78-3
Technical Report
Effects of Low Ambient Temperature on the Exhaust Emissions
and Fuel Economy of 84 Automobiles in Chicago
October, 1978
by
Wayne Heinmiller
Technology Assessment and Evaluation Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air, Noise and Radiation
-------
Abstract
This report describes the results of a project in which pairs of tests
were conducted on 84 in-use passenger cars, once under low temperature
conditions (16°F to 57°F), and again under standard laboratory condi-
tions. Each sequence included the 1975 Federal Test Procedure (exhaust
emissions only), the Highway Fuel Economy Test and three short cycle
tests. The vehicles were obtained randomly from private owners in the
Chicago area and represented the 1972-1977 model years. They were
tested in an "as-received" condition. The work was performed during the
period from January through March, 1978 using a test cell exposed to
prevailing outdoor ambient conditions.
Results show that HC and CO are most sensitive to cold temperature,
while NOx is affected only slightly. At an average of 23°F, vehicles in
the fleet produced approximately 50% more HC, 75% more CO, and 5% less
NOx when compared to results under normal conditions. Fuel economy
suffered by an average of 7%. The initial (cold transient) phase of the
FTP is the most negatively affected by colder temperatures. The final
(hot transient) phase is least sensitive to low temperature operation.
Vehicle fleets from manufacturers which use different control tech-
nologies were found to behave considerably differently at low temperatures.
-------
Introduction
"~~rt
Serious attempts have been made to reduce air pollution created by motor
vehicles in the past several years. Emission standards for new vehicles
have been set by the federal government to try to control this problem.
Tests are performed to determine if new vehicles meet these standards.
However, these tests are performed under a narrow range of environmental
conditions, while vehicles in actual use operate under more varied
conditions. To more fully evaluate success in controlling total emis-
sions, the contribution of motor vehicles must also be evaluated under
environmental conditions that differ from the narrow range used for
certification.
One of the larger variations that a vehicle must undergo in-use is
temperature. To determine if pollution control efforts now in use are
reducing emissions across the entire spectrum of temperatures normally
encountered, testing at various temperatures is necessary. Certification
tests are run between 68° F and 86° F, towards the higher end of the
temperature range in which most vehicles usually operate. Of interest
then, is vehicle emission performance at lower temperatures.
Objective
The objective of this program was to gather information on how exhaust
emissions from typical passenger cars are affected by low ambient
temperatures.
Program Design
General Test Design - The goal of the project was to provide two sets of
emissions results on each vehicle, one set under conditions of low
ambient temperature and the other under standard test conditions. Each
vehicle underwent a test sequence at a low temperature followed by a
test sequence under normal laboratory temperatures. Tests performed
included the 1975 Federal Test Procedure, Highway Fuel Economy Test,
Federal Short Cycle Test, Two-Speed Idle Test, and the Federal 3-Mode
Test.
Selection of Test Vehicles Vehicles for low-temperature testing were
selected from the 433 vehicles used in the FY77 Emission Factors program
in Chicago. These vehicles were of the 1972-1977 model years and were
obtained from randomly-selected local residents. A vehicle's parti-
cipation in the low-temperature portion of the program was determined by
the testing schedule and the forecast of the outside temperature at the
time it was scheduled for its first test.
In all, 84 vehicles were selected for low temperature tests. Origi-
nally, 150 vehicles were scheduled, but the rapid onset of warm weather
prevented testing all of them. The original 150 vehicles were chosen to
be representative of the national fleet in age, make and type. Since
-------
TABLE 1: Fleet Breakdown by Make and Medal Year
American Motors
Buick
Cadillac
Chevrolet
Chrysler
Dodge
Ford
Lincoln
Mercury
Oldsmobile
Plymouth
Pontiac
Volkswagen
Totals
77 76
Model Year
75 74 73
72 Total
1
2
2
6
-
2
9
—
1
4
1
2
1
1
2
1
6
—
1
4
—
—
3
-
1
—
-
2
1
3
1
1
3
1
—
2
-
-
•
—
1
-
-
-
1
1
—
-
1
1
—
••
-
1
-
2
-
1
2
-
—
1
1
—
~
—
1
-
2
-
1
1
-
1
-
-
1
""
2
9
4
19
1
6
20
1
2
11
3
4
1
31 19 14
84
TABLE 2: Fleet Breakdown by Number of Engine Cylinders
Number
of
Cylinders
8
6
4
77
24
4
3
76
16
2
1
Model
75
12
1
1
Year
74
3
2
0
73
7
1
0
72
6
1
0
Totals
68
11
5
-------
-2-
not all of the 150 vehicles were tested, the actual sample is not as
representative as originally planned. A breakdown of the test fleet by
year, make and number of cylinders is given in Tables 1 and 2.
Test Facilities Testing was performed during January, February and
March of 1978 at the facilities of Automotive Testing Laboratories in
Bensenville, Illinois, a suburb of Chicago. This facility consisted of
three separate areas (See Figure 1). One area was used for both the
cold soak and low-temperature dynamometer tests. The large garage doors
in this area were left open in an attempt to maintain temperatures in
the dynamometer and test area equal to those outside. The second area
contained another dynamometer for normal laboratory testing, as well as
all analyzer equipment for both the cold and warm testing areas. The
third section was used for soaking the vehicle which would be tested
inside the laboratory. Vehicles undergoing the cold weather test were
soaked outside.
There were several differences between the equipment set-up in the cold
test area and the normal set-up for performing basic emission testing.
To allow the dynamometer power absorber unit (PAU) to operate in below-
freezing temperatures, anti-freeze had to be added to the water in the
system. The system was set up to recycle the fluid rather than to
operate on a once-through basis. After passing through the PAU, the
fluid was run through a heat-exchanger and then into a 55 gallon drum
used as a reservoir. From the drum, it returned to the PAU. The
flywheels on the dynamometer were completely enclosed and the bearings
were heated by radiation from incandescent lights. The exhaust tubing
was wrapped in insulation and passed through the wall to the constant
volume sampling unit and analyzer in the warm test area.
Test Procedures Vehicles were allowed to soak under outdoor conditions
between 12 and 24 hours before testing.
Cold tires inflated to the pressure necessary for dynamometer operation
were expected to produce problems with tire breakage, so the drive tires
and wheels were removed from the vehicle and stored in a warm area while
the vehicle was soaking outside. The vehicles were tested in an "as-
received" condition. No maintenance work or adjustments were performed
before testing.
For the cold testing, the vehicle was tested with its own tank fuel.
For the warm tests, the tank was drained and refilled with appropriate
Indolene fuel. No heat builds were performed on the fuel tank and no
evaporative emission tests were performed.
-------
FIGURE 1: Layout of Test Facilities
MAIN
SOAK
AREA
(WARM)
Doorway
(closed)
WARM TESTING AREA ._ ,_,
(Outside door open)
Constant
Volume
Sample
Dynamo
ter
LOW-TEMPERATURE
TESTING AREA
(Doorway S
covered P
by canvas]?
Dynamometer
O G^OO (tail-pipe probe]
ooooo
Power Absorber
Unit
nsulated Cold Soak
Tubing to CVS Area
(Outside door open)
Offices
L
Offices
© Warm Soak for
Drive Wheels
(3) (5) and Tires
-------
-3-
General Results
Different methods of quantifying the data to allow analysis and com-
parison were investigated. Ratios of cold emissions to warm emissions,
differences between warm and cold emissions and absolute levels of
emissions were all examined. Statistical analysis showed that all
figures generated for comparison are strongly correlated to a vehicle's
performance on the standard FTP, which depends on the condition of the
engine and the degree to which it is tuned to specifications. Because
these variables are not quantifiable, it is not possible to directly
compare results from different vehicles. However, it is possible to
note general trends in the vehicle population tested.
For each of the six model years of vehicles tested, the mean composite
hydrocarbon CHCl and carbon monoxide (CO) levels are significantly
higher in the low-temperature FTP than in the normal FTP(See figures 2
and 3, and table 3). These means are likely to be exaggerated by out-
liers. For the model years 1975-1977, the majority of the vehicles
tested lie below the mean (see table 4). Overall, 86% of the vehicles
produced higher HC levels in the low-temperature FTP, while 87% produced
more CO. The significance of the differences in oxides of nitrogen
(NOx) caused by temperature varies with model year. With model years
1972-74 grouped together, and using a signficance level of .80, it may
be concluded that the cold and warm NOx emissions do not differ. To
make the same conclusion with 1977 vehicles, a much tighter significance
level of .02 must be used.
As the test sequence progressed and the vehicle warmed up, the HC and CO
emissions from the vehicles in low-temperatures generally improved in
relation to the normal-temperature test emissions. In many cases, the
final sample bags from low-temperature tests showed fewer emissions than
the same sample bags in the warm tests.
Results of FTP measurements
Three kinds of comparisons were made. First, a qualitative comparison
between the level of emissions on the warm and cold tests can be made.
Secondly, a quantitative comparison of these figures can be made by
calculating the ratio of cold emissions to warm emissions. This is the
technique which has been used frequently by other researchers investigating
low-temperature emissions. Finally, it is possible to compare the
number of vehicles which produced lower emissions in the warm tests
(compared to their cold test results) to the number of vehicles which
produced lower emissions in the low-temperature tests.
-------
Mean Composite Emission Levels — Low-Temperature and Normal FTP
61
M 2
[x
\
X
X
I
77
(31)
X
X
X
\
76
(19)
"— ^
X
X
X
X
X
X
X
X
f
X
X
X
X
X
X
\
X
1
X
X
X
X
X
\
\
tBB^
X
X
X
X
X
X
X
\
75 74 73 72
(14) (5) (8) (7) .
FIGURE 2: Hydrocarbon (HC) Emissions
Cold
Normal
100
80-
r-I
|60-
to
240-
H
60
20-
0 •
\
X
N
z
/,
X
X
X
?
/
\
X
X
X
X
X
X
^"7**
^
y,
X
X
X
X
X
— M
%
V,
//
2
\
X
X
X
X
x
f 4
y
y
//
X
\
X
X
X
X
f y
y
V
V
y
y
Model Year: 77 76 75 74 73 72
No. of Vehicles: .(31) (19) (14), (5). .(8) (7)
FIGURE 3: Carbon Monoxide (CO) Emission
grams /mile
O !-• NJ W -P
i . t • i
r-ri
X
\
X
\
X
X
X
\
-
\
y
\
\
X
X
X
X
\
X
Model Year: 77 76 75 74 73 72
No. of Vehicles: (31) (19) (14) (5.) (8) (7)
FIGURE 4: Oxides of Nitrogen (NOx) Emissions
-------
TABLE 3
Mean Composite Emission Levels
Cold and Normal FTP
(in grams per mile)
1977
1976
1975
1974
1973
1972
All
(32)
(19)
(14)
(5)
(8)
(7)
Years
COLD
NORMAL
COLD
NORMAL
COLD
NORMAL
COLD
NORMAL
COLD
NORMAL
COLD
NORMAL
COLD
NORMAL
HC
2.
1.
2.
1.
5.
2.
5.
3.
3.
2.
4.
4.
3.
2.
7
7
4
7
3
0
4
9
9
7
4
0
3
2
CO
39.
22.
42.
25.
88.
37.
78.
60.
71.
40.
79.
70.
57.
33.
NOx
8
5
6
0
8
2
0
8
3
8
0
4
1
1
2
2
2
2
2
2
3
2
3
3
4
3
2
2
.3
.1
.7
.4
.3
.4
.1
.9
.5
.8
.0
.7
.7
.6
TABLE 4
Number of Vehicles Below or Equal to Mean FTP Emissions
HC
CO
NOx
1977
1976
1975
1974
1973
1972
(32)
(19)
(14)
(5)
(8)
(7)
Cold
20
12
9
2
4
2
Normal
18
13
8
3
3
3
Cold
- 19
13
8
3
4
4
Normal
18
13
9
2
4
3
Cold
17
13
8
2
4
3
Normal
21
11
8
3
5
4
-------
FIGURE 5
Distribution of Temperatures for Low-Temperature Tests
20
18
16
<4-l W
o > 3
o w
C o)
0) t-i
3 0)
cr a
a) s
t-i (U
fe H
14
12
10
8
6
4
2
0
Temperature Range (CF)
-------
-4-
Overall Fleet Emissions and Fuel Economy
The average composite FTP emissions of the entire fleet when tested at
low temperatures were higher than when tested under normal conditions
(see table 3). However, 11 cars produced less HC in the low temperature
test than they did in the warm test, 12 cars produced less CO in the
cold test, and 34 produced less NOx in the cold test. Hydrocarbon
emissions in low temperatures ranged from one half to twenty times the
emission levels under normal temperatures. Carbon monoxide ratios ran
from one half to eighty times normal levels, and oxides of nitrogen went
from a quarter to twice normal levels.
The entire fleet obtained 7% fewer miles per gallon on the FTP in cold
temperatures than in warm. Still, 14 cars or 16% got better fuel
economy in the low temperature FTP. The largest increase in fuel
economy in low temperatures was 9%, while the greatest loss was 36%.
Temperature and Composite Emissions
To examine the effect of different temperatures on emissions, groups of
vehicles in 3 temperature ranges were examined. The first group con-
sisted of 14 vehicles tested at temperatures from 16° to 25° F (average
23°), the second group of 26 vehicles were tested from 26° to 35° F
(average 32°) and the third group of 13 vehicles were tested from 46° to
55° (average 50°). Each group had vehicles from different manufacturers
and model years.
Hydrocarbon emissions showed an increase as temperature decreased. The
fleet of cars around 23° showed a 75% increase in hydrocarbons, the
fleet at 32° showed a 54% increase, and the fleet at 50° showed a 12%
increase (see figure 6).
Carbon monoxide emissions showed a greater sensitivity to temperature
decreases. The 23° group of vehicles showed a 82% increase in CO, the
32° group increased CO emissions by 74%, and the group at 50° produced
21% more emissions (see figure 7).
Oxides of nitrogen were not very sensitive to temperature differences.
The 23° and 32° group showed 7% and 6% increases, respectively (see
figure 8). The group at 50° showed no signficant increases in NOx after
an analysis of differences between paired tests.
The test groups at 25° and 32° each showed a 9% loss of fuel economy
(see figure 9). The group at 50° showed only a 1% loss in fuel economy.
-------
Mean Emissions at Various Temperatures Normalized to Standard Conditions
(Composites)
FIGURE 6
z.u •
1 C .
1 . J
0.5 •
n -
Hydrocarbons (HC)
/
/
/
/
/
,/
/
/
/
/
/
/
/
/
23 32 50 Normal
Temperature Group (°F)
2.0i
1.5-
1.0-
0.5-
FIGURE 7
Carbon Monoxide (CO)
23 32 50
Temperature Group
f
Normal
2.0
1.5
1.0-
0.5
FIGURE' 8 •
Oxides of Nitrogen (NOx)
23 32 50
Temperature Group (°F)
f
Normal
2.0-
1.5-
1.0-
o.:
FIGURE 9
Fuel Economy (FE)
y~
171
23 32 50
Temperature Group (°F)
Normal
-------
-5-
FTP: Bag by Bag
The greatest increase in the HC and CO emissions in the low-temperature
tests occurred in the cold transient phase. For the entire fleet, HC
and CO emissions during this phase of the low-temperature FTP were
approximately 3 times the level of the emissions under standard con-
ditions. Over the entire fleet, HC and CO emissions were 27% higher in
the stabilized phase, and about the same (within 5%) in the hot transient
phase (see table 5). Low-temperature NOx emissions were about the same
percent greater than normal test emissions (3 to 6%) in all three phases.
The greatest loss in fuel economy occured during the cold transient
phase when the fleet fuel economy dropped 16% compared with the normal
FTP results. Fuel economy was reduced by 7% in the cold stabilized
phase of the low-temperature tests, and remained the same (within 2%) in
the hot transient phase.
While overall NOx emissions during the cold transient phase were greater
in cold temperatures, the majority of vehicles tested (approximately
60%) actually produced lower NOx emissions during this phase in low
temperatures. Very few vehicles (4-5%) produced fewer HC or CO emis-
sions during the cold transient phase in low temperatures, and all
vehicles had lowered fuel economy (see figure 10).
During the stabilized phase of the low-temperature FTP, a significant
number of vehicles (about 40%) had lower emissions on each HC, CO, and
NOx compared to their results at normal temperature.
A majority of vehicles produced fewer emissions on each HC and CO (60%
and 55% respectively) during the hot transient phase of the low-temper-
ature FTP when compared to the same phase on the normal FTP. The
percent of vehicles producing fewer NOx emissions during this phase
dropped to 31%. Some vehicles (30%) achieved better fuel economy during
this phase in cold weather than they did in tests at normal temperature.
Temperature and Bag-by-Bag Emissions
The results from the groups tested around 23°F and 32°F are very simi-
lar. On the cold transient phase, HC and CO emissions for both fleets
were about three times higher in cold weather than under normal test
conditions (see figures 11 and 12). Both groups suffered a 20% loss in
fuel economy during this phase (see tables 6 and 7). Each of the two
groups had large percentages of vehicles (66-72%) that produced less NOx
in cold weather on this phase (see figures 17 and 18).
-------
Figure 10
Bag by Bag Profile of 84 Vehicles Given Low-Temperature FTPs
Bag 1 Bag 2 Bag 3
Stabilized^ Hot Transient Q
,20
40
•60
80
100
80-
60-
40'
20'
n-
77
77
HC CO NOx FE
HC CO NOx FE
Hatched Area — represents those vehicles with fewer emissions or
better fuel economy on the low-temperature test
Solid (or Cross-Hatched) Area — represents those vehicles with the
same performance on both the low-temperature and warm tests
Blank Area — represents those vehicles with fewer emissions or
better fuel economy on the warm test
Table 5
Bag by Bag Mean Emissions for 84 Vehicles
Low-Temperature and Normal FTPs
(emissions in grams per mile)
COLD
NORMAL
Ratio
HC '
8.5
3.2
2.68
BAG
f CO
142.3
46.9
3.03
NOx
3.5
3.3
1.06
FE
10.6
12.6
0.84
HC
2.3
1.8
1.27
BAG
CO
41.0
32.3
1.27
2
NOx
2.0
2.0
1.03
FE
12.8
13.7
0.93
HC
1.8
1.9
0.96
BAG
CO
23.8
22.8
1.04
3
NOx
3.2
3.1
1.04
FE
14.9
15.3
0.98
-------
Mean Bag by Bag Emissions by Temperature
Values Normalized to Standard Conditions
3.0
2.0J
1.0
»eg
FIGURE 11
Hydrocarbons (HC)
3.0
2.0
1.0
FIGURE 12
Bag 1; Carbon Monoxide (CO)
7
23 - 32 50 Normal
Temperature Group (°F)
23 32 50 Normal
Temperature Group (°F)
FIGURE 13
FIGURE 14
Bag 2; Carbon Monoxide (CO)
^.u-
1.0'
0 •
/
/
/
/
•e> •
/
/
/
/
7"
/
/
7~
/
Z.U-
1.0-
n J
- -
/
/
/
f
/
/
/
/
' ' /
/
ir-
/
/
23 32 50 Normal " 23 32 50 Normal
Temperature Group (°F) Temperature Group (°F)
2.0i FIGURE 15
Bag 3; Hydrocarbons (HC)
l.(
7
2.0
1.0--
FIGURE 16
Bag 3; Carbon Monoxide (CO)
23 32 50 Normal
Temperature Group (°F)
^3 32 50 Normal
Temperature Group (°F)
-------
—6~
On the stabilized phase, in the 23° and 32° groups, between 22% and 36%
of the vehicles in each group achieved lower emissions in each HC, CO,
and NOx. Total HC and CO emissions were higher: 29% and 35% in the 23°
group, and 14% and 20% in the 32° group, respectively (see figures 13
and 14). NOx emissions in both groups were also greater, but by only 7-
10%. Both groups suffered a 6% loss of fuel economy at these temperatures.
For the hot transient sample, the 23° group produced 9% less HC, and 6%
more CO. 70% of the vehicles in the group produced less HC in cold
temperatures and 57% produced less CO. The 32° group produced 5% to 6%
less of both HC and CO, but fewer vehicles produced less HC and CO (55%
and 48% respectively) at cold temperatures than in the 23° group. Both
groups had a 7-10% increase in NOx, with only about 30% of the vehicles
doing better or as well on the cold tests as they did in the warm.
The 50° group had results different than either the 23° or 32° groups.
For the cold transient phase, only a doubling of HC and CO was produced
in the cooler temperatures, (see figures 11 and 12) and only half (54%)
of the vehicles produced less NOx (see figure 19).
In the stabilized and hot transient phases, total HC and CO emissions
for the entire 50° group were reduced between 9% and 16% from the
normal levels. Total NOx emissions were also slightly reduced for both
these phases. Of the thirteen cars in this group, in both the sta-
bilized and hot transient phases, 7 produced less HC, the same number
produced less CO and 8 produced less NOx in the colder temperatures. No
major change in the fuel economy for the entire group occured in either
phase.
FTP Results - Comparison of Two Manufacturers
The differences in low-temperature emissions that exist between vehicles
of different manufacturers is of interest. It is desirable to know if
wide variations in low-temperature emissions exist between manufacturers
or if fleets from different manufacturers would behave similarly. For
this analysis, 16 Ford cars from the 1975-77 model years were chosen for
comparison with 15 Chevrolets from the same years. Average temperature
for the low-temperature tests on the Fords was 42°F and 32°F for the
Chevrolets.
Since the Chevrolets were tested at a lower average temperature, it was
expected that the ratio between composite HC and CO low-temperature and
normal emissions would be higher than for Ford. This was not the case.
HC and CO composite emissions for the Chevrolet cars were only about one
and one half times as high at low temperatures as compared to normal
temperatures, while the Fords produced 2 times their normal HC and
nearly 3 times as much CO as at normal temperatures. The Chevrolets
-------
0)
o
co a
V 03
o C
•H O
JS M-l
111
PL,
100!
80-
CO
0) O ,
4-1 4J O 40-
i-l
O W O
3 CD CD
ftJ (D
31-1 O
i-h
60
CO l-ti
rt O
HC CO NOx FE
HC CO NOx FE
80
LlOO
FIGUR1 17
Bag by Bag Profile
14 Vehicles
(Cold Tests between
16° and 25°F,
average 23°F)
100 Bag l
±. \J\J*
(U
o
co c
£ | 80-
0 M
•rl O 4J
J3 M-l 0)
Q) M 0)
> « H 60-
M-l T3
O t-i rH
QJ O
4-1 4J U
C 4J AO
Q) 0) C 4U
U PQ O
1-1
a) x
p , 4J
g 20
^\
S S
//
/ /
s .
f /
/ /
//
/ s
1
u HC CO NOx FE
Bag 2
Bag
vm
HC CO NOx FE
.0
20
40
60
80
rt T3
3" (D
i-l
W O
ESg
o> n>
o
3
-, ^ -
H CD <
CD ^ CD
CO hh 3*
rt O H-
H O
S CD
O
fD
FIGURE 18
Bag by Bag
26 Vehicles
(Cold Tests between
26° and 35°F,
average 32°F)
HC CO
FE
100
100-
OJ
80
•H O 4J
J2 M-l 0)
Q) H
O-i
M-l T3
O I-l rH
0) O /n
4J 4-1 U 40
C 4-1
CU Q) C
U PQ O
M
0) -C on
p., 4J 20
0J
Bag 1
Bag
HC CO NOx FE
HC CO NOx FE
-0
on
20
40 s:
3* CD
ri
C)^ O
(D CD
rt 3
rt rt
60 {?
CO
rt O H-
HC CO NOx FE
80
100
PI CD
3 co
o
CD
FIGURE 19
Bag by Bag Profile
(Cold Tests between
46°and 55°F,
average 50°F)
-------
TABLE 6
Composite and Bag by Bag Emissions for 14 Vehicles, Low-Temperature and Normal FTP
(Low-Temperature Tests from 16° to 25°F, average 23°?)
(All figures are mean emissions in grams per mile)
COLD
NORMAL
Ratio
HC
10.2
3.6
2. 84
BAG
CO
154.5
45.9
3.37
1
NOx
4.0
3.8
1.07
FE
10.0
12.5
0.80
HC
2.4
1.8
1.29
BAG
CO
55.5
41.0
1.35
2
NOx
2.0
1.8
1.07
FE
12.7
13.5
0.94
HC
1.5
1.7
0.91
BAG
CO
23.1
21.8
1.06
3
NOx
3.6
3.4
1.07
FE
14.9
15.3
0.97
HC
3.7
2.1
1.75
COMP(
CO
67.0
36.8
1.82
ISITE
NOx
2.8
2.7
1.07
FE
12.5
13.7
0.9l|
TABLE 7
Composite and Bag by Bag Emissions for 26 Vehicles, Low-Temperature and Warm FTPs
(Low-Temperature Test from 26° to 35 F, average 32°F)
COLD
NORMAL
Ratio
HC
8.7
3.2
2.73
BAG
CO
158.4
46.2
3.43
1
NOx
3.7
3.8
0.96
FE
10.2
12.7
0.80
HC"
2.3
2.0
1.14
BAG
CO
37.1
30.9
1.20
2
NOx
2.4
2.1
1.10
FE
13.0
13.9
0.94
HC
2.1
2.2
0.95
BAG
CO
21.1
22.5
0.94
3
NOx
3.7
3.4
1.10
FE
15.0
15.2
0.98
HC
3.5
2.3
1.54
COMPO
CO
56.9
32.7
1.74
SITE
NOx
3.0
2.8
1.06
FE
12.7
13.9
0.911
TABLE 8
Composite and Bag by Bag Emissions for 13 Vehicles, Low-Temperature and Warm FTPs
(Low-Temperature Tests from 46 to 55 F, average 50 F)
COLD
NORMAL
Ratio
HC
4.8
2.8
1.71
BAG
CO
93.4
43.3
2.15
1
NOx
3.6
3.4
1.06
FE
11.5
12.5
0.91
HC
1.8
2.0
0.90
BAG
CO
33.7
37.0
0.91
2
NOx
2.0
2.2
FE
13.6
13.6
Q.91J 1.00
HC
1.6
1.8
0.87
BAG 3
CO I NOx
21.8 3.0
26.0 J3.1
1
0.8410.97
FE
15.3
15.0
1.01
HC
2.4
2.1
1.12
COMPO!
CO
42.8
35.3
1.21
5ITE
NOx
2.6
2.7
0.97
FE
13.5
13.7
0.99|
-------
-7-
produced 11% more NOx, while the Fords produced 5% less. Both fleets
suffered about the same reduction In fuel economy (7-8%) (see table 111.
In a bag-by-bag analysis, the fleet of Fords produced higher ratios of
cold to warm HC and CO emissions in all bags compared to the Chevrolet
fleet. However, the Ford fleet produced lower ratios of NOx in all bags
compared to the Chevrolets. Bag by bag losses in fuel economy were
about the same Cby percent) for the two fleets (see tables 9 and 10).
In the cold transient phase (bag 1) 80% of the Fords produced less NOx
in cold weather, while only 53% of the Chevrolets produced less NOx (see
figures 20 and 21)..
The Ford fleet continued into the stabilized phase (bag 2) with a high
percentage (80%) of its vehicles producing less NOx, while the same
figure for the Chevrolets had dropped to 20%. However, in bag 2, 46% of
the Chevrolets produced lower HC levels in cold weather, while 66%
produced less CO, compared with 20% and 12% for the Fords.
In the hot transient phase, all but one Chevrolet produced less HC in
cold weather (93% of the fleet) and 10 of the 15 produced less CO. In
the Ford fleet, 31% of the vehicles produced less HC and 38% produced
less CO. Half the Fords produced the same or lower NOx levels, compared
with only 20% for the Chevrolets.
There could be any number of reasons for these differences but the major
cause could be the difference in control technologies. Ford vehicles
generally use air pumps and monolithic catalysts while General Motors
cars do not usually employ air pumps and have pelleted catalysts.
Thermostats in vehicles by different manufacturers are set to operate at
different engine temperatures to meet emissions at standard temper-
atures. These different settings between manufacturers probably affect
emissions performance at low temperatures. Carburetion and ignition
timing are also set differently between manufacturers to maximize the
effects of their emissions contrql systems.
-------
FIGURE 20
Bag-by-Bag Performance of 16 Ford Cars On FTP
Percent of vehicles with better performance on warm test
100, Bag 1 Bag 2
^
K
\
»
j
I
r \J
20
40
•60
80
. in
80 -
60
40
20
0 .
HC CO NOx FE HC CO NOx FE HC CO NOx FE
Percent of vehicles with better performance on cold test
FIGURE 21
Bag-by-Bag Performance of 15 Chevrolet Cars on FTP
Percent of vehicles with better performance on warm test
80
60
40
20'
n.
^//
//
\
Bag 2
I
Bag 3
0
20
40
60
30
HC CO NOx FE HC CO NOx FE HC CO NOx FE
Percent of vehicles with better performance on cold test
100
NOTES:(1)"Better performance" means fewer emissions or better fuel
economy (FE).
(2) Cross-hatched areas indicate percent of vehicles with same
performance on warm and cold test.
-------
TABLE 9
Bag by Bag Emissions for 16 Catalyst-Equipped Ford Vehicles
Low-Temperature and Normal FTPs
(average temperature of cold tests is 43°F)
(all figures are mean emissions in grams per mile)
COLD
NORMAL
Ratio
HC
10.5
2.3
4.53
BAG
CO
153.9
26.1
5.88
1
NOx
2.9
3.0
0.94
FE
10.5
12.6
0.83
HC
1.7
1.4
1.25
BAG
CO
20.3
15.5
1.30
2
NOx
1.9
2.0
0.91
FE
13.2
13.7
0.96
HC
1.6
1.5
1.06
BAG
CO
12.6
12.6
0.99
3
NOx
2.8
2.9
0.98
FE
15.0
15.1
0.991
TABLE 10
Bag by Bag Emissions for 15 Catalyst-Equipped Chevrolet Vehicles
Low-Temperature and Normal FTPs
(average temperature of cold tests is 32°F)
COLD
NORMAL
Ratio
HC
8.3
2.7
3.07
BAG
CO
118.9
45.5
2.61
1
NOx
3.4
3.2
1.06
FE
10.7
13.0
0.82
HC
1.8
1.9
0.96
BAG
CO
40.2
41.6
0.96
2
NOx
1.7
1.5
1.14
FE
13.7
L14.3
0.95
HC
1.2
1.6
0.73
BAG
CO
20.3
21.4
0.94
3
NOx
3.0
2.7
1.12
FE
15.6
15.8
0.98
TABLE 11
Composite Emissions for Low-Temperature and Normal FTPs
16 Ford Vehicles COLD
(average temperature NORMAL
of cold tests was 43°F) Ratio
15 Chevrolet Vehicles ' COLD
(average temperature NORMAL
of cold tests was 32°F) Ratio
HC
3.5
1.5
2.34
3.0
2.0
1.50
CO
45.7
15.2
3.02
50.8
36.9
1.38
NOx
2.3
2.4
0.95
2.4
2.2
1.11
FE
12.8
13.9
0.92
13.4
14.4
0.93
-------
-8-
Conclusions
1. In general, a vehicle with a cold engine will produce significantly
higher emissions of HC and CO in a low-temperature cold-start test
than in a similar cold-start test under standard conditions.
2. The major contribution to higher low-temperature emissions on the
FTP occurs during the cold transient phase (bag 1). Bag 1 HC and
CO emissions increase more greatly with a decrease in temperature
than do HC and CO emissions in other phases.
3. Changes in temperature only have minor effects on NOx.
4. Differences in emissions produced from a warm engine do not differ
greatly with ambient temperature changes.
5. In the test sequence as run, a vehicle gets worse fuel economy in
low temperatures than in warm. The greatest loss of fuel economy
occurs during the first bag (cold transient) of the FTP. By the
final bag of the FTP (hot transient), there is little difference in
fuel economy compared to standard temperatures.
6. There appears to be a range of temperatures (around 46° - 55°) when
vehicles produce fewer emissions on the stabilized and hot tran-
sient phases without loss in fuel economy.
7. Vehicle fleets from different manufacturers can vary widely in
their low-temperature performance.
-------
-9-
Comparisons with Other Works
Two other major studies have been performed to investigate the effect of
low temperatures on emissions. One was performed by the U.S. Bureau of
Mines , the other by Fisheries and Environment Canada .
The Canadian study concluded, as does this one, that HC and CO increase
sharply at low temperatures, and that low temperatures have their
greatest effect on the cold transient phase. It concluded that NOx
increases only slightly with a drop in temperature, and that NOx is .
affected by about the same amount in each phase. This report supports
this conclusion, but goes on to add the number of vehicles producing
more NOx may change significantly from one phase to the next. The
Canadian study also concluded that converters warm up faster at lower
temperatures, and that engines warm up slower the colder it is. This
study supports the conclusion that the colder it is, the longer it takes
the engine to warm up (compare Figs. 9.1, 9.2, 9.3). For converters,
however, this report may only conclude that it is likely that converters
from different manufacturers warm up differently. This study agrees
fairly well with the Canadian study in regards to the magnitude of HC,
CO, and NOx increases caused by temperature.
The Bureau of Mines study also concluded that HC and CO are most sen-
sitive to temperature, and that the cold transient phase is the most
sensitive part of the FTP. As has been already pointed out, this study
supports these conclusions.
Recommendations
Low-temperature testing can be valuable in providing information about
motor vehicle emissions in colder weather. This information may be
useful in determining how vehicles contribute to air pollution at dif-
ferent times of the year. It could help to determine which emission
controls are best at reducing emissions across the entire temperature
spectrum in which a vehicle may be expected to operate.
In order to provide this information, further tests must be designed to
provide data from which specific conclusions can be made. Ideally, a
test cell should be used in which humidity and atmospheric pressure can
be kept in a fairly narrow range while temperature can be made to vary.
For the best data base, vehicles tested should be tuned to manufac-
turer's specifications, and tested several times at both normal and low
temperatures. This procedure would eliminate the variables of pressure
and humidity. With all vehicles set to manufacturers specification, the
variation due to engine adjustment may be discounted.
-------
-10-
Multiple tests on a vehicle provide more accurate data than a single
test. However, testing in this manner is time consuming. This can be a
difficulty when the vehicles used belong to the general public. Also,
the facilities best suited to this type of testing are expensive and
take time to construct.
A more realistic arrangement would use the same testing facilities and
test procedure as described in this report. Selection of vehicles would
be different. Combinations of engine configurations, emissions control
systems, and vehicle inertia weights of interest would be identified,
and vehicles having these characteristics would be noted. When ob-
taining vehicles for testing, the goal should be to obtain several
vehicles of each configuration. Both a low-temperature and a normal
test sequence should be run after the vehicle has been tuned to manufac-
turer's specifications. By obtaining more than one pair of tests for a
particular configuration, enough data may be obtained to compare per-
formances between configuration groups. The goal would be to gather
enough information on any configuration to be able to predict their
performance at different temperatures. A second phase would be to
determine how these same configurations perform in the general vehicle
population with vehicles in different states of tune. Much of the data
from the tests performed for this report could be used for this analysis.
The information gathered could be used for two purposes. First, those
configurations that are most effective at reducing emissions across a
broad spectrum of temperatures could be identified. Second, by noting
the distribution of the various configurations in different vehicle
populations (city, state, national), it may be possible to better
estimate total emissions from a given population at a given temperature.
References
1) Eccleston, B.H. and Hurn, R.W.;"Ambient Temperature and Vehicle
Emissions", U.S. Bureau of Mines, EPA-460/3-74-028
2) Ostrouchov, Nicolaus;"Effect of Cold Weather on Motor Vehicle
Emissions and Fuel Economy", Fisheries and Environment
Canada, SAE-780084
------- |