11
The HC/CO cutpoint pairs are then used to look up the failure rates for each
model year group in Table 4 of Reference 1. The failure rates are then
weighted by the registration fraction of each model year group. Note that the
registration fractions listed in Table 4 must be normalized to account for the
absence of the 1981 and later vehicles.
Mean Registration
Vehicle model year Failure rate fraction
pre-1968 . 52.0 .027/.783 = .035
1968-1971 38.9 .084/.783 = .107
1972-1974 34.5 .152/.783 = .194
1975-1979 37.1 .447/.7S3 = .571
1980 35.2 .073/.783 = .093
The failure rates are then multiplied by their respective registration
fractions and summed to yield the total theoretical failure rate for pre-1981
vehicles:
Mean failure Registration
rate fraction
52.0 x .035
38.9 x .107
34.5 x .194
37.1 x .571
35.2 x .093
Total failure rate 37.12
Therefore, a stringency of 37 will be used for pre-1981 vehicles in the
MOBILE2 program, and the default ID rate of 50% will be used for 1981 and
later vehicles. Comparison of the CO composite emission factor from this
MOBILE2 run with the CO composite emission factor from a "no-I/M" run will
enable the user to calculate the CO emission benefit for this particular
program. (The HC emission factors from these MOBILE2 runs are meaningless,
and should be discarded.)
-------
12
5.0 CONCLUSIONS
CO cutpoints are sufficient to obtain essentially all of the normal CO
emission reduction benefit possible from an I/M program, and a CO-only program
operating at a given failure rate will deliver a greater CO benefit than an
HC/CO program operating at that same failure rate. Cutpoints needed for a
desired initial program failure rate can be chosen from the recommendations in
this report, or can be determined from analysis of the results of a local
trial I/M program. Cutpoints for future program years may need to be altered
to insure a constant failure rate, and cutpoint selection should be based on
results from the first year of the program. CO emission benefits can be
calculated from MOBILE2, but not in the same manner as for an HC/CO I/M
program. The chosen CO cutpoints for each pre-1981 model year group are
matched to hypothetical HC cutpoints from the recommended locus pairs in
Reference 1; these pairs are used to calculate a total failure rate for
pre-1981 vehicles which is input as the program stringency in MOBILE2.
Effectiveness of the I/M program for 1981 and later vehicles is represented in
the input by an ID rate of 50 or 70 percent, depending on the type of emission
test used. CO emission factors from MOBILE2 are used to calculate emission
benefits, and the corresponding HC emission factor outputs are ignored.
-------
13
References
[1] U.S. EPA. January, 1981. Recommendations Regarding the Selection of Idle
Emission Inspection Outpoints for Inspection and Maintenance Programs.
EPA-AA-IMS/81-1.
[2] U.S. EPA. April, 1981. Update on the Fuel Economy Benefits of Inspection
and Maintenance Programs. EPA-AA-IMS/81-10.
[3] U.S. EPA. April, 1981. Low-Cost Approaches to Vehicle Emissions
Inspection and Maintenace. EPA-AA-IMS/81-7.
[4] U.S. EPA. November, 1980. Derivation of 1981 and Later Light Duty
Vehicle Emission Factors for Low Altitude, Non-California Areas.
EPA-AA-IMS/80-8.
------- |