UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
SUBJECT: Impact on Existing Standards Due to Proposed Instru- DATE: September 18, 1974
mentation Changes in the Heavy Duty FTP
F*ROM*
rKUl-u William B. Clemmens
^J/M^^-
TO
Gary W. Rossow
Part of the proposed interim changes in the heavy duty regulations
includes changes in instrumentation. This involves updating the method
of measuring unburned hydrocarbons (HC) for heavy duty gasoline engines
and utilizing a more accurate instrument for measuring oxides of nitrogen
(NOx) for both heavy duty gasoline and Diesel engines.
These new instruments measure quantities of constituents that the
present instruments may or may not measure due to their inherent limitations.
A close look at the correlation between old and new instruments is in order
such that the impact on the standards of this newer technology may be
assessed.
In terms of unburned hydrocarbons from heavy duty gasoline engines
the present NDIR analyzer will be replaced by a heated FID which is
similar to the unit used in the heavy duty Diesel FTP. This instrument
measures total carbon in ppm C of most forms of unburned hydrocarbon
emitted by an 1C engine, whereas the NDIR is more selective in the types
of hydrocarbons that it detects. For instance, the NDIR instrument will
not detect HC emissions of the Benzene ring type; and FID will. This
type of ..pollutant occurs in significant quantities in the exhaust of
gasoline engines and therefore one can expect the amount of HC emissions
measured by an FID to be greater than that measured by an NDIR analyzer.
This is indeed the case, and the current heavy duty FTP uses ari-.FID/NDIR
"volume correction ratio of 1.8 to account for the difference.
New data indicate that this factor underestimates the difference between
the two instruments. S.R. Krause 1 ran 168 tests following the FTP. Four
nine mode cycles were run on 7 gasoline engines with a resultant FID/NDIR
volume ratio of 2.30 and a standard deviation of 0.14. Data from South-
west Research 2 when corrected for the traditional^FID/NDIR volume ratio
of 1.8 have a mean FID/NDIR ratio of 2.56 and a standard deviation of .27
for 12 tests that were conducted following the 9-mode FTP on 9 different
gasoline engines. From this information, the impact of adding the FID
can be assessed by comparing the previous correction factor to the new
correction factor. Essentially this effectively removes the old correction
factor from the federal calculations and substitutes the new one. The
difference between the factors is 1.27 to 1.42. Leaning toward the
conservative side of the data spread, because of the fact that S.R. Krause
had more data, I recommend the correction on a mass basis (i.e. standards
are in mass) of 1.27. Thus substitution of an FID using the present FTP.
calculations for mass of pollutant will give a 27% higher value than an
NDIR hydrocarbon analyzer using the same FTP.
EPA fbirn 1320.6 (fev. 6-72)
-------
-2-
Accurate measurements of oxides of nitrogen emissions (NOx) have
always been difficult. The present NDIR analyzers used in both the heavy
duty gasoline and Diesel FTP measure only NO and not NO + N0£ (NOx).
The NDIR system is very susceptible to interference from water vapor, CO,
C02 and propane. In addition, the water vapor tends to condense inside
the instrument creating an equipment malfunction. Therefore, in order to
alleviate the probability of a malfunction, a drier is installed in the
sample line upstream of the instrument. However, if the dryer is a water
trap a sizeable portion of the N0£ is scrubbed out of the sample. If the
dryer is a dessicant like "Drierite", the dessicant reacts with the sample
to alter the NO readings.
More accuracy is available when using a chemiluminescence (CL) oxides
of nitrogen analyzer, mainly because it is not subject to the interferences
that the NDIR instrument is, and it can be made to measure total NOx with
the addition of a thermal conversion. The fact that the CL analyzer is
more specific in its detection indicates that the two types of instruments
produce different values.
o
S.R. Krause in a later paper J assesses the relationship on Diesel
engines with equation (1). This was the result of 6 engines tested on
a modified 13-mode cycle using SAE instrumentation techniques as required
by the FTP. A dryer (Aquasorb) upstream of the NO (NDIR) analyzer was
included. No mention was made of the procedure or hookup of the N02
instrument (NDUV), other than it was used.
A more thorough study was conducted by Scott Research Labs in conjunc-
tion with TRW Systems Group. The report ^ included test data from 450
automobiles with gasoline engines as part of the CAPE-13-68 program. All
tests utilized the 1975 FTP or the Federal Short Cycle test procedure.
The instrumentation (FIG 1) consisted of an NDIR (NO) analyzer with Drierite
and a downstream CL (NO or NOx) analyzer, both parallel to an NDUV (N0£)
analyzer. Parallel to this group of instruments was still another CL (NOx)
analyzer for cross correlation.
-------
-3-
The results indicate that equation (2) describes the relationship
between CL and NDIR with all interferences for the NDIR instrument
•i
present including Drierite. However, neither Heavy Duty Gasoline nor
Diesel measure NOx in the FTP. All that is measured is nitric oxide
(NO). Bag samples from reference ^ indicate the (NO) relationship in
equation (3) between a CL analyzer without drierite and an NDIR in-
strument with drierite.
Eq (1) (NOx) NDIR + NDUV = .994 (NOx) CL +75.6 ppm
Eq (2) (NOx) NDIR + NDUV = 1.34636 (NOx) CL - 13.43882 ppm
Eq (3) (NO) NDIR = 1.12184 (NO) CL + 8.36667 ppm
This last comparison is probably the most valid approach since it
compares directly to the instruments in use - an NDIR with drierite
measuring (NO). However, the expected use of the chemiluminescence
analyzer will be in the (NOx) mode. Remembering that (NOx) = (NO) +
(N02)> evaluating the relationship expressed in equation (3) provides
half of the answer. The relative change in response between NDIR and a
CL in the (NOx) mode should then be the change in response due to
measuring (NO) on the CL compared to the NDIR response plus the amount
of N02 in a typical sample.
In general the amount of N02 formed during a typical internal
combustion process can be characterized as 2-5% of the total NOx formed.
This remains fairly true for a raw sample of the combustion products.
Putting equation (3) into the following form
(4) (NO)^T =N°NDIR - 8.36667 ppm
CL 1.12184
and evaluating this equation (4) for a typical range of NO values which
are presently being measured by NDIR, Table I is constructed.
Table I
Relative
Response of
ppm (NO) Calculated CL compared to
by NDIR ppm NO from CL CL/NDIR (%) NDIR (%)
500 438.24 87.64 - 12.35%
1000 883.93 88.39 - 11.61%
2000 1775.33 88.76 -11.23%
Stating once again that NOx = NO + N02 it can also be said that
(5) ANOx = ANO + AN02
-------
-4-
Substituting the relative response of the CL analyzer from Table I as
the change in NO measurement (ANO) and the conservative value of the
previously unmeasured N02 as the change in NC>2 measurement (AN02), the
total effect of substituting a chemiluminescence analyzer operating in
the NOx mode, can be assessed.
ANOx = A NO +AN02
ANOx = -12% + 2%
The net change in NOx measurement is approximately a negative 10%.
The recommended adjustment factor for substitution of the chemiluminescence
system becomes .90 times a standard based on the existing heavy duty FTP.
Bibliography
1. "Effect of Engine Intake-Air Humidity, Temperature and Pressure on
Exhaust Emissions," S.R. Krause, Oct. 1971, SAE 710835
2. Letter from Southwest Research to J. Bascunana (EPA) dated Feb.
11, 1974.
3. "Effect of Inlet Air Humidity and Temperature on Diesel Exhaust
Emissions, " S.R. Krause, January 1973, SAE 730213.
4. "A study of Mandatory Engine Maintenance for Reducing Vehicle
Exhaust Emissions," Vol VI, year end report, Scott Research & TRW
Systems Group, July 1972.
-------
£}.= EQ//2
g^ EQ//3
•EQ//1
FIG 2
L_
500
PPM CL
Correlation Between CL and NDIR, NDUV
1000
------- |