Emission Laboratory
        Correlation Study Between EPA
          and Audi - NSU Auto Union
                      by

              Richard E. Lowery

                  June 1974
       Environmental Protection Agency
 Office of Air and Waste 'Management Programs
Office of Mobile Source Air Pollution Control
     Emission Control Technology Division
  Regulations Development and Support Branch
             Ann Arbor, Michigan

-------
                         Emission Laboratory
                    Correlation Study Between EPA
                     and Audi - NSU Auto Union
                              ABSTRACT
     This report presents the results of a laboratory correlation study
between EPA and Audi - NSU Auto Union of Germany.   An attempt is made
to isolate equipment, procedure, and ambient differences which affect
the degree of emission result correlation.

     A research vehicle tested at both Audi and EPA revealed a good HC
emission correlation, but significant discrepancies in measurements of
CO and NOX.  Differences in dynamometers and ambient conditions seemed
to be the most important factors affecting the degree of emission cor-
relation.

-------
                          Emission Laboratory
                     Correlation Study Between EPA
                      and Audi - NSU Auto Union
Introduction:
     In an attempt to ensure a good laboratory correlation between Audi-
NSU and EPA, a program was initiated to exchange information with Audi
and relate that information to the emission test results of a research
vehicle tested at both laboratories.  The findings of that program, in-
cluding conclusions regarding the factors affecting the degree of cor-
relation, are the substance of this report.

Organization and Sequence of Testing;

     Because Audi's request for correlation information was made during
peak demand for EPA certification work, a small scale testing program was
a necessity.  Also, since OMSAPC's Test and Evaluation Branch had already
scheduled the Audi vehicle for evaluation testing, the correlation testing
had to be arranged within the limitations of the vehicle's availability.
Under these restrictions, EPA and Audi representatives agreed on the fol-
lowing program:

          (1)  Audi and EPA would exchange information relating to
          their facilities, equipment, and procedures.

          (2)  EPA would conduct three valid hot-start emission tests
          on the Audi vehicle on a certification site.  Audi would
          provide data of similar tests at the Audi - NSU laboratory.

          (3)  Audi would submit static gases to EPA for analysis
          and comparison with values generated at the Audi laboratory.

     Steps 1 and 2 of the program were initiated simultaneously on May 31.
The gases required for Step 3 arrived at EPA on June 17, and their analyses
was completed on EPA's Train - 15 on June 24.

Teat Procedures;

     The vehicle emission tests conducted for this study were hot-start
tests over the LA-4 cycle using only one set of collection bags.  Sample
collection was begun upon vehicle ignition as is done in 1975 federal
tests.  CVS flow rates were adjusted to accommodate the usage of only
one sample bag.  These vehicle tests were run at EPA on certification
site #5.

-------
                               -2-
Test Vehicles;

     The vehicle used for correlation testing was a modified 1974 Audi
Fox equipped with an oxygen - sensor system which controlled air-fuel
ratio.  The following data apply to this vehicle.

          Model:  Audi Fox
          V.I.D:  8442000011
          C.I.D:  97.0
          Inertia Weight:    2500 pounds
          Transmission:      4-speed manual
          No. of Cylinders:  4

In testing this vehicle, dynamometer horsepower was set at 7.0 at 50 mph,
a value different than Federal Register specifications.  This was done
because Audi had supplied appropriate data to EPA's Certification Branch
such that their 1975 certification fleet will be tested at that setting.
                          N
Analysis of Data:

     The analyses of Audi - NSU facilities data reveals the following
comparative differences with the EPA facility.

                                   EPA                  Audi - NSU

Dynamometer

  Drive Type                       Direct               Belt
  Roll Spacing                     17.25 in.            20.25 in.

Driver's Aid

  Chart Speed                      4 in./min.           4.72 in./min.
  Resolution                       60 mph=5.00 in.      60 mph=4.56 in.

Facility Altitude (ft.)            897.5                1276.3

Analyzers

  Span Gases (%Full Scale)         75-96                47-97.5
  # Calibration Points
        HC and NOX                   6                    2
        CO and C02                   6                    8

Complete Audi - NSU and EPA facility data may be found in the Appendix.

-------
                                -3-
     In general, average emission levels measured on the Audi correlation
vehicle were higher at the Audi facility than at EPA.  Specifically, com-
paring average levels of three tests at each facility showed Audi 3.8%
higher on HC, 19.3% higher on CO, and 56.6% higher on NOX.  No C02 values
from the tests at Audi were available, so no comparison of CC>2 levels
was possible.  Complete vehicle emission test data are summarized in
Table 1.

     The static gas checks compared measurements of four span gases at EPA,
Volkswagen, and two Audi analysis sytems.  The largest discrepancies were
revealed in HC (C3H8) measurements, where Audi's values were 5.2% and 4.3%
lower than the EPA measurement.  All CO and C02 values obtained at Audi and
Volkswagen were lower than correponding readings at EPA, the maximum dif-
ference being 3.1%.  The NO correlation was quite good, with Audi's Train
II measuring the largest deviation from EPA's value, 2.3%.  A summary of
these calibration gas data is presented in Table 2.

Discussion:

     Significant CO and NOX emission level differences between EPA and
Audi-NSU were observed in this program, and several factors were documented
which could possibly contribute to this problem.  The most important factor
is probably the dynamometer roll-spacing difference, Audi's 20.25 inch span  .
vs. EPA's 17.25 inch.  This difference significantly affects the tire-roll
friction, especially since the test vehicle was equipped with relatively
small tires (13" rims;  approximately 11" inch rolling radius).  The result
of this difference is that the work necessary to complete the LA-4 driving
cycle is significantly higher on the rolls of wider spacing.  Thus the effect
of the wider roll span would probably be higher levels of all emittants,
the magnitudes of which would depend on the particular vehicle's character-
istics.  The Audi-EPA NOX discrepancy, although larger than would be ex-
pected because of roll spacing difference, may have been largely attributable
to this factor.  Perhaps the Audi vehicle was particularly sensitive to
the effects of increased loading.

     Another factor which may have significantly affected the correlation
results is the difference in barometric pressure (facility altitude) between
the two laboratories.  Although Audi's oxygen - sensor emission control
system should compensate for barometric differences, its effectiveness is not
certain.  A trend toward higher CO and HC measurements at Audi than at EPA
would be a result of Audi's lower barometric pressure if the altitude com-
pensation was not effective.

     The influence of analyzer calibrations could also have had a significant
effect on the emission correlation.  The static checks showed that Audi
would measure higher NOX levels (approximately 2%), and lower HC concentrations

-------
                                -4-
(4-5%) than EPA, which would account for a minor portion of the NOX emission
discrepancy.  However, it would increase the HC emission difference.   The
effect of different HC and NOX analyzer techniques might also be.important,
since Audi's two point calibrations are not as accurate as EPA's six -
point technique.

     One other factor may have contributed significantly to the NOx emission
differences, that being the test cell relative humidity.  Audi typically tests
their vehicles at humidity near 75 grains of water per pound dry air, while
the EPA test cell measured an average of about 50 grains during testing.
If the Audi vehicle was substantially less NOX sensitive than t^he correlation
factor assumes, then the NOx factor over - corrected EPA results.   Other
correlation studies with American manufacturers have indicated that NOX
measurements at EPA were frequently over - corrected.

Conclusions:

     A comparison of laboratory test equipment and procedures and vehicle
emission levels has indicated correlation problems between EPA and Audi-NSU.
The most significant problem was revealed by Audi's average measurement of
56.6% higher NOX on the correlation vehicle.  Audi's large dynamometer roll
spacing is the most probable cause of the NOX discrepancy.

     The difference in barometric pressures between the EPA and Audi - NSU
laboratories did not seem to have a major effect on the degree of correlation
on the vehicle tested, but it should be realized that this factor could be
of greater importance on other control system configurations.

     Differences between EPA and Audi - NSU analysis systems were discovered,
but their effect on the emission discrepancies were minimal.  However,
Audi - NSU should be encouraged to compare the effects of six - point
HC and NOX instrument calibrations to their current two - point technique.

     Effects of ambient test cell humidity may have contributed slightly
to the NOX correlation problem.  EPA should continue its efforts to improve
the consistency of its humidity controls and to bring the ambient condition
of its test cells nearer the 75 grains of moisture specification.

-------
                           TABLE 1
                       AIIDT CORRELATION
Vehicle
Audi - Fox 8442000011
Test
1972 Hot - Start
Test
Laboratory
Audi - NSU
3 Dynamometers
•3 Drivers

EPA
1 Dynamometer
1 Driver
Test
No.
1
2
3
AVE

1
2
• 3
AVE
HC
1.19
1.11
0.94
1.08

1.01
1.03
1.07
1.04
CO
13.89
14.41
11.81
13.37

11.25
11.23
11.16
11.21
NOx
3.66
3.33
3. 30
3.43

2.22
2.13
2.23
2.19
C02




•
346
343
344
345
Fuel
Consumption





24,2
24.3
24.3
24.3
Barometric
Pressure





28.78
28.80
28.84
28.81
NOx
Factor





0.8872
0.8792
0.9289
0.8984
   Emission Values in grams/mile
   Fuel Consumption  in miles/gallon
   Barometric Pressure in  inches  of  mercury

-------
    TABLE 2




Audi Correlation


Calibration
Gas Data
Facility Readings
Deviation from
Cyli
100
200
100
100
nder No.
098
178
030
053
Contents EPA Reading
£3^8 116.19 ppm
CO 1874.96 ppm
NO 122.7 ppm
C02 3.538 %
VW
114
1825
122
3.43
%
-1.
-2.
-0.
-3.

7
7
8
1
Audi
Train
111
1855
124.5
3.43
I
-4
-1
+1
-3
and
EPA
%
.3
,1
.5
.1
Percent
Reading
Audi
Tra i n
110
1862
125.5
3,47

II
-5
-0
+2
-2

%
.2
J
.3
.0

-------
APPENDIX

-------
                         LABORATORY CORRELATION



                              DATA SUMMARY



TEST FACILITY .  EPA #6	



A.  Dynamometer



    Manufacturer    Clayton	Model  CTE - 50	Absorber HP  50
    Torque Bridge   Strain Gage   Inertia Style  Direct Drive	Auto



    Loader  Not Used	Roll Diameter  8.65 in.     Roll Spacing  17.25 in.



    Calibration Procedure    Coast - downs at 4,8,  and 14 indicated	



     horsepower used to calculate actual absorbed horsepower vs.  indicated



     horsepower for each inertia weight.







B.  Driver's Aid



    Manufacturer   Varian	Control  Manual	Scaling  5 inches =



     60 mph, 4 inches/minute	Speed Signal Origin    Rear Roll	



C.  Ambient Monitors



    Facility Altitude  897.5 feet	



    Location of:



         Barometer  W&11 of Test Cell .	



         Humidity Sensors  1) Center of Cell - Ambient, 2) Dilution Air Inlet





    Nominal Test Cell Humidity   See Test Results	



D.  Analysis System



    1.  HC Analyzer



        Manufacturer    Beckman                 Model          400
                    H2 / N2                                A4
        Fuel Used                            Zero Gas      Alr

-------
    Ranges Used     50/100
    Span Gas (es)   41/89.5



    Curve Fit       Linear
    # Cal. Points     6



2.  CO Analyzers



    Manufacturer   Bendix
                                Hi   98.05
    Ranges Used   500/1000
    Span Gas (es)   427/923
    // Cal. Points_




    Curve Fit	




    Zero Gas
                               Hi
                   3 rd Order
    Manufacturer	Beckman




    Range Used	




    Span Gas
                        2500
                        2112
3.  NOX Analyzer




    Manufacturer  Teco




    Zero Gas        2
    Range (s) Used    10°




    Span Gas (es)	




    # Cal. Points
                        95.8
                               Hi
4.
    C02 Analyzer




    Manufacturer_




    Zero Gas
                     Beckman
    Ranges Used_
                       3.3%
    Span Gas (es)
                       2.5%
    # Cal. Points
                               Hi
                                      470
                                      91.2
     Lo  10'24
                                               Model  . -  85°
   Lo
                                                          94
                                              Model
                                                        315A
                                             Model
                                                        10A
                                            Curve Fit  Linear
 Lo
      8.5
                                          Model
   315A
                                          Curve Fit   3rd  Order
                                      3-33
Lo
        .78

-------
E.  CVS System
    Manufacturer  Air Monitoring, Inc.     Model     705-G
    Calibration Date  5/17/74
    Calibration Technique    Laminar Flow Element & Propane Injection.
    Results of Last 3 Propane Injections:
    Date 5/20/74	         5/21/74                   5/31/74
    % Error  -1.92	         0.326	      	0.946
     P        8.25               8.20                      5.10

-------
                         LABORATORY CORRELATION




                              DATA SUMMARY




TEST FACILITY Audi NSU Ingolstadt #3




A.  Dynamometer




    Manufacturer  Clayton          ModelEC 200 modified Absorber HP  50
    Torque BridgeSpring loaded coikiertia Style  mech - V - Belt	Auto




    Loader  none	Roll Diameter   8.65"      ROH Spacing   20.25"



    Calibration Procedure   Coast down at 4,  8 and 12 indicated hp used	




     to calculate actual hp vs.  indicated hp  for 2000,  2500 and 3000 Ibs
B.  Driver's Aid




    Manufacturer with°ff	Control manual       Scaling 12° mm=100 km/hr




    (63 mph),  2 mm/sec	Speed s±gnal Origin   rear ro11	




C.  Ambient Monitors




    Facility Altitude 12?6-3 feet	




    Location of:




         Barometer    Wal1 of Test Lab	




         Humidity Sensors  Test cel1 near dilution air inlet	







    Nominal Test Cell Humidity   50% ± 5	
D.  Analysis System   Train II




    1.  HC Analyzer




        Manufacturer   IPM                      Model    RS 5
        Fuel Used          2                 Zero Gas       Air

-------
    Ranges Used    20° PPm
    Span Gas (es)
                    94
    Curve Fit       Linear
    # Cal. Points      2	Hi      X  98	Lo	?1


2.  CO Analyzers


    Manufacturer     H & B                     Model _   URAS I]
    Ranges Used      75° PPm
    Span Gas (es)	


    # Cal. Points         8    Hi         745       Lo       74'5
    Curve Fit   Linearized
                   N
    Zero Gas        2
    Manufacturer     H & B                    Model   URAS II
    Range Used     3000 ppm
    Span Gas	1895
3.  NOX Analyzer


    Manufacturer       Teco                  Model      10A
    Zero Gas    -                           Curve Fit  linear
    Range (s) Used _ 100
    Span Gas (es) _ 97.5
    # Cal. Points      2	Hi       97.5      Lo       48.75



4.  C02 Analyzer


    Manufacturer    H & B                 Model   URAS 2T
    Zero Gas   N2                         Curve Fit  Linearized
    Ranges Used    5.^
    Span Gas (es)  3-65%
    // Cal. Points       8      Hi    5'0%        Lo    °'5%

-------
E.  CVS System




    Manufacturer    Audi	Model    Mk II	




    Calibration Date  last; 4-30-74




    Calibration Technique  Laminear flow element and propane injection
    Results of Last 3 Propane Injections:




    Date  8-9-73	         1-11-74	          4-30-74




    % Error    0.97              1.02	          0.97
     P         0.5"              0.5"                    0.5"

-------