EPA-420-R-73-104
Calspan
On November 17, 1972 Cornell Aeronautical Laboratory (CAL) changed its name to Calspan Corporation and converted to
for-profit operations. Calspan is dedicated to carrying on CAL's long-standing tradit on of advanced research and development
from an independent viewpoint. All of CAL's diverse scientific and engineering programs for government and industry are being
continued in the aerosciences, electronics and avionics, computer sciences, transportation and vehicle research, and the environ-
mental sciences. Calspan is composed of the same staff, management, and facilities as CAL, which operated since 1946 under
federal income tax exemption.
A UTOMOBILE EXHA UST EMISSION SUR VEILLANCE
P. J. Donovan
Calspan Report No. NA-5194-D-4
Prepared For:
ENVIRONMENTAL PROTECTION AGENCY
DIVISION OF CERTIFICATION AND
SURVEILLANCE
ANN ARBOR, MICHIGAN
NOVEMBER 1973
FINAL REPORT
Calspan Corporation
Buffalo, New York 14221
-------
TABLE OF CONTENTS
Section Title Page No.
TEXT . 1
1 SUMMARY AND CONCLUSIONS 1
1.1 Summary 2
1.2 Conclusions. 4
2 RESULTS AND DISCUSSION. . 10
2.1 Results of the Great Plains Surveillance Program ... 12
2.1.1 1968-1969 Surveillance Results 13
2.1.2 Vehicle Maintenance Results 14
2.1.3 Evaluation of Short Test Cycles 16
2.1.3.1 Hot-Idle Mode 17
2.1.3.2 Clayton KEY MODE Diagnostic Test ... 18
2.2 Results of the National Surveillance Program 21
2.3 A Study of Emissions from Light Duty Vehicles in
Six Cities - 1957-1971 Model Year Survey 22
2.4 Rental Vehicle Surveillance Program 24
2.5 In-House Vehicle Surveillance Program 25
2.6 N0X Baseline Program . 28
2.6.1 Background 28
2.6.2 Analysis 30
TABLES 31
FIGURES 41
ii
-------
Tab
1
2
3
4
5
6
7
8
9
10
11
32
33
34
35
36
37
37
38
38
39
40
LIST OF TABLES
Title
Summary of 7-Mode and CVS Test Programs .......
Summary of Emissions (Arithmetic Means) for 1968 and
1969 Model Vehicles
Summary of Emissions (Arithmetic Means) for 1970 Model
Vehicles
Summary of Emissions (Arithmetic Means) for 1971 Model
Vehicles
Summary of 7-Mode Emission Tests - 1968-1971 Model
Vehicles
Composite Emission Levels as Determined by 1972 Test
Procedures (Excluding Denver)
Composite Emission Levels as Determined by 1972 Test
Procedures (Denver Only)
Composite Emission Levels as Determined by 1975 Test
Procedures (Excluding Denver) .
Composite Emission Levels as Determined by 1975 Test
Procedures (Denver Only)
Comparative Mean Emission Levels - Rental Vehicle vs
1968-1969 Great Plains Surveillance Programs
Comparative Emission Levels - NOx Baseline Program vs
Six-City Program
iii
-------
LIST OF FIGURES
Figure Title Page No.
1 Histogram of Discriminant Factors for KEY MODE
DataPass Seven-Mode/Fail Seven-Mode 42
2 1975 CVS Emission Levels vs Cumulative Percentage of
Test Vehicles for Pre-Control, 1968-1969, and 1970-1971
Model Years (All Cities Excluding Denver) 43
3 Hydrocarbon Emissions vs Mileage - Five 1969 Model
In-House Surveillance Fords . 44
iv
-------
1. SUMMARY AND CONCLUSIONS
The purpose of this report is to coordinate information from several
programs dealing with the measurement of emissions from light-duty vehicles and
to make this information available in concise form. The report supplements the
previously issued publication entitled, "Automobile Exhaust Emission Surveil-
lance: A Summary." The two documents can be used to estimate the impact of
light-duty vehicle emissions on air quality and to assess the degree to which
vehicles in use conform to the standards under which they were certified.
The previously published report, hereinafter referred to for the
sake of convenience as APTD-1544, summarized the findings and results of three
major emission surveillance programs conducted by the Environmental Protection
Agency (EPA):
1. The Great Plains (Two-City) Surveillance Program, 1968-
1969 Model Year Surveys (actual testing was conducted from
October 1969 to October 1971).
2. The National Surveillance Program, 1970 (Six-City) and
1971 (Four-City) Model Year Surveys (actual testing was
conducted from November 1970 to January 1972).
3. A Study of Emissions from Light-Duty Vehicles in Six
Cities, 1957-1971 Model Year Survey (actual testing was
conducted from July 1971 to June 1972) .
The present report combines the findings of the above surveillance
programs with three additional sources of information:
4. The Rental Vehicle Surveillance ProgTam, March 1968 to
January 1970 (actual testing conducted from March 1968
to January 1970).
5. The In-House Vehicle Surveillance Program, January 1967
to March 1970 (actual testing conducted from January 1967
to March 1970).
1
-------
6. Determination of Exhaust Emissions from 1971 Model
Vehicles (commonly referred to as the NO^ Baseline data)
(actual testing conducted from March 1971 to April 1971).
Except for 3 and 6, both of which utilized the Constant Volume Sam-
pling (CVS) Federal Test Procedure, all test programs employed the Federal
Seven-Mode Test Procedure. Although correlation factors have been developed
by EPA to relate the 7-mode and CVS Federal Test Procedures, they are based on
average results of a sales-weighted sample of many different vehicles and are
inappropriate for converting test results foT individual vehicles. Consequently,
for the purposes of this analysis, it was felt more appropriate to treat the
results of tests made by the two procedures separately, without endeavoring to
translate these results from one procedure to another.
1.1 SUMMARY
Hydrocarbon and carbon monoxide emissions for the vehicles tested in
three major surveillance programs (Programs 1, 2, and 3) were assessed by com-
paring the mean emission levels measured for a given model year with applicable
Federal standards. It is emphasized that such comparison is consistent with
the goals of pre-1972 regulations, which were aimed at insuring that average
vehicle emission levels meet standards. In general, hydrocarbon and carbon
monoxide emissions as measured by the CVS test procedure showed a significant
downward trend for late model vehicles; NO^ emissions, which weri; not subject
to control during the model years covered by the program, tended to show a
corresponding increase. Hydrocarbon and carbon monoxide emission measures
for post-1968 model years showed approximately 50% and 35% reduction, respec-
tively, from emissions measured for pre-1968 model year pre-control vehicles.
On the same basis, NO emissions from post-1968 vehicles showed an increase
of approximately 35% over pre-control vehicles. It should be noted that these
test programs were conducted on as-received vehicles and that the test results
may reflect to some extent the influence of such variables as state of vehicle
maintenance and repair.
2
-------
In an effort to assess the extent to which local climate, terrain,
driving practices and other geographically differentiated factors affect emis-
sions, vehicles were sampled in several cities, these cities being selected to
span the range of such factors. Only small differences were observed in the
emissions measured in the cities included in the survey, the only notable
exception being Denver. Higher carbon monoxide and hydrocarbon emissions and
lower NO emissions were observed in Denver than in the other cities, presum-
ably because of the effect of altitude on air-fuel ratios.
The mean emission levels from 1968 and 1969 Great Plains vehicles
in use by the general public were compared to emission levels from similar
vehicles operated by rental vehicle companies to assess whether emissions
from fleet operated vehicles differ significantly from general use vehicles
and to provide an additional data source to complement the Great Plains sur-
veillance data. Mean emission levels from the rental vehicles were consistently
lower than comparable 1968 and 1969 vehicles in use by the general public.
This difference, however, is consistent with the age and the associated aver-
age mileage difference between the two classes of vehicles.* The rental vehi-
cles, with an average mileage of 7,700 miles, were tested during 1968 and 1969
while the Great Plains vehicles, with an average mileage of approximately
30,000 miles, were tested from 1969 to 1971.
Finally, the effect of maintenance was also investigated by comparison
of two sets of emission data--"as-received" low mileage vehicles and low mile-
age vehicles which were tuned prior to testing. Based upon these comparisons,
no significant maintenance effect was observed for HC and CO, as was expected,
because these vehicles had very low mileage. A small but significant reduction
in NO emissions, however, was observed on the maintained vehicles.
x
In summation, the Environmental Protection Agency has recognized
that any realistic assessment of Federal air pollution regulations requires
the monitoring of emissions from production vehicles in use by the motoring
public. Consequently, this report presents a summation of six emissions pro-
grams carried out from 1967 to 1972. Although no attempt has been made to
present all the data generated during the past several years, this report
does present synopses of major programs.
* "Automobile Exhaust Emission Surveillance -- A Summary", (APTD-1544)
3
-------
1.2 CONCLUSIONS
Automobile exhaust emission levels depend upon a variety of complex
factors which vary from vehicle to vehicle. Such variables as a vehicle's
accumulated mileage, state of tune or driving history all contribute to the
wide dispersion in emission levels commonly exhibited by vehicles of any
particular category. Consequently, two categories of vehicles, such as those
tested in two different cities, may show considerable overlap of their statis-
tical distributions even though the mean emissions for the two categories are
appreciably different. Generalizations with regard to make, city or other
categories of interest, therefore, are often not applicable to comparison of
individual vehicles or small subsets of vehicles drawn from the two categories.
On the other hand, care should be exercised in interpreting results drawn from
the application of statistical techniques to categories with large sample sizes.
By combining large quantities of data, it is possible to label as statistically
significant a small difference in mean emission levels despite the fact that
the magnitude of this difference may be too small to be of practical importance.
Finally, many of the test programs were conducted on as-received
vehicles whose statistical distribution of emission levels is frequently skewed
toward the high end of the distribution curve. If a hypothetical emissions
distribution were symmetrical and fifty percent of the vehicles met the standard,
the mean of all these vehicles would also meet the standard. This relationship
does not apply, however, with a skewed distribution. If an indication of total
mean emissions is desired and the emissions distribution in question is skewed,
the mean emission level of a group of vehicles must be looked at independently
of the percent of these vehicles which conform to the standard.
In what follows, conclusions are numbered to conform to the programs
from which they were primarily drawn. The numbering sequence conforms to the
previous listing of the several surveillance programs encompassed by this
report. With these guidelines in mind, the results of the automobile exhaust
emissions programs summarized in this report reveal the following:
4
-------
1. In the Great Plains Surveillance tests of 1968 and 1969 model
year vehicles in Kansas City and Houston, the following obser-
vations were made, based upon the 7x7 Federal Test Procedure.
a. 46% of the 1968 model year vehicles with an average mile-
age of approximately 34,000 miles individually complied
with applicable standards for HC; that is, 46% of the
vehicles exhibited emission levels at^ or below the level
allowable by the standard. Similarly, 35% were at or
below standards for CO, and 23% were at or below standards
for both HC and CO. Mean emission levels including Volks-
wagens for HC, CO and N0xc were 352 ppm, 2.14% and 1275
ppm, respectively, where NO refers to oxides of nitrogen
xc
corrected for humidity. It should be noted that Volks-
wagens were subject to standards different from those
applicable to other vehicles during model years 1968 and 1969.
b. 53% of the 1969 model year vehicles with an average mile-
age of approximately 24,000 miles were at or below appli-
cable standards for HC, 42% were at or below standards
for CO, and 30% were at or below standards for both HC
and CO. Mean emission levels including Volkswagens for
HC, CO and N0xc were 303 ppm, 1.86% and 1453 ppm,
respectively.
c. The mean emission levels for both the 1968 and 1969 model
year vehicles failed to meet the applicable standards of
275 ppm HC and 1.50% CO.
d. Based upon 192 tests, engine adjustments resulted in
reduction of 110 ppm in mean HC levels in Houston but no
change in Kansas City. A 0.2% reduction in mean CO was
observed in both cities. Replacing dirty air filters
resulted in no significant change in HC levels but reduced
CO emissions by 0.3% in both Houston and Kansas City.
5
-------
Results of the maintenance survey indicate that spark
plugs and breaker points were the most frequent items of
maintenance.
e. Analysis of two short test cycles, the Chew Hot-Idle Mode
and the Clayton KEY MODE Diagnostic Test, revealed the
following:
(1) Based upon simple linear correlation, neither short
test cycle is a reliable predictor of the emission
levels from the cold start seven-mode test procedure.
(2) Based upon a discriminant analysis approach on 2,038
Hot Idle tests on 1968-1969 Great Plains vehicles,
83% pass/fail agreement with the seven-mode procedure
was achieved. Of the 17% incorrect decisions, the
Hot-Idle test failed 10.4% that should have passed
and passed 6.6% which should have failed.
(3) Based upon a similar discriminant approach on 1749 KEY
MODE tests on 1970 model National Surveillance vehi-
cles, 83.5% pass/fail agreement with the seven-mode
procedure was achieved. Of the 16.5% incorrect
decisions, the KEY MODE test failed 11.6% that should
have passed and passed 4.9% which should have failed.
These conclusions, however, do not necessarily apply to the
current 1972 and 1975 Federal Test Procedures.
2. In the National Surveillance tests of 1970 and 1971 model year
vehicles, the following observations were made, based upon the
7x7 Federal Test Procedure.
a. In the surveillance tests of 1970 model year vehicles with
average mileages of approximately 10,000 miles in Kansas
City, Houston, Los Angeles, Detroit, Denver and Washington,
the following observations were made:
6
-------
(1) With Denver vehicles excluded, 32% of the vehicles
tested were at or below the applicable standard for
HC, 35% at or below the standard for CO, and 19%
were at or below standards for both HC and CO. In
Denver, the corresponding percentages were 5%, 3%,
and 0.3%.
(2) Mean emission levels for HC, CO and NO for all
xc
cities except Denver were 2.83 gms/mi, 35.3 gms/mi
and 4.69 gms/mi, respectively. The corresponding
values for Denver were 4.25 gms/mi, 65.1 gms/mi,
and 3.83 gms/mi.
(3) In general, the mean emission levels of the vehicles
tested failed to meet the applicable standards of 2.2
gms/mi and 23.0 gms/mi for HC and CO, respectively.
b. In the surveillance tests of 1971 model year vehicles in
Houston, Los Angeles, Detroit and Denver, the following
observations were made.
CI) Low-mileage (less than 60 odometer miles) and stab-
ilized engines showed distinct differences in the
percentage of vehicles tested which were at or below
applicable standards for HC, CO and both HC and CO.
Percent at or Below
Standards for
HC
CO
Both
87
60
58
59
49
38
36
2
1
29
10
6
Total Excluding Low Mileage
Denver: Stabilized
Denver Only: Low Mileage
Stabilized
(2) Mean emission levels in gms/mi for HC, CO and NO
XC
for low mileage and stabilized engines were as follows:
HC CO NO
xc
Total Excluding Low Mileage 1.47 23.7 3.16
Denver: Stabilized 2.12 29.0 3.74
7
-------
Denver Only: Low Mileage 2.63 64.0 2.86
Stabilized 3.16 45.2 3.19
(3) Applicable standards of 2.2 gms/mi HC and 23.0 gm/mi
CO were, for the most part, exceeded by the higher
mileage and correspondingly older stabilized engines,
which would more nearly represent engines operating
in vehicles in general use than do the low mileage
vehicles.
Tests performed in 1971 on light-duty vehicles, model years
1957 through 1971, show a consistent downward trend in HC and
CO emissions from 1957 through 1971. The trend is highlighted
by the following average emission levels based upon the 1975
CVS Federal Test Procedure for pre-control and controlled
vehicles, excluding Denver and for Denver alone.
Emission Levels (gm/mi)
1975 CVS Test Procedure
HC CO NO
xc
Total Excluding Denver:
Pre-Control 8.74 86.5 3.54
Controlled 4.37 53.9 4.92
Denver Only:
Pre-Control 10.16 126.9 1.89
Controlled 6.46 92.2 2.67
These results are indicative of the effect of emission controls
on air quality.
Results from the Rental Vehicle Surveillance Program, based upon
the 7x7 Federal Test Procedure, reveal that mean emission levels
for 1968 and 1969 model year rental vehicles are, in general,
consistently lower than similar vehicles tested during the Great
Plains program and in use by the general public.
Seven-Mode tests from the In-House Surveillance Program on 1968-
1969 foreign and domestic vehicles indicate that manufacturers1
recommended emission control maintenance practices were, in
8
-------
general, adequate to control emissions for 50,000 miles. In
addition, no evidence was found that manufacturers critically
tuned 4,000-mife certification vehicles to meet Federal emis-
sion standards.
6. Based upon tests using the 1972 CVS Federal Test Procedure, a
comparison of baseline data oti 1971 model vehicles tuned
prior to testing and comparable data from ''as-received." 1971
model light-duty vehicles fron the Six-City Study reveals no
significant difference in HC or CO levels. This is not sur-
prising because both groups of vehicles were tested in a
relatively low mileage condition. Moreover, the same conclu-
sion can be drawn from other EPA. studies. Therefore, this
conclusion should not be generalized to maintenance effects
on the total population of in-use vehicles because Iok mileage
v. , N
vehicles represent a small percentage of the vehicle population.
A small but significant reduction in NO- emissions, however,
xc
was observed in the maintained vehicles.
9
-------
2. RESULTS AND DISCUSSION
This section of the report is a discussion of the findings and
results of the six test programs discussed in this report. In the aggregate,
these programs represent tests on over 6,500 light duty vehicles in eight
cities. For convenience, the programs are listed below.
1. The Great Plains (Two City) Surveillance Program - 1968-1969
Model Year Surveys.
2. The National Surveillance Program - 1970 (Six City) and 1971
(Four City) Model Year Surveys.
3. A Study of Emissions from Light-Duty Vehicles in Six Cities -
1957-1971 Model Year Surveys.
4. The Rental Vehicle Surveillance Program - March 1968 to
January 1970.
5. The In-House Vehicle Surveillance Program - January 1967 to
March 1970.
6. Determination of Exhaust Emissions from 1971 Model Vehicles
(commonly referred to as the N0x Baseline Data).
Table 1 presents an overview and summary of the objectives, procedures, test
locations and number of vehicles tested for each of these programs.
In order to keep these programs in perspective, a brief summaryjof
the salient features of each test program is presented below. For clarity,
the programs will be referred to not only by their names, but also by the
number associated with each program in the above list.
The Great Plains (1) and National Surveillance (2) programs represent
the major EPA surveillance programs for the model years 1968 to 1971. Each of
these programs employed the 7-mode Federal Test Procedure and represents a
significant attempt to monitor the emission levels of in-use vehicles and to
determine the presence of any mileage or city effects. Similarly, the Six
10
-------
City 1957-1971 Model Year Survey (3) also represents a large scale program
aimed at determining the emission levels of in-use vehicles. Unlike the first
two programs outlined above, however, the Six City program utilized the 1972
and 1975 Constant Volume Sampling (CVS) Federal Test Procedures and was directed
mainly toward determining a nationwide emissions inventory for pre-control and
controlled vehicles. All three of these programs are discussed in detail in
APTD-1544 and will be discussed only in summary form in this report.
The final three programs, not discussed in the previous report,
represent smaller scale programs aimed at providing additional insights into
the analysis of emissions from light duty vehicles. Although these programs
were not of the same magnitude as the three major surveillance programs, they
have provided information useful in such areas as the assessment of the cer
tification process and the establishment of baseline N0x levels as required
by the Clean Air Act of 1970.
The fourth program is the Rental Vehicle Surveillance Program (4).
This program, which employed the 7-mode Federal Test Procedure, represents a
relatively large scale surveillance program (705 test vehicles) undertaken to
determine emission levels from in-use vehicles and to establish guidelines
for future EPA large scale surveillance programs.
The fifth program of interest, the In-House Vehicle Surveillance
Program (5), represents a unique program carried on from January 1967 to March
1970, using the 7x7 Federal Test Procedure. One of the earliest EPA Surveil-
lance programs, the in-House program was aimed primarily at gaining better
insight into the Federal Certification Procedures and any problems inherent in
them. By operating a fleet of prototype, "production and emission durability
vehicles in a manner similar to the Federally required procedures used by the
automobile manufacturers, EPA attempted to determine if prototype vehicles were
critically tuned by the manufacturers to meet Federal standards, if surveillance
vehicles operated by EPA differed markedly from manufacturers' durability vehi-
cles and whether the manufacturers' recommended emissions maintenance practices
were adequate to provide acceptable emissions levels over the 50,000-mile vehi-
cle lifetime.
11
-------
The final test program which will be discussed in this report is the
NO^ Baseline Program (6). Under the Clean Air Act of 1970, a 90% reduction
from 1971 NO^ emission levels was called for by 1976. Consequently, the estab-
lishment of the NO^ baseline level representing the average emissions from a
typical well maintained 1971 model vehicle was necessary. The data gathered
from 243 vehicles in this program were obtained using the 1972 CVS Federal
Test Procedure. Data from those vehicles not designed to meet California
NO standards represent the emissions baseline for 1976 standards.
X
The above paragraphs and Table 1 present a general overview of the
test programs to be discussed in this report. In the following sections, a
more detailed discussion of each program will be presented. For easy reference,
the number of each of the following sections corresponds to the identification
number of the six programs. It is hoped that such a numbering system will
enable the reader to clearly differentiate each program and to determine from
which program conclusions were drawn.
Although numerical results appearing herein have been drawn liber-
ally from contractors* final reports for each of these programs, additional
and different types of data analyses have been performed to provide additional
perspectives and insights into the results and, in some cases, tOi broaden or
refine the application of statistical methodology. In the formulation of
conclusions, caution has been exercised to consider inherent sampling biases,
confounding factors and other limitations that are associated with the data
base. The results of data analysis have been interpreted pragmatically from
the standpoint of the magnitudes of the emissions measurements as well as from
the standpoint of the statistical significance levels associated with the
results.
2.1 RESULTS OF THE GREAT PLAINS SURVEILLANCE PROGRAM
Contained in this section of the report are the results and findings
of three major areas investigated during the Great Plains Surveillance Program.
These three general areas are:
12
-------
a. 1968-1969 7-mode suxveillance results
b. Tests to determine efficacy of idle adjustments in
reducing emissions and results of maintenance survey
c. Evaluation of short test cycles.
The seven-mode surveillance results were covered in detail in APTD-
1544 and the results presented here will only highlight the important conclusions.
The efficacy of idle adjustments and the evaluation of short test cycles such
as the Hot Idle Mode and the Clayton KEY MODE diagnostic test were not discussed
in APTD-1544, but will be covered in sane detail in this report.
2.1.1 1968-1969 Surveillance Results
This section of the report represents a brief summary of the 1968-1969
surveillance data. Composite exhaust emissions data taken from each vehicle
tested were based on the seven-mode, cold start test procedure as stipulated by
the Federal Register appropriate to the model year tested (see Table 1). The
measurement of N0x (NO), which was not specified in the Federal Register, was
performed using NDIR (non-dispersive infra-red). In all cases, an attempt was
made to select the test vehicles from the motoring public according to a
strategy that yielded a sample as nearly representative as possible of the
vehicles that were in use (or, in the case of new models, projected to be in
use).
Exhaust emissions data for each of the three principal effluents
determined from measurements on 1968-1969 model vehicles tested during the
Great Plains Program are shown in Table 2. These data consolidate results
from the three phases of this program during which time each vehicle, on the
average, accumulated approximately 15,000 miles. Included in these tables are
vehicles produced by American Motors, Chrysler, Ford, General Motors and Volks-
wagen representing, in total, a statistical sampling of 90% of the sales for
these model years. The oxides of nitrogen are expressed as NO , the subscript
xc
"c" denoting the correction for humidity.
13
-------
Although the results are expressed both in terms of concentration
and grains per mile, these results should be assessed against the certification
standards applicable to these models (275 ppm HC and 1.50% CO, except for VW,
for which the standards are 410 ppm HC and 2.30% CO). The grams-per-mile
figures are primarily presented to facilitate cross year comparisons with 1970
and 1971 National Surveillance data discussed in the next section. It is to be
noted, however, that the change in basis from concentration to grams per mile
is primarily one of data processing since the basic direct measurement for
1968-1971 model year vehicles was in terms of concentration. For certification
of the 1970 and 1971 model year vehicles, a functional relationship was devel-
oped to calculate the vehicle exhaust volume per mile with vehicle inertia
weight and transmission type as independent variables. By means of this rela-
tion, grams-per-mile values are calculable to afford a comparison of 1968, 1969,
1970 and 1971 model vehicles on a common basis. Although there were no standards
for the oxides of nitrogen for these model years, the NO^ data were nonetheless
measured in the surveillance studies and are included.
The major conclusions derived from the analysis of 1968 and 1969
Great Plains surveillance data are the minimal practical differences not only
between emissions from Kansas City and Houston, but also between model years
1968 and 1969. Neither of these conclusions is surprising, however, in view
of the geographic and climatic similarities between both cities and the emis-
sion control equipment utilized during these model years.
2.1.2 Vehicle Maintenance Results
In addition to data described in the previous section, data were also
collected as part of the Great Plains surveillance effort to establish the re-
duction in emissions attributable to minor engine adjustments and to determine
the maintenance practices of the motoring public. As the number and complexity
of emission control devices increases, the effects of vehicle maintenance be-
come increasingly vital to the furtherance of national air quality goals.
Involved in the maintenance study discussed in this section are maintenance
characteristics such as type, frequency and effectiveness in reducing emissions.
It must be emphasized that the study reported here evaluated the effects of
14
-------
maintenance on vehicle emissions using the hot start seven-mode test procedure.
It is now believed that use of the 1972 or 1975 CVS Federal Test Procedures is
necessary in order to relate measured emission values to their impact on air
quality. Thus, the data presented here on the effects of vehicle maintenance
on emissions are principally of historical interest. Results of more recent
EPA studies using the 1972 and 1975 Federal Test Procedures should be used to
estimate the impact on overall emissions and air quality which would result
from improved vehicle maintenance practices. With these limitations in mind,
the maintenance findings are summarized below. For reference purposes, Phase
1 refers to the initial test point at around 25,000 miles, while Phases 2 and
3 refer to two subsequent tests after the accumulation of approximately 4,000
and 8,000 miles, respectively, from the initial Phase 1 test.
As a part of the surveillance program of 1968-1969 model vehicles,
basic ignition timing, dwell angle and idle RPM data were measured for each
vehicle. During Phase 1 of this program, readjustment of these engine param-
eters was effected whenever the manufacturer's specifications were exceeded by
-2° in timing or -75 rpm for idle speed. These corrected settings were made
following the nominal cold start, 7-mode 7-cycle test and a hot start test was
then made with emissions data collected during two 7-mode hot cycles. The
efficacy of the engine adjustments was assessed by comparing composite emission
levels for the hot cycles (6 and 7) of the cold start and hot start tests.
Based on an overall total of 192 tests, the mean HC levels were reduced by
110 ppm in Houston, but were relatively unchanged in Kansas City. On the
other hand, a 0.2% reduction in mean CO was observed in both cities.
During Phases 2 and 3 attention was given to testing carburetor air
filters for restricted air flow. A rejection criterion was established which
failed approximately 20% of the units tested. When a filter was rejected, a
new unit was installed and the effect on emissions evaluated in the identical
manner described above. Tests on 123 vehicles indicated no appreciable effect
on mean HC levels and a decrease of approximately 0.3% in CO. Neither engine
adjustments nor air filter replacements had any significant influence on N0XC
levels.
15
-------
In addition to the efficacy of idle adjustments described above,
information on the maintenance practices of 1968-1969 model vehicle owners'in
Houston and Kansas City was recorded on approximately one-half (i.e., about
one thousand) of the vehicles tested. Ten maintenance items were reported in
the survey: ignition timing, breaker points, spark plugs, ignition cables,
carburetor, PCV valve, air filter, fuel filter, valve repair and engine tune-up.
On a frequency-of-reporting basis, maintenance of spark plugs and
breaker points led the list, being reported approximately 60% of the time.
Ignition timing, air filter and tune-up maintenance occurred at about a 40%
level. No significant differences in maintenance practices between the two
reporting cities was demonstrated. There was evidence to show that the fre-
quency at which maintenance operations were performed tended to decrease as
the vehicles accumulated mileage, i.e., the newer the vehicle, the more fre-
quent the maintenance.
2.1.3 Evaluation of Short Test Cycles
Achievement of national air quality goals as affected by automobile
emissions is dependent upon the owner providing proper maintenance for his
vehicle as well as for the newly manufactured vehicles meeting the certifica-
tion standards. With approximately 100 million passenger vehicles currently
registered in the United States, there is great interest in identifying a
short, reliable tiest procedure that yields pass/fail results consistent with
those of the lengthy, composite Federal test procedures. Evaluation data
gathered during the Great Plains and National Surveillance programs on two
likely candidates are summarized below. It must be emphasised that thc-
results presented here consider the ability of short tests to evaluate seven
mode FTP results. Subsequent adoption of the 1972 and 1975 CVS Federal Test
Procedures as standard measures of emissions limits the applicability of
these studies to illustrating the types of analyses shich m&y be performed.
16
-------
2.1.3.1 Hot-Idle Mode
Based on her investigations, Marian F. Chew* concluded that vehicular
emissions as Measured during the hot-idle node constitute a reliable prediction
of the pass/fail discrimination afforded by the entire seven-mode Federal test
procedure if appropriate "decision values" were used (which differed from the
"policy values" associated with the complete seven-mode test). In the present
program, 2,038 tests were performed on 1968-1969 model vehicles in which the
hot-idle mode pass/fail results were compared with those of the complete test
procedure. For policy values of 275 ppm HC and 1.50% CO, the best correlation
was realized when decision values of 205 ppm HC and 2.85% CO were applied to
the hot-idle mode data. Agreement on pass/fail decisions at these levels was
achieved on 83% of the trials. Of the disagreements, the hot-idle test failed
10.4% that should have passed; the remainder, 6.6%, were passed whereas they
should have failed.
In addition to the discriminant approach discussed above, an attempt
was made to predict the seven-mode composite data for each emittant based upon
the mathematical relation:
Calculated 7-Mode Composite = a + b (Hot Idle)
where a and b are constants determined by least squares analysis. Results of
this analysis, however, were not encouraging since the limits of the data were
too broad. More specifically, 95% of the actual 7-mode values fell within the
following limits from the predicted 7-mode values in the region of best
estimation;
HC = -269 ppm
CO » -1.67%
N0*c ° "980 ppm
The dispersion of these results and a comparison of the relative magnitude of
this dispersion with the 1968 to 1969 HC and CO standards of 275 ppm and 1.5%
indicate that a simple linear relation is not adequate to accurately predict
7-mode composite emission levels from hot-idle data.
+
"Auto Smog Inspection at Idle Only," SAE Paper No. 690505, May 1969.
17
-------
Although the success of the linear regression approach described
above was limited, the results of the discriminant approach were encouraging
insofar as the efficacy of employing short test cycle procedures in predicting
the pass/fail seven-mode test results is concerned. It should be noted, however,
that because of the specifics of the data collection process, the idle test data
constitute an integral part of the composite (seven-mode) data. The two sets
of data are therefore correlated and, as a consequence, the final assessment
may have been more favorable than would have been the case if the data had
been obtained from independent test series.
2.1.3,2 Clayton KEY MODE Diagnostic Test
The Clayton KEY MODE diagnostic test comprises the collection of
steady-state, stabilized emission data (on a concentration basis) during three
operational modes: idle, low-speed cruise, and high-speed cruise. The cruise
speed levels are quantified in relation to vehicle inertia weight according to
the schedule below:
Inertia Low-Speed High-Speed
Weight, lbs Idle Cruise, mph Cruise, mph
3800 - up (Automatic 32/35 48/50
2800 - 3750 Transmissions 29/32 44/46
2000 - 2750 in Drive) 22/25 36/38
The vehicle is operated on a dynamometer whose inertia flywheels are
decoupled and the power absorption unit adjusted to an indicated 30 road-load
horsepower at 50 mph. Alternate settings for vehicles with lower inertia
weights are 24 hp at 46 mph or 15 hp at 38 mph.
Using results from over 2000 tests on 1970 model National Surveillance
vehicles, the test contractor made a comparison of KEY MODE and seven-mode com-
posite data for each emittant based on the following mathematical relation:
Calculated 7-Mode Composite = a + bKM^ + cKM2 + dKM^
In the relation, KM^, KM^, KM^ represent the three emittant concentrations from
the KEY MODE low speed cruise, high speed cruise and idle modes while a, b, c,
18
-------
and d are constants determined by the method of least squares. This technique
is analogous to the unsuccessful linear regression on hot-idle data discussed
in the previous section. Again, the analysis indicated a wide dispersion in
the data with 95% of the actual 7-mode values falling within the following
limits from the predicted 7-mode values in the region of best estimation:
HC = ±216 ppm
CO = ±1.0%
N°xc= ±729 ppm
As in the case of the hot-idle data, a comparison of the relative
magnitude of these dispersions with the 1968 to 1969 HC and CO standards of
275 ppm and 1.5% indicate that a linear combination of KEY MODE data has
limited value as an accurate predictor of emission levels as determined by the
Federal Seven-Mode Test Procedure.
KEY MODE tests were also performed during the program comparing emis-
sions from low mileage and stabilized engines in 1971-model vehicles during the
National Surveillance Program. With the negative results from the earlier
data in mind, a comparison was made between the data from the individual modes
of the KEY MODE test with the composite data of the seven-mode test. While
sane of the individual modes gave better agreement for one emittant than for
another, on an overall basis the low cruise mode best represented the results
of the seven-mode test. This conclusion, however, does not alter the fact
that the Clayton KEY MODE test is not a reliable predictor of emission levels
as determined by the seven-mode seven-cycle test procedure.
In addition to the regression approach, a linear discriminant analysis
similar to the Chew technique discussed in the previous section was also per-
formed on 1749 non-Denver 1970 National Surveillance vehicles. The objective
of this analysis was to determine the performance of the KEY MODE data in pre-
dicting the pass/fail results of the seven-mode test procedure. For the pur-
poses of this analysis, two groups of data were identified338 vehicles
passing both HC and CO seven-mode Federal standards and 1411 vehicles failing
these standards. A detailed discussion of the discriminant approach was pre-
19
-------
sented in APTD-1544 in Appendix III. Briefly, the technique involves selecting
a set of weights which when applied to several groups of data produce minimum
scatter within the groups and maximum separation between the groups. The
objective of this type of analysis is to highlight the differences between
selected groups of data.
Based upon this linear discriminant approach, the following formula
for classifying vehicles into pass/fail categories was determined.
FACTOR = -.02024 x HC - ppm High Cruise
+12.08 x CO - % High Cruise
+.00178 x NOX - ppm High Cruise
+.1446 x HC - ppm Low Cruise
+4.802 x CO - % Low Cruise
-.00103 x NOX - ppm Low Cruise
+.03364 x HC - ppm Idle
+15.14 x CO - % Idle
+.01214 x NOX - ppm Idle
By thresholding this factor to minimize the total number of incorrect
decisions, the following decision rule was established:
If FACTOR ^45. Vehicle will pass 7-mode
^ 45. Vehicle will fail 7-mode
Employing this technique on the 1749 test vehicles resulted in 83.5% correct
pass/fail decisions. Of the 16.5% incorrect decisions, 4.9% passed but should
have failed and 11.6% failed but should have passed.
Figure 1 presents a histogram of the discriminant factors calculated
for each of the vehicles in the two groups. In figures of this type, the non-
overlap of the two histograms is a measure of the success of the linear dis-
criminant approach in predicting the pass/fail results of the seven-mode test
procedure.
20
-------
The 83.5% correct decisions achieved using the KEY MODE data is vir-
tually identical to the 83.0% level achieved vising the slightly different Chew
technique on the hot idle data. Although no data was available from the pro-
grams covered in this report to substantiate the ability of any short test
cycle to predict the pass/fail results of the CVS test procedure under which
vehicles manufactured after 1972 were certified, the relative success demon-
strated by the KEY MODE and hot idle tests in predicting seven-mode pass/fail
results is encouraging insofar as the efficacy of employing short test cycles
as an inspection test is concerned.
2.2 RESULTS OF THE NATIONAL SURVEILLANCE PROGRAM
This section of the report presents a brief compilation of the impor-
tant conclusions and summary tables for the National Surveillance Program. Once
again, the reader is referred to APTD-1544 for a more detailed discussion of
these results.
Exhaust emissions data for each of the three principal effluents
determined from measurements on 1970 model vehicles in six cities (Kansas
City, Houston, Los Angeles, Detroit, Washington- and Denver) are shown in
Table 3. Examination of this table reveals that only Denver departs markedly
from the other cities in the study. HC and CO emission values in Denver are
significantly higher than in the other five cities in the program. N0xc
emissions, which were not subject to control during this model year, exhibited
correspondingly lower levels in Denver.
Table 4 presents similar data for 1971 model vehicles tested in four
cities (Houston, Los Angeles, Detroit and Denver). Data in this table, how-
ever, have been divided into two groups, Phase 1 and Phase 2. Phase 1 data
represent vehicles tested in low mileage engine conditions (less than 60
odometer miles), while Phase 2 data represent emission levels on the same
vehicles with "stabilized engines" after the accumulation of approximately
5,000 miles. The principal effect of this stabilization on emissions is
believed to be the result of "wearing-in" or "seating" of valves and piston
rings as well as a stabilization of combustion chamber deposits.
21
-------
As in the case of 1970 model vehicles discussed above, only emission
levels in Denver depart significantly from levels in the other cities. Although
it is often hazardous to ascribe differences observed in test results to one
particular effect or another, special unique considerations such as Denver's
high altitude can be advanced as a substantive constraint affecting the com-
pliance of Denver with emissions standards.
A summary of the seven-mode surveillance results obtained during the
Great Plains and National Surveillance Programs is presented in Table 5. Con-
tained in this table are N, the number of vehicles tested, the mean mileage of
the vehicles tested and the arithmetic and geometric mean and standard devia-
tion of the data. The interpretation and use of these figures is discussed in
detail in Section 2.1 of APTD-1544.
2.3 A STUDY OF EMISSIONS FROM LIQ4T DUTY VEHICLES IN SIX CITIES -
1957-1971 MODEL YEAR SURVEY
This section of the report briefly summarizes the results from a
single test program devoted to the measurement of exhaust emissions from 1957
through 1971 model light-duty vehicles in six metropolitan areas. Constant
Volume Sampling (CVS) test procedures were employed throughout the program and,
consequently, the data presented in this section are not readily comparable to
those discussed in Sections 2.1 and 2.2.
A brief overview of the program was previously presented in Table 1.
The principal function was the collection of data from which average emission
factors could be formulated in order to define the contribution of the auto-
mobile population to the nation's air pollution burden. To achieve this
objective, the best available methodology and technology were employed to
accurately determine mass emissions under vehicle operating conditions repre-
sentative of road use. Since 1957-1971 model vehicles comprised more than 95%
of the population as of 1971, a statistically-representative sample of this
population was tested in each of the six cities which were chosen to maximize
variations in climate, terrain, and urban development. Despite a somewhat
limited sample for the early-year models (1957-1960), the overall sample is
very representative of the overall vehicle population on the road. Data from
22
-------
this program are particularly useful, therefore, in estimating the overall
impact of emissions on air quality. The reader is referred to the previous
report, APTD-1544, for a more detailed description of the vehicle selection
procedures and test techniques. For summary purposes, it is sufficient to say
that all vehicles were tested for exhaust emissions in an as-received condition.
Cold start tests were performed in accord with the 1972 (CVS-C) and 1975 (CVS-CH)
Federal Test Procedures (FTP). In actuality, however, only one test per vehi-
cle was made, since diluted exhaust emissions were bagged in such a manner that
vehicle mass emissions could be calculated according to both 1972 and 1975
specifications.
Mass emissions data, expressed as arithmetic and geometric means as
well as their respective standard deviations, are summarized in Tables 6 and 7
for the 1972 FTP, and Tables 8 and 9 for the 1975 FTP. In addition, the per-
centage of vehicles with emission levels at or below the reference levels of 3.4 gins/
mi HC, 39.0 gms/mi CO, and 3.0 gms/mi NC>xc is also presented. Although these
reference levels do not represent emission standards under which the vehicles
were certified, they are useful in examining time trends for vehicles of vari-
ous model years in which no standards or differing standards were in effect.
The results are aggregated by year and by all cities except Denver and Denver
only.
The differences that are inherent in the 1972 and 1975 procedures
reflect the fact that not all trips made by a vehicle originate from a "cold-
start" (defined in the Federal Register as a start preceded by a 12-hour, no-
use soak period). The 1972 FTP determines mass emissions from a driving
schedule (LA-4) that comprises two portions, a cold-transient and a cold-
stabilized. The 1975 FTP uses the same driving schedule but also adds a "hot-
S
start" transient portion. In this latter case, emissions levels represent a
weighted average of the cold start and hot start test results. Since a large
fraction of the composite HC and CO emissions are generated during the fuel-rich,
engine warm-up phase of the cold-transient portion of the driving schedule, the
1975 data (as compared with the 1972 data) will show lower levels of emission
for these two effluents. The data of Tables 6 through 9 bear out this conclu-
sion. On the other hand, N0xc levels remain substantially uniform, the observed
differences probably being attributable to experimental sources.
23
-------
Trends by model year in mean emission levels and percentage of vehi-
cles below the reference levels are well defined for all effluents. HC and CO
levels show continuing decreases and reflect the impact of increasingly more
stringent emission controls. Consistent with this decrease, there is observed
an increase in N0XC levels which, during this period of time, were not subject
to control. Figure 2 which presents 1975 CVS emission levels vs the cumula-
tive percentage of test vehicles, emphasizes these model year trends for the
three pollutants.
In addition to the surveillance tests discussed above, two additional
kinds of data were gathered during this program:
a. Modal emissions data for 32 accel/decel modes and five
steady states.
b. Evaporative emissions data utilizing the SHED technique
(SAE J171).
Since these data are not easily synopsized, the reader is referred to APTD-1544
for a complete discussion of these results.
2.4 RENTAL VEHICLE SURVEILLANCE PROGRAM
Exhaust-emission tests using the 7-mode Federal Test Procedure were
performed on 1968 and 1969 model year rental vehicles representing 26 differ-
ent engine classes. This program was conducted to ascertain the effectiveness
of Federal exhaust emission standards on in-use vehicles. In addition, although
not an original objective of the program, data from this program were used to
determine whether fleet operated vehicles receiving minimum maintenance and
rapidly accumulating mileage significantly differ in emission levels from
vehicles in general use. Table 10 presents Great Plains surveillance data
and comparable Rental vehicle surveillance data for engine classes consistingg
of more than 15 vehicles. With the exception of 1968 Chevrolet 327 CID engines,
the data are presented for both model years lumped. Examination of this table
reveals that the rental vehicles tested exhibit consistently lower mean emis-
sion levels than comparable vehicles in general use. These differences in
mean emission levels are in part explainable by the age and associated mileage
24
-------
difference between the two classes of vehicles. The rental vehicles tested
from 1968 to 1969 had an average mileage of 7,700 miles whereas the Great
Plains vehicles had an average mileage of approximately 30,000 miles and were
tested from 1969 to 1971. In addition, the rapidly accumulating mileage
experienced by the rental vehicles and the as-received nature of the Great
Plains vehicles further confound the difference in emissions experienced by
these two groups of vehicles.
2.5 IN-HOUSE VEHICLE SURVEILLANCE PROGRAM
The In-House Vehicle Surveillance Program was established by the
Division of Motor Vehicle Pollution Control (DMVPC) of the National Air Pollu-
tion Control Administration in 1968 to provide additional HC and CO emissions
data for prototype and production vehicles. The DMVPC operated a fleet of 1968
prototype and 1969 production vehicles (foreign and domestic) for the purpose
of accumulating mileage under actual road use conditions. The major objectives
of this program were:
a. To determine whether prototype vehicles weTe critically tuned
by manufacturers to meet Federal exhaust emission standards.
The prototype vehicles in question were prototype vehicles
from manufacturers1 certification fleets and were driven on
an accelerated mileage schedule for 4,000 miles and tested
for exhaust emissions.
b. To determine whether exhaust emissions from vehicles in the
prototype surveillance fleet differ substantially from the
exhaust emissions from vehicles in the test track durability
fleet. The prototype surveillance vehicles referred to above
represent certification vehicles chosen by the DMVPC to corres-
pond to vehicles operated by the manufacturers in their dura-
bility fleets.
c. To determine whether manufacturer's recommended maintenance
practices keep production surveillance fleet vehicles within
Federal emissions standards for 50,000 miles of operation.
25
-------
The production surveillance vehicles were actual production
vehicles chosen on the basis of three criteria. First, vehi-
cles were chosen to give the widest representation of vehicles
marketed in the U.S. Second, vehicles were chosen with engine
control system combinations not in the manufacturer's durability
fleets. Third, vehicles were chosen that were equivalent to
those which performed marginally in the prototype surveillance
fleet.
The testing and maintenance procedures applied to these vehicles are
summarized below. The test track durability fleet vehicles were tested for
exhaust emission levels every 4,000 miles by the manufacturer. Maintenance of
these vehicles was performed in conformance with the specifications given in
the Federal Register. These specifications allow for a major tune-up after
24,000 miles, but little other maintenance.
Prototype surveillance fleet vehicles were tested for exhaust emis-
sions using the Federal coldstart emission tests at intervals of 3,000-4,000
miles. These vehicles were scheduled to be serviced in accordance with the
maintenance schedule supplied in the manufacturer's Application for Certifi-
cation.
In reviewing the maintenance data on these vehicles, it was found
that carburetors, distributors, air injection system components, timing, and
idle specifications were oftentimes changed by the manufacturer or dealer to
incorporate latest production vehicle components or specifications into the
prototypes. These changes were made to correct vehicle drivability problems
reported by Motor Vehicle Compliance Section personnel, or to correct excessive
vehicle emissions. Changes made to prototypes were incorporated into all
vehicles manufactured. Data on run changes were submitted by the manufacturers
to the Certification Branch before the change was made, if it were one which
would affect emissions. Maintenance was performed according to manufacturer's
recommendations. In addition to scheduled maintenance, vehicles were usually
serviced whenever their emissions exceeded Federal standards. Complete Federal
7-mode, 7-cycle exhaust emission tests were performed in the EPA laboratories
before and after both scheduled and unscheduled engine maintenance.
26
-------
Production vehicle surveillance fleet vehicles were tested every
3,000-4,000 miles using the Federal cold start test procedure as stipulated in
the Federal Register. All vehicles were maintained in accordance with the main-
tenance schedules outlined in the "Vehicle Owner's Manual" supplied with each
car. In addition to scheduled maintenance, all vehicles were serviced whenever
their emissions exceeded Federal standards. Furthermore, dealers or manufacturers
were allowed to perform any maintenance necessary to put a vehicle back within
standards, short of a major engine overhaul. Complete Federal coldstart exhaust
emission tests were performed both before and after any engine maintenance.
Although the substantive conclusions which can be drawn from the In-
House Surveillance Program are somewhat limited because of the small sample
size present (122 vehicles) and the variety of maintenance practices employed
during the program, several conclusions arise from this data. First, it appears
that for the vehicles tested during this program, manufacturers did not criti-
cally tune 4,000-mile certification vehicles to meet emission standards. Sec-
ond, vehicles in the prototype surveillance fleet did not perform as well, with
respect to exhaust emissions, as did vehicles in the manufacturer's test track
durability fleet. Finally, tests done during this program indicate that vehicle
maintenance, if correctly performed and backed up by emissions measurements,
will keep properly manufactured vehicles below Federal standards for exhaust
emissions. It should be emphasized that the above conclusions must be tempered
by the limited nature of the In-House data as well as the complicating effects
introduced by maintenance practices employed.
One interesting and significant point which arises from the In-House
data is illustrated in Figure 3 which presents hydrocarbon emission levels for
five relatively homogeneous prototype surveillance vehicles operated over a
32,000-mile period. The significant point to be mentioned about this figure
is the nonlinear trend in HC emissions with increasing mileage. A similar
phenomenon was also observed for carbon monoxide emissions over this period.
The decrease in HC emissions as mileage increased above 16,000 miles could
seriously affect the linear regresssion techniques used to determine deteriora-
tion factors in the certification procedure. This phenomenon will be further
investigated in a future EPA contract dealing with the certification process.
27
-------
2.6 NO BASELINE PROGRAM
x
2.6.1 Background
The nitrogen oxides .CNO^) baseline program was conducted to acquire
emission data necessary so that the EPA could implement required legislation.*
These data were collected during the period from March 5, 1971 to April 17,
1971, in Detroit, Michigan. The program involved testing a randomly selected
sample of vehicles by the 1972 CVS-C Federal Test Procedure. The number of
each engine size and make tested represented, as closely as possible, the
estimated percent of this group of vehicles in the population of 1971 vehicles.
Emissions of hydrocarbon, HC, carbon monoxide, CO, carbon dioxide, C02, and
NO^c were determined on a mass basis (gm/mile) for the 243 vehicles in this
program.
Prior to testing, each vehicle was inspected and set to manufacturer's
recommended specifications. Parameters used were ignition cam dwell, idle RPM,
and basic timing. The automatic choke operation was checked and spark plugs
were replaced if misfire was detected. The idle mixture was set to specifica-
tion by manufacturer's recommended procedures. One percent of vehicles inspected
were rejected since evidence of tampering with the emission control equipment
was apparent. After the tune-up, the vehicle was soaked for a minimum of 12
hours prior to testing.
All studies reported in the original summary report, APTD-1544,
were done on vehicles tested in as-received condition. The NO^ Baseline
Study, on the other hand, recorded exhaust emission values of vehicles
which received maintenance prior to testing. Since it was not known,
however, if these vehicles were maintained by their owners before they were
brought in for the test, the long term effects of good maintenance could
not be deduced. These data can, however, be used to determine the effect
of pre-test maintenance on emission results. The comparison of these
data with in-use vehicle data obtained on an as-received basis should
detect those classes of vehicles which have high emissions because
they are out of adjustment. Since the N0x baseline study was done using the
* Public Law 91-604, "The Clean Air Act," December 1970.
28
-------
CVS procedure, it seemed appropriate to compare these data with 1971 vehicles
from the Study of Emissions from Light Duty Vehicles in Six Cities. Before
making comparisons, several assumptions were made. Since 1971 California
vehicles were subject to NO^ standards, some manufacturers chose to build two
versions of given models of vehicles, one version for California only and one
version for the other 49 states. These two groups of vehicles are referred
to as one-state vehicles and 49-state vehicles, respectively. Other manu-
facturers chose to build one version of a model for all 50 states. This
group of vehicles is known as 507state vehicles. The NO^ Baseline Study
tested both 49-state vehicles and 50-state vehicles, depending on the model.
All vehicles were from the Detroit area. The Six-City Study tested one-state
vehicles, 49-state vehicles, and 50-state vehicles depending upon the model
vehicle and the-test location. Since Denver based vehicles exhibit exhaust
emissions unlike those from the other five cities,* these vehicles were elim-
inated from the comparison with NO^ Baseline Study. Also, the assumption was
made that all models tested in California which could have been one-state
models were one-state vehicles. This is probably reasonable since one-state
models were intended to be sold in California only and all vehicles tested
were at most one year old and probably still in the hands of the original
owners. These one-state vehicles were eliminated for comparative purposes.
The two studies tested different numbers of each make and engine
size. This fact was considered in the choice of analysis. Also, the mean
mileages for the two groups differed. The N0x Baseline vehicles had a mean
mileage of 7,100 miles while the Six-City vehicles had a mean mileage of 15,600
miles. A summary of the comparative emission levels for the N0x Baseline and
Six-City programs is given in Table 11. Presented there are the number of
vehicles tested, mean mileage, percent of the vehicles tested with HC, CO, and
NOxc emission levels at or below the reference levels of 3.4, 39.0 and 3.0 gm/mi,
respectively, and the mean emission levels themselves.
"Automobile Exhaust Emission Surveillance--A Summary" (APTD-1544).
29
-------
2.6.2 Analysis
Inasmuch as variation in mileage could contribute to variation in
emissions, it was thought necessary to adjust both groups to a common mileage
point as a prerequisite to assessment of a maintenance effect. Such adjust-
ment can be accomplished by means of regression analysis. Upon application
of the regression technique to selected make/engine-size combinations of data,*
however, it was found that in all cases the slopes of the regression lines
were not significantly different from zero and that adjustment for mileage
differences could accordingly be ignored.
The six city and NO^ baseline data were compared to see whether a
significant difference existed which could be attributed to maintenance. As
was expected, no significant maintenance effect occurred for hydrocarbons or
carbon monoxide in any of the subgroups of low mileage vehicles considered.
However, a significant reduction in emissions of oxides of nitrogen occurred
when vehicles were tuned to specifications before testing. The above obser-
vations should be tempered by noting the small sample sizes present which
were frequently fewer than ten vehicles. In addition, both groups of vehicles
were tested in a relatively low mileage condition. Consequently, the improve-
ment in emission levels due to maintenance would tend to be smaller than if
the vehicles had been older. Finally, the magnitude of the observed main-
tenance effect may have been reduced since some of the "as-received" Six-City
vehicles may well have received tune-ups sufficiently close to the time they
were tested.
k
It was not considered appropriate to compare groups of data across make
since the data were unbalanced and arbitrary weighting of various models could
influence results.
30
-------
TABLES
31
-------
Table 1
SUMMARY OF SEVEN-MODE AND CVS TEST PROGRAMS
SUMMARY OF SEVEN-MODE TEST PROGRAMS
NATIONAL SURVEILLANCE PROGRAM
1W1MGT. PLAINS
1MB SIX CITY
' 1t71 FOUR CITY
RENTAL VEHICLE
1IM40USE.SURVEILLANCE
OBJECTIVES:
EVALUATE MILEAGE ANO CITY EFFECTS;
DETERMINE CONFORMITY WITH CERTIFICATION
STANDARDS
DETERMINE CONFORMITY WITH CERTIFICATION
STANDARDS; EVALUATE CITY EFFECTS
DETERMINE CONFORMITY WITH CERTIFICATION .
STANDARDS; EVALUATE CITY EFFECTS.
COMPARE LOW MILEAGE ANO STASILI2E0 ENGINE -
EMISSIONS (PAIRED TESTS).
DETERMINE CONFORMITY OF IN
USE VEHICLES WITH CERTIFICA-
TION STANDARDS. PROVIDE
PLANNINO DATA FOR FUTURE
SURVEILLANCE PROGRAMS.
DETERMINE IF PROTOTYPE VEHICLES ARE CRITICALLY
TUNEO TO MEET FEOERAL STANOAWDS: DETERMINE «F
PROTOTYPE SURVEILLANCE VEHICLES 0IFPER IN
EMISSIONS FROM DURABILITY VEHICLES FLEET. DETER-
MINE IF MANUFACTURERS RECOMMENDED EMISSIONS
MAINTENANCE PRACTICES ARE ADEQUATE OVER
U.0B0 MILE VEHICLE LIFETIME
BASIC TEST PROCEDURES:
7X7 FTP IN ACCORDANCE WITH FEDERAL
REGISTER. VOL. 31. NO. tl (PART III. MARCH
JO. mt; EACH VEHICLE TESTED 3 TIMCS.
INITIALLY ANO AT EACH Of TWO 4,000
OLE INTERVALS.
7 X 7 FT? IN ACCORDANCE WITH FEDERAL REG<
ISTER. VOL. 33. NO. 10i. JUKI 4. 1M (PART 111;
EACH VEHICLE TESTED ONCE (MINIMUM MILEAGE
REQUIREMENT 4,000 MM.
7 X 1 FTP IN ACCORDANCE WITH FEDERAL REGISTER.
VOL. 33, NO. 10S. JUNE 4. IKS (PART III; TWO PAIRED
TESTS PER VEHICLE: LOW MILEAOE ENGINE (SO HILCS)
ANO STABILIZED ENGINE 14.0007.000 MILES).
7 ¦ 7 FTP IN ACCORDANCE WITH
FEOERAL REGISTER. \K)L. IT.
NO. SI. MARCH X IMS. EACH
RENTAL VEHICLE TESTED ONCE
IN "AS RECEIVED" CONDITION.
7 m 7 FTP IN ACCORDANCE WITH
FEOERAL REGISTER. VOL. 31. NO. SI.
MARCH 10. IMS. IF EMttSION*
EXCEEOED ALLOWABLE LEVELS
VEHICLE TUNED ANO RETESTCD.
INSTRUMENTATION;
N0N-DI9ERSIVE INFRARED (NDIIM WAR'
A1U3.
N0N-0<7ERStVl INFRARED (KOIRI AW AHA
TUS.
NON-OISPCRSIVE INFRAHEO (ND1H1 APPARATUS.
NON-OISHRSIVE INFRARED
APPARATUS(NOIRI
NON-OlSPf RSIVt INFRAREO APPARATUS
INDIA)
CITIES:
KANSAS CITY AND HOUSTON
KANSAS CITY, HOUSTON. LOS ANGELES,
DETROIT. DENVER. WASHINGTON
HOUSTON. LOS ANGELES. OETROIT ANO DENVER
LOS ANGELES, DETROIT
OCTROIT
EMISSION STANDARDS:
276 pf*n HC. 1.50% CO <410 HC. S.30X
CO FOR VW)
JJ QMMMI. HC. 2X0 OAAMS/HI CO
. 2.3 GRAMSMI. HC, 13.0 GRAMSMI, CO.
» n nm hc. i.sM co wio
mt HC. 2.30* FOR VW)
9 TS PPM HC. IMk CO WTO PPM HC
2.W& CO FOR MR
TOTAL VEHICLES TESTED:
IMS
J1H
706
48 PROTOTYPC
in S3PROOUCTION
«S DURABILITY
o, SUMMARY OF CVS TEST PROGRAMS
K>
SIX-CITY SURVEILLANCE PROGRAM- 1957-1971 MODEL VEHICLES NOx BASELINE PROGRAM -1971 MODEL VEHICLES
OBJECTIVES:
DETERMINE CONTRIBUTIONTO AIR POLLUTION BY VEHICLE
POPULATION RESPONSIBLE FOR 95% OF VEHICLE EXHAUST
EMISSION (CIRCA 1971); COLLECT EMISSIONS DATA DURING
STEADY-STATE MODES, ACCEL/DECEL DRIVING CYCLE;
EVALUATE CITY EFFECTS; COLLECT EVAPORATIVE
EMISSIONS DATA.
ACQUIRE BASELINE NOx DATA FOR IMPLEMEN-
TATION OF PUBLIC LAW 91-604, 'THE CLEAN
AIR ACT," DECEMBER 1970.
BASIC TEST PROCEDURES:
1972 FTP (CVS-C) AND 1975 FTP (CVS-CH). IN ACCORDANCE WITH
FEDERAL REGISTER. VOL. 35, NO. 219, NOV. 10. 1970; VOL. 36,
NO. 55. MARCH 20. 1971; VOL. 36, NO. 128, JULY 2, 1971. EACH
VEHICLE TESTED ONCE (EXCEPT FOR SUBSET OF REPLICATES).
EVAPORATIVE TESTS BY SHED TECHNIQUE (SAE J71). VEHICLES
TESTED IN "AS-RECEIVED" CONDITION AND ANY MEASUREMENT
OF IDLE SPEED TIMING. ETC. MADE ONLY AFTER VEHICLE
TESTED.
1972 FTP (CVS-C). IN ACCORDANCE WITH
FEDERAL REGISTER, VOL. 35, NO. 219, NOV.
10, 1970; VOL. 36, NO. 55, MARCH 20. 1971,
VOL. 36, NO. 128, JULY 2, 1971. EACH VEHICLE
TESTED ONCE AFTER BEING TUNED TO MANU-
FACTURERS SPECS. IF EVIDENCE OF TAMPERING
WITH EMISSION CONTROL SYSTEM FOUND.
VEHICLE WAS REJECTED.
INSTRUMENTATION
HC - FLAME IONIZATION ANALYZER
CO, C02 - NON-DISPERSIVE INFRARED (NDIR)
NO. NOx - CHEMILUMINESCENT ANALYZER
HC - FLAME IONIZATION ANALYZER
CO. C02 - NON-DISPERSIVE INFRARED (NDIR)
NO. NOx - CHEMI LUMINESCENT ANALYZER
CITIES:
HOUSTON. LOS ANGELES, DENVER, CHICAGO. ST. LOUIS, AND
WASHINGTON. D. C.
DETROIT.
TOTAL VEHICLES TESTED:
1020
243
-------
Table 2
SUMMARY OF EMISSIONS (ARITHMETIC MEANS)
FOR 1968 AND 1969 MODEL VEHICLES
HYDROCARBONS
CARBON MONOXIDE
NOx
c
ppm
gm/mi
Percent
gm/mi
ppm
gm/mi
1968 MODEL VEHICLES
KANSAS CITY
HOUSTON
320
385
3.81
4.48
2.04
2.25
46.3
51.3
1389
1145
4.69
4.50
1969 MODEL VEHICLES
KANSAS CITY
299
3.65
1.81
42.4
1562
5.34
HOUSTON
306
3.67
1.92
44.1
1337
5.33
68-69 STANDARDS: HC-275 ppm, CO-1.5% (HC-410 ppm, CO-2.3% FOR VOLKS)
33
-------
Table 3
SUMMARY OF EMISSIONS (ARITHMETIC MEANS)
FOR 1970 MODEL VEHICLES
HYDROCARBONS
gms/mi
CARBON MONOXIDE
gms/mi
NO_
*c
gms/mi
KANSAS CITY
2.66
33.9
4.63
HOUSTON
3.00
40.5
4.69
LOS ANGELES
2.46
32.5
5.04
DETROIT
3.08
39.4
4.27
WASHINGTON
2.99
29.8
4.82
DENVER
4.25
65.1
3.83
TOTAL EXCLUDING DENVER
2.83
35.3
4.69
70-71 STANDARDS: HC-2.2 gm/mi. CO-23 gm/mi
34
-------
Table 4
SUMMARY OF EMISSIONS (ARITHMETIC MEANS)
FOR 1971 MODEL VEHICLES
NOv
HYDROCARBONS
CARBON MONOXIDE
c
gms/mi
gms/mi
gms/mi
LOW MILEAGE ENGINES - PHASE 1
HOUSTON
1.58
27.1
3.35
LOS ANGELES
1.29
19.7
3.07
DETROIT
1.77
28.5
3.08
DENVER
2.63
64.0
2.86
STABILIZED ENGINES - PHASE 2
HOUSTON
2.41
33^7
4.27
LOS ANGELES
1.68
25.2
3.41
DETROIT
2.77
31.5
3.80
DENVER
3.16
45.2
3.19
70-71 STANDARDS: HC-2.2 gm/mi, CO-23 gm/mi
35
-------
Table 5
SUMMARY OF SEVEN-MODE EMISSIONS TESTS 1968-1971 MODEL VEHICLES
MANU-
FACTURER
N
MEAN
MILES
% BELOW
STANDARDS
HYDROCARBONS gm/mi
CARBON MONOXIDEgm/mi
NOy
Ac
gm/mi
IKI
HC
CO
BOTH
ARITHMETIC
MEAN SD
GEOMETRIC
MEAN SD
ARITHMETIC
MEAN SD
GEOMETRIC
MEAN SD
ARITHMETIC
MEAN SD
GEOMETRIC
MEAN SD
68 TOTAL
967
34.3
46
35
23
4.13
3.16
3.66
1.56
49.0
29.2
41.1
1.82
4.60
1.94
4.11
1.70
69 TOTAL
982
23.2
53
42
30
3.66
139
3.38
146
43.3
23.0
37.9
1.69
5.33
1.83
4.95
1.53
70 EXCLUDING
DENVER
1829
11.0
32
35
19
2.83
1.53
2.60
1.48
35.3
21.9
29.2
1.89
4.69
1.73
4.36
1.49
70 DENVER
352
10.1
5
3
.3
4.25
2.39
3.92
1.45
65.1
29.3
59.1
1.56
3.83
1.51
3.53
1.52
71 PHASE 1
EXCLUDING
DENVER
71 PHASE 1
DENVER
225
144
0.25
0.21
87
38
60
2
58
1
1.47
2.63
0.74
1.03
1.34
2.46
1.55
1.46
23.7
64.0
15.0
27.6
19.5
58.5
1.89
1.54
3.16
2.86
1.10
1.42
2.99
2.54
1.39
1.63
71 PHASE 2
EXCLUDING
DENVER
71 PHASE 2
DENVER
225
144
5.4
5.2
59
29
49
10
38
6
2.12
3.16
1.22
4.01
1.86
2.71
1.68
1.66
29.0
45.2
20.0
22.8
23.0
39.8
2.02
1.71
3.74
3.19
1.52
1.41
3.46
2.88
1.49
1.61
36
-------
Table 6
COMPOSITE EMISSION LEVELS AS DETERMINED BY 1972 TEST PROCEDURES (EXCLUDING DENVER)
MEAN
% 8ELOW
LEVEL t
HYDROCARBONS gm/mi
(K)
NOv
ARITHMETIC
GEOMETRIC
ARITHMETIC
GEOMETRIC
ARITHMETIC
GEOMETRIC
YEAR
HC
CO
Xc
MEAN
SD
MEAN
SD
MEAN
SO
MEAN
SO
MEAN
SO
MEAN
SD
PRE 68
57
20
80.0
0
10
40
7.07
2.32
6.72
1.39
80.3
33.2
73.5
1.57
3.89
2.15
3.24
1.98
58
20
86.7
5
5
50
10.74
10.65
8.58
1.81
85.2
38.6
76.5
1.65
3 70
2.44
2.92
2.14
59
21
79.7
0
19
33
11.37
8.70
9.19
1.89
81.6
45.0
70.7
1.74
4.47
2.39
3.71
1.99
60
16
65.4
13
13
50
9.53
8.48
7.40
2.01
86.8
44.8
75.0
1.80
394
2.37
3.26
1.94
61
23
67.6
17
17
61
6.49
3.64
5.69
1.67
84.6
59.8
66.5
2.06
3.12
2.25
2.53
1.97
62
38
72.7
5
8
40
9.57
6.87
7.83
1.88
85.5
38.8
77.5
1.60
3.34
1.57
2.93
1.75
63
55
77.3
0
2
47
10.26
8.11
8.85
1.62
105.8
41 8
98.1
1.49
3.63
2.15
3.07
1.81
64
64
72.1
0
3
42
8.14
3.70
7.49
1.49
91.0
41.4
82.3
1.60
3.61
1.71
3.19
1.70
65
80
62.9
1
4
53
12.17
13.29
9.24
1.89
97.9
40.7
90.0
1.52
3.27
1.66
2.87
1.70
66'
67
61.8
2
3
45
9.20
6.18
8.07
1.61
103.3
43.2
94.3
1.56
3.44
1.71
3.01
1.75
67-
54
54.6
0
2
46
8.19
3.61
7.51
1.51
103.5
47.7
94.5
1.54
3.26
1.45
2.90
1.71
TOTAL PRE-68*
458
68 6
2
5
46
9.56
8.06
8.02
1.70
95.2
43.5
85.6
1.61
3.51
1.87
3.01
1.79
CALIFORNIA
66
16
65.7
19
19
38
8.72
8.64
6.62
2.00
78.1
38.3
70.2
1.61
3.23
1.44
2.91
1.64
67
17
56.4
12
6
53
6.22
3.52
5.52
1.63
81.4
38.0
74.6
1.52
3.30
1.45
2.98
1.61
CONTROLLED
1968
84
46.4
20
25
35
6.25
7.10
5.13
1.70
78.6
59.0
63.8
1.86
4.22
1.88
3.75
1.69
1969
89
39.5
12
10
11
5.96
4.72
5.19
1.59
73.8
36.2
66.0
1.62
532
2.01
4.93
1.50
1970
86
28.7
30
33
9
4.41
2.18
4.08
1.47
56.5
27.2
50.2
1.66
4.93
1.64
4.66
1.41
1971
101
15.6
61
46
13
3.44
1.38
3.21
1.45
47.5
27.2
40.4
1.80
4.75
1.73
4.42
1.48
TOTAL 68 71
3G0
31.9
32
29
17
4.96
4.50
4.27
1.62
63.5
40.9
53.5
1.80
4.81
1.85
4.43^
1.54
CARBON MONOXIDE gm/mi
NOy gm/mi
EXCLUDING CALIFORNIA 6&67
HC
CO
NOXe
3.4 gm/mi
39-0 gm/mi
3.0 gm/mi
Table 7
COMPOSITE EMISSION LEVELS AS DETERMINED BY 1972 TEST PROCEDURE (DENVER ONLY)
YEAR
N
MEAN
MILES
-------
Table 8
COMPOSITE EMISSION LEVELS AS DETERMINED BY 1975 TEST PROCEDURES (EXCLUDING DENVER)
YEAR
MEAN
MILES
IK)
% BELOW
LEVEL ~
HC
CO
NOv
HYDROCARBONS gm/mi
ARITHMETIC
MEAN SO
GEOMETRIC
MEAN SD
CARBON MONOXIDE gm/mi
ARITHMETIC
MEAN SD
GEOMETRIC
MEAN SO
NO
y gm/mi
Ac
ARITHMETIC
MEAN SD
GEOMETRIC
MEAN SD
PRE 68
57
58
59
60
61
62
63
64
65
66*
67*
TOTAL PRE-68*
CALIFORNIA
66
67
CONTROLLED
1968
1969
1970
1971
TOTAL 68-71
20
20
21
16
23
38
55
64
80
67
54
458
16
17
89
86
101
360
80.0
86.7
79.7
65.4
67.6
72.7
77.3
72.1
62.9
61.8
54.6
68.5
65.7
56.4
46.4
39.5
28.7
15.6
31.9
40
45
38
56.
61
49
53
39
49
42
48
46
31
41
26
10
11
15
15
6.63
10.03
10.80
8.79
5.94
8.87
9.43
7.28
11.18
8.26
7.3B
8.74
7.84
5.33
5.54
5.19
3.90
3.06
4.37
2.32
10.20
8.24
8.09
3.27
6.84
7.76
3.29
12.62
5.48
3.28
7.63
8.34
3.52
7.07
4.26
1.95
1.26
4.30
6.25
7.93
8.70
6.82
5.20
7.09
8.04
6.68
8.36
7.21
6.75
7.26
5.81
4.60
4.45
4.53
3.60
2.85
3.75
1.42
1 84
1.88
1.98
1.68
1.94
1.65
1.50
1.92
1.63
1.52
1.73
2.03
1.70
1.73
1.56
1.46
1.44
1.61
81.4
78.2
77.3
81.6
79.7
78.0
96.5
81.7
87.9
91.0
93.6
86.5
65.2
67.2
67.8
61.7
48.2
40.1
53.9
30.7
36.8
45.9
43.4
56.8
36.6
38.4
38.8
36.8
38.9
44.9
40.3
36.6
37.0
57.5
31.0
24.7
24.5
37.9
75.6
69 7
65.8
70.0
61.8
69.6
89.2
73.1
80.8
823
85.1
77.4
568
59.7
52.7
55.0
42.0
33.5
44 4
1.50
1.67
1.79
1.83
2.12
1.67
1.50
1.64
1.51
1.60
1.55
1.63
1.72
1.63
1.96
1.63
1.72
1.86
1 86
3.83
3.62
4.49
3.94
3.06
3.33
3.64
3.66
3.37
3.57
3.28
3.54
3.40
3.42
4.34
5.45
5.05
4.81
4.92
2.19
2.31
255
2.48
2 35
1.57
2.15
1.77
1.66
1.82
1.46
1.91
1.54
1.50
1.92
202
1.67
1.78
1 88
3.17
2.88
3.61
3.18
2.43
2.91
3.08
3.24
2.97
3.13
2.93
304
3.04
3.08
3.85
5.06
4.78
4.47
4.52
EXCLUDING CALIFORNIA 6fr67
HC
CO
NOXc
3.4 gm/mi
39.0 gm/mi
3.0 gm/mi
Table 9
COMPOSITE EMISSION LEVELS AS DETERMINED BY 1975 TEST PROCEDURE (DENVER ONLY)
YEAR
N
MEAN
MILES
-------
Table 10
COMPARATIVE MEAN EMISSION LEVELS
RENTAL VEHICLE AND 1968-1969 GREAT PLAINS SURVEILLANCE PROGRAMS
PKUGRA.H
* *
* rtAi^uFAC rUkfcR *
~
*
*
MAKt
*
* CliJ *
* «
* - fc
I-!Ea N *
Ll S*
( K ) *
ii t. L G .*
i> f AKT.'Ak'jS
*
*
v
*
* HC * Cu * tiCTH *
' C
Mc.
PPM
*
$
*
*
C 1
RENTAL
GT.PLAINS
CHKY C0«P
225
20
75
14.2
27.*
95
55
55
*0
55
2 k
215
i It
1 .71
1 . ? 7
RENTAL
GT.PLAINS
CHRY CORP
318
28
80
8. 1
31.9
82
39
86
33
75
21
240
342
1.18
1. 92
RENTAL
GT.PLAINS
RENTAL
GT.PLAINS
FGMGCU
FQMJCU
302
390
95
79
72
80
10.0
30.2
12.5
31.6
37
1 5
75
51
80
t>8
76
53
37
11
64
40
314
413
252
286
1.23
1.37
1 .3b
1.57
RENTAL
GT.PLAINS
CHEVROLET
307
36
81
12.0
27.9
2 5
20
44
31
16
9
355
*14
1.56
1.87
RENTAL
gt.plains
19t>6 CHtV
3^7
*8
80
17.0
32.5
5h
13
42
8
29
1
2 89
440
1.76
3.2 6
RENTAL
GT.PLAINS
RENTAL
GT. PL A I,MS
RENTAL
GT.PLAINS
PONTIAC
PONT I AC
VULKS
350
400
91
21
75
21
7tt
30
80
11.2
25.8
7. 1
2 b. 0
11.5
20. 9
86
59
100
8 7
o3
43
36
43
48
37
80
59
34
35
4c
35
53
35
215
2 57
177
239
391
5c^
l.Gl
l.o2
1.50
1.61
2.1 7
2.37
«»~ APPLICABLE S T AiNOAKOS J HC-2 7 5 PP* CG-1.5i ( 4 lvj Pf M, 2 . 5 % h L'.k \$) ~ **
-------
NOX
BASE
9
69
31
IT
21
87
62
22
3
3
6
5
I
6i
36
8
10
7
2
6
Table 11
COMPARATIVE EMISSION LEVELS
NOX BASELINE PROGRAM VS SIX CITY PROGRAM
~
HC
*
CU
NOX
HVORUCAABONS GN/MI
CARdCM MONOXIDE GM/MI
NUX
GM/MI
ME AM
HllfcS
X BELOW 3.4
t BELOW 39.
* BELOW
3.0
*
GM/Ml
GN/MI
GN/MI NOX
bASE
SIX
CITY
NOX
BASE
SIX
CITY
NO A
BASE
*
SIX
CITY
iU
NOX
SIX
NOX
SIX
NOX
SIX
NOX
SIX
6
; irv
BASE
city
BASE
CITY
BASE *
CUT
BASE
CITY MEAN
SO
MEAN
so
MEAN
* SO
MEAN
SO
*
MEAN
SO
*
MEAN
* SO
49
STATE VEHICLES
3
6. 1
10.2
69
100
67
100
78
0 2.73
0.70
1.99
0*42
36.2
14.7
23.9
9.3
2.43
0.80
4.16
0.05
a
6.6
17.3
43
25
31
12
32
0 3.67
0*65
4*34
1.37
50.9
21.0
61.2
16. 1
3*51
1.45
6.06
1.55
4
6*6
11.6
56
25
43
25
35
0 3.38
0.61
4*67
1.64
44.2
21.6
60.5
21. 1
3.34
1.31
6.39
1.90
3
6.0
24.2
47
66
29
0
24
0 3.60
0*81
3.46
0.55
51.3
19.7
54.7
13.0
3*42
1*35
5*60
1.50
1
7.0
IS. 1
19
0
14
0
33
0 4.15
0.75
5.69
60.4
17.9
63.5
3.65
1.73
6.13
20
6.9
16.2
25
SO
32
45
9
5 4.02
0.95
3.65
1.21
49.1
17.9
50.2
33.5
4.55
1.40
6.03
1.88
15
6.0
16.5
24
60
29
47
10
7 4.08
0.96
3.55
1.30
50.5
17.4
43.3
20. 1
4.51
1.42
6.24
1.60
4
7. 1
14.9
23
25
41
50
9
0 3.98
0.94
3.56
0.53
45.6
10.5
45.0
11.8
4.70
1.42
5.96
2.75
I
6.7
17.0
67
0
33
0
0
0 3.12
. 0.62
5.41
45.5
27.3
170.5
-
4.13
1*04
3.12
50
STATE VEHICLES
2
5.7
15.1
67
100
67
100
100
0 3.18
0.67
2.83
0*34
34.6
6.5
27.9
2.4
2.31
0.73
3.04
0.89
5
7,9
10.6
33
80
17
60
100
20 3*74
0.75
3.86
3.56
49.4
13.4
30*5
16.5
2.10
0.43
4*03
1.51
5
6.1
18.6
40
80
20
60
100
20 3.66
0.81
3.86
3.56
47.6
14.1
30.5
16.5
2.03
0.45
4.03
1.51
0
6.a
0
0
loo
4.14
50.8
2*43
*8
9.0
15. 1
60
58
52
40
44
15 3.39
1.28
3.32
1.16
41.5
22.0
49.0
27.2
3.16
0.97
4.44
1.48
2T
7.1
14. e
64
63
69
44
50
19 3.25
1.05
3.14
1. 12
34.9
2O.0
45.6
25.2
3.02
0.67
4. 34
1.45
8
10.9
18.3
25
25
38
25
36
12 4.51
1.47
4.09
0.94
56.6
31. 1
55.7
30.7
2.81
0.95
4.49
1.52
6
9.4
16.4
70
83
30
50
40
0 3.33
1.59
2.82
0.65
42.2
14. 1
37.2
14.7
3.68
1.52
5.00
0.08
5
7.8
12.6
57
40
29
20
29
20 J.50
1.11
4.11
1.49
54.2
20.7
82.4
28.2
3.62
0.99
3.87
1.75
2
6.2
7.4
100
100
0
50
50
0 1.42
0.24
2.16
0.02
51.6
4.6
40.2
15.2
2.77
1.00
5.35
3.25
i
9.2
17.3
100
3)
33
33
10J
100 2.61
0.45
4.90
1.46
44.6
9.7
56.6
20. 3
1.46
0.43
4.09
0.43
0A« BASED UPON 19J2 CVS IEST PRUCEOUfcE
-------
FIGURES
41
-------
FACTOR DISTRIBUTION FOR KEY MQDE DATA-VEHICLES FAILING 7 MODE
Id
/1 vo u it.* 11/ it-.c ics °i ir -z o l- ^ ; : i
EACH,»"cilUALS "» V
P.I INI S
172
16a
164
160
*
* * *
* » *
150
... 152
1*8
1*0
136
* * * *
* * « _ .
132
12*
. _12Q
» *
* »' » * *
116
112
108 **~**
104 .*****
too
9b
*«***
42
68
*
*¦»«**«***
84
eo
»*»*»
*****
76
72
», » * * * * * -«
* * ****»
68
64
** **«»*»
»****»»#'
60
56
52
48
* * * *,* * * »
* * ~ * * * * +
44
40
* * * * * * . *
36
_J*_ . . _
28
Zh
20
16
»****~**
~ * *
****»**««»
********» +
***«~*»*
12
8
4
«
~
« * * ~ ' ~ * ~ * « »
. * * * * * i$ * « * *
INTERVAL Ijj
FACTOR
FREtiUtNCY
bJ. 1«J3« 15u. si.,.
DISTRIBUTION FOR KEY MQDE DATA-VEHICLES PASSING 7 MODE
h 1 L-j > i **2 1 / 3 1 \J i. i C o 0 ».''* w
EACH,* EJJA'U, 2
PUINTS
80
78
76
7*.
*
72
7U
* *
* *
68
66
* *
64
62
*
* *
60
58
*
56
54
*
52 .
50
.
~
48
46
*
4*
42
*
~ * *
40
38
36
34
*
~ ~ ~
32
30
* * » *
* * *
28
26
~ ~
* * *
24
22
*
9
~
*
20
LB
*
*
16
14
*
*****
*
12
10
*
* * * »
8
6
~
*
*«*»»
***«*
4
2
***.
* * * * * * *
INTERVAL IS* lub. 1st* lvu. 2i;3.
Figure 1 HISTOGRAM OF DISCRIMINANT FACTORS FOR KEY MODE DATA -
PASS SEVEN MODE/FAIL SEVEN MODE
42
-------
68-69 X J
70-71
0 20 40 60 80 100
VEHICLES - CUMULATIVE %
200
0 20 40 60 80 100
VEHICLES - CUMULATIVE %
E
I
X
o
0 20 40 60 80 100
VEHICLES - CUMULATIVE %
Figure 2 1975 CVS EMISSION LEVELS VS. CUMULATIVE PERCENTAGE OF TEST
VEHICLES FOR PRECONTROL, 68-69 AND 70-71 MODEL YEARS
(ALL CITIES EXCLUDING DENVER)
43
-------
350
300
250
5
a.
6
I
o
-fe.
¦t*
200
150
a
o
o
A
G
A
GB
G
A
A
G
<3>
~
G
A
%
8
6
o
GALAXIE
429 CID
~
GALAXIE
390 CID
o
FAIRLANE
250 CID
A
FAIRLANE
351 CID
A
MUSTANG
302 CID
a
LTD
302 CID
a
a
A
o
O
A
A
O
~
*
A
&
100
_i_
_i_
12 16 20
MILEAGE - K MILES
24
28
32
Figure 3 HYDROCARBON EMISSIONS vs MILEAGE - FIVE 1969 MODEL IN-HOUSE
SURVEILLANCE FORDS
------- |