EPA-AA-SDSB-80-15
Technical Report
"Independent Coastdown Road Load Power Determination for Ten
Diverse Production Vehicles"
by
Terry Newell
August 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 tech-
nical information and to inform the public of technical developments
which may form the basis for a final EPA decision, position or re-
gulatory 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
As part of the continuing effort by EPA to verify the integrity and
accuracy of alternative determinations of dynamometer power absorber
(PAU) settings, a test program involving ten various production vehicles
was conducted. The objective of this program was to determine the
appropriateness of the dynamometer PAU settings, submitted to EPA by the
vehicle manufacturers as part of the 1979 model year certification
process, for production vehicles.
II. Discussion
Ten production vehicles were selected for testing in this program.
The three major domestic manufacturers were represented in the set of
test vehicles, which consisted of six General Motors products and two
vehicles each manufactured by the Ford Motor Company and the Chrysler
Corporation. The inertia weight classes of the test vehicles ranged
from 2000 to 4500 pounds. More detailed descriptions of these ten
vehicles appear in Table 1.
There were two important criteria by which the vehicles to be
tested were chosen. The foremost consideration was the use of dynamo-
meter PAU settings in the 1979 model year (MY1979) certification pro-
cess that were lower than the range of values typically submitted for
other similar vehicles. Secondarily, selection was based on the sales
volume of the vehicles chosen. While there may have been other vehicles
for which manufacturers submitted apparently low PAU settings, the
relatively high sales volume of the ten vehicles chosen increased the
potential impact on fleet-wide emissions and fuel economy of any dyna-
mometer underloading during the certification process.
In selecting the specific production vehicles to represent each of
the ten models chosen, care was taken to avoid vehicles that had been
altered in any way that could have led to unrepresentative coastdown
times. Internal alterations avoided in selection included any modifi-
cations to the engine, drivetrain, or braking system. Unacceptable
external alterations included the addition of roof racks, air dams, or
non-standard mirrors. All tires used in the coastdowns were either
original equipment or directly comparable to the originals. The vehi-
cles were all checked for correct wheel alignment before testing. These
precautions should have eliminated any effects of unrepresentative
frictional drag, aerodynamic drag, or inertial effects.
This testing program was conducted by Automotive Testing Labora-
tories (ATL) as a task order under a contract with EPA.I/ The coast-
downs were performed on the track at the Transportation Research Center
of Ohio (TRC) in East Liberty, Ohio. Test vehicles were procured by ATL
from local sources, such as private owners and area car rental agencies.
All vehicles and tires used in the program were required to have accumu-
lated a minimum of 100 km C62 mi) before testing commenced. This require-
ment was exceeded by a considerable margin for all of the test vehicles.
A value for road load horsepower was determined for each vehicle in
accordance with the EPA Recommended Practice for Determination of Vehi-
-------
-3-
cle Road Load^2/ In this method of determining road load the vehicle is
accelerated to a speed greater than 60 mph, the transmission is shifted
to neutral, and the vehicle is allowed to freely decelerate (coastdown)
until the speed has dropped to less than 20 mph. Vehicle speed-versus-
time data is collected during each of a minimum of five pairs of alter-
nating-direction coastdowns. This road coastdown data is subjected to
analysis by a computer program; the program output includes values of
the acceleration coefficients a and a~, corrected to the standard
ambient air temperature of 68°F°(20°C) and barometric pressure of 29.00
in Hg. From this road acceleration data, the force acting on the vehi-
cle at 50 mph, or any other velocity within the range spanned by the
coastdown, can be computed. A "target" 55-45 mph coastdown time is then
calculated for reproducing this force on the dynamometer at the test
inertia weight. The correct dynamometer PAU setting is determined when
the 55-45 mph coastdown time on the dynamometer matches the ambient-
and mass-corrected 55-45 mph "target" coastdown time. This matching
reproduces, for operation on the dynamometer, the 50 mph force exper-
ienced by the vehicle on the road.
The results of the independent determination of dynamometer PAU
settings for the test vehicles are listed in Table 2, under the heading
"ATL Value." The values submitted by the manufacturers are shown in the
table under the heading "Mfr. Value."
As can be seen in Table 2, the dynamometer PAU settings submitted
by the manufacturers are generally lower than those independently deter-
mined by ATL. The sole exception among these ten vehicles was the Dodge
Omni, for which the two values were equal.
A significant difference was observed between the dynamometer PAU
settings supplied by the manufacturers for certification, and those in-
dependently determined by ATL. For this reason, the total force acting
on the test vehicles at the match-point speed of 50 mph was examined.
The existance of a force offset would indicate differences between test
and production vehicles, while an offset only between the dynamometer
PAU settings would provide evidence for differences in the tires used,
the degree of tire wear, in tire behavior on the track and the dynamo-
meter, or in dynamometer calibration between the dynamometers of the
manufacturers and ATL.
The total road force on the test vehicles, operating on the track
or on the dynamometer at 50 mph, can be computed using Newton's Second
Law and the coastdown data. Computer analysis of the coastdowns results
in the derivation of ambient-corrected, vehicle-specific 55-45 mph
coastdown times. The force is then computed, using the equation:
F = ma = m (~)
Where:
F = total force at velocity of 50 mph.
m = inertial weight + driving rotating equivalent.
-------
-4-
Av = change in velocity (10 mph) during 55-45 coastdpwn.
At = time (seconds) required for 55-45 mph coastdown.
-— = approximation of acceleration at 50 mph.
Results of the computations of total force at 50 mph appear in
Table 3. The force calculated for the production vehicle is in the
column headed "ATL Force," and the force calculated for the certification
vehicle is shown under "Mfr. Force." The 55-45 mph coastdown times are
also listed in this table. These times are the "target" coastdown
times, intended to reproduce the total 50 mph road force experienced by
the vehicle on the dynamometer at the test inertia weight, as calculated
by the manufacturers and by ATL.
III. Results
The manufacturer-submitted dynamometer PAU settings were found to
be lower than those independently determined in this test program, with
the exception of the Dodge Omni, for which the two values were equal.
AT"L HP MFR HP
The percentage difference, defined as x 100, ranged from
Mr K Hr
zero (for the Omni) up to 36 percent (for the Monza). The mean differ-
ence was 12.3 percent.
The Corvette and the Monza dynamometer PAU settings submitted by
General Motors were derived by determining the reference frontal areas
of these vehicles, and using the equation in 40 CFR 86.129-79. The
other eight vehicles had the manufacturer-submitted PAU settings deter-
mined by the coastdown method, which was used by ATL for all ten vehi-
cles. Thus, the manufacturer PAU settings for the Monza and Corvette
are not directly comparable to the ATL-determined values.
When these two vehicles are deleted from the computation of percen-
tage difference in Table 2, both the mean and the range are decreased.
The mean difference drops from 12.3 to 8.7 percent, while the range is
narrowed to zero-to-24.8 percent.
Since the Monza and Corvette used the frontal area method to deter-
mine PAU settings, no target dynamometer coastdown times were calculated
for these vehicles. Since this target time, At, is necessary for these
computations of the total road force acting on the vehicle at 50 mph,
there are no entries in several of the columns of Table 3 for these two
vehicles.
Calculation of the force on the test vehicles at 50 mph shows that
the forces averaged nearly the same for both the production and the
prototype certification vehicles. The percentage difference, defined as
ATL FORCE - MFR FORCE
MFR .FORCE x 100, ranged between -7.1 percent and +30.3
percent. The mean difference in total 50 mph force was +2.8 percent.
Figure 1 presents a plot of percentage difference in dynamometer
PAU settings (vertical axis) against the percentage difference in total
-------
-5-
force at 50 mph (horizontal axis). These are eight data points on the
graph, representing each vehicle except the Chevrolet Corvette and
Monza.
The dashed line in Figure 1 represents the situation of the percen-
tage difference in dynamometer PAU settings being equal to the per-
centage difference in total force at 50 mph. If a systematic difference
exists between the test vehicles used in determining PAU settings as
part of the MY 1979 certification process, and the production vehicles
that were used in this program, then the data points would be expected
to cluster near this dashed line. Since the data points do not cluster
in any linear fashion, it can be concluded that the discrepancies in the
PAU settings are not attributable to any systematic differences between
the certification and production vehicles. Possible reasons for the
observed discrepancies include (i) differences in the degree of wear of
the tires used for, or in tire behavior during, coastdowns conducted by
the manufacturers and by ATL, or (li) calibration differences between
dynamometers of ATL and the manufacturers.
The results for three of the test vehicles appear anomalous, rela-
tive to the test fleet as a whole. As indicated earlier, the dynamo-
meter PAU settings submitted by the manufacturer for the Corvette and
the Monza were determined using the frontal area equation. These vehi-
cles represent two of the three greatest percentage differences observed
in PAU settings. The third case is the Granada, which was unique among
vehicles tested in that it is represented in Figure 1 as the only point
that is both near to the dashed line and significantly displaced from
the origin.
The following is a brief discussion of several possible reasons for
the apparently anomalous results for the aforementioned vehicles.
(a) Erroneous test program data. It is possible that errors
occurred during the test program in either the collection or transcrip-
tion of vehicle and coastdown data. For example, the weight of a vehi-
cle is an important parameter in computing the force acting on that
vehicle. The test weight of 3,660 Ib (1,660 kg) reported by ATL for the
Monza appears to be excessive for this vehicle. By comparison, World
Cars 1979 3/ lists the same Monza (2 + 2 Coupe1, 196 CID 6-cylinder
engine) as weighing 2,775 Ib (1,259 kg), and the inertia weight class
used for certification (based on curb weight plus 300 Ib) was only
3,500 Ib (1,588 kg)..
(b) Advantageous use of options in road load power determination.
This refers particularly to the use of the frontal area equation for
determination of the PAU settings for the Monza and the Corvette. Since
EPA approval of the coastdown method of road load power determination,
the majority of domestically produced vehicles have been certified with
PAU settings determined in this way. It is safe to assume that the
manufacturers will submit PAU settings for a particular vehicle deter-
mined by the method that results in the lowest value for the PAU set-
ting. Particularly the Corvette, with its unusual body geometry, may
benefit considerably from use of the frontal area equation.
-------
-6-
(c) Anomalous vehicles. Despite the precautions exercised in
selection of the ten production vehicles tested, the possibility exists
that a vehicle unrepresentative of its model line was used in the pro-
gram. This possibility seems most likely in the case of the Granada,
for which the ATL target coastdown time and calculated total force at 50
mph appear to be inconsistent with the vehicle's general body shape and
weight class. Alternately, the prototype Granada used in certifica-
tion may have been unrepresentative of the production vehicles. In any
case, there are wide discrepancies in the data on the Granada.
Three of the ten test vehicles, the Corvette, Granada, and Monza,
exhibited the largest discrepancies in dynamometer PAU settings. As
previously noted, no comparison of the values for the total road force
at 50 mph were made for the Monza and the Corvette. The Granada also
exhibited the greatest discrepancy between total road forces at 50 mph
as calculated from the manufacturers and ATL data. Since these three
vehicles appear anomalous in one fashion or another, it is useful to
consider the remaining set of seven vehicles.
The dynamometer PAU settings for those seven vehicles differed by a
mean of only 6.4 percent. The range of the percentage difference is
then zero to 13.6 percent. This mean percentage difference is barely
half that of the entire ten vehicle test fleet.
The calculated total road force for the seven vehicles differs by a
mean of -1.1 percent, with the range being -7.1 to +6.3 percent. The
scattering of the force percentage difference for these seven vehicles,
and the much lower mean difference in PAU settings suggests that overall
these models were certified experiencing total force that was repre-
sentative of their respective production vehicles.
The results from this test program give some indication of an
effect based on odometer mileage. Of the ten vehicles tested, four had
more than 10,000 miles on the odometer, five had between 4,000 and
10,000, and the odometer mileage was not recorded for one vehicle.
Comparing the rightmost columns of Tables 1 and 3 reveals some corre-
lation between vehicle odometer mileage and percentage difference in
total 50 mph force, where greater mileage is associated with lower,
generally negative, percentage differences. For all vehicles with less
than 10,000 odometer miles, total force calculated from the ATL coast-
down data was greater than that calculated from manufacturer supplied
data. The reverse was true of vehicles with greater than 10,000 odo-
meter miles. A least-squares regression of percentage difference in
road force against odometer mileage yielded a value of -0.71 for the
correlation coefficient, and an r-squared value of 0.50. Thus 50
percent of the variation in the percentage difference in total force
can be attributed to variation in odometer mileage, with increases in
odometer mileage associated with decreases in road force.
This odometer mileage related offset has been observed previously
in relation to fuel economy. Several studies have shown that the fuel
economy of in-service vehicles is generally below that of EPA certi-
fication vehicles when odometer mileage is low, and that this offset
is gradually reduced and in some cases reversed with increased odometer
-------
-7-
A reduction in the total (50 mph) force experienced by
vehicles as odometer mileage increases, as observed in this program, is
consonant with the more well-known increase in fuel economy.
This information indicates that a consistent offset in both dynamo-
meter PAU setting and calculated road force might have been observed in
this program if all the production vehicles selected had low accumulated
mileage, since the road force offset appears to diminish with accumu-
lated mileage. It may be advisable to set an upper bound on the accumu-
lated mileage of all road load determination vehicles in the future.
IV. Conclusions
1. Most of the vehicles (7 out of 10) appear to have experienced
total force at 50 mph on the certification dynamometers that was repre-
sentative of the road force experienced by the corresponding production
vehicles.
2. Since total force was representative of production vehicles,
the discrepancies in dynamometer PAU settings were likely the result of
differences in vehicle tire behavior, and/or slight calibration differ-
ences between the dynamometers at ATL and the manufacturers' facilities;
no data acquired in this test program indicated consistent differences
between certification and production vehicles.
3. For the two vehicles in this program (Corvette, Monza) that
had frontal area determined dynamometer PAU settings submitted to EPA by
the manufacturer, those values were considerably lower than values
independently determined by the coastdown method. The test weight of
the Monza reported by ATL appears to be quite high, which may have
caused the discrepancy for that vehicle; while the Corvette, which has
an unusual body configuration, may benefit more than many other vehicles
through the use of the frontal area equation.
4. The PAU setting and calculated road force for the Granada, as
determined by ATL in this program, seem to be anomalous when compared to
the data on the entire test fleet. It would appear that the specific
Granada used in the program may have been an atypical vehicle. Con-
versely, it may have been that the Granada prototype used in certifi-
cation was unrepresentative of the production vehicles.
V. Recommendations
While results of this program indicated that the road forces cal-
culated from data submitted by the manufacturers are generally appro-
priate for and representative of production vehicles, three recommen-
dations should be considered.
1. In future test programs of this type, concerned with comparing
road-versus-dynamometer vehicle performance, at least one contractor-to-
EPA correlation vehicle should be included. This would resolve the
question of possible dynamometer differences.
2. When in similar test programs a vehicle appears to yield
anomalous results, the cause of such behavior should be determined,
-------
-8-
and additional vehicles tested if necessary.
3. Due to the possibility of decreased total 50 mph force re-
sulting from increases in odometer mileage, it may be desirable to set
an upper bound on the mileage of road load determination vehicles. A
mileage bound consistent with certification vehicle requirements would
be approximately 4,000 miles, with some acceptable tolerance in excess
of that figure.
-------
-9-
References
I/ Task Order No. 5 for Contract No. 68-03-2693, "Coastdown Road
Load Power Determination of Ten Diverse Production Vehicles," March
1, 1979.
2J "EPA Recommended Practice for Determination of Vehicle Road Load,"
Attachment I to OMSAPC Advisory Circular No. 55A, February 8, 1978.
3/ World Cars 1979, edited by the Automobile Club of Italy, Herald
Books, Pelham, NY, 1979, p. 268-269.
<4/ Murrell, Dillard, "Passenger Car Fuel Economy: EPA and Road,"
Draft of Report to the Congress, April 1980, p. 66-68.
-------
-10-
Table 1
Ten Diverse Production Vehicles
Mfr.
Ford
Chrysler
CMC
Ford
CMC
Chrysler
CMC
CMC
CMC
CMC
Make & Model
Ford Fiesta
Dodge Omni
Chevrolet Monza
Ford Granada
Pontiac Firebird
Chrysler Lebaron
Inertia
2000
2500
3500
3500
4000
4000
Chevrolet Corvette 4000
Cadillac Eldorado
Pontiac TransAm
Oldsmobile '98'
4000
4000
4500
Wgt.
Ib.
Ib.
Ib.
Ib.
Ib.
Ib.
Ib.
Ib.
Ib.
Ib.
Engine
98
105
196
250
301
318
350
350
402
402
CID (L)
(1.6)
(1.7)
(3.2)
(4.1)
(4.9)
(5.2)
(5.7)
(5.7)
(6.6)
(6.6)
A/C
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Odom.
12412
10873
8651
N/A
12417
8716
4478
8220
18449
8443
Table 2
Dynamometer Power Absorber (PAU) Settings, in HP
Vehicle
Mfr. Value
ATL Value
% diff.
Fiesta
Omni
Monza
Granada
Firebird
Lebaron
Corvette
Eldorado
TransAm
Olds '98'
7.3
7.8
8.3 (1)
10.1
8.8
. 10.8
8.0 (1)
9.6
9.5
11.6
8.05
7.8
11.3
12.6
10.0
11.6
9.4
10.0
9.9
12.2
mean % difference =
10.3
0.
36.1
24.8
13.6
7.4
17.5
4.2
4.2
5.2
12.3
(1) Manufacturer submitted PAU setting as derived from vehicle frontal
area and Federal Register equation.
-------
-11-
Vehicle
Table 3
Total Force at 50 mph, in Newtons
Manufacturer ATL
55-45 At
Force
55-45 At
Force
diff.
Fiesta
Omni
Monza (1)
Granada
Firebird
Lebaron
Corvette (1)
Eldorado
TransAm
Olds '98'
10.12 s
12.22 s
-
15.84 s
16.63 s
15.21 s
-
15.14 s
16.16 s
15.46 s
408.4
422.1
-
456.2
496.7
543.1
-
545.4
510.9
601.0
mean
10.32 s
13.16 s
13.61 s
12.16 s
17.30 s
14.66 s
14.24 s
14.25 s
17.17 s
15.35 s
percent difference
400.6
392.1
531.2
594.3
477.3
563.5
579.5
579.7
481.0
605.4
=
- 1.9
- 7.1
-
+30.3
- 3.9
+ 3.8
-
+ 6.3
- 5.9
+ 0.7
+ 2.8
(1) Manufacturer submitted PAU setting determined using frontal area. No
dynamometer target coastdown times calculated.
Figure 1
Percent Difference in 50 mph Force (from Table 3)
plotted against Percent Difference in Dynamometer
PAU Settings (from Table 2)
PAU Setting
% diff.
-Q-
0
-10'
ox
20
10
10
20
Key
Fiesta O
Granada Q
Eldorado A
TransAm Q
Omni •
Firebird •
Lebaron A
Olds '98 4
FORCE
% diff.
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
DATE:
August 8, 1980
SUBJECT. Distribution of Technical Report "Independent Coastdown Road Load Power
Determination for Ten Diverse Production Vehicles"
TO:
FROM: Terry NewelP
Standards Development and Support Branch
See Distribution List
The attached report describes the results of a program to independently
verify the appropriateness of the dynamometer PAU settings submitted to
EPA by the manufacturers, as part of the MY79 certification process, for
production vehicles. The test fleet was selected from vehicles that had
PAU settings that appeared to be relatively low in comparison with other
similar vehicles, and had relatively high sales volumes. The report
concludes that there was no evidence supporting the existence of any
systematic differences between certification and production vehicles.
Attachment
Distribution List
M. Walsh
C. Gray
R. Maxwell
K. Hellman
R. Stahman
G. Reschke
G. Thompson
C. Bramson
A. Kaupert
M. Korth
J. Marzen
D. Murrell
T. Hudyma
G. Dana
W. Duncan (JAMA)
D. Schwentker (AIA)
H. Weaver (MVMA)
EPA Form 1320-6 (R.». 3-76)
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
DATE: AugUSt 8, 1980
SUBJECT: Release of Technical Report "Independent Coastdown Road Load Power
Determination for Ten Diverse Production Vehicles"
FROM: Terry Newell*
Standards Development and Support Branch
TO: Charles L. Gray, Jr., Director
Emission Control.. Technology Division
The attached report describes the results of a test program to inde-
pendently verify the appropriateness of certification dynamometer PAU
settings, submitted by the manufacturers, for production vehicles. The
test fleet was selected from vehicles with seemingly low PAU settings
and relatively high sales volumes. Dynamometer PAU settings were deter-
mined using the coastdown procedure outlined in OMSAPC Advisory Circular
No. 55B.
The independently determined PAU settings were lows?? than those submitted
by the manufacturers for nine of the ten vehicles tested; the mean dif-
ference was about 12%, or 1.1 hp. The total force acting on the vehi-
cle, operated on the track and the dynamometer at 50 mph, was calculated
in order to determine whether .the observed discrepancies in PAU settings
were the result of the systematic differences between certification and
production vehicles. The forces were found to differ by a mean of only
2.8 percent, with some vehicles experiencing greater force on the
dynamometer and some on the track.
The report concludes that none of the data acquired in this test program
support the existence of systematic differences between certification
and production vehicles. The observed offsets in PAU settings for seven
of ten tested vehicles may have been the result of differences in tire
behavior between certification testing and this program, and/or calibra-
tion differences between dynamometers of MVEL and manufacturers. The
report recommends more immediate investigation into apparently deviant
test results and the inclusion of a correlation vehicle in test programs
investigating road/track versus dynamometer vehicle performance. Due to
an observed decrease in road force with increased odometer mileage, an
upper bound on the accumulated mileage of road load determination
vehicles is suggested.
Attachment
Date
Charles L. Gray^3r(^, Director
Emission ContrxrL TecnTiulogy-Division
EPA Form 1320-6 (R.y. 3-76)
------- |