LDTP  78 - 10
                           Technical Report
      "Comparison of Fuel Economy Values from the Preconditioning
      and Measuring Cycles of the Highway Fuel Economy Test Procedure"
                                  by

                            Myriam Torres
                              May, 1978
                                NOTICE

Technical reports do not necessarily represent the final EPA decision.
They are intended to present a technical analysis of an issue and
recommendations resulting from the assumptions and constraints of that
analysis.  Agency policy constraints or data received subsequent to the
date of release of this report may alter the recommendations reached.
Readers are cautioned to seek the latest analysis from EPA before using
the information contained herein.
               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

-------
                               -1-
Introductlon

     The Highway Fuel Economy Dynamometer Procedure consists of driving
the test vehicle over two highway driving cycles with fifteen seconds of
idle between them.  The first cycle is for preconditioning purposes and
the second cycle is used for the fuel economy determination.  This
procedure is designed to follow the Federal Emission Test Procedure,
known as the FTP.  However, in the event that it cannot be scheduled
within three hours of the FTP, the vehicle will be re-driven over one
cycle of the EPA Urban Dynamometer Driving Schedule, known as the LA-4,
before running the two highway driving cycles (see Appendix B). JY

     Since the vehicles are driven either over a FTP or a LA-4 prior to
the highway fuel economy test, the vehicles may already be in a stabilized
condition during the first cycle.   Therefore, a study was conducted on
fifteen light-duty certification vehicles during the regular certification
highway fuel economy tests to determine if it is necessary to run the
second highway cycle.  For this study, fuel economy values were determined
during both cycles and were then compared.  All fifteen vehicles had
smaller fuel economy values during the preconditioning cycle than during
the measuring cycle.  This report presents summary statistics of the
absolute and relative (percent) differences between the fuel economy
determinations of the two cycles and compares these differences to the
time elapsed between the end of the FTP (or LA-4) and the start of the
highway test.  An additonal comparison of fuel economy values determined
during a three highway cycle study is presented and discussed.

-------
                              —2-
Discussion/Analysis

     The sampling procedure and recording during  the preconditioning
cycle were conducted in exactly the same manner as during  the measuring
highway cycle.  Additional recordings were:   (1)  whether an  FTP  or  LA-4
were run previous to the highway  fuel economy  test (HFET)  and  (2) the
soak time between the FTP (LA-4)  and the HFET.  The purpose  of recording
these parameters was to determine if either of these two factors have  an
effect on the difference seen between the two  fuel economy values.

     For each vehicle, the fuel economy values during  the  preconditioning
cycle were less than those determined during  the  measuring cycle.   This
supports the results of a one vehicle study conducted  by Austin, Hellman
and Paulsell _2/ in which they found that during the first  highway cycle
the vehLcle was not yet stabilized.  After running a vehicle over an FTP
and allowing a one hour soak time, they found  that the fuel  economy
determinations during the first cycle were 98% of the  highway fuel
economy values from a fully warmed-up vehicle.  If we  assume that the
vehicle is fully warmed-up during the second  cycle, then the 98% compares
favorably with the 96% found for  the average  of the eight  vehicles  in
this study that had an FTP run just prior to  the  highway tests with an
average of one hour and 26 minutes of soak time.  The  decrease in the
percentage may possibly be due to the increase in the  average soak  time.

     Table 1 shows the mean values for the soak time and the mean absolute
and relative differences for the  fifteen vehicles differentiating by FTP
and LA-4.  The relative differences do not seem to be  affected significantly
by the procedure that was used prior to the highway test.  However, a
direct relationship between soak  time and the percent  differences was
found when these two variables were plotted against each other.  This
plot is given in Figure 1 together with the regression line  running
through the points.   The purpose of the regression line is to more
clearly visualize; the direct relation between  the variables  shown by the
positive slope oi: the line.   The  point that is circled seems to  have a
percent difference too large for a small elapsed  time  of 52 minutes.
Since this point corresponds to a manual transmission  vehicle, its
highway driving trace was examined to see if  there was any discrepancy
between the preconditioning and measuring cycles' shift patterns.   The
shift speeds were actually different and the preconditioning acceleration
line was substantially below the  trace with the vehicle being driven at
wide-open throttle.   All the traces for the other manual transmission
vehicles in the sample were also examined and no  differences were found
between their two cycles'  speed time traces.  A description  of the
vehicles and their corresponding  fuel economy determinations can be
found in Appendix A.

-------
* OIFF
 7.0000
 6.0000   *
 5.0000   *
 4.0000   +
 3.0000   *
 2.0000  . *
 l.OOUO   *
 0.
        0.                   50.000               loO.00              150.00
                  2S.OOO               75.000               125.00               175.00

                                           TIME(MlfJ)
            Figure 1  - Plot of percent difference  between  the  fuel economy
           values versus the soak time preceding the  highway  fuel economy  test

-------
Procedure Preceding Sample
  Highway Test	  Size
FTP

LA-4

Overall
 8

 7

15
 -3-
  Table 1

Mean Soak
  Time
  (min)

   86

   31

   60
 Mean F.E.
Difference
   (mpg)

    .9

   1.1

   1.0
                                       Mean %
                                     Difference*
4.1

3.9

4.0
        CERT F.E. - PREC F.E.
              CERT F.E.
          x  100
     The results of one of the tests were not included in the analysis
because of a change in the test procedure.  During the second cycle run,
it was discovered that the sampling pump was off.  Lab personnel then
decided to complete the cycle and sample during a third cycle to obtain
the certification fuel economy values.  This procedure was considered to
be in agreement with the Highway Fuel Economy Test Procedure since the
second cycle could serve the purpose of preconditioning the vehicle
followed by a cycle driven for the fuel economy measurement.  Furthermore,
the underlying assumption was that the vehicle is already in a stabilized
condition during the second cycle and, therefore, the fuel economy
values determined during a third cycle should not increase significantly.
Any change seen is assumed to be purely by chance so that at times a
small increase can be seen and at other times a small decrease.  This
assumption was based on the study conducted by Austin, Hellman, and
Paulsell that was mentioned earlier.  However, their results were based
on testing only one vehicle under two different testing conditions.

     The results for the certification vehicle that was run over three
cycles were 20.5 mpg during the first cycle and 22.0 during the third, a
6.8% difference.  Since this difference was the largest for all the
tested vehicles, it was considered a possibility that if the second
cycle were sampled, it's fuel economy value would be less than 22.0 mpg.
Since valid three cycle tests have been run in the past, it was considered
important to verify if the fuel economy values do continue to increase
after the second cycle.  Therefore, a small test program was conducted
on three light-duty non-certification vehicles during which the vehicles
were soaked for at least 12 hours, were run over an FTP and allowed to
soak for one hour before running three highway cycles and sampling from
all three.  The results for the three vehicles are in the following
table:

-------
                          -4-
                           Table 2

            Fuel Economy Determination from Three
                       Highway Cycles
Manufacturer
GM
GM
Chrysler
Mean
^ CYCLE 2 F
1975
1976
1976
.E.
Model
Cycle 1 FE
(mpg)
Nova
Chevette
Aspen
- CYCLE 1
F
15.
27.
20.
.E.
8
5
8
                                  Cycle 2 FE
                                    (mpg)

                                     16.1
                                     27.6
                                     21.2
  Cycle 3 FE
      16.3
      27.7
      21.4
% Diff
1 to 2*

  1.9
   .4
  1.9

  1.4
% Diff
2 to 3**

  1.2
   .4
   .9
                                                                         .8
          CYCLE 2 F.E.
                               x 100
CYCLE 3 F.E. - CYCLE 2 F.E.
        CYCLE 3 F.E.
             x 100
     It is clear from the table that vehicles are not yet stabilized and
will therefore obtain slightly higher fuel economy values during the
third highway cycle.  It is to the manufacturers' advantage when the
fuel economy values published for their 'vehicles were determined during
the third cycle.  In addition, note that the percent difference between
the first and second cycle fuel economy values for these vehicles -are
much smaller than for the fifteen vehicles tested during regular certifi-
cation testing.  The reason for this discrepancy is not known but may
result from different warm-up characteristics and emission control
devices.

     From the table we can see a tendency for the percent difference
between the second and third cycles to increase as the difference between
the first and second cycles increases (see Figure 2).  If we assume that
this increase will be linear we can then extrapolate the line that goes
through the means of the Y-axis values in Figure 2.  Therefore, 2.0% is
an estimate of the difference between the second and third cycle for the
average case of a 4.0% difference between the first and second cycle for
certification vehicles.   The difference between the first and third
cycle for this average certification case would be less than the sum of
2.0% and 4.0%, that is,  less than 6.0% (this is mathematically proven in
Appendix C).   Since the certification vehicle eliminated from the
analysis had a difference of 6.8%, it seems to be an above average
vehicle, and therefore,  the estimate of 2.0% appears to be not only
reasonable but may be conservative.

Summary/Recommendations

     1.   Fuel economy values determined during the preconditioning
cycle of the Highway Fuel Economy Test will be lower than those determined
during the measuring cycle.

     2.   There is a direct, but not very strong relation between the
percent difference of the fuel economy values from both cycles and the
soak time between the FTP (or LA-4) and the highway 'test.

-------
 J
 o
3
u)

V?
       /.O
       0.5-
          0.
/.O
/.S'
^.0
2.S
3.0
3.5"
                        Figure 2 - Plot of percent  difference  between second and  third cycle

                        fuel economy values versus  the  difference  between  those from the first

                        and second cycles.

-------
                              -5-
     3.   The mean absolute difference found between the fuel economy
values was 1.0 mpg and the mean percent difference (—
                                                         CERT F.E.
x 100) found was 4.0%.

     4.   Vehicles are not yet stabilized during the second highway
cycle since the fuel economy values continue to increase when a third
cycle is run.  For an average difference between the first and second
cycle fuel economy values of 4.0%, the difference between the second and
third cycle values is estimated to be approximately 2.0%.

     It is recommended that only data collected during the second HFET
cycle be used for calculation of a vehicle's fuel economy.  Using data
from the first cycle would penalize the manufacturer by lowering his
corporate average fuel economy (CAFE).  Using data from the third cycle
would give the manufacturer an advantage over other manufacturers by
raising his CAFE and his published fuel economy values.

-------
                           -6-
                          References

1.    "Fuel Economy Testing; Calculation and Exhaust Emissions Test
     Procedures for 1977-1979 Model Year Automobiles",  Federal Register,
     Vol.  41, No.  177,  Friday, September 10, 1976.

2.    Austin, Thomas C., Karl H.  Hellman, C. Don Paulsell,  "Passenger
     Car Fuel Economy During Non-Urban Driving", SAE 740592,  New York,
     New York, 1974.

-------
    APPENDIX A

Vehicle Description
        and
 Fuel Economy Data

-------
                                       Table A-l




                                  Vehicle Description
Test No.
79-0024
79-0046
79-0044
79-0075
79-0106
79-0103
79-0104
79-0148
79-0149
79-0203
79-0186
79-0150
79-0284
79-0151
79-0348
Manufacturer
Ford
Ford
Ford
Ford
Checker
AMC
Chrysler
Avanti
Chrysler
Chrysler
Toyota
GM
CM
AMC
Ferrari
Model
Pinto
Thunderbird
Wagon
Ford
Ford Sedan
Marathon
Pacer Wagon
Plymouth
Avanti II
VW Rabbit
Chrysler
Corolla 4DR
Sedan
Monza Sta
Wagon
Regal Sport
Coupe
Pacer Wagon
Ferrari
Vehicle ID
8E2-2.3-F-24
8S1-400-F-280
8A1-351M-F-276
8A1-400-F-275
8C8
D76-84C(D)
A163
2261A
A005T
A197
78-FE-2
84B2-288G
84E3-48170F
D76-84C(D)
13492
Curb Wt.
(Ibs)
2548
4365
4500
4501
4094
3625
2210
3793
2218
4929
2012
3032
3377
3625
3240
Trans
M-4
A- 3
A- 3
A-3
A- 3
A-3
M-4
A-3
A-3
L-3*
M-4
M-4
A-3
A-3
M-5
Tire Size
A78X13
HR78X15
HR78X15
HR78X15
G78X15
D78X14
P165/75R13
195-15
P155/80R13
JR78X15
6.15S13
BR70-13
P205/70R14
E78X14
205/70VR14
*L = lock up/auto

-------
        Table A-2




Fuel Economy Measurements
Test No.
79-0024
79-0046
79-0044
79-0075
79-0106
79-0103
79-0104
79-0148
79-0149
79-0203
79-0186
79-0150
79-0284
79-0151
79-0348
Manufacturer
Ford
Ford
Ford
Ford
Checker
AMC
Chrysler
Avanti
Chrysler
Chrysler
Toyota
GM
GM
AMC
Ferrari
Test Date
01-05-78
01-05-78
01-05-78
01-06-78
01-06-78
01-09-78
01-09-78
01-11-78
01-11-78
01-16-78
01-16-78
01-18-78
01-18-78
01-18-78
01-19-78
Curb Wt.
(Lbs)
2548
4365
4500
4501
4094
3625
2210
3793
2218
4929
2012
3032
3377
3625
3240
Time
(min)
8
27
115
163
5
10
30
148
18
7
52
160
55
125
3
CERT FE
(mpg)
35.2
19.2
20.1
19.1
14.5
17.9
39.7
18.4
32.6
16.7
44.7
23.8
25.8
17.5
18.3
PREC FE
34.8
18.3
18.9
18.2
1A.2
17.3
37.5
17.5
31.4
16.1
41.7
23.0
24.9
16.7
18.0
Diff.
(mpg)
0.4
0.9
1.2
0.9
0.3
0.6
2.2
0.9
1.2
0.6
3.0
0.8
0.9
0.8
0.3
% Diff
C-P
( ^ vi n n
V r, JAJ.UU
C
1.1
4.7
6.0
4.7
2.1
3.4
5.4
4.9
3.7
3.6
6.7
3.4
3.5
4.6
1.6
Proc.
LA-4
FTP '
FTP
FTP
LA-4
LA-4
LA-4
FTP
FTP
LA-4
LA-4
FTP
FTP
LA-4
FTP

-------
            APPENDIX  B

Section 600.111-78 Federal Register
   Friday, September 10, 1976

-------
                                    Federal Register Description of
                                        Highway  Fuel Economy  Test
S MO. 11 I-To  T<:.-t jii-o.-cduro--.
  ui) The test procedures to be followed
for  Generation of the city fuel economy
data, are  those  prescribed in  ;H So. 1-7
throush liCJ.inS of this  chapter, as ap-
plicable.  (.The evaporative loss  portion
of the ttit  procedure may be omiucd
unices  specifically required by  tht; Ad-
ministrator.)
  i.b) The test- procfrjuye.s to bo followed
for  generation of the highway fuel  econ-
omy data are thc:-e specified  in  5 tSOO.-
111-78 ib) through (h) inclusive.
  (I)  The   Highway  Fuel  Economy
Dynamometer Procedure  consists of a
preconditioning   highway  driving  se-
quence and a measured  h'jlv.vay driving
sequence.
  (it) The highway fuel economy test Is
designated to simulate non-;r.etropoHu:.:i
driving  with nn  average  speed or 13.6
mph and  a  maximum ipeed of 60 mph.
The cycle is  10.2  miles Ion? with 0.2 stops
per mile and consists of warmed-up vehi-
cle  operation on & chassis dynamometer
through a specified driving cycle. A pro-
portlonnl  part of  the  diluted e.xhaas-:.
emissions is coHeL-rod conUnu continuously  analyzed
for  hydroo.irbons using  a  heated sam;.>ji;-
line and analyzer.
  '3) Except in cases ot  component. r,:sl-
fxinction  or  failure, all  emission convroi
systems installed on or incorporated  in
a new motor vehicle must  be functioning
during all procedures in this subpart. The
Administrator may authorise  mainten-
ance to correct component malfunction
or failure.
  ic) Transmissions—The provisions  c:i
5 tlbML'fi of  this  chapter  apply  for ve-
hicle transmission operation during Ws.'i-
way fuel  economy testing  under thi-;
.stibpart.
  'd> Road lc:ui po <•,-•;• and inert;;! we.'ahi.
deter:r.;naiio::—5 85.120 of thl; chnprer
applies  for  determination of rond !cac>
power  ana  i'.-.ertia  ve;?!ii for  h.gh?:;v;-
fuel economy testnir:.
  I e)   Vehicle   precondition;:';?—T;'..
H:;;hsvay  Fuel  Economy  DyEar.ionj Engine  starting and restarting—
  (1) If the engine  Is  not running at
the initiation of tne highway fuel econ-
omy  test (pivcondltlonlng cycle), the
start-up procedure must  be  according
to the manufacturer's recommended pro-
cedures.
  (2) False starts" and stalls during the
preconditioning cycle  must  be treated as
in paragraphs (d)  and (e)  of 5 88.136 of
this chapter, If the vehicle  stalls during
the  measurement cycle  of  the highway
fuel economy test, the test is voided, cor-
rective  action rr.t'.y be  taken  according
to  5 86.077-25 of this chapter, and the
vehicle  may be rescheduled for test,. The
person taking the corrective action shall
report the notion so that the test records
for the vehicle contain  a record  of the
action.
   (h) Dynamometer Test Run—The fol-
lowing steps rr.v.sc be taker, for each test:
   CD Place the drive wheels ci the re-
hide on  the dynamometer. The vehicio
rnfty be driven onto the dynamometer.
  (2)  Open the vehicle engine compart-
ment  cover and position  the  cooling
fam's) required. Manufacturers may re-
quest  the use of additional cooling fans
for additional  engine  compartment or
under-vehicle cooling and for controlling
high tire or brake temperatures during
dynamometer operation.
  (3)  Preparation of the "CVS must be
performed before the measurement high-
way driving cycle.
  (4)  Equipment preparation—The pro-
visions of paragraphs  (b)  (3)  through
(5) inclusive of § 86.137 of this chapter
apply for highway fuel  economy test ex-
cept that only  one exhaust sample  col-
lection bag and one dilution air sample
collection bag  need be  connected to the
sample collection systems.
  (5)  Operate  the vehicle over one High-
way Fuel  Economy Driving  Schedule
cycle  according  to  the  dynamometer
driving schedule specified in  paragraph
(b> of 3 600.109.
  (C)  \Vhen' the  vehicle reaches 'zero
speed at  the end  of the preconditioning
cycle, the driver has 13 seconds to  pre-
pare for the emission measurement cycle
of  the test. Reset and enable the roll
revolution counter.
  (7)  Operate the vehicle over one High-
way Fuel  Economy Driving  Schedule
cycle  according  to  the  dynamometer
driving schedule specified in  paragraph
(b> of §  600.109 while  sampling the ex-
haust gas.
   (8)  Sampling must begin two seconds
before beginning the first acceleration of
the fuel economy measurement cycle and
must  end two seconds  after the end  of
the deceleration to zero. At the end  of
the deceleration to zero speed,  the roll
or shaft revolutions must be recorded.

-------
        Appendix C




Mathematical Computations

-------
                          Mathematical Proof
Problem;

Prove that the percent difference between the first and third cycle is
less than or equal to the sum of the percent difference between the
first and second cycle and the percent difference between the second and
third cycle.
     Let X = F.E. value for first cycle
         Y = F.E. value for second cycle
         Z = F.E. value for third cycle
     Prove that

         Z-X    Z - Y    Y - X
           Z   - .   Z        Y


Proof:

         Z-X =  Z - Y + Y - X
           Z   ~        Z"

                  Z - Y .  Y - X
                                 Y _ V   V _ Y
               Since Z > Y, then   „   <   v
     80  ?_JL_X = ?_JL-! + Y ~ x   2 - Y   Y - X
         "Z     "  Z       Z   -   Z       Y
     therefore

         Z-X   Z-Y.Y-X

-------