EPA-AA-SDSB-80-7
Technical Report
The Effects of Tire Rolling Resistance
on Automotive Emissions and Fuel Economy
by
Randy Jones
and
Terry Newell
May 1980
NOTICE
Technical Reports do not necessarily represent final EPA decisions
or positions. They are intended to present technical analysis of
issues using data which are currently available. The purpose in
the release of such reports is to facilitate the exchange of
technical information and to inform the public of technical devel-
opments which may form the basis for a final EPA decision, position
or regulatory action.
Standards Development and Support Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air, Noise and Radiation
U.S. Environmental Protection Agency
-------
EPA-AA-SDSB-80-7
Technical Report
The Effects of Tire Rolling Resistance
on Automotive Emissions and Fuel Economy
by
Randy Jones
and
Terry Newell
May 1980
NOTICE
Technical Reports do not necessarily represent final EPA decisions
or positions. They are intended to present technical analysis of
issues using data which are currently available. The purpose in
the release of such reports is to facilitate the exchange of
technical information and to inform the public of technical devel-
opments which may form the basis for a final EPA decision, position
or regulatory action.
Standards Development and Support Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air, Noise and Radiation
U.S. Environmental Protection Agency
-------
-2-
I. Introduction
Tires are an important factor in the energy required to
operate a motor vehicle, since approximately 20-30 percent of the
resistive forces experienced by a vehicle in motion are due to the
rolling resistance of the tires. Consequently, tires with lower
rolling resistance result in less vehicle fuel consumption.^/
Exhaust emissions are affected by the load placed on the
engine. An increase in engine load will result in an increase in
oxides of nitrogen emitted from the vehicle, and often will result
in increased total hydrocarbon and carbon monoxide emissions.^/
Since tires affect the load placed on the vehicle, they would
logically be expected to have an effect on vehicle exhaust emis-
sions.
This study was conducted to quantify the effects of tires on
vehicle exhaust emissions and fuel consumption. The test program
involved one test vehicle with four different sets of tires; three
types of radials and one type of bias ply tires.
In general the study was conducted in the manner in which a
vehicle manufacturer would observe tire effects in the EPA certi-
fication process. A series of road coastdowns were performed, and
the dynamometer power absorber adjustment was determined inde-
pendently for each vehicle-tire combination by matching dynamometer
coastdown characteristics to road coastdown characteristics.
The tires and rolling resistance data were provided to EPA by
General Motors. The tires and their measured rolling resistance
coefficients are listed in Appendix A. The following sections of
the report discuss the test methods, test procedures, and results
of the study.
II. Summary of Results
\
Highly significant correlations were found between rolling
resistance, expressed in terms of the rolling resistance coeffi-
cient (RRC) of the tires, and fuel consumption; and between RRC and
NOx emissions. Confidence that the observed relationships reflect
the actual relationships, and were not simply the result of chance
variation in the tests, is greater than 90 percent for RRC and NOx
emissions. In the case of RRC and fuel consumption, this confi-
dence approaches certainty, 100 percent.
Results of a comparison between RRC and CO emissions show a
similar tendency: for higher levels of CO emission to be associated
with higher RRCs. However, these results cannot be stated with as
high a level of confidence as can those concerning NOx or fuel
consumption.
Note: All references in this report are shown by /.
-------
-3-
The relationships between HC emissions and RRC were weaker.
Implications found in the analysis were contradictory: results
of the FTP tests show a tendency for greater emissions of HC to
correspond to higher tire RRCs, while results from the HFET tests
reveal a slight tendency for lowered HC emissions to be associated
with higher RRCs. In both the FTP and HFET results, the tendencies
were very weak, and of little statistical significance.
III. Discussion
The rolling resistance of tires, quantified by a rolling
resistance coefficient (RRC), has a direct effect on the load under
which a vehicle is operated. Since exhaust emissions and fuel
consumption increase with vehicle load, they would logically be
expected to increase with tire RRC. The purpose of this study was
to quantify the relation between tire RRC, and exhaust emissions
and fuel consumption.
The study was conducted similar to the EPA certification
process, in which vehicles are tested for emissions and fuel
economy. The results, therefore, are representative of the tire
effects a vehicle manufacturer would observe in the certification
process.
In order to have a dynamometer simulate the total road load of
a vehicle, the dynamometer power absorber must be adjusted to
reflect the road load characteristics of the vehicle. Currently,
most certification vehicles are tested using dynamometer power
absorption values obtained according to the methods in the "EPA
Recommended Practice for Determination of Vehicle Road Load.'^S/ In
this method the basic concept is to perform a series of road or
track coastdowns with the vehicle. Coastdowns are then performed
on the dynamometer at different power absorber settings, and the
dynamometer power absorber adjustment is determined when the
vehicle dynamometer coastdown time matches that of the road coast-
down . ...
The important steps in this study were therefore:
1. The determination of .a dynamometer power absorber adjust-
ment for each vehicle-tire combination by matching road and dyna-
mometer coastdown characteristics.
2. Fuel economy and emissions testing, based on the standard
Federal Test Procedure (FTP) and Highway Fuel Economy Test (HFET).
These steps are discussed in detail in the following sections of
the report. '.'.
A. Determination of Power Absorber Adjustment
The vehicle used in this study was a 1979 Chevrolet Nova. The
vehicle was tested with four different sets of tires: a bias ply
type, designated as tire "D", and three different radials, desig-
-------
-4- ' ,
nated as tires "A", "B", and "C". The manufacturers and brand
names of the tires tested are withheld by agreement with General
Motors. Detailed vehicle and tire descriptions are given in
Appendix A.
The road coastdown tests were conducted at the Transportation
Research Center of Ohio. Trials were conducted on the straight,
smooth, north-south section of the high speed oval test track.
The dynamometer portion of the study took place at the EPA
Motor Vehicle Emission Laboratory in Ann Arbor, Michigan. All
dynamometer testing was conducted in dyno cell D207.
The road coastdowns provided the speed vs. time characteris-
tics of the vehicle when freely decelerating. The vehicle was
accelerated to a speed slightly greater than 60 mph and allowed to
stabilize for several seconds. The transmission was then shifted
to the neutral postion, and speed vs. time data were collected on
the strip chart recorder until the vehicle speed had dropped to 20
mph. Seven pairs of opposite direction coastdown trials were
conducted for each vehicle-tire combination. A detailed descrip-
tion of the road coastdown procedure is listed in Appendix D.
The speed was read from the strip charts at five second
intervals. Typical 60 mph to 20 mph coastdowns lasted between 60
and 90 seconds, resulting in 12 to 18 speed data points.
The data" "was analyzed to extract the acceleration versus
velocity information from the speed versus time data points* The
result was the calculation of a rolling resistance force coeffi-
cient, and an aerodynamic drag force coefficient for the decel-
erating vehicle. The force coefficients were then corrected to
standard ambient conditions of 68.0"F, 29.0 in. Hg, and zero wind
speed; and a dynamometer 55-45 mph coastdown time was calculated
for the appropriate vehicle inertia weight class. Results of the
coastdown calculations for each set of tires are shown in Appendix
B and summarized in Table 1.
The road coastdown tests for the "B" and "D" sets of tires
were conducted on the same day, and thus under similar ambient
conditions. The coastdowns for the "A" and "C" sets of tires
were likewise conducted on the same day, but about two weeks after
the "B" and "D" tire tests and hence, under different ambient
conditions. The most valid pairwise comparisons are therefore
between the "B" and "D" sets of tires, and between the "A" and "C"
sets of tires.
The vehicle was then tested to determine the vehicle-dyna-
mometer coastdown characteristics and an appropriate dynamometer
power absorber setting for representing the road experience of the
vehicle. Several coastdowns were conducted and the 55-45 coastdown
time was measured at different power absorber settings for each
vehicle tire combination. The result was a mathematical expression
-------
of the form of equation (1), relating dynamometer AHP setting to
dynamometer 55-45 coastdown time. The dynamometer coastdown
procedure is described in Appendix E.
__
AT = b0 + bj(AHP) (1)
where:
AHP = dynamometer 50 mph actual horsepower,
AT = dynamometer 55-45 coastdown time in seconds,
bg,b] are regression coefficients.
The characteristic coastdown time vs. 50 mph dynamometer power
adjustment curves resulting from the dynamometer coastdown tests
are shown in Appendix F. The 50 mph AHP setting for emissions and
fuel economy testing was then determined for each vehicle-tire
combination. The 55-45 coastdown time calculated from the road
coastdown tests is substituted in the appropriate mathematical
expression, yielding the AHP setting. The results of this proce-
dure are listed in Table 2.
B . Emissions and Fuel Economy Tests
The emissions tests in this study were conducted according to
standard EPA testing methods. One day of testing consisted of a
cold start Federal Test Procedure and a Highway Fuel Economy Test.
Evaporative emissions were not measured. The procedure is outlined
in Appendix D.
A total of five FTP-HFET sequences were conducted for each
tire set used in the study. The five test sequence was originally
planned to be conducted on consecutive test days before beginning
testing on another set of tires. After completing the five test
sequence on the bias-ply construction tires, it was decided to
randomize the test order for the remaining tire sets. Randomi-
zation of test order was introduced to prevent the occurrence of
any change in vehicle or dynamometer behavior over the course of
the study being misinterpreted as an effect of tire type. Addi-
tional tests on the bias-ply tires were then conducted before the
conclusion of the study.
IV. Results
This section of the report describes the data obtained during
the dynamometer portion of the test program, the analysis of that
data, and the results obtained from the analysis.
A. Data and Data Analysis
The standard EPA computer analyses of the results of FTP and
HFET tests formed the basis of data used in the statistical
analysis. From these computer outputs, gram-per-mile emission
-------
-6-
rates for hydrocarbons, carbon monoxide and oxides of nitrogen were
taken, in addition to fuel consumption data measured in cnrVkm.
These figures were reduced to obtain a mean value for each tire
tested, for each of the test cycles used. This resulted in four
mean data points for each variable (HC, CO, NOx, fuel consumption),
for each of the drive cycles used.
For each variable studied, under the conditions of each
driving cycle, an analysis of variance was performed using the
unreduced data. This analysis- tests the null hypothesis of
equality of the means for each tire, against the alternative
hypothesis, that for at least two of the tires tested the means are
unequal. Rejection of the null hypothesis is evidence that vari-
ation in HC, CO, NOx, or fuel consumption is based on the tire
used. The significance of rejecting the null hypothesis is stated
in terms of the probability of being incorrect by doing so.
The mean values of each variable for each tire-test combina-
tion are graphed, against RRC, in figures I to VIII. Linear
regressions were fitted to these mean data points, and are shown in
the figures. These regressions represent the closest linear
approximation to the functional relationship of each variable to
rolling resistance. There was very little evidence of a non-linear
component in any of these functional relationships. The correla-
tion coefficients of these regressions are an indication of how
closely the regression describes the data; a correlation coeffi-
cient of +1.0 represents a perfect fit.
B. Results of Analysis
1. RRC and Fuel Consumption
There is very strong evidence for concluding that the fuel
consumed by the test vehicle is a direct function of the tires that
are used. This relationship was observed with consistent strength
in both the FTP and HFET driving cycles.
Information gathered in this test program pertaining to fuel
consumption and RRC is summarized in Table 3. These mean values
are graphed, and shown with their respective regression lines, in
Figures I and II.
Results from the analysis of variance firmly support the
rejection of the null hypothesis, that the tire used has no effect
on the fuel consumption of the test vehicle. In the case of the
FTP tests, the significance attached to this rejection is 0.0004;
in other words, there is less than one chance in 2000 that the
observed differences in fuel consumption are due to chance varia-
tion or other random-effects errors. The significance of rejecting
the null hypothesis as applied to the HFET results is zero to four
decimal places, meaning that the observed differences in the means
are virtually certain to have been an effect of the rolling resis-
tance of the tires tested.
-------
-7--
Th e equations of the regression lines shown in Figures I and
II are listed below. Equation (2) describes the relation of RRC to
fuel consumption (FC) for the test vehicle over the FTP driving
cycle, and equation (3) does the same for the HFET cycle. The
correlation coefficients, r, of the regressions are also given.
FC (in cm3/km) = 951.1(RRC) + 133.2 [r=.95] (2)
FC = 1276(RRC) + 83.8 [r=.91] . (3)
These equations are vehicle-dependent, and may not be appli-
cable to other makes or models without verification. Interpola-
tions from these equations to predict the fuel consumption of the
test vehicle will be reasonably accurate within the range of RRCs
used to compute the regression, approximately 0.0095 to 0.0145.
The predictive ability of the equations will decline when rolling
resistance coefficients outside of that range are used.
The average speed of the FTP cycle is slightly below 20 mph.
Under the conditions of that cycle, the regression projects an
increase in FC of almost 2 cnrVkm for each increase of 0.0020 in
the RRC of the tires used. For the test vehicle, this translates
to a fuel economy penalty of about 0.2 MPG with an increase of
0.0020 in the RRC. In the HFET cycle, with an average speed of
nearly 50 mph, the same 0.0020 increase in RRC causes a projected
increase of about 2.6 cnH/km in fuel consumption. The equivalent
fuel economy penalty is approximately 0.6 MPG for the test vehi-
cle-, with its'average highway fuel economy of 24 MPG.
These differences in fuel consumption are quite significant.
The difference in the RRC of the highest and lowest rolling re-
sistance tires used in this program was 0.0046, which is more than
twice the increment used in the above projections. For the test
vehicle, the fuel economy penalties resulting from the use of tires
having an RRC of 0.0142, rather than 0.0096, range from 0.5 MPG in
urban driving to 1.5 MPG in highway-rural driving. These fuel
economy penalties are based on the test vehicle fuel economy
averages of 16 MPG city and 24 MPG highway.
Fuel economy penalties of this magnitude, based on tires, have
major implications for manufacturers and consumers alike. Con-
sumers generally have no way of knowing whether the tires being
considered for purchase are fuel-efficient, beyond the general rule
that radials deliver better fuel economy than do nonradials.
Manufacturers striving to meet increasingly stringent CAFE stan-
dards u»e the most fuel-efficient tires available, but consumers
have no way of guaranteeing that equally fuel-efficient tires will
be supplied to them when they purchase aftermarket replacement
tires. The total dollar cost of even a 1 MPG penalty over the
useful life of a set of tires is considerable, and rising steadily.
2. RRC and NOx Emissions
Of the emissions studied in this test program, those of NOx
-------
-8-
showed .,the closest correlation to rolling resistance. As in the
case of fuel consumption, a direct linear dependence of NOx emis-
sion levels on the RRC of the tires used describes the data well.
A summary of the mean gram-per-mile NOx emission rates, by
tire and driving cycle, appears in Table 4. Figures III and IV
display these data graphically, with the computed regression lines
also shown. .
Performing an analysis of variance on the data allows rejec-
tion of the null hypothesis, for both driving cycles. It can be
stated at the 90 percent level of confidence that tire rolling
resistance has a significant effect on NOx emissions over the FTP
cycle. In the highway driving cycle, the effect of RRC on NOx is
even more pronounced. The probability that a test program would
show the variation in mean NOx levels for different tires that was
observed, when in reality no such relationship existed between RRC
and NOx, is less than one in twenty thousand.
Computation of least-squares regressions on these data results
in the equations below. Equation (4) is derived from the FTP
tests, and equation (5) from the HFET cycles.
NOx (gm/mi) = 1.161 + 14.62(RRC) [r = .78] (4)
NOx = 0.521 + 78.12(RRC) [r = .87] (5)
The correlation-coefficients of the above equations, while not as
high as those for the equations relating fuel consumption to RRC,
are still great enough to allow interpolations to be made from them
with certain limitations; the equations are vehicle-dependent, and
were determined from results of tests with tires having RRCs in
the O.Q095 - 0.0145 range.
The projected difference in the NOx emission rate of the test
vehicle, between using tires with very low and very high rolling
resistances, is considerable. In urban driving as simulated by FTP
cycles, the increase in NOx emissions for tires with RRC = 0.0170,
as compared with tires having RRC = 0.0090, is projected as about
0.12 grams-per-mile. Extended high-speed driving, as simulated in
HFET cycles, is far more seriously affected: Using the two hypo-
thetical tires mentioned above causes a projected 0.62 gram-per-
mile difference in NOx emissions.
3. Irregularity in the Data
The results already discussed were beginning to be evident
early in the course of the program. Preliminary data analyses,
undertaken after data became available from at least .one test of
each of the tires, showed higher fuel consumption and greater NOx
emissions associated with the bias-ply tires than any of the radial
tires. Early results concerning CO and HC emissions were mixed,
with no clear indications of the relation of rolling resistance to
these emissions.
-------
-9-
As computer output of the standard EPA analyses of FTP and
HFET cycles became available, generally a few days after the
tests were conducted, information on fuel consumption and emissions
rates was added to the existing data. An abrupt and unexpected
change in this information was noted, beginning with the tests
conducted on Monday 28 January 1980. This change was most readily
apparent in the carbon monoxide figures, which showed a consider-
able drop from the range anticipated on the basis of the initial
results. As additional results became available, continuing
investigation revealed that the rate of emission of HC and NOx had
also declined in the course of the test program, while fuel con-
sumption appeared to be slightly increasing.
When possible reasons for this shift in the data were con-
sidered, most were quickly discounted. No significant variation
was found to have occurred in any of the important control vari-
ables of the test program. The standard requirements for FTP and
HFET testing were observed throughout the program. Tire inflation
pressure was always checked, and consistently measured at 45.0
psi. Pre-test vehicle preparation consisted of running the vehicle
through one LA4 driving cycle unless tests were scheduled on
consecutive days, in which case one day's testing served as vehicle
prep for the following day. No changes in the internal performance
characteristics of the test vehicle were noted during the program.
The same test driver, experienced in FTP and HFET cycle driving,
was used in all of the tests in this program, thus eliminating
another potential source of variability in the results.
The single isolable event was a change in instrumentation.
The roll speed sensor on the dynomometer used for all testing (site
D207) was replaced by a new unit on January 28.
When the data on fuel consumption and NOx emissions are
divided by test date, into tests conducted with the old roll speed
sensor and those conducted with its replacement, the results
already discussed are virtually unchanged. The results concerning
correlations of CO emissions to tire rolling resistance, and HC
emissions to RRC, are considerably different when test date is
taken into account. For this reason, the results and interpre-
tation are presented in two parts in the following sections on RRC
and CO emissions, and RRC and HC emissions.
The nature of the shift in the data can be seen in Table 7.
This table summarizes the data from tests of the bias-ply tires
over the FTP cycles. The decreased standard deviations in the
means of each variable indicate greater repeatability in measure-
ments taken after January 28. Proportionately similar changes
occurred in data from HFET cycles for this set of tires. While
there is some evidence that similar reductions in emissions and
increases in fuel consumption took place for the various radial
tires tested, the unequal number of tests conducted before and
after January 28 preclude the use of those data to illustrate this
shift.
-------
-10-
4. RRC and CO Emissions
There is evidence to suggest that CO emissions, as well as
those of NOx, are partially dependent on the tire used. When all
of the data on CO emissions is analyzed, this evidence is incon-
clusive, and can best be described as a trend toward increased CO
emission rates being associated -with higher RRCs. No conclusions
can be drawn from the analysis of variance of the complete data
set. There is approximately a 40 percent probability of being in
error by rejecting the hypothesis that the mean gram-per-mile
emission rate of CO is equal regardless of the tire used.
If only the results of tests conducted on or after January 28
are used in the analysis, the aforementioned trend is considerably
strengthened. In this case, the results are nearly as significant
as those concerning fuel consumption or NOx emissions drawn from
the full set of data.
The analysis of variance of the post-January 28 data yields
much more information on the effect of tire rolling resistance on
CO emissions. The hypothesis that the RRC of the tires has no
effect on the rate of CO emissions can be rejected at the 90
percent confidence level for the HFET cycles, and can be rejected
at the 95 percent level of confidence for the FTP cycles.
Consideration of all of the CO data gave only very slight
evidence for concluding that tire RRC affects CO emissions.
Consideration-of only the post^-January 28 data gave strong evidence
for that conclusion. The fact that the standard deviations of the
CO measurements decreased after January 28 lends added weight to
the latter interpretation.
The results of all of the tests are summarized in Table 5.
Graphs of mean CO emission rates against tire RRC, again using the"
full set of data, appear as figures V and VI. Regressions on these
mean data points are drawn on the graphs.
Two things became apparent when least-squares regressions were
fitted to these data. A relation of linear dependence of CO
emission rates on tire RRC describes the HFET cycle data more
closely than such a relation describes the FTP data. In both
cases, linear regressions can be fitted much more closely to the
means of the post-January 28 tests than can be fitted to the means
of all of the tests.
The equations below are those of the regressions shown in
figures V and VI. Equation (6) is the regression on the FTP cycle
results, while equation (7) corresponds to the HFET cycle results.
CO (g/mi) = 11.384 + 225.3(RRC) [r = .44] (6)
CO = 0.371 + .27.KRRC) . [r = .88] (7)
-------
-li-
lt is interesting to note how these equations differ from the
equations of the regressions on the mean data points computed using
only tests conducted after the replacement of the speed sensor. As
noted previously, when the results of tests conducted before
January 28 are deleted, the mean rate of CO emission drops for each
of the tested tires, and the associated standard deviations also
decreased. When regressions are fitted to these "new" means, the
constant terms decrease, and the effect of changes in tire RRC is
heightened.
CO = 8.922 + 331.7(RRC) [r = .74] (6A)
CO = 0.119 + 42.9(RRC) [r - .99] (7A)
Equations (6A) and (7A) are those of the regression to the
data collected after Janaury 28 on CO emissions. Equation (6A)
is derived from results of FTP cycle testing, and thus should be
contrasted to equation (6). The same connection exists between
equations (7A) and (7), which are derived from the HFET cycle
results.
Note that the changes in the rates of CO emission projected
to occur with a given change in tire RRC are greater for the
post-Janaury 28 data. This, together with the higher correlation
coefficients of equations (6A) and (7A), implies that the depen-
dence of CO emissions on tire RRC is stronger than is indicated by
analysis of the complete set of data from the test program.
5. RRC and HC Emissions
There appears to be very little evidence to suggest that tire
RRC has a direct effect on HC emission rates. The analysis of
variance on all data suggests that chance, error, and experimental
noise had more to do with the observed variations in mean HC
emission rates than did the tires being tested. Restricting the
analysis to the results of tests conducted after January 28 im-
proves this situation only slightly; the hypothesis that the mean
HC emission rate is the same for each tire still cannot be rejected
with adequate statistical confidence.
Table 6 contains a summary of the data collected on HC emis-
sions. The mean rate of HC emissions for each tire tested is
plotted against the RRCs in figures VII and VIII.
The "scattered" nature of the data can be seen clearly in the
table and the graphs. Regressions on these data were computed, and
are shown on the graphs for reference. Very little information
about HC emissions with different tires can be obtained from these
equations. The correlation coefficients are very low, r = 0.45 for
the FTP data and r = -0.39 for the HFET data. The FTP cycle testing
showed a slight tendency for greater HC emissions to be associated
with higher RRCs, while the HFET cycle testing showed a weak
indication that the opposite relation holds.
-------
-12-
When the results of tests conducted before January 28 are
deleted and new mean rates of HC emissions computed, the change is
in the direction of a stronger tendency for higher HC emissions to
be associated with higher RRCs. The regression on the post-January
28 FTP data in this case has a correlation coefficient of r =
0.60. While this is not an excellent fit, it is a better fit than
the regression on all data. The slight tendency toward lowered HC
with higher RRC.that was shown in all HFET data is further weakened
in this case, r = -0.39 for all data and r = -0.31 for the partial
set. . ..
' *'.-
While no strong conclusions about HC emissions and tire RRC
are possible from the results of this test program, there are
indications that the effect of tires with higher rolling resis-
tance, if any, is toward higher HC emission rates over the FTP
eyele.
V. Conclusions
1) A strong correlation exists between automotive tire
energy dissipation, quantified by the rolling resistance co-
efficient (RRC) of the tires as determined in general accordance
with the SAE recommended procedure kf, and vehicle fuel consumption
rates as measured during the EPA dynamometer test procedures. In
this test program, the vehicle experienced a 0.5 MPG fuel economy
penalty on the FTP cycle and a 1.5 MPG penalty on the HFET cycle
when the relatively high rolling resistance bias-ply tires were
used, instead of "the lowest rolling resistance radials.
2) The gram-per-mile emission rates of NOx correlate strong-
ly with tire RRC, with- greater NOx emission rates associated with
the use of tires having greater rolling resistances. 'This effect
was seen to be more pronounced over HFET cycles than over FTP
cycles. . . ...;. ; . , - . . '-...-.
3) Emissions of CO are: affected by the rolling resistance of
the tires used. Data from HFET cycles show strong, significant
evidence for increases in CO emission rates with increases in the
RRC of the tires. Data from FTP cycles show a tendency toward
higher CO emissions with higher RRCs, but not with the significance
of the HFET cycle results. ;
4) In the case of HC emissions rates and tire .RRCs, the
analysis gave some indication that higher tire rolling resistance
resulted in increased rates of HC emissions over the FTP cycles;
however, the observed relationship was rather weak. Emissions of
HC over the HFET cycle did not seem to be dependent on tire RRCs.
-------
-13-
References
.' ' " ..
\J "Tire Related Effects on Vehicle Fuel Economy," Yurko, John,
EPA Technical Report, SDSB 79-27, July 1979. - .*;
2/ Patterson, D.J., Henein, N.A., Emissions from Combustion
Engines and Their Control, Ann Arbor Science Publishers, Inc.
1979.
3/ "Determination and Use of Alternative Dynomometer Power
~ Absorption Values," EPA OMSAPC Advisory Circular 55B, December
6, 1978.
4/ "Rolling Resistance Measurement Procedure for Passenger Car
~~ Tires," SAE J1269, and "The Measurement of Passenger Car Tire
Rolling Resistance," SAE J1270, Society of Automotive Engin-
eers, 1979.
-------
-14-
Table 1
Target Dynamometer
Coastdown Time from Road Coastdown Tests
Tire
Tire "A"
Tire "B"
Tire "C"
Tire "D"
Vehicle
Nova
Nova
Nova
Nova
Inertia
Weight Class
3750 Ibs.
3750 Ibs.
3750 Ibs.
3750 Ibs.
Target 55-45
Coastdown Time
13.79 sec.
14.10 sec.
13.23 sec.
12.42 sec.
Table 2
Proper AHP Setting for Dynamometer Testing
Tire
Tire "A"
Tire "B"
Tire "C"
Tire "D"
Vehicle - Inertia Weight
Nova - 3750 Ibs.
Nova - 3750 Ibs.
Nova - 3750 Ibs.
Nova - 3750 Ibs.
50 mph AHP Setting
10.6
9.9
10.3
12.9
-------
-15-
Table 3
RRC and Fuel Consumption
Fuel Consumption
Tire
Tire "A"
Tire "B"
Tire "C"
Tire "D"
RRC
.0096
.0109
.0119
.0142
N
5
5
5
8
FTP
Mean
142.6
142.8
145.2
146.6
(cm3/km)
HFET
N Mean
5 97.2
5 96.7
5 98.0
10 102.6
Table 4
Tire
Tire "A"
Tire "B"
Tire "C"
Tire "D"
RRC
RRC
.0096
.0109
.0119
.0142
and NOx Emissions
N
5
5
5
8
NOx Emissions
FTP
Mean
1.322
1.288
1.342
(g/mi)
HFET
N Mean
5 1.364
5 1.266
5 1.418
1.375 10 1.677
-------
-16-
Table 5
RRC and CO Emissions
Tire
tire "A"
Tire "B"
Tire "C"
Tire "D"
Tire
Tire "A"
Tire "B"
Tire "C"
Tire "D"
RRC
.0096
.0109
.0119
.0142
RRC
RRC
.0096
.0109
.0119
.0142
CO Emissions
FTP
N Mean
5 14.474
5 12.700
5 13.844
8 15.015
Table 6
and HC Emissions
HC Emissions
FTP
N Mean
5 0.998
5 0.936
5 0.976
8 1.019
(g/mi)
HFET
N Mean
5 0.654
5 0.626
5 0.706
10 0.761
(g/mi)
HFET
N Mean
5 0.1062
5 0.1010
5 0.1044
10 0.1029
-------
-17-
Table 7
Summary of the Shift in Data After 1/28/80
FTP Test Date
HC (g/mi)
CO (g/mi)
NOx (g/mi)
Fuel cons.
Before
Mean
1.095
16.182
1.445
145.75
1/28/80 (N=4)
St. Dev.
0.060
1.905
0.047
0.50
On-After
Mean
0.943
13.847
1.297
147.50
1/28/80 (N=4)
Std. Dev.
0.038
0.923
0.025
0.58
-------
FUEL
T
eraJ/km
103.0
102.0
101.0
,100.0
99.0
98.0
97.0
g/mi
-IIhI1I-
.0090 .0120 .0150
Fig it ''FUEL CONSUMPTION vs RRC (HFET cycle)
FUEL
3
cm /km
147.0
146.0
145.0
144'.0
143.0
142.0
4th-4-
Fig II:
.0090 .0120 .0150
FUEL CONSUMPTION VS RRC (FTP cycle)
RRC
RRC
1.85
1.75
1.65
1.55
1.45
1.35
1.25
Fig III:
NOx
g/tni
1.38
1.36
1.34 4-
1.32
1.30
1.28
II 1IIII
.0090 .0120 .0150
NOx EMISSION VS RRC (HFET cycle)
HII1-
0090 .0120 .0150
Fig IV: NOx EMISSIONS VS RRC (FTP cycle)
RRC
oo
I
RRC
-------
-19-
0.62
.0090 .0120 .0150 .0090 .0120 .0150 RRC
Figure V: CO EMISSIONS VS. RRC (HFET) Figure VII: HC EMISSIONS VS. RRC (HFET)
12.75
1
4 H
.0090 ' ' .0120 .0'150 .0090 ' ' .0120 " .0150
Figure VI: CO EMISSIONS VS. RRC (FTP) Figure: VIII: HC EMISSIONS VS. RRC (FTP)
-------
-20-
Appendix A
Test Vehicle
1979 Chevrolet Nova Sedan
250 CID/ 1 bbl.
Model 350 turbo-hydramatic automatic transmission
Test Weight: 3,745 - 3,790 Ibs.
Test Tires
Designation
Tire "D"
Tire "C"
Tire "B"
Tire "A"
Construction
Bias-ply
Radial
Radial
Radial
Size
E78-140
P215/70 R14
P195/75 R14
P195/75 R14
RRC*
.0142
.0119
.0109
.0096
* The reported rolling resistance coefficients are the ratio of
the transverse spindle force to the normal load.
-------
. 3
f,
<<
3
9
11
Vci-:iCLF: Ui i::-ii-AT:rji. IT
':;::; v" :vijf: u ::; v?
rvoAi; I:O.'I.-:T com! TUT.T
i:avi-. no .ciu.-'Lii'
*A" -
Appendix B
-21-
o.:<::-;:.."3V-1EIOO
d.-!201IMOO
0.;6 0.64j'jSJ.9E-02 (Hl/'HR-SEC)
A2 = 0.1649777E--03 (I-IR/IU-SEC) 0.1042721E -04 CIIR/ill-SEC)
PAIR 1 REJECTED DUE TO STANDARD DEVIATION
NUMSEF: (IF PACKS nSHAIMIriCi 13 ' '
KGS'J!.TS APTEfi STAMiA^B DEVIATION REJECTIONS
AVERSE C':I;-;:MCEN'I-;; STAMBARD DEVIATIONS
PAIR 4 REJECTED lil't TO 3TAKBARD DEVIATION
NUMi:.!;r< OF PA:RS Rii-iAiwr:;?, is 3
RESULTS AFTER STAMDARD DEVIATION REJECTIONS
AV't-IRAC't COEFF J nilMTS
AO == 0.3-!7iS?iir-:TOO (HI/HS-SEC)
A2 = 0.1S32343E-03
OBS'.ERVED Cfl.'.ST TIME AFTER REJECTIOK
RESULTS AFTER WIND UUKKECTIONS
AVERAGE COSFFICEHrS
AO ~ 0.3471S70EtOO (HI/HR-BEO
62 = 0.1382J43U-03
WIND CORRECTED COAST TIME 13.470
UilATHER CORRECTED RESULTS
UEATHER COP::-^.-"! ED COAST TIKE 11,010
CAT A Ur.ATHSIK r.CRREuTEO TO 63 r> 27.0' HG
coEFF-icsirrs
A2 o. O.1344036E-03 (rlR/KI-SECS
MASS CORRECTED REiiL'LTS
TOTAL SVSFE.H UEICHT 3381. L.BS (TEST UT +
FOSKilLA I.VERTIA3 Ai:E i;ril!'-'. USED
STANDARD DEVIATIOtfS
0.69:-7773EvOO (LJiS)
O.RS1381AE-03 (I.S3)
13.444
STANDARD DEVIATIONS
0.3931814E-02 (lil/HR-SEO
0.45099 l-iE-03 li-:UlI.-. UP.IEMT CI.A5:i
3.-i.r/. l.SS. MASS
HO-40 itfll) 13.7,10
CATLOU:5Y
COAST - DOllil TJilE
3'JOO
3VJO
4000
i2.esr,
13. ?..«/
13.7CS
-------
r.-:,; 0 i
T':LK;: i:r;s.;:!.:';iit:: :r:-;::;V
NOv'A J10 'J
PT'.i'.l-:- 1'-' - 00
Appendix B (cont'd)
-22-
Al
J -:'.
_ t..-
IV
21
VI
V.J
IS
u
'17
70
89
VI
Vc,
97
iiV '!£;": i i '.::-. r > I 3735. (TOTAL)
! A'.V_£ i-;;'i:.:.-;rr o.
NTS 1'tt TERMS OF AOyA2 < OPSERVCtO
0.17713-VJ7E-03 0.319C-/-:-rOO
c.
0 >
o.
G »
0.
0 .
. iV/'V.'.7'^ ^ Oil 1-00
. ':;,:.'.".''.';»' : ;-oo
. '.M : ' ''.:: ; r-'tv : oo
.::; i .:;..,:,:; :";:ioe
:::: v: o';.-'^t-:oD
->J . v' ':.!. ':.':. i CO
2:i i.n-ny~..:c:>
oo'C:C'0'_/'..:'"*!:-': vO
22Vvr;i::V;:::i"C
vVj c'1 r" o J1 tj ij L r ') "
0. 159o,-.7ooE:>?.
o. JO^SJ./;-.:.E-OV;
o. iifir-i2-"/:;::-E>;:
0.15o5122c>K-03
0.77503776E--05
0.10JJ01772E-21
A AT TEST CON31TIONS (NO nOSHiECT
0.3 c. 0 ^ u. v 0
0.2SS7'1vO<>
0.2V.'1 !!.: .>'»
0 l?..-'.w3iiv'J- ^
0 » 3. :/r> ';'£:.'; './i'
0 (H277II-! 00
0.27
0.&379727E-0-". (HR/M1-SEC)
0.144600 IE 03 (Hft/MI-SEC)
3 HAS BEEN REJECTED CUE TO i-i;:GM is MS
PAIR 7 HAS t'EEN RF.JECTEE DUE TO HIGH SI1S .
DESEKV-:^ r.iATA AT TEST CONDI"ONS (NO CORRECTIONS BUT-AFTER KMS REJECTIONS:
OMSEiJ'.'EO Cr--.3T TlisE 1-1.502
STANDARD DEVIATIONS
0.13c>S2'.u2LI-01 < i-il/HR-SEC )
0.2Bo;'907t:-rOO (MI/HR-SEO
0.1rJ17-USrr--03 (rlS/ill-SEC)
oss'irivbi'i cciA'nT r:a'E :;!-'ii~:v REJECT JON
RFSULT'-: i':r'Ii":: WtNC CORRECTIONS
O.nS-MOoOE-i-O"/ (MI/KR-SEC)
AC '=
WINti CCiRRECTEO COAoT TTii;- 1.4. 557
tiEATHEk C'CilRrCTED RESULTS
W£AT!-!E!--: CCiRRECTE-1" COAST 7 I.ME 14.4A7
ii.'.TA u;:".!': n-i£K rORRECTEK TO 63 F» 27.0" l-ICi
0 . S49656VE-05 i i 1R/M1-SEC )
STANDARD DEU I AT I Oi-iS
0. 1375B25i:-01 (MI/HR-SEC5
0 . 5-*9&369c -OS ( 1-iR/HI -SEC >
0.1o235'1S!£-03 (KR/MI-SEC)
MASS CORRECTED RESULTS
TOTAL SYSTEM HEIGHT 2717. ' LBS (TEST UT
FtlRMUL'i Ii!EF:TIAU ARE H-EIiiu USED
CR I VIM C7 -RQIATI^JCj ^auivALSNT 08.
UGN-rftlv'iN.:; r.-OlATIi-iG EtUilVALEriT o-U
INCK'I li>: ^EIC1 :T CLASS 3750. L2S.
V',i.;~. c;c;-;!:'EC". u;;;u TG 08 IG. LBS. MASS
UY?!.'i:-:CiM£l :.';'- COAST TIME (55- -V.") i-if'H) 1-1.100
CAfi£7iOK-i
ROTATING EQUIVALENCES>
COAST-HQ'.JN TIME
-'
-j.
>
>
>
>
>
107
1 Of--
10?
110
111
112
113
3375
3500
3i25
3/50
3:J7'.",
V>00
12.716
13 . 177
I3.o39
M.100
1-1 . 562
1 5 . 02-1
4250
-------
>
3
1
3
6
0
9
U
12
13
11
IS
19
21
30
31
32
33
5
J.'.
60
> 61
> A3
> 65
> 67
71
7A
7G
80
62
fl.5
3;.
87
BV
91
93
9A
90
100
102
104
106
103
110
112
lit
116
119
121
123
125
126
12U
129
13?
TEST
TIRL tiis'.::;;i--- n>.-.:.
TEST CCNDITIil-15
A.M3IEUT TCM?, -r
MO
Gil CKEVKCL1T
11 OG 77
rf.U.K- 2-1.00
TIFIE "C"-RADIAU
NOVA ilQ SILVER
REAR- 24.00
Appendix B (cont'd)
-23-
57.
r9.10
3/30.
o.
WT. (AFTER T£SI'>: 3730. (TOTAL)
DRIVEN AXLE i::r:r,:ir:
NO OF RUN PAIf.3: 7
COEFFICIENTS IN TERM? OF AO.A2 (OBSERVED)
AO A2 PiiS ERROR
0.32£i27:3EfO? 0.17i6390?E-0."S 0.-1220E + 00
J 39j Vjit
0.
0.18R43490E-03
o.iV7y£!SS'Jt:-o:i
o. J7-U417SK:>.;
0.29K4E+00
0.2962E:-00
O.329SE+00
o.
OBSE-.::VSfJ lii'.TA
o.i;
o.u
3T CUNajTIONfi I
9y?lE:-03 0.217AE-i-00
CORRECTICHfO
(MI/HR-SEC)
STAN1V.RH CEVIATIi.liJS
0.-10-KS3HOF-01 ;i1l/MR-SEC>
0.203a07V£;-0'1 (HR/III-SEC)
0.3190f/o-C-E-:-00.
0.17c',?o9E:: -03
i HAS pr.r;;^ INJECTED r».i[-: TO HiOri RHS ,
1 PrtIF;(3; GF r-;'JMS y.t-.VZ SEEi! REJCCTtril .
OBr,C';y£i! c.M'A AT ICST "o.itiinoiiS 0. 1304569E-02 (LBS)
PAIR 7 REJECTEB DUE 10 STAriPARD DEVIATION
NUSPER OF PAIRS RF.riAIifJNG IS'- <
RESULTS AFTER 3TAKP:H:l" tEVIATIOM REJECTIOrfS
AWERAPh: cn::?r-ic.i;:ir^ STANUARO
AO = o.'j;<:-.-:.joi?:foo R'IMAlr-IJ.riG la j
KESULFb A:"1'I:'R 3 iiVrlZXii! iitlVIATICi--1 REJECTIONS
AVE.K'ACf
AO -
O.33A9A-UE-K'0 (HI/Hfi-SEC)
A3 -- O.17-5oo29E-0:i (KR/MI -EEC)
OBSUKVEIi CUASV TTH' AFTER REJECTION
RESULTS AfTL'R UJ1ND CORRECTIONS
AVtRAGE CCIEFFICFIITS
AO - O.33C&770E-fOO (HI/HR-SEC)
A2 = O.17-16429E-03 (HR/HJ-SEC)
WIND CORRECTED COAST TIKE 13.GA4
UEATHCR CORRECTED RESULTS
UEATKEK Cfln-IVECrED COAST TIHE 13.5«0
DATA U'E.IIIICR CORRECrED TO AS FF 29.0' HO
COEFFICEIUS
AO - O.3132172E+00 (MI/HR-SEC)
A2 = 0.1704343E-03
-------
Riji'.E MASr DOU:i-: Ttiii ,
TEST ID: :!K-
V£i--;.c:!.!i:: (::! C:MI-:V..O';J::T "OVA&IO SILVER
ST p.'.rt:: jo :uv >*;;
££:;;>:,?:; ,: KG;-; r-: :;i.oo REAR- 24.00
;C:F-i !:;:(t TIRE "D" - BIAS-PLY .
Appendix B (cont'd)
-24-
.:>
\L
IV
"24
2S
26
0-7
. .":^.
62
64
.-'..'.
.'...^
70
77
87
a?
91
S'tt
100
105
107
103
111
115
117
lift
; 119
> 120.
> 121
12-1
125
Aii:;:?..NV TEMF-.V-F: 65.
BArUJt'iETER IH !!,T.: 29.100 ', .
U r . (AFTER 1CST ', :. 3745 . . < VUT'1
MO cr uuN-'-pAUVvi:1 /' -. :. . -. ',,:'' ':
,- Cirif'F'iCICNTS ih'-TERiVS OF;:AO»>.2! (OBSERVED) . . v ' -
:: .'AO ;. . -.' ' .1-, . '" .'. .A2-'::"i-.' . - .Ri-iS .ERROR
'0.37
/.' '"-O. ;T3VS.3a«>?-*'"-;^3 . .' "0+'3lii-Vr)E'4'0() --V-
. 0..3179E+00 .
5E-03 -.''. 0 ..387.9E+OO
0.'lc>C'7o024E-'03 ','":0 .313CJE'!-00.
FAIR ' -'A
' PA-IK-''":''. T.
2 T "A I r*C'. o )
OESc.RVEC
. ': ;? ' 0.915S708E-03 . (HR/'HI-SEC)
CUE ,to'v HIGH '"'RMS '.;:'<-> .''' ' '' ' "' ' :
DUF ' TO- .
r;~
'lA-AT TQ)T.'COI\'rilTIONS (NO CORRECTIONS BUT AFTER RMS REJECTIONS).
AVERAUF: CIOEF-ICEHTS ;'-''" - '..-' .'"'. '"G.TAHrjMRji'DEVIATIONS
AO - ' 'o.331o77GEiv>0 (HI/HR-SEC) G. 177300AE-01 (.i-il/HR-SEC)
A2 «'--;-. 0.1694-022E-03- '(HR/MIrSEC) - 0.86J5294E-05
'.
PAIR 6 'REJIECTED E-UE TO STANUARn 'PEVIATIOH.- ; -. .
NM^BER OF fAIR;:; REMAlNINa IS ' ' 4"- ' . . '-./'.
RES-'jL'!S >:FTE!» STAi-iDAR.O DiiVIATlGN REJECTIONS "
AVSSAi-iE -'.'OEF? ICE. NTS ..
AO =. . ' . 0.3SB>:301E-)00 (KI/HR-SEC)
A2 = oil.65642iE-.C3 "(HR/MI-SEC)
OBSERUEr.i COAST TIME .AFTER REJECTION
RESULTS 'AFTER WIND CORRECTIONS
AVERAGE COEFFICEN7S " . ....- '''
AO = 0.:if:i84301E-!-00-
-------
-25-
Appendix C
Dynamometer 55^-45 Coastdown Data
Tire: "C" - Radial
Test Weight: 3,790 Ibs.
Date: 01/24/80
Target Dyno Coastdown Time: 13.23 sec.
55-45 Coastdown Times (sec.) ,
50 mph Horsepower Setting: 11 AHP 10AHP 12AHP 9AHP
(Trial)
1 12 54 13.42 12.09 14.20
2 12.63 13.51 12.16 14.38
3 12.63 13.57 12.15 14.43
4 . 12.70 13.59 12.21 14.44
T 12.63 13.52 12.15 14.36
Tire: "D" - Bias-ply
Test Weight: 3,745 Ibs.
Date: 01/04/80
Target Dyno Coastdown Time: 12.42 sec.
-"--'' ---.- 55-45 Coastdown Times (sec.)
50 mph Horsepower Setting: 10.9AHP 10AHP 11.9AHP 9AHP 12.9AHP
(Trial)
1 . 13.50 14.36 13.02 15.57 12.44
2 13.56 14.22 13.03 15.54 12.50
3 13.61 14.51 13.09 15.56 12.51
13.56
14.36
13.05
15.56
12.48
-------
-26-
Appendix C (con't)
Tire: "B" .- Radial
Test Weight: 3,785 Ibs.
Date: 01/04/80
Target Dyno Coastdown Time: 14.10 sec.
55-45 Coastdown Times (sec.)
50 mph Horsepower Setting: 10.9AHP 1QAHP 11.9AHP 9AHP 13AHP
(Trial) .
1 13.07 13.87 12.57 - 12.08
2 13.12 13.84 12.66 14.91 12.11
3 13.18 13.90 12.69 14.95 12.14
T 13.12 13.87 12.64 14.93 12.11
Tire: "A" - Radial
Test Weight: 3,750 Ibs.
Date: 01/03/80
Target Dyno Coastdown Time: 13.79 sec.
55-45 Coastdown Times (sec.)
50:friph Horsepower Setting: 10.9AHP 10AHP 11.9AHP 9AHP
(Trial)
1 13.63 14.30 12.98 15.36
2 13.61 14.38 12.85 15.25
3 13.61 14.37 12.87 15.30
₯ 13.62 14.35 12.90 15.30
-------
-27-
Appendix D
Road Coastdown Procedure
a. Tires inflated to manufacturer's recommended pressure.
b. Vehicle driven over warm-up cycle consisting of steady speed
operation at 50 mph for about 45 minutes.
c. Vehicle accelerated to stable speed of slightly greater than
60 mph. ,
d. Transmission shifted to neutral position and strip chart
recorder activated.
e. Coastdown terminated at 20 mph.
f. Coastdown repeated immediately in opposite direction.
g. Seven pairs of opposite direction coastdown trials for each
tire set.
h. Ambient wind speed, wind direction, barometric pressure, and
temperature recorded before and after each set of coastdowns.
i. Vehicle weighed before and after each set of coastdowns.
-------
-28-
Appendix E
Dynamometer Coastdown Procedure
a. Vehicle mass adjusted to corresponding mass for road coastdown
trials.
b. Tires inflated to 45 psig. -
c. Dyno inertia weight set at 3750 Ib. ,
d. Vehicle loosely secured on dynamometer.
e. Vehicle operated over 2 HFET driving cycles for tire-vehicle
warm-up.
f. Vehicle accelerated to 62 mph.
g. Transmission shifted to neutral.
h. Timer which sensed speed from the front roll (roll coupled to
inertia weights and power absorber) recorded 55 mph - 45 mph
free deceleration time interval.
i. Coastdown repeated 3 times for 4-5 different power absorber
settings.
j. Surface tire temperatures monitored so temperature did not
exceed 200°F.
-------
-29-
Appendix F
Figure I
)yno AHP Settings vs. Dyno 55-45 Coastdown Time
Tires "A" and "C" (Both Radials) .
Legend^;
£]-Tire "A": 1 = .00419 (AHP) + .0277
AT
0-Tire "C": 1 = .00432 (AHP) + .0309
AT
13.0
12.0--
11.0
AHP
10.0--
9.0--
12.0
13.0
14.0
15.0
55-45 Coastdown Time (sec.)
16.0
-------
-30-
.Appendix F
Figure II
Dyno AHP Settings vs. Dyno 55-45 Coastdown Time
Tires "B" (Radial) and "D" (Bias-ply)
Legend; .':-
Q-Tire "D": 1 = .00385 (AHP) + .0332
AT
"B": 1 = .00399 (AHP) + .0293
AT
13.0--
12.0- -
HP
11.0-
10.0--
9.0--
12.0
13.0
14.0
15.0
55-45 Coastdown Times (sec.)
16.0
-------
-31-
Appendix G
Emissions Test Procedure
a) Tires changed when necessary and inflated to 45 psig.
b) Fuel tank topped every two days of testing.
c) Vehicle mass adjusted to corresponding mass when road coast-
downs were conducted. .,. ,
d) One day test sequence:
j
1. 1 cold start FTP
2. 1 HWFET
3. 3 quick check coastdowns after HWFET
e) One days testing serves as prep for next day;
f) or 1 LA-4 driving cycle serves as prep.for next day.
g) 12-24 hour soak between prep and tests.
h) Propane injection diagnostics performed on sampling equipment
every day of testing.
i) Same driver for all tests.
* US. GOVERNMENT PRINTING OFFICE: 1980- 651-112/0260
------- |