HDV 78-06
                          Technical Report


                             June,  1978
       A Preliminary Examination of the  Repeatability  of
       the Heavy-Duty Transient Dynamometer  Emission Test
                                by

                       William B.  Clemens
                                 NOTICE

Technical Reports do not necerssarily 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  develop-
ments which may  form  the basis for  a  final EPA decision,  position
or regulatory action.
               Standards Development and Support  Branch
                 Emission Control Technology Division
             Office of Mobile Source Air Pollution Control
                  Office of Air and Waste Management
                 U.S. Environmental Protection Agency

-------
     The purpose of this report is  to  briefly  examine  the repeat-
ability of  the  new Heavy-Duty  (HD)  transient  test  procedure based
upon the limited test  experience gained to date.

     The first two HD  1969  baseline  engines have been tested on the
proposed 1983 MY transient  test procedure.  This procedure involves
the use of a modified  DC  electric dynamoiater  at the EPA Motor
Vehicle Emissions  Laboratory  in Ann  Arbor,  Michigan.    Emissions
were  analyzed  by  the CFV-CVS  (constant  flow  venturi -  constant
volume sampler)  technique.  The CVS  was rated at 1500 SCFM.

     The heavy-duty transient  procedure  is a  new  test procedure,
and a preliminary estimate of  the accuracy of  the test  results
gathered to  date (April 28,  1978)  were  of paramount  interest  in
evaluating the test procedure.  The  accuracy of these results would
be subject  to  two basic types of  errors.   Errors  due to  lack  of
precision,  commonly referred to as repeatability, and errors due  to
offset, usually  considered bias errors.  As shown  in Table 1, the
precision  errors appear to be  quite good  for the  complexity  of
the test procedure, and the  amount of experience  with  the equip-
ment.  The bias  errors are  mainly equipment calibration errors, and
are not quite  as good as the precision errors.   However, the
accuracy of the data presented is considered  adequate  for the
initial stages  of the baseline program.   An   ongoing  effort  will
attempt to minimize  both precision and bias errors  during the
baseline program.
                            Table  1

               Average Coefficient of Variation

                HD Tests           Typical LDV
               (g/BHP-hr)          REPCA  (g/mile)

       HC        13.4%                 5.7%
       CO         6.5%                 6.0%
       NOx        4.4%                 4.4%
The  specific coefficients of variation are listed in Table  2.
As  shown,  the variability of the HD  transient  procedure  compares
favorably  with  results obtained on  the LDV  transient  procedure,
which  is  quite  pleasing  this early  in the program.  However,  it
should be emphasized that  these values  are only average values, and
in some instances the variation for an  individual engine  or vehicle
can be quite high.

     In order to  investigate the  source  of the variability, the
coefficients  of variation for the hot  start portion  of  the heavy-

-------
                                -2-
                              Table  2

                       Coefficient of Variation
                               (s/x)

A.               HD Cold Soak Emission Test  (g/BHP-Hr)
                       (cold start + hot start)
                   BHP-hr     BSHC      BSCO      BSNOx
1. 225 CID
  Dyno cal. "A"     1.5%      9.6%      5.5%      4.1%
  Dyno cal. "B"     3.9%      4.8%      6.4%      4.9%

2. 392 CID
  Dyno Cal. "B"     6.2%     25.9%      7.7%      4.2%

B.                  HD Hot Portion (g/BHP-Hr)
                         (hot start  only)
                        BSHC     BSCO      BSNOx
1. 225 CID
  Dyno cal. "A"         9.8%     5.1%      4.5%
  Dyno cal. "B"         3.6%     5.6%      4.9%

2. 392 CID
  Dyno Cal. "B"        27.6%     9.2%      3.5%

C.                       LDV FET  (g/nile)

                         HC       C0_        NOx
1. Production Catalyst Vehicle
     Dyno A              5.9%    47.3%      3.3%
     Dyno B              6.1%    26.2%      4.1%
     Dyno C             10.0%    31.1%      5.1%

2. REPCA Non-Catalyst Vehicle
     Dyno A              5.6%    5.9%       4.1%
     Dyno B              6.3%    5.5%       4.6%
     Dyno C              5.3%    6.5%       4.4%

-------
                               -3-
duty test were computed.  These  data (Table  2A and B)  indicate  that
there is essentailly no difference in terms  of emission variability
between  the  hot portion  and  the  total test.   This  result is not
surprising since  the weighting  factor for  the  hot  start portion
for  these  tests was  0.875 versus  0.125  for the cold  start  por-
tion.*   Therefore,  it  is  assumed  that most of the variability can
be attributed  to. the  repeatability  of each segment.   It is  sus-
pected that  the starting  segment  of each  portion has the  greatest
variability.    A more  detailed  analysis  may be  undertaken in the
future to verify this assumption.

     The emission measurement  accuracy discussed  so far, deals  only
with errors  due to  random variability for a given  set of condi-
tions.   Another source of errors in emission results  are bias
errors.  An  example of  a  bias error  would be the testing  of an LDV
with an  incorrect  power  absorber  setting.   It  would be  expected
that the emission results  would  have equivalent variability regard-
less of  the  power  absorber  setting, but,  only  one  group of  data
would be correct,  the data taken  with the  correct  power  absorber
setting.

     This analogy applies  to the heavy-duty  transient  test  as well.
Only, the  heavy-duty test has  many more parameters  that must be
measured to assure the correct power setting.  As in  the light-duty
test, most bias errors can be  corrected  with  proper calibration.

     During these tests, two different calibrations were used.  The
difference between calibration "A" and calibration "B" was  that the
engine operating speed was increased from  the values  in calibration
"A"  to values slightly  over the reference speed  in the test cycle.
The  control  system was  adjusted in  this manner because calibration
"A"  generally  had  a negative error in speed which  resulted   in a
measured  cycle  horsepower-hour  approximately  10% below the refer-
ence  horsepower-hour.   The speed  adjustment  was overcompensated
slightly, and the change  resulted  in  calibration "B"  operating the
engine  approximately  14% above  the  reference   or  correct horse-
power-hour.  Although this change did  not significantly affect the
repeatability,  it  did  affect the emission levels  (see  Table 3).

     Based on  the  difference between calibration "A" and "B" in
inegrated BHP-Hr  over cycle,  the following  observation can be
stated;  for  the tests  on  the  225 CID Chrysler engine,  every 1%
change  in  cycle BHP-Hr resulted  in approximately a  2%  change in
*  Note:  Since  this report  was prepared,  the  hot/cold weighting
factor has  been  changed based upon  final  analysis  of the CAPE-21
data  base  -  see  EPA Technical  Report  No. HDV  78-04,  "Transient
Cycle Arrangement  for Heavy-Duty Engine  and  Chassis  Emissions
Testing", by C. France, June 1978.

-------
                               -4-
                             Table  3

                           Test Results
                    Cold  Start Weighting = .125
                     Hot  Start Weighting = .875

                       Cold Soak Emission Test

                   BHP-hr    BSHC      BSCO      BSNOx
1.   225 CID
  Dyno Cal. "A"    10.95      6.68      52.24     9.66
  Dyno Cal. "B"    13.96      3.96      47.59     8.58

2.   392 CID
  Dyno Cal. "B"    19.39      11.83     206.17     3.91

-------
                                -5-
BSHC,  a 0.375% change in  BSCO,  and a  0.5% change in  BSNOx.
Although no similar LDV data are available  for  REPCA at MVEL in Ann
Arbor for comparison, the variation in emissions relative to BHP-Hr
changes from this HD engine seem acceptable.

     The International  Harvester  392  CID engine was operated only
with  speed calibration "B".   However, the  IHC engine had  more
difficulty in following  the  torque  cycle than  the Chrysler engine.
Thus, the cycle power developed by the IHC  engine was approximately
6% below the reference power.   Because the IHC engine did not meet
the statistical  criteria for torque,  it is most  likely that that
engine will be retested.

     From this  discussion  of bias  errors,  it  is obvious that bias
errors  do  occur during  testing.   The effect  of these bias errors
influence   the  emission  test results in  two  important  ways.
Initially,   the  engine is  operated  at an incorrect  power setting.
The magnitude and  direction of the error  in  the engine emissions
due  to  operating the  engine at an incorrect  power  setting would
generally be unknown.  Possibly more  important though, is that the
test  results  in  grams/horsepower-hour would  be  computed with  a
horsepower-hour value that is different than the reference or cycle
horsepower-hour.   The magnitude  and  direction of  this  effect  is
easily  computed  by knowing  the  operating  horsepower-hour  and the
reference horsepower-hours.

     To some  degree  bias erros due to  operating  the  engine  at an
incorrect  power setting, and  bias errors due  to  dividing  by a
different horsepower-hour value must be accepted.  No machine will
ever be perfect.  However, errors  of  this  type can be minimized by
continuing  to  emphasize accurate  calibration of  the  equipment.

     The actual  emission results  from each individual  test  on the
two engines are given in Tables 4 and  5. Additional examination of
the  repeatability  of  the  HD  transient  dynamometer  emission  test
will be performed as more test  data become  available.

-------
                              Table 4
Engine Description;
MFC:   Chrysler
CID:   225
Type:  L-6
Cold Start Weighting
Hot Start Weighting
         S/N:  FW225R2994032
   Rated BHP:  94
   Rated RPM:  3556
.125
.875
Maintenance Requirements/Problems;
Tune-up prior to BLT-1
Rebuild Carburetor After BLT-7
10-Hour Cold Soak Test Results:

Test No.
BLT-1
BLT-2
BLT-4
BLT-5
BLT-7
BLT-8
BLT-9
BLT-11

x"
s

Dyno Cal. HC
"A" 6.08
"A" 6.31
"A" 6.77
"A" 7.54
"B" 5.12
"B"* 3.74
"B"* 4.05
"B"* 4.08
A B*
6.68 3.96
.64 .19
4-Hour Cold Soak Test Results
Test No.
BLT-3
BLT-6
BLT-1 0
Dyno Cal. HC
"A" 6.17
"A" 6.83
"B"* 4.23
(g/BHP-Hr)
CO
55.61
49.04
50.95
53.34
50.00
46.33
45.50
51.04
A B*
52.24 47.59
2.86 3.02
•
•
CO
48.11
51.75
49.74
(Ib /BHP-Hr)
NOx
10.00
9.74
9.23
—
9.12
9.00
8.57
8.16
A B*
9.66 8.58
.39 .42

NOx
9.19
9.36
9.20
BSFC
.821
.655
.650
.627
.597
.574
.554
.576




BSFC
.640
.618
.591
BHP-Hr
10.767
10.894
10.967
11.154
13.551
14.488
14.011
13.394




BHP-Hr
11.016
11.110
14.140
* plus carburetor rebuild.

-------
                                -7-
                              Table 5
Engine Description;
MFC:   IHC
CID:   392
Type:  V-8
Cold Start Weighting
Hot Start Weighting
             S/N:  V392 658417
       Rated BHP:  159
       Rated RPM:  3527
    .125
    .875
Maintenance Requirements/Problems;
Tune-up prior to BLT-1
Replaced faulty coil wire after BLT-4
10-Hour Cold Soak Test Results:
Test No. Dyno Cal.  HC
BLT-1
BLT-3
BLT-5
BLT-7
BLT-8
c
s
"B"
"B"
"B"
"B"
"B"


15.96
—
12.22
8.93
10.22
11.83
3.07
          (g/BHP-Hr)
              CO
           200.30

           191.86
           203.87
           228.64

           206.17
            15.81
                 (Ib/BHP-Hr)
             NOx     BSFC     BHP-Hr
             3.68    .795     21.153
4.02
4.04
3.89
3.91
.17
.825
.819
.868


18.941
18.958
18.506


4-Hour Cold Soak Test Results:
Test No. Dyno Cal.
 BLT-2      "B"
 BLT-4      "B"
 BLT-6      "B"
 HC
8.80

8.41
   CO
172.10

193.90
NOx
4.02

4.29
.817
BHP-Hr
19.805

19.098

-------