LDTP  -  78  - 06
              Technical Report
  Investigation of the Requested Alternate

    Dynamometer Power Absorption for the

            Ford Mercury Marquis
                     by

              Glenn D. Thompson
                March,  1978
  Standards Development and Support Branch
    Emission Control Technology Division
Office of Mobile Source Air Pollution Control
     Office of Air and Waste Management
    U.S. Environmental Protection Agency

-------
Abstract

Concern about the EPA fuel economy measurements has focused greater
attention on the dynamometer and the dynamometer adjustment.  Specifi-
cally the alternate procedures for determining the dynamometer power
absorption to simulate the vehicle road experience affords an oppor-
tunity for both greater precision and possible abuse.  Because of the
possibility for abuse it was decided to occasionally check the appro-
priateness of the alternate dynamometer power absorptions requested by
vehicle manufacturers.

The question of the representativeness of the requested dynamometer
power absorption for the Mercury Marquis was first raised during the
summer of 1977.  This report collects and summarizes the pertinent
available data which have been generated by the EPA Emission Control
Technology Division, the EPA Certification Division, and by Ford Motor
Company.

It is concluded that a dynamometer adjustment of 11.4 horsepower is
representative of typical, recommended vehicle use.  The dynamometer
adjustment requested by Ford, 8.8 horsepower is certainly inappropri-
ately low.  Some of the discrepancy between the original requested power
absorption and the EPA results occurred because of the increased tire
pressures used by Ford during their road tests.  These tire pressures
are considered unrepresentative of anticipated consumer use of the
vehicle because they are significantly higher than the tire pressures
used by Ford dealers in preparation of the vehicles for consumer use.
Even with the increased tire inflation pressures the recent tests indi-
cate a dynamometer adjustment of 10.8 horsepower is necessary to sim-
ulate the road experience of the vehicle.

The EPA exhaust emission certification and fuel economy measurements of
the Ford Mercury Marquis were unrepresentative because of the inappro-
priately low dynamometer adjustments used during these tests.  The
unrepresentative nature of these tests resulted in reduced NOx emission
measurements for the vehicle and resulted in significantly increased
vehicle fuel economy.  If the vehicle were tested at the representative
dynamometer power absorption, it is predicted that the NOx emissions
would increase approximately 0.05 gm/mi.  The predicted decrease in the
urban fuel economy would be about 0.25 mi/gal, while the highway fuel
economy would be expected to decrease about 1.5 mi/gal.  The composite
urban-highway fuel economy would be expected to decrease approximately
0.6 mi/gallon, or about 3 percent.

-------
                                -1-

Purpose

The question of the representativeness of the test vehicle and the
requested dynamometer power absorption for the Mercury Marquis was first
raised during the summer of 1977.  This report collects and summarizes
the pertinant available data which have been generated by the EPA Emission
Control Technology Division, the EPA Certification Division, and by
Ford Motor Company.

Background

During the summer of 1977, ECTD personnel became aware of several requests
for alternate dynamometer power absorption which appeared to be anomalously
low.  One of the most outstanding of these was the 8.8 horsepower reques-
ted by Ford Motor Company for the Mercury Marquis.  This request was 34%
below the 13.4 horsepower specified by the Federal Register, for testing a
vehicle of the inertia weight category of the Marquis.  The concern over
the appropriateness of this requested dynamometer adjustment resulted in
an ECTD test program which was conducted in July 1977.  The report of
this program concluded that the alternate dynamometer power adjustments
requested for this vehicle did not appropriately reflect the road exper-
ience of the vehicle as reported by Ford.  In addition, this report
questioned the representativeness of the certification vehicle with
respect to the production vehicle. The report of July 1977, which is
attached as Appendix A, has been the basis for the additional investiga-
tions which have been conducted.

Discussion

This report first considers all of the available data, assuming the road
data submitted by Ford is correct.  The second section of the report
challenges the road data, and the final section documents the effects
the indicated descrepancies in dynamometer adjustment would probably have
on the fuel economy and exhaust emissions of this vehicle.

A.   Analysis of the Dynamometer Data assuming Validity of  the Ford Road
     Data

In response to the questions raised by the first EPA report, Ford Motor
Company conducted a series of road and dynamometer tests on three
production Mercury Marquis.   The results of these tests are given in
Table 1 of Appendix B and are summarized in Table 1.

                              Table 1

                   Ford Production Marquis Tests
                           November 1977

          Average Corrected Road        Average Dynamometer Power
             Coastdown Times                    Absorption
          	(sec)	         	(horsepower)	

                   16.72                         10.67

-------
                                   -2-
For comparison the data presented by Ford in their request for an alter-
nate dynamometer power absorption is given in Table 2.
                               Table 2

              Data Presented by Ford in Support of the
          Requested Alternate Dynamometer Power Absorption

             Corrected Road          Dynamometer Power
             Coastdown Time             Absorption
            	(sec)	          (horsepower)

                  16.98                     8.8
In the request for the alternate dynamometer adjustment Ford reported a
corrected road coastdown time 0.26 seconds or 1.5%, longer than the
production vehicle coastdown time.  This would indicate the production
vehicles had slightly greater actual road load than the prototype vehicle.
However, this degree of consistency of the Ford road data indicates the
original road data is probably representative of the vehicle road exper-
ience for the test conditions used.  In contrast, the difference between
the requested dynamometer adjustment of 8.8 horsepower and the November
1977 test results of 10.7, was 1.9 horsepower, or 21 percent of the
requested dynamometer adjustment. The early EPA results of 10.5 horse-
power tend to corroborate the November results from Ford.

In October 1977, the 1978 model year vehicles became available in the
rental car fleet.  EPA subsequently rented a Ford Mercury Marquis, and
performed road coastdowns on this vehicle at the Transportation Research
Center of Ohio (TRC) test track.  The vehicle was then brought to EPA
for the dynamometer coastdown tests.  The resulting test data are pre-
sented in Table 2 of Appendix B, and average values are plotted in
Figure 1.  Matching the original Ford road coastdowns time of 16.98
seconds with this vehicle required an average of 10.6 horsepower.
Matching the road coastdown time obtained by Ford from the production
vehicle, 16.7 seconds, required an average dynamometer power absorption
of 10.9 horsepower.  All of the test results, based on Ford road coast-
down times, are summarized in Table 3.

All of the data show reasonable agreement, except the alterante dynamo-
meter power absorption requested by Ford for the certification vehicle.

-------
                                   -3-
                              Table 3

       Summary of Dynamometer Power Absorption Determination
                      for the Mercury Marquis

                                             Average Dynamometer Power
     Test Series                             	(horsepower)	

Ford Certification Request                              8.8

EPA Tests on the Certification
Vehicle, July 1977 using the Ford                      10.5
corrected road coastdown time of 16.98 seconds.

Ford Tests on Production                               10 7
Vehicles, November 1977

EPA Tests on a 1978 Rental
Vehicle, January 1978 using the                        10.6
Ford corrected road coastdown time of
16.98 seconds

EPA Tests on a 1978 Rental Vehicle
January 1978, using the Ford production                10.9
corrected road coastdown time of 16.7 seconds
Throughout this extended investigation, Ford indicated the differences
between the original request and the subsequent tests were a result of
the variations among dynamometers.  Reasonable estimates of the dyna-
mometer variability can be obtained from the multiple dynamometer tests
which were conducted.  The Ford confirmation tests of November 1977
consisted of six dynamometer coastdown tests on two different dynamo-
meters.  The two Ford dynamometers differed on the average by about one
horsepower.  Even including variations in the road test times, the
twelve determined dynamometer adjustments were between 9.72 horsepower
and 11.86 horsepower.  In the case of the EPA production vehicle coast-
downs, tests were conducted on all four of the EPA certification dyna-
mometers which were available.  At the original Ford road coastdown time
of 16.98 seconds, the test results ranged only from 10.2 to 10.9 horse-
power.  Thus, there is no evidence dynamometer-to-dynamometer variations
could cause the extremely low value for the dynamometer adjustment
originally requested by Ford.

B.   Comparison of the Submitted Ford Road Data
     with EPA Track Results

The major purpose of the road tests conducted by EPA on the rented
production Mercury Marquis was to verify the road coastdown data sub-
mitted by Ford.

The road test portion of the program was conducted by TRC personnel.
The vehicle was first driven for about 250 miles for drive train com-
ponent brake-in.  The vehicle system was then allowed to equilibrate to

-------
                                         -4-
                 Average Vehicle Dynamometer Coastdown  Time
                                   versus
                        Dynamometer Power Absorption
       13  --
S-i
0)
g
p-
M
 fi
 o
•H
4J
 cx
 VJ
 o
 w
       12 ••
                                          Each point is the average
                                          of the results obtained
                                          from four EPA light-duty
                                          vehicle certification
                                          dynamometers.
11 ••
 0)
 §
 PH
 0)
 4J
 
-------
                                   -5-
ambient temperatures over night.  Prior to the vehicle warm-up for  the
coastdown tests the vehicle tires were adjusted to the recommended  cold
inflation pressures.  The vehicle was then warmed up for approximately
one half hour at 50 mph.  Twenty coastdowns were subsequently conducted,
ten in each direction of the TRC track.  Ten of the coastdowns, five in
each direction, were started at approximately 60 mph.  The remainder
were started at approximately 40 mph.  It was necessary to divide the
coastdowns into these two speed ranges because of the relatively short,
1 km, section of constant grade track available on the TRC skid pad.

The data analysis was conducted in the manner described in the data
analysis section of the EPA Recommended Practice for Road Load Deter-
mination except that a Av/At approximation was used for the vehicle
deceleration during the coastdown.

A two term model of the acceleration versus velocity was chosen, that
is:

                              A = aQ + a2v2                      (1)

where:

          A  = the calculated deceleration of the vehicle
          v  = the vehicle velocity
          a_ and a_ are coefficients to be fitted by the
          regression analysis.

Additional terms were added to equation 1 to account for the directional
dependent effects caused by track grade and wind.  The grade effect was
assumed to be independent of velocity while the wind effect was assumed
to be linearity dependent on the vehicle velocity.

The an term of the regression will contain a constant term introduced by
the ambient wind.  This correction to still air conditions was made
using the measured value for the ambient wind.  In addition, since  the
a~ term represents the aerodynamic drag, an air density correction  was
applied to this term to correct to the standard ambient conditions  given
in the EPA recommended practice.  The corrected coefficients which  were
obtained are:

                    a*  = 0.284 mi/hr-sec
                                                                 (2)
                    a*  = 0.000132 hr/mi-sec

The coefficients of equation 2 were used to calculate the total road
force on the vehicle from the vehicle mass and the estimates of the
rotational inertias of the rotating components of the vehicle.  The 55
to 45 mph dynamometer coastdown time interval necessary to reproduce
this force was then calculated by correcting for the differences between
the total effective vehicle mass during the road coastdowns and the
dynamometer simulated mass plus the rotational inertia of the drive
train components.  The final, dynamometer "target" coastdown time obtained
from the EPA track measurements was:

                              AT = 16.31 seconds                 (3)

-------
                                  -6-
This coastdown time gave an average dynamometer adjustment of 11.4
horsepower based on the EPA dynamometer tests of December 1977, plotted
in Figure 1.  This value is significantly greater than the average
dynamometer power absorbtion of 10.6 horsepower obtained from these data
using the coastdown time of 16.98 seconds reported by Ford for the
prototype vehicle.  It is also greater than the dynamometer power absor-
ption of 10.9 horsepower which was obtained from the EPA dynamometer
data based on the coastdown time of 16.72 seconds reported by Ford for
their production vehicle tests.

The test data were reveiwed to attempt to locate the reason for this
discrepancy.  An obvious difference observed between the Ford and EPA
test conditions was the tire inflation pressures.

The Ford road tests were conducted at tire inflation pressures of 30 and
32 psi, front to rear respectively, while the EPA tests were conducted
with inflation pressures of 26 and 28 psi.  This difference in tire
pressure would theoretically be expected to cause a change of about 0.5
horsepower in the vehicle road power.  Since the differences between the
Ford and EPA results from production vehicles were approximately this
amount, these differences were attributed to the difference in the tire
inflation pressure during the road tests.

TRC personnel had been instructed to conduct the road tests at the
normal tire inflation pressures recommended by the manufacturer, there-
fore the basis for the test pressure difference was investigated.  TRC
personnel reported that the vehicle was prepared by the local Ford
dealer who had inflated the tires to approximately 26-28 psi.  In addi-
tion the consumer safety information contained in the vehicle glove box
was based on the tire inflation pressure of 26-28 psi.

The vehicle owners manual did not specifically recommend any tire infla-
tion pressures but referred the reader to the decal on the jamb of the
passenger door.  This decal recommended tire inflation pressures of 26-
28 psi for best vehicle ride and 30-32 psi for maximum fuel economy.

Since either inflation pressure might be considered as recommended by
the manufacturer, a brief telephone survey of local Ford dealers was
conducted to ascertain which inflation pressure might be considered
normal for the vehicle, at least as originally delivered to the customer.
All dealers responded that the vehicles were normaly prepared for the
consumer with tire inflation pressures of 26-28 psi or lower.  It was
therefore concluded that tire inflation pressures of 30-32 psi are not
typical for this vehicle even at the time of dealer preparation.  Con-
sequently, the high tire inflation pressures cannot be considered typical
for the normal use of this vehicle.

The dynamometer adjustment of 11.4 horsepower is therefore more repre-
sentative of typical, recommended vehicle use than the power absorption
values obtained from data collected at the evaluated tire pressures.
The dynamometer adjustment requested by Ford, 8.8 horsepower is certainly
inappropriately low.

-------
                                   -7-
C.   The Emission and Fuel Economy Effects
     of the Dynamometer Adjustment

In December 1977 the Certification Division conducted exhaust emissions
and fuel economy measurements as a function of dynamometer power absorp-
tion.  The vehicle used was a Ford LTD which was tested at the appropri-
ate weight category to represent the Mercury Marquis.  This vehicle was
used because it is a full sized Ford vehicle, as is the Marquis, and it
was available at the EPA laboratory.  The data from the Certification
Division tests are given in Appendix C.  Regression analyses of these
data conclude that the vehicle fuel economy significantly decreases with
increasing dynamometer power absorption, the NOx emissions increase with
increasing dynamometer power absorption, and the HC and CO emissions
tend to decrease with increasing power absorption.  The regression
statistics indicate there is little confidence that the slopes of the HC
and CO regression lines are different from zero.  Therefore the statistically
significant effects are the increase in NOx emissions and the degradation
of the vehicle fuel economy with increasing dynamometer power absorption.
The fuel economy and NOx emissions data, together with the regression
lines are plotted in Figures 2 and 3.

The plots of the data demonstrate the strong dependence of the vehicle
fuel economy on the dynamometer power absorption.  The statistical
confidence in the effects of the dynamometer power absorption on the
vehicle fuel economy was, as expected, very high.  The confidence in the
prediction of the effects on NOx emissions, was however, somewhat weaker.
The statistical uncertainty in the effects on NOx emissions was primarily
caused by the, possibly outlier, datum from the test at the lowest dyna-
mometer power absorption.

The regression lines can be used to predict the effect on NOx emissions
and fuel economy if vehicles had been tested at the higher, more repre-
sentative, dynamometer power absorptions.  When predicting this effect
it should be noted that Mercury Marquis were not selected as EPA exhaust
emissions certification vehicles.  Because of the extreme low road load
requested for these vehicles the test vehicles selected to represent the
Mercury engine family categories were Ford LTDs.  If the more representative
power absorption of 11.4 horsepower had been requested for the Marquis,
these vehicles would presumably have been selected.  The alternate
dynamometer adjustment requested for the LTD was 9.5 horsepower.  The
discrepancy in the dynamometer adjustment for the vehicle during the
applicable emissions tests was therefore 9.5 horsepower versus 11.4 for
vehicles without air conditioning and 10.5 versus 12.5 horsepower for
air conditioned vehicles.  The predicted increase in NOx emissions would
be approximately 0.05 gm/mi for both of the two cases.

Including the predicted NOx increase of 0.05 gm/mi most of the certification
vehicles tested to represent the Mercury Marquis would still have met
the NOx standard.  There was, however, one which would have equaled the
2.0 gm/mi standard but would have been certified because of round-off

-------
                                      -8-
                            Vehicle Fuel Economy
                                   versus
                        Dynamometer Power Absorption
  TEST  VEHICLE:   Ford LTD  8A1-351W-F-64
  INERTIA WEIGHT CATEGORY:  5000 pounds
n)
60
o
o
a
w
<1)
iH
O
25


24

23 f


22


21

20 .


19 -


18 '

17 .


16


15
             Highway Fuel Economy
             Regression Line
                  Urban Fuel Economy
                  Regression Line
                   -i-
                        -i-
                                          -t-
-t-
                                                          •4-
-4-
               6     7     8    9   10   11  12   13   14   15   16

                  Dynamometer Power Absorption (horsepower)
                                  Figure 2

-------
                                       -9-



                        Nitrogen Oxide Emissions

                                  versus

                      Dynamometer Power Absorption
 TEST VEHICLE:  Ford LTD   8A1-351W-F-64

 INERTIA WEIGHT CATEGORY:   5000 pounds
co
C
o
•H
CO
CO
•H
e
w


-------
                                 -10-
tolerances.   It should be noted that this analysis strongly depends on
the slope of the NOx versus dynamometer power absorption regression,
shown in Figure 3.  If, for example, the datum at the lowest dynamometer
power absorption were judged to be an outlier and deleted, the slope of
the regression would increase and failure of these vehicles with high
NOx emissions would probably be predicted.  Considering this sensitivity
of the regression and the normal test variability, the success or failure
of this particular vehicle, if tested at a representative dynamometer
power absorption cannot be accurately predicted.

The Mercury Marquis was tested as a fuel economy vehicle.  Therefore, in
this case the horsepower discrepancy between the dynamometer power
absorption during the test and a representative value was 8.8 versus
11.4  horsepower for vehicles without air conditioning and 9.7 versus
12.5 horsepower for vehicle equipped with air conditioners.  The pre-
dicted decrease in fuel economy would be 0.24 and 0.26 mi/gallon respectively
for the urban cycle.  The predicted highway fuel economy decrease would
be 1.5 and 1.6 mi/gallon respectively.  The composite urban-highway fuel
economy would be expected to decrease approximately 0.6 mi/gallon, or
about 3 percent for vehicle both with and without air conditioning.
Based on current sales predictions, such a decrease in the fuel economy
of this vehicle would be expected to lower the Ford Corporate Average
Fuel Economy for 1978.

Conclusions

It is concluded that the EPA tests of the Ford Mercury Marquis were
unrepresentative because:

1.   The dynamometer adjustment requested by Ford was inappropriate
     for the road data submitted by Ford;

2.   The road data submitted by Ford is unrepresentative of anticipated
     consumer use of the vehicle because the tire pressures used during
     these road tests were significantly higher than the tire pressures
     used during the dealer preparation of the vehicle for consumer use.

It is further concluded that the unrepresentative nature of these tests
resulted in reduction of the NOx emissions measurements for these vehicles.
Most of the vehicles tested to represent the Mercury Marquis would still
be expected to meet the exhaust emission standards if these vehicles were
tested at the higher, more representative dynamometer power absorption.
One test vehicle was, however, sufficiently close to the 2.0 gm/mi NOx
standard that the success or failure of this particular vehicle cannot
be accurately predicted.

The unrepresentative nature of the EPA fuel economy tests resulted in a
significant increase in the vehicle fuel economy for the Mercury Marquis.
This increase could effect the 1978 Ford Corporate Average Fuel Economy.
  C. Larson, discussion, February 1977.
2
  J.D. Murrell, discussion, February 1977.

-------
      APPENDIX  A
EPA Report of July 1977

-------
Abstract:

     Concern about the EPA fuel economy measurements has focused greater
attention on the dynamometer and the dynamometer adjustment.  Specifi-
cally the alternate procedures for determining the dynamometer adjust-
ment to simulate the vehicle road experience affords an opportunity for
both greater precision and possible abuse.  Because of the possibility
for abuse it was decided to occasionally check the coast down time of
the test vehicle on the dynamometer, when the dynamometer is adjusted to
the requested power setting, versus the road coast down times reported
for the vehicle by the manufacturer.  While this procedure assumes the
validity of the submitted road coast down data, it does provide an easy
and convenient check that the dynamometer experience of the test vehicle
is similar to the reported road experience.

     The Ford Mercury Marquis was selected for such a "quick check"
since the alternate dynamometer adjustment requested by the manufac-
turer, 8.8 horsepower, was considered unusually low.  This requested
alternate power adjustment is 4.6 horsepower below the 13.4 horsepower
specified by the current Federal Register table and 4.2 horsepower below
the results of EPA measurements on a 1975 Mercury Marquis.  This report
concludes that the requested dynamometer power adjustement of 8.8 horse-
power, when used in conjunction with the certification test vehicle,
does not accurately reflect the road experience of the Mercury Marquis,
as reported by Ford.

-------
-2-
  Background:

       Concern about the EPA fuel economy measurements  has focused g'reater
  attention on the dynamometer and the dynamometer adjustment.   Specifi-
  cally the alternate procedures for determining the dynamometer adjust-
  ment to simulate the vehicle road experience affords  an opportunity for
  both greater precision and possible abuse.   Because of  the possibility
  for abuse it was decided to occasionally check the coast down times of
  the certification test vehicle on the dynamometer, with the dynamometer
  adjusted to the requested power setting, versus the road coast down
  times reported for the vehicle by the manufacturer.   While this proce-
  dure assumes the validity of the submitted  road coast down data, it does
  provide an easy and convenient check that the dynamometer experience of
  the test vehicle is similar to the reported road experience.   It insures
  that vehicle components such as tires and drive train lubricants are
  effectively the same on the EPA test vehicle as on the  vehicle which was
  road tested by the manufacturer.

  Discussion:

       The Ford Mercury Marquis was selected  for such a "quick check"
  since the alternate dynamometer adjustment  requested  by the manufac-
  turer, 8.8 horsepower, was considered unusually low.  This requested
  alternate power adjustment is 4.6 horsepower below the  13.4 horsepower
  specified for this 5000 pound vehicle in the current  Federal Register
  table.  This larga^difference seemed particularly questionable since
  there is no obvious reason, such as exceptional aerodynamics, to expect
  this large a discrepancy for a conventional front engine, rear axle
  drive vehicle.

       EPA measurements on a 1975 Mercury Marquis estimated the appropriate
  dynamometer adjustment to be 13.0 horsepower.   In response to requests
  for reference frontal area information,  Ford submitted  the same area data,
  26 square feet, for both the 1975 and the 1978 model.  It is therefore
  concluded that no significant reduction in  the size of  the Mercury Marquis
  has occurred between 1975 and 1978.   Using  the submitted frontal area data,
  the proposed road load equation for 1979 model year predicts a dynamo-
  meter adjustment of 13.0 horsepower for this vehicle.

       It was originally intended to coast the Mercury  Marquis from 55 to
  45 mph immediately following the highway fuel economy test.  In this
  manner -the highway schedule would be used as the vehicle warm-up, and
  the actual test horsepower adjustment would be confirmed versus the road
  measurements.  Unfortunately this could not be arranged because of
  scheduling problems.  At the time the test  could be scheduled, the
  vehicle was placed on the dynamometer and warmed up for approximately 30
  minutes.  Coast downs, from 55 mph to 45 mph were then  performed for
  four horsepower settings.  The total dynamometer absorbed power; the
  indicated dynamometer absorbed power and the coast down times are given
  in Table 1.

-------
-3-

                                  Table 1

                         Dynamometer No.  1,  7/9/77

  Total Dynamometer             Indicated Dynamometer          Coast Down
   Power at 50 mph        .         Power at  50 mpn"           Time Interval
   (horsepower)                      (horsepower)                 (sec)

       8,8                                6.2                    19.2
       8.8                                6.1                    19.4
       9.8                                7.0                    18.5
       9.8                                7.0                    18.5
      10.8                                8.0                    17.5
      10.8                                8.0                    17.6
      11.8                                9.0                    16.7
      11.8                                9.0                    16.8
       The coast down times are plotted against total dynamometer power at
  50 mph in Figure 1.  A "reasonable" fit line is also drawn in Figure 1.
  Ford reported a road coast down Lime of 15.8 seconds for  the road test.
  After correction to standard ambient conditions and adjustment to the
  dynamometer simulated inertia weight Ford calculated the  target dynamo-
  meter coast down time to be 16.98 seconds.   The dynamometer power
  adjustment corresponding to this dynamometer coast down time was re-
  ported to be 8.8 forsepower.  A verticle line corresponding to 17
  seconds is shown on Figure 1.  This line intersects the curve at 11.5
  horsepox^er.

       The Ford representative was asked by Certification Division person-
  nel for his explanation of this 2.7 horsepower difference.   The response
  attributed the difference to variations in the vehicle warm-up procedure
  on the dynamometer.  The EPA procedure recommends 30 minutes warm-up at
  50 mph while the Ford procedure calls for 30 minutes warm-up with the
  speeds slowly varying between 70 mph and 40 mph.   While it seemed very
  unlikely this minor variation in warm-up procedure would  cause a 2.7
  horsepower difference, it was decided to re-run the dynamometer coast
  downs on the vehicle utilizing the Ford method of preconditioning.

       The results of these coast downs are given in Table  2 and are
  plotted in Figure 2.  In this instance the dynamometer adjustment
  corresponding to a 17 second coast down is 9.9 horsepower.

-------
-4-
      PS
      W
      S
      O
      PH
      W
      H
      W

      g
      s
      Pi
      O
      W
      M
      H
      O
      H
                           DYNAMOMETER NO. 1


                                 7/9/77


                    VEHICLE FUEL TANK NEARLY EMPTY
      w
      §   111
      g
10..
              11.5
             15
              16          17          18

                     COAST DOM TIME (SEC)
19
20
                                    FIGURE  1

-------
-5-
      w
      ^
      o
      pu
      w
      Pi
      w
      3:
      o
      p-l
      o
      CO
      w
      H
      W


      §
      H
      O

      H
                            DYNAMOMETER NO. 6




                                 7/16/77




                         VEHICLE FUEL TANK FULL
             15
16          17           18




  COAST DOWN TIME  (SEC)




        FIGURE 2
19
20

-------
-6-

                                 Table 2

                       Dynamometer No. 6, 7/16/77

 Total Dynamometer        Indicated Dynamometer         Coast Down
   Power at 50 mph           Power at 50 mph            Time Interval
    (horsepower)              (horsepower)                  (sec)

      7.8                        5.4                       19.0
      7.8                        5.4                       19.0
      8.8                        6.4                       17.9
      8.8                        6.4                       18.1
      9.8                        7.3                       17.2
      9.8                        7.3                       17.1
     10.8                        8.3                       16.2
     10.8                        8.3                       16.1
     11.8                        9.3                       15.2
     11.8                        9.3                       15.3
     10.0                        7.5                       16.8
     10.0                        7.5                       16.8
      Since  this power adjustment is significantly different from the
 previous results of 11.5 horsepower, efforts were made to resolve this
 1.7 horsepower discrepancy.  Three parameters were identified which had
 changed between the two measurements; the warm-up procedure, the test
 dynamometer and the quantity of fuel in the vehicle tank. It was con-
 sidered improbable that either the warm-up procedure or the difference
 in the quanity of fuel in the vehicle tank could have this large an
 effect.  Therefore it was decided to investigate the possible effect
 caused by the change in the test dynamometer.  The first step was to
 verify the previous results on the first dynamometer.  However in this
 case the fuel tank was filled to capacity and the Ford warm-up procedure
 was used.  These data are presented in Table 3 and plotted in Figure 3.

                                 Table 3

                       Dynamometer No. 1, 7/16/77

 Total Dynamometer        Indicated Dynamometer         Coast Down
  Power at 50 mph           Power at 50 mph            Time Interval
   (horsepower)              (horsepower)                 (sec)

      11.5                       8.7                      16.6
      11.5                       8.7                      16.6
      10.8                       8.0                      17.2
      10.8                       8.0                      17.4

-------
-7-
      PS
      W
      13
      O
      P-i
      PQ
      Pi
      O
      CO
      w
      H
      a
      8
                                    DYNAMOMETER NO. 1


                                        7/16/77


                               VEHICLE  FUEL TANK FULL
          124-
              11.1
      5   11

      W

      {3

      g
10
 8 ••
                                      •4-
                                         -4-
             15
                16          17           18


                        COAST DOWN  TIME (SEC)


                             FIGURE 3
19
20

-------
-8-
     The dynamometer adjustment obtained was 11.1 horsepower.  While
similar to the previous results, there is a difference of 0.4 horse-
power.  The most recent tests were conducted with a full tank of fuel,
while the first coast downs were conducted with a nearly empty fuel
tank.  Checking the fuel records showed approximately 20 gallons, or
about 100 pounds of fuel, were added to the vehicle between tests.  Tire
power consumption is very nearly proportional to the vertical load on
the tires, therefore this increase in load would increase tire losses
and therefore reduce the dynamometer power adjustment necessary to
achieve the same coast down time interval.  The rear axle load on the
vehicle was measured to be 2040 pounds with a full tank of fuel.
Consequently the tire losses would be expected to be 5.3 percent greater
with the full tank of fuel than with an empty fuel tank.  The submitted
Ford data gives a computed total corrected road power of 18.23 horse-
power.  Therefore, if the dynamometer adjustment necessary to reproduce
the same coast down time was 11.5 total dynamometer horsepower, 6.7
horsepower was being dissipated in the vehicle tires.  After the vehicle
fuel tank was filled, the tire power dissipation would be about 5.3
percent greater, or 0.34 additional horsepower.  Therefore to maintain
the same total horsepower, the expected dynamometer adjustment would be
11.16 horsepower.

     The dynamometer adjustments, after correction for differences in
the rear axle loads, are in very good agreement.  The difference, 0.06
horsepower may be attributed to the differences in the vehicle warm-up
procedure or may simply be the limits of the test precision.  In either
case it is not considered significant.  This indicates the different
preconditioning methods did not produce the 2.7 horsepower difference as
suggested by Ford.

     The difference between tests two and three, 1.2 horspoewer is
considered significant.  This difference was attributed to the differ-
ences between the dynamometers.  It was hypothesized that one of the two
dynamometers might yield an anomalous result, while all other EPA
dynamometers might be in good agreement.  To test this hypothesis the
measurements were repeated on certification dynamometer No. 3.  This
dynamometer gave an intermediate value of 10.5 horsepower for the same
coast down time.  The data from these measurements are presented in
Table 4 and are plotted in Figure 4.

                                Table 4

                      Dynamometer No. 3, 7/18/77

Total Dynamometer        Indicated Dynamometer         Coast Down
 Power at 50 mph           Power at 50 mph            Time Interval
  (horsepower)              (horsepower)                 (sec)

     11.1                       8.5                      16.4
     11.1                       8.5                      16.3
     10.3                       7.8                      17.3
     10.3                       7.8                      17.2

-------
—9—
                                   DYNAMOMETER NO. 3



                                        7/16/77



                                 VEHICLE FUEL TANK FULL
          12 -
       w
       :s
       o
       FU
       W
       en
       Pi
       O
       o
       P4
11
    10.5
10,,
       o
       W
       PQ
       H
       W
       s
       H
       O
       H
           9,.
           8.
                         -4-
             15
                16         17          18


                  COAST DOM TIME INTERVAL  (SEC)


                             FIGURE 4
                                                              19
20

-------
-10-

       The  results  of  the  test  series  are summarized in Figure 5 where all
  measurements,  conducted  under similar vehicle conditions are plotted.
  These measurements were  conducted with the  same  test personnel,  the same
  instrumentation,  the vehicle  fuel tank was  filled prior to each  test and
  the vehicle warm-up  was  the same in  each  instance.  The dynamometer is
  the only  parameter known to vary between  the tests.  The range of  the
  observed  variations,  1.2 horsepower, is within the range of dynamometer
  variations reported  in the recent EPA Technical  Support Report "Compari-
  son of Dynamometer Foxier Absorption  Characteristics and Vehicle  Road
  Load Measurements."   It  is, however, unknown if  this variation is  the
  extreme of the variations which might be  observed since only dynamo-
  meters No. 1,  No.  3  and  No. 6 were used.

       The  mean  of  the three EPA measurements conducted under similar
  conditions is  10.5 horsepower.  These measurements were conducted  with
  a full tank of fuel;  however,  EPA certification  and fuel economy mea-
  surements are  conducted  with  the fuel tank  40 percent filled.  Using the
  weight method  previously described,  correcting to the situation  of a 40
  percent full fuel tank,  the estimated mean  dynamometer power is  10.7
  horsepower.  The  difference between  this  result  and the Ford result of
  8.8 horsepower is definitely  significant.   If a  dynamometer adjustment
  of 8.8 horsepower is used for the certification  vehicle, this will not
  represent the  road experience of the vehicle tested by Ford.  It is also
  very unlikely  that a dynamometer adjustment of 8.8 horsepower would
  represent the  road experience of the certification vehicle.

       It may be informative to speculate how Ford may have obtained a
  dynamometer adjustment of 8.8 horsepower  for their test vehicles.  The
  problem,  in addition to  the dynamometer to  dynamometer variations  is the
  question  of vehicle  selection.  The  Ford  test vehicle had reported axle
  loads,  when dynamometer  tested, of 2722 and 2281 pounds front to rear
  respectively.   The axle  loads measured at EPA for the Certification
  vehicle were 2760 and 2040 respectively,  with a  full fuel tank.  There
  is the possibility of a  difference in driver weights, and in the possi-
  bility for scale  inaccuracies, however the  front/rear weight distribu-
  tions are 54.4/45.6  for  the Ford vehicle, but 57.5/42.5 for the  certi-
  fication  vehicle.

       It appears that about 250 pounds of  ballast were added to the trunk
  of.the Ford test  vehicle in addition to a full fuel tank.  The rear axle
  load on the vehicle  tested by Ford was about 300 pounds greater  than
  would be  anticipated for the  rear axle load of the certification vehicle
  with a 40 percent full fuel tank.  This would cause abnormally high tire
  power dissipation and result  in a reduction of the necessary dynamometer
  power adjustment  to  match the road coast  down time.  In support  of this
  speculation, the  data submitted by Ford includes an operator note  that
  one test  sequence was aborted because of  tire failure.  Using the  pre-
  vious method for  estimating the changes in  the tire power dissipation
  with changes in axle load, a  300 pound increase  in the axle load would
  decrease  the dynamometer adjustment  by about one horsepower. If  such a

-------
-11-
                          COMPOSITE DYNAMOMETER MEASUREMENTS




                                        7/16/77




                                VEHICLE FUEL TANK FULL
                                               DYNAMOMETER NO.  1
                                               DYNAMOMETER NO.  3
                                               DYNAMOMETER NO.  6
           15
16
17
18
19
20
                                       FIGURE 5

-------
-12-

 vehicle was coast down tested on a dynamometer which gave low power
 absorption results, such as EPA dynamometer No. 6, a dynamometer adjust-
 ment of about 8.8 horsepower could be expected.

      The original EPA recommended practice for road load determination
 did recommend ballasting the vehicle to match the dynamometer simulated
 inertia weight.  The current, revised version, does not recommend bal-
 lasting, and requires the front/rear axle weight distribution of the
 test vehicles to agree within +_ 2% to the certification vehicle.  It is
 assumed that Ford is well aware of the effects which result from changes
 in rear axle loads since the Ford coast down procedure calls for the
 dynamometer tests to be conducted with vehicle engine operating from an
 auxiliary fuel  tank.

 Conclusions:

      1)   A dynamometer adjustment of 8.8 horsepower, as requested by
 Ford, does not  result in an accurate dynamometer simulation of the road
 experience of the vehicle tested by Ford when the certification vehicle
 is operated on  the dynamometer.

      2)   The vehicle road tested by Ford was significantly different
 from the vehicle provided for certification testing.

 Recommendations:

      1)   It is recommended that the vehicle discrepancy be resolved.
 That is, which  vehicle best represents the vehicle Ford intends to sell?
 The road experience of the intended sales vehicle should be determined,
 then the dynamometer adjustment necessary to simulate the road experi-
 ence of the intended sales vehicle should be used for certification and
 fuel economy testing.  The weight and weight distribution of the certi-
 fication vehicle should be adjusted, if necessary, to simulate the
 intended sales  vehicle.

      2)   The requested alternate dynamometer adjustments of other Ford
 vehicles should be examined more thoroughly.  These vehicles should be
 tested if any anomolies seem apparent.

      3)   The apparent 1.2 horsepower difference between dynamometers
 needs to be resolved.  Some of this variation may be attributed to the
 differences in  dynamometer calibration while some may be caused by para-
 meters not currently considered in this dynamometer calibration process.
 Dynamometer to  dynamometer calibration differences are currently be-
 lieved to have  a range of about 0.6 horsepower.  Examples of parameters
 which are not currently considered are the vehicle restraint, the dif-
 ferences in dynamometer rear roll bearings and changes in roll bearing
• friction when under vehicle supporting loads.  In any case, it should be
 noted that the  range of the dynamometer adjustment specified in"the
 current Federal Register is only 6.2 horsepower.  If there are dynamo-
 meter to dynamometer differences of 1.2 horsepower in the simulation of

-------
-13-
   the  vehicle  road  load,  this  is almost  20% of the range of  this para-
   meter.  The  simulation  of  the vehicle  road  load is possibly  the most
   important  parameter, with  respect  to composite fuel economy, in the EPA
   simulation of  the vehicle  road experience.  Clearly such a large random
   effect  in  this  parameter is  undesirable.

-------
APPENDIX  B

-------
                               Table B-l

                Ford Submitted Data for Confirmation of
            Alternate Dynamometer Power Absorption Request
                            November, 1977

VEHICLE:  1978 Mercury 4-Door Sedans -
INERTIA WEIGHT CLASS:  5000 pounds
ROAD TIRE PRESSURE:  Front-30 psi; Rear-32 psi
DYNO TIRE PRESSURE:  Rear-45 psi
Vehicle
Number
313-T-823
313-T-823
313-T-825
313-T-825
313-T-826
313-T-826
313-T-823
313-T-823
313-T-825
313-T-825
313-T-826
313-T-826
Corrected
Coastdown
Time
(sec)
16.79
17.09
16.98
(70-20)
17.04
16.43
(70-20)
16.00
(70-20)
16.79
17.09
16.98
(70-20)
17.04
16.43
(70-20)
16.00
(70-20)
Dynamometer
Cell
Number
2
2
2
2
2
2
6
6
6
6
6
6
Total
Dynamometer
Power
(horsepower)
9.88
9.85
9.72
10.06
10.24
10.95
11.10
11.09
11.04
11.05
11.22
11.86
AVERAGE
16.72
10.67

-------
              Table B-2




   EPA Dynamometer Coastdown Data




Mercury Marquis (TRC Rental Vehicle)
Nominal
Actual
Power (Hp)
12.95
11.95
10.90
9.85
7.75
6.75
Dyno
15.25
15.93
16.93
17.76
20.17
21.58
Mean Coastdown Time
(sec)
//I Dyno #2 Dyno #3
15.42
16.28
16.98
17.98
20.37
21.73
14.66
15.61
16.22
17.31
19.49
20.96
Dyno #4
—
15.87
16.61
17.70
19.81
21.14
Average Coastdowi
Time for All Dym
(seconds)
15.11
15.92
16.69
17.69
19.96
21.35

-------
                APPENDIX  C




Effects of Dynamometer Power Absorption

-------
                               Table C-l
               Vehicle Emissions and Urban Fuel Economy
                                versus
                     Dynamometer Power Absorption
   Dynamometer
Power Absorption
      (HP)
               Carbon    Nitrogen  Carbon
Hydrocarbons  Monoxide    Oxides  Dioxide
   (gm/mi)     (gm/mi)   (gm/mi)   (gm/mi)
   Urban
Fuel Economy
  (mi/gal)
6.3
9.5
9.5
12.7
12.7
15.9
0.75
0.82
0.90
0.73
0.75
0.81
10.5
12.0
13.7
11.0
11.0
11.1
1.85
1.65
1.50
1.83
1.65
2.00
552
552
566
579
577
583
15.5
15.5
15.0
14.8
14.9
14.7
TEST VEHICLE:  Ford LTD  8A1-351W-F-64
INERTIA WEIGHT CATEGORY:  5000 pounds

-------
                              Table C-2

                        Highway Fuel Economy
                               versus
                    Dynamometer Power Absorption

              Dynamometer             Highway Fuel
             Power Absorpton              Economy
                  (Hp)                    (mi/gal)

                  6.3                     24.8

                  6.3                     25.0

                  9.5                     23.7

                  9.5                     23.0

                  9.5                     23.2

                 12.7                     21.1

                 12.7                     21.4

                 12.7                     21.1

                 15.9                     19.7

                 15.9                     19.3
TEST VEHICLE:  Ford LTD  8A1-351W-F-64
INERTIA WEIGHT CATEGORY:  5000 pounds

-------