EPA-AA-SDSB-30-22
Technical Report
Summary & Analysis of Comments to the Draft Recommended
Practice for Measurement of Gaseous and
Particulate Emissions from Heavy-Duty
Diesel Engines Under Transient Conditions
by
Daniel Reiser
November 1980
NOTICE
Technical Reports do not necessarily 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
technical information and to inform the public of technical devel-
opments which may form the basis for a final EPA decision, posi-
tion or regulatory action.
Standards Development and Support Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air, Noise and Radiation
U.S. Environmental Protection Agency
-------
-2-
EPA will soon publish a Notice of Proposed Rulemaking (NPRM)
for the control of heavy-duty diesel particulate emissions for 1985
and later model year engines. Because the early establishment of a
test procedure is essential for obtaining test data in response to
the proposal, a draft particulate test procedure, entitled "Draft
Recommended Practice for Measurement of Gaseous and Particulate
Emissions from Heavy-Duty Diesel Engines Under Transient Condi-
tions" was distributed to interested parties in May 1979. The
document was accompanied by a request for comments and suggested
modifications. Two heavy-duty diesel manufacturers, Cummins and
Caterpillar, submitted comments addressing the draft particulate
test procedure. A review of these comments and recommendations for
changes to the draft test procedure follows.* The comments have
been grouped into five general categories which are shown below.
In each category, quotations of the comments are presented followed
by analyses and recommendations.
I. Single vs. Double Dilution.
II. Equipment Specifications.
III. Temperature and Residence Time Requirements.
IV. Engine-Related Requirements.
V. Filter Weighing Procedure.
* Only those comments addressing particulate measurement will be
considered here. Comments addressing gaseous emission testing have
been analyzed elsewhere.
NOTE: All references referred to in this report are shown as /.
-------
-3-
I. Single vs. Double Dilution
A. Cummins
1. Comment (§86.1310-83(b)(2)(a)(i) and (ii))
"It is inconsistent to require 125°F maximum at the sampling
zone for single dilution while allowing 125°F maximum immedi-
ately ahead of the filter for double dilution. This allows
considerably more cooling by heat transfer for the double
dilution technique with the result that a lower overall
dilution is required."
2. Analysis
Both the single-dilution system and the double-dilution system
are designed to cool diesel exhaust by adding dilution air so that
the sampling zone temperature never exceeds 125°F. By cooling the
exhaust with dilution air, both the single- and double-dilution
processes better simulate actual environmental conditions than a
system which cooled the exhaust via convection (e.g., with a heat
exchanger). The two processes differ in that the single dilution
system cools the diesel exhaust with a one-step dilution process
while the double dilution system cools the diesel exhaust through a
two-step dilution process. Along with this cooling by dilution
air, both systems inevitably allow some cooling by convection
through the walls of the sampling system.
In reality, the double dilution system can allow more cooling
by convection than the single-dilution system due to the low flow
rates occurring in the particulate transfer tube and the secondary
dilution tunnel. It is generally believed that the equilibrium
between the gaseous hydrocarbons and those on the particulate is
affected by the method of cooling (specifically cooling by dilution
versus cooling via convective heat transfer), as well as by the
final absolute temperature of the diluted exhaust. Directionally
speaking, cooling the exhaust by convective heat transfer should
increase the amount of hydrocarbons associated with the particulate
compared to cooling by dilution. This would cause an increase in
the mass of particulate formed and measured.
The "Draft Recommended Practice" did attempt to limit the
degree of heat loss from the particulate transfer tube by re-
stricting its length to 35 inches. However, manufacturers can
design the length of the particulate transfer tube to be much
shorter than 35 inches to reduce the convective heat loss even
further. Convective heat loss can also be reduced by insulating
both the transfer tube and the secondary dilution tunnel. The use
of insulation is allowed under the existing requirements of the
"Draft Recommended Practice."
Tests of Mercedes-Benz and Peugeot light-duty diesels using
the heavy-duty measurement system have shown only very small
-------
-4-
differences in particulate measurements between single and double
dilution systems, when the sample temperature is 125°F or less
in both cases.I/ This shows at least in this case that the
extra convective cooling for the double dilution system is not
important and that use of the two-stage dilution process does not
affect results. This also appears to confirm that the maximum
sample zone temperature of 125°F is the controlling factor for
accurately measuring particulate emissions and that this tem-
perature should remain as a requirement in the proposed test
procedure.
In a related development discussed in Section II below,
it is recommended that the 2 second residence time requirement
for the secondary dilution tunnel be reduced to 0.25 second.
This change should allow a further reduction in the amount of
convective cooling occuring with double dilution and reduce
even further any differences seen between double and single
dilution. Thus, the 125°F temperature requirement should re-
main as the primary dilution controlling specification for both
systems.
B. Caterpillar
1. Comments (§86.1310-83(b)(1)(A))
"Has accurate measurement of gaseous emissions been demon-
strated with a diesel engine for the high-dilution ratios
necessary for the "single-dilution method" of measurement?
Because of this concern and to minimize equipment size, we
intend to use the "double-dilution method" of measurement."
2. Analysis
At the present time no test data are available on the measure-
ment of gaseous emissions using the single-dilution method for
heavy-duty diesel engines. However, concentrations of gaseous
emissions have been measured in the primary dilution tunnel of a
double dilution system.^/ If these primary dilution tunnel con-
centrations are simply divided by the dilution factor from primary
to secondary tunnel, then the resulting concentrations should be
those that would occur in a single dilution system for the same
test, as the overall dilution ratio should be the same in both
cases. Thus, the ability to measure gaseous emissions using the
single dilution method can be estimated from test results obtained
using the double dilution system.
The concentration of gaseous emissions in the primary dilution
tunnel of a double dilution system can be estimated from heavy-duty
diesel test data taken at Southwest Research. One example of such
test results is shown below for a hot start of a Cummins NTC-350
diesel engine.
-------
-5-
HC CO CO 2 NOx
Bag (ppm) (ppm) (percent) (ppm) Cycle
1 16 30 0.30 27.1 New York Non-Freeway
2 19 23 0.45 41.3 Los Angeles Non-Freeway
3 22 23 1.23 174.9 Los Angeles Freeway
4 17 18 0.29 24.9 New York Non-Freeway
The dilution ratio (dilution air/exhaust) in the secondary tunnel
is about 4:1. If the above test were performed under a single
dilution system, the resulting concentrations should then be about
one-fourth of these values. Thus, most of the above values for HC,
CO, and NOx would be in the 0-10 ppm range, and some values may
even be in the 0-5 ppm range for HC and CO. The C02 concen-
trations would also be very low, on the order of about 0.10 percent
(except for Bag #3, which would be about 0.30 percent). These HC,
C02, and NOx levels should not be a problem to measure accurately
with analyzers presently used for measuring these emissions.^/
These analyzers have the ability to measure several different
ranges of emission concentrations, with a few range selections
covering the low concentrations shown above. Also, background
levels, which are presently measured for these three pollutants,
are on the order of 0-5 ppm for HC and NOx, and 0-1.0 percent for
C02.
However, the CO concentrations discribed above would be
difficult to measure accurately with the instruments normally
used to measure CO concentrations. For example, background CO
levels are usually assumed to be zero because most analyzers
presently used cannot measure CO concentrations as low as ambient
levels. However, these low concentrations can be measured with
instruments of higher sensitivity. For example, commercially
available nondispersive infrared (NDIR) analyzers can detect CO
concentrations down to about 0.5 to 1.0 ppm^4/ The degree of this
sensitivity is directly proportional to the length of the cells,
the electronic amplification, and operating pressures.^/ The cost
of a system that measures CO concentrations as low as 0.5 to 1.0
ppm would be about $3000 more than a system used to measure minimum
CO concentrations of 10 ppm._3/ This extra cost is not large in
comparison to the overall cost of equipment needed for particulate
measurement.
Thus, low concentrations of HC, C02, and NOx emissions
can be measured accurately in a single dilution system with
analyzers currently used while low CO emissions can be measured
accurately if high sensitivity analyzers are used. Given this,
the single dilution method should still be considered an accept-
able method for measuring gaseous as well as particulate emis-
sions .
-------
-6-
II. Equipment Specifications
Tunnel Diameter
A. Cummins
1. Comments (§86.1310-83(b)(6)(ii))
"It is our understanding that the 18" primary tunnel diameter
is not a requirement. Cummins plans to use tunnels of 12 to
14 inches in diameter."
B. Caterpillar
1. Comments (§86.1310-83(b)(6)(ii))
"Why must the tunnel be "at least 18.0 inches in diameter?"
What is wrong with 12, 13, 14, etc. as long as turbulent flow
and complete mixing are achieved? We are using a 12 inch
diameter tunnel with our PDP-CVS with no apparent problem."
2. Analysis
A minimum primary dilution tunnel diameter of 18 inches is
specified in the "Draft Recommended Practice" and is a requirement.
The use of a tunnel diameter smaller than 18 inches would be
violating this requirement as it currently exists.
A minimum tunnel diameter and a distance from exhaust inlet to
particulate probe has been prescribed to insure that adequate
residence times and mixing occur for both the single and double
dilution tunnel systems. The question of a minimum tunnel diameter
will be examined here, while the question of tunnel length will be
examined in the discussion of the following comment.
In general, the dilution tunnel diameter should be specified
to insure that differences in tunnel diameter do not produce
differences in measured particulate emissions. Since the effective
tunnel diameter in-use is quite large, EPA's policy has been to
rely on the data taken using relatively large tunnel diameters.
Smaller tunnel diameters have only been allowed after their equiva-
lency with larger diameter tunnels has been demonstrated. To date,
no such work has been done on heavy-duty diesel testing systems.
Only the work performed on light-duty diesels is available.^/
There, the minimum tunnel diameter with demonstrated equivalency
was 8 inches.J>y Since the tunnel length specification is being
relied upon to provide adequate mixing, the residence time is the
parameter of interest here and should be used to scale this 8 inch
diameter up to a heavy-duty diesel system. For the light-duty
diesel tests in question, the CVS flow would have been around 600
cubic feet per minute. Assuming that the recommended tunnel length
of ten tunnel diameters was used, the residence time would then be
at least 0.23 seconds, or roughly a quarter of a second.
-------
-7-
The residence time required for measuring particulate emis-
sions from heavy-duty diesel (HDD) engines should then also be at
least 0.25 seconds. From this residence time, the minimum tunnel
diameter can be determined if the tunnel length and the flow rate
are known. Looking first at a HDD single dilution system, it is
again assumed that the tunnel length will be 10 tunnel diameters,
the same as that recommended in the LDD test procedure and the
"Draft Recommended Practice." The tunnel flow rate is assumed to
be 6000 CFM. The minimum tunnel diameter, then, must at least be
17.7 inches for a residence time of 0.25 seconds. A diameter of
18 inches would assure this proper residence time. This agrees
with the minimum tunnel diameter of 18 inches stated in the "Draft
Recommended Practice." Thus, this specification should remain for
the single dilution system.
For the double dilution system, the residence time has been
specified to be at least 2 seconds in the secondary dilution
tunnel alone. However, the need for this residence time should be
reexamined in light of the conclusion on residence time shown
above. Also, Cummins has expressed concern that a residence time
of 2 seconds or more in the second tunnel may lead to propor-
tionality problems due to time delays in the secondary tunnel. On
the basis of the light-duty data mentioned above, a residence time
of 0.25 seconds should be adequate for interaction between par-
ticulate matter and hydrocarbons. However, this residence time
should occur after the final dilution step to assure enough time
for particulate and hydrocarbon interaction at exhaust sampling
concentrations. For this reason, a residence time providing
adequate particle-gas interaction is not a concern in the primary
tunnel of the double dilution system.
As it turns out, the residence time for particle-gas inter-
action is the only consideration in the secondary tunnel. Mixing
is not important in the secondary tunnel since the filter covers
the whole area of the tunnel, collecting all particulate matter
flowing through this tunnel. Thus, a residence time of 0.25
seconds minimum in the secondary dilution tunnel should be rec-
ommended for the proposed test procedure.
As we have seen, the tunnel diameter (and length) of the
primary tunnel does not depend on a time necessary for particulate
and hydrocarbon interaction. However, the primary dilution tunnel
diameter must be large enough to allow for good mixing and to
accomodate both the incoming exhaust and dilution air flow rates,
so that high pressure drops do not occur along the tunnel. High
pressure drops may result in air leaks into the transfer tube and
unrepresentative conditions at the particulate sampling zone.
Also, the tunnel diameter should be large enough to provide ade-
quate space for the orifice plate and other equipment associated
with the primary dilution tunnel (such as sampling probes, transfer
tube, etc.). For accommodating raw exhaust only, the diameter of
the primary dilution tunnel should be at least the same as the
diameter of the tubing from the engine exhaust system to the
-------
-8-
entrance of the primary dilution tunnel. This diameter is usually
5-6 inches. The primary dilution tunnel diameter should then be at
least 8 inches to prevent high pressure drops and to provide space
for the orifice plate and other equipment. More importantly, a
primary dilution tunnel of 8 inches has been the smallest diameter
used for testing light-duty diesels,^/ where accurate measure-
ments have been demonstrated. A smaller absolute diameter should
not be allowed until equivalency has been demonstrated with
smaller diameters.
An initial attempt has been made to look closer into the
necessary tunnel diameter (and length) for good-mixing between
particulate emissions and surrounding dilution air. A model was
developed to describe the radial diffusion of particulate matter in
a tunnel with constant turbulent bulk flow in the axial direction.
The model consists of the following equation:
_, 32C Ed 3C 3C
Ed — + =
3r2 r 3r 3t
(1)
where:
Ed = eddy diffusivity
C = concentration of particulate in dilution air
r = radius at any point in the tunnel
t = space time, or Al/v
A = area of tunnel, tfr22, 1= tunnel length, v= average
gas velocity in tunnel
Initial conditions and boundary conditions must be specified
so that the above equation can be solved. The following initial
condition was assumed:
C(r, t = 0) = C(r) for 0 <^ r <^ r2 (2)
C(r) is further described by:
C = Co at 0 <_ r <_ ri, t = 0
C = 0 at TI <^ r jC r2, t = 0
where:
TI = exhaust pipe radius
*2 = tunnel radius
The following boundary condition was specified:
j *-»
-j— = 0 for all t at r = r2
The rigorous details needed to obtain the final solution will
not be given here. The final solution was obtained by using the
separation of variables technique along with a Bessel function
solution.6/ The formal solution to this boundary value problem
is:
-------
-9-
r2
C(r, t) = — " J° (ot(J)r) exp(-o(j)2 • Ed • t) / sf (s)Jo( a( j )s)ds (3)
0
The Bessel function Jn(x) can be found in typical numerical
tables of Bessel functions «2/ The numbers a(j) are square
roots of the eigenvalues associated with the eigenfunctions of
equation(3). The variable s is a variable of integration.
A significant problem with the solution to the above problem
is that it neglects the effect of the orifice plate at the tunnel
entrance. Such a mixing orifice is allowed as a mixing enhancer.
The effect of the orifice should be large enough to void the
results of any model which ignored its effect. Thus, further work
is needed to incorporate the effect of the orifice before analyti-
cal results can replace the empirical.
In summary, the dilution tunnel for a HDD single dilution
system should have a diameter of 18 inches minimum as specified in
the "Draft Recommended Practice." For the primary tunnel of the
double dilution system, the diameter should be at least 8 inches.
Also, the required residence time for the secondary tunnel should
be reduced from 2 to 0.25 seconds.
Tunnel Length
A. Caterpillar
1. Comments ( §86. 1310-83(b) (8) (b)(i)(A)( ))
"Why must particulate probe be "approximately 10 tunnel
diameters downstream of the point of where the exhaust enters
the primary-dilution tunnel1 if good mixing has been achieved?
The tunnel length could be quite short if mixing vanes and a
smaller tunnel diameter were allowed. Centrifuging of the
particulate is unlikely because of its very small size.
2. Analysis
The proposed tunnel length is necessary for two reasons.
First, it provides adequate time for gaseous phase hydrocarbons to
come to equilibrium with particulate matter before sampling takes
place in the case of single dilution. Second, sufficient distance
is needed for thorough mixing of exhaust and dilution air. Even if
interaction between particulate and gaseous emissions were not
important, mixing vanes could not be used to achieve quick mixing
since it has been shown that a measureable decrease in particulate
emissions from diesels can occur due to particulate deposition on
mixing exhancers, such as baffles. 8/ Thus, the use of mixing vanes
should still be prohibited.
-------
-10-
Although mixing vanes should be prohibited, the "Draft Rec-
ommended Practice" does allow the use of a mixing orifice at the
engine exhaust inlet to the primary dilution tunnel (see Figures
83-3 and 83-4). This orifice causes the exhaust and dilution air
flow to become more turbulent, thus enhancing the mixing. Par-
ticulate deposition should not occur with the mixing orifice as the
exhaust inlet plane is located at the mixing orifice plane or
just downstream of the plane. The mixing orifice is the only
mixing enhancer allowed in the "Draft Recommended Practice".
The mixing length of 10 diameters is generally accepted as a
length for good mixing in a tunnel with turbulent flow. In some
cases, mixing may also be achieved by using a shorter tunnel length
than the recommended 10 tunnel diameters. At present, data are not
available showing the extent of mixing for tunnel lengths less
than 10 tunnel diameters. If good mixing is demonstrated with data
(i.e., tunnel transverse study), then the acceptance of a shorter
mixing length would be considered. Any reconsideration of length
would also have to take into account the effects on residence time
in the case of single-dilution (see previous comment, Tunnel
Diameter). In the absence of the data, the recommended tunnel
length should remain at 10 tunnel diameters.
Transfer Tube
A. Caterpillar
1. Comments (§86.1310-83(b)(8)(b)(i))
"(D) The available literature supports a length greater than
35 inches for a 0.5 inch inside diameter line. Besides,
shouldn't the length be a function of the inside diameter of
the line being used since only a minimum diameter is specified
in the preceding paragraph? (E) An example of a "sharp bend1
would be helpful."
2. Analysis
The "Draft Recommended Practice" states that the particulate
transfer tube should not be longer than 35 inches^. A length
greater than 35 inches will allow additional cooling by convection
(see Cummins' comment in Section I, Single vs. Double Dilution).
This may cause additional interaction of gaseous hydrocarbons on
the particulate surface which could result in a higher particulate
measurement.
The diameter of the transfer tube is also required to be at
least 0.5 inches. The use of a diameter less than 0.5 inches is
prohibited as this could cause high pressure drops along the
transfer tube which could result in a great variation of the
transfer tube flow rate as a function of temperature. Also, the
use of a smaller diameter could bring about particulate deposition
on the tube walls. While the use of a larger diameter is allowed,
-------
-11-
th is could also result in more cooling by increasing the resi-
dence time in the transfer tube, assuming that the flow rate
would remain constant. If a larger diameter were coupled with
a longer transfer tube, the residence time would increase even
further. More heat would be lost by convection and the possi-
bility of increased particulate readings would be much greater
than increasing either the diameter or the length alone. Thus,
the maximum length of the particulate transfer tube is at least
necessary to help prevent a possible increase in paritculate
readings.
It is not known what available literature supports a length
greater than 35 inches for a 0.5-inch inside diameter. Thus,
the particulate transfer tube length of 35 inches and diameter of
0.5-inch minimum should remain as a requirement unless supporting
data show otherwise.
The requirement, "Free of sharp bends," does appear to
be difficult to define and should therefore be deleted. This
requirement was made so that particulate deposition would be
minimized. The requirement should be revised to read, "Designed
to minimize deposition of particulate." For example, this may
mean making bends as gradual as possible and eliminating obstruc-
tions such as sensors. This requirement should be added to the
proposed test procedure description.
Heat Exchanger
A. Caterpillar
1. Comment (§86.1310-83(b)(8)(b)(iv)(B) and (viii))
"We assume that these gas temperatures can also be controlled
with standard type heat exchangers as long as they are
not located in the mixture stream ahead of the particulate
filter. This would be like the SwRI secondary dilution
system."
2. Analysis
That is correct.
Microgram Balance
A. Cummins
1. Comment (§86.1312-83(b))
"A 100mm Pallflex filter weighs about 300mg. If we expect a
lOmg loading, then O.Olmg weighting accuracy should be suf-
ficient. Microgram readability is not required."
-------
-12-
B. Caterpillar
1. Comments (§86 . 1312-83(b) )
"Measurement capability to one microgram is more than is
needed. A 10 microgram capability would be more than ade-
quate. Also, can the readability and precision both be one
microgram?"
C. Analysis
Before analyzing the necessary measurement capability of the
weighing balance, the terms readability, accuracy, and precision
should be clarified as their meanings seem to be confusing in these
comments and in the "Draft Recommended Practice." Readability is
defined as the closeness with which the scale of the instrument may
be read.9j Accuracy is defined as the deviation of the reading
from a known output, usually expressed as a percentage of full
scale reading. 9/ Precision is defined as the ability to produce a
certain reading within a given accuracy,^/ or the closeness of
repeated measurements to one another for measurements of the same
quant ity._10/ An example of the distinction between accuracy and
precision can be shown for a particulate sample that has a true
weighing of say 100 micrograms, but the microgram balance shows
weighings of 102, 103, and 104 micrograms in three different
weighings. From these values it can be seen that the weighing
balance could not be depended on for an accuracy better than
+ b percent (4 micrograms) while a precision of +^1 percent is
indicated since the maximum deviation from the mean is only 1
microgram. This example also shows that the accuracy can be
improved to but not beyond the precision of the instrument of
calibration.
The "Draft Recommended Practice" states that the microgram
balance should have a precision (standard deviation) of one micro-
gram. The term "precision" should be revised to "accuracy"
because the deviation from the true weighing of a particulate
sample is of primary concern here. The above definitions and
example show that if precision only is specified, then the accuracy
may be poor or even unknown. Thus, the term "precision," as stated
in the "Draft Recommended Practice," should be replaced with
"accuracy" in the proposed test procedure.
The necessary measurement capability, or accuracy, of the
weighing balance can be determined through an uncertainty anal-
ysis.^/ The general equation for measuring uncertainty of an
experiment can be expressed as follows:
WR = [(4-.W1) + (JJ|- W2> + ... + (JL^Wn)] (1)
3X1 + 9X2 3xn
-------
-13-
Where:
X£ = input variables.
W£ = uncertainty in the input variable, Xj[.
R = the final measurement, a function of the input variables
KI, X2, X3, ..., Xn.
WR = uncertainty in the result, R.
This uncertainty analysis can be applied to the weighing of
diesel particulate test samples. This can be accomplished in three
major steps. First, the total weight of the particulate as a
function of individual measurements must be determined. Second,
the uncertainty of each independent variable must be established.
Third, the particulate weight function and the uncertainties of
each independent variables can be substituted into equation (1) to
determine the uncertainty of the total sample weight. After these
steps have been performed, the necessary accuracy of the balance
can be discussed.
The weight of a particulate sample can be determined from the
following equation:
P = (flp - flc) + (f2p - f2c) (2)
Where:
P = total particulate weight.
f}p = weight of primary filter plus particulate.
^Ic = weight of primary filter.
f2p = weight of back-up filter plus particulate.
f2c = weight of back-up filter.
The uncertainties for each of the four weighings flp, flc»
f2p, and f2c are ej^, e^c, e2p, and e2c, respectively. These
uncertainties are due to weighing balance error only. The errors
due to dust, humidity, faulty handling, etc., are not included.
If equation (2) is substituted into equation (1), the partial
derivatives of P with respect to fjp and P with respect to f2p
become 1. The partial derivatives of P with respect to f^c and P
with respect to f2c are equal to -1. Thus, equation (1) would be
revised to:
WR = [(elp)2 + (-elc)2 + (e2p)2 + (_e2c)2]l/2 (3)
If the uncertainty in each weight due to the weighing balance
accuracy is 10 micrograms, as suggested by Caterpillar and Cum-
mins, then equation (3) would yield the square root of 400, or
20 micrograms. Thus, for weighing balance errors only, the uncer-
tainty of any particulate weighing is 20 micrograms. Since other
-------
-14-
errors are involved with the weighing of particulate, such as
faulty handling of the filter or contamination from dust, the
uncertainty of each particulate sample weight due to the weighing
procedure alone is greater than 20 micrograms.
In addition to the accuracy of the weighing procedure, the
overall accuracy of transient test procedure results for parti-
culate measurement is affected by many other factors. For example,
accurate measurements of temperature and flow rates are necessary
to assure that proportional sampling occurs during a transient test
cycle. Because these and other factors contribute to overall
accuracy, good measurements are required for each of these factors.
For the weighing balance, it should be reasonable to require an
accuracy of _+_! percent of the particulate sample weight when
considering the other factors involved with determining the overall
accuracy.
An accuracy of +_! percent for the weighing balance would
require the most sensitive detection for filter loadings of engines
with low particulate emissions. Low filter loadings can occur with
smaller engines, such as the IHC DTI-466B engine. Present South-
west Research testing data show that this engine emits about 4.33
grams of particulate during a hot or cold start cycle.J/ With a
particulate transfer tube flow rate of about 1.6 CFM and an
overall flow rate of 2000 CFM*1J_/, the filter loading for this
engine is currently about 3.5 mg. A +^1 percent change of this
filter loading would require that the uncertainty in each weighing
is no worse than +_ 35 micrograms. A weighing balance with an
accuracy of 10 micrograms and consequently an uncertainty of +^ 20
micrograms should provide the necessary accuracy for this partic-
ulate measurement. It is also likely that a weighing balance with
an accuracy of ^10 micrograms could be used to accurately measure
particulate emissions from other engines that are currently
marketed.
Future reductions could easily bring the filter loadings for
the IHC DTI-466B engine down to 2 mg or less. A 1 percent change
of this filter loading would mean that the accuracy must be better
than +20 micrograms. However, it has been shown above that the
uncertainty of a particulate sample weight is greater than 4^20
micrograms if the weighing balance has an accuracy of 10 micro-
grams. Therefore, the accuracy of the weighing balance should be
better than 10 micrograms.
The recommended heavy-duty diesel test procedure also states
that if a change of more than +_! percent of the nominal filter
loading occurs in the weight of the reference filter during the
conditioning period, then all filters in the process of being
stabilized must be discarded and any test repeated. In the case of
the IHC DTI-466B engine this nominal loading could be 2 milligrams
with future reductions in particulate emissions. Once again, a 1
* For this engine, Southwest Research used a flow rate of about
2000 CFM. This is not to be confused with the assumed flow rate of
3000 CFM used in other sections of this report.
-------
-15-
percent change of this filter loading would mean that the accuracy
must be better than +20 micrograms during the conditioning period.
The change of weight in the reference filter is determined by two
weighings, one for the weight of the filter before and one for the
weight of the filter after the conditioning period. If the weigh-
ing balance has an accuracy of 10 micrograms, then the overall
uncertainty due to the weighing balance alone would be the square
root of 200, or 14 micrograms (see equation 3 above). This alone
would provide the necessary accuracy for the weighings of the
reference filter. However, as mentioned above, many other sources
of errors are involved during the weighing period, such as faulty
handling or changes in humidity. These other sources of errors
could decrease the overall accuracy of the weighing procedure from
+ 14 micrograms to over +_20 micrograms. Therefore, an accuracy of 10
micrograms for the weighing balance may not provide a safe enough
margin for accurate weighings of the reference filter and this may
result in many needless filter rejections.
Accuracy is at best equal to readability, and information on
microgram balances shows that nearly all microgram balances
currently sold have an accuracy equal to readability.12/13/14/15/
16/ For this analysis, it will be assumed that accuracy and
readability are the same. For engines with no particulate control,
a balance with an accuracy and readability of 10 micrograms should
be adequate. However, for future engines with particulate control,
both the accuracy and the readability should be better than 10
micrograms. Present literature on microgram balances does not show
balances having a readability between 1 microgram and 10 micro-
grams •JJ^/J^3_/14VJ_5/ For example, a 5 microgram readability and
accuracy may be sufficient in many cases for engines with par-
ticulate control, but available information only shows balances
with readability and accuracy either to the nearest 1 microgram or
to the nearest 10 micrograms. A one microgram accuracy and read-
ability is thus necessary for future engines with particulate
control to assure proper measurements and this requirement should
be included in the proposed test procedure.
The cost of a microgram balance with an accuracy and reada-
bility of 1 microgram is approximately $1500 more than a balance
with an accuracy and readability of 10 ug.16/ This cost should be
small compared to the overall cost of the test equipment particu-
larly considering that one balance can service a number of test
cells.
Air Filter - Cyclonic Separator
A. Cummins
1. Comments (§86.1310-83, Figure N83-3 and 4)
"The schematic drawing in Figure N83-4 shows a dilution air
filter which appears to be constructed with three different
types of elements.
-------
-16-
Because of the very large size of a filter system anticipated
to be required for handling 10,000 SCFM, it is necessary to
have a more explicit description of dilution air filtration
requirements. The amount and type of filtration required will
determine size, cost, facility requirements, etc.
The use of a cyclonic separator is also specificed presumably
for particulate removal. This concept is effective where
relatively large particles such as catalyst pellets, etc., are
concerned but will not be effective in removing the submicron
particulates found in diesel exhaust. It is recommended that
the requirement for the cyclonic separator be deleted for
diesel engine testing."
2. Analysis
The three divisions of the air filter in Figure 83-4 have no
significance. Thus, the illustration of the air filter in the
proposed test procedure should not show these divisions.
In general, an air filter is illustrated in Figure 83-4 to
show the need to filter ambient air before it enters the primary
dilution tunnel. If the air inlet is left unfiltered, suspended
particulate could enter and cause an increase in the particulate
measurement of the diesel engine exhaust. The size of this in-
crease can be estimated if the particulate concentration of the
dilution air is assumed to be the National Ambient Air Quality
Standard of 75 micrograms per cubic meter. For a double dilution
system with a primary dilution tunnel ratio similar to that used by
Southwest Research, or about 4:1, a CVS flow rate assumed to be
3000 CFM, and a secondary dilution ratio of 3:1 the equivalent
weight of the particulate originating from dilution air is about
0.38 gtn for a 20 minute hot or cold start cycle. The suspended
particulate in the dilution air of the secondary tunnel may
represent as much as 75 percent of this effect. This particulate
weight should have the greatest effect, in terms of grams per brake
horsepower-hour (gm/BHP-hr), on the emission test results for
engines performing the least amount of work over a transient test
cycle. For example, the IHC DTI-466B only produces 12.04 BHP-hr
over the transient cycle.jZy The emission results from tests on
this engine could increase by as much as 0.03 gm/BHP-hr with an
equivalent amount of approximately 0.02 gm/BHP-hr originating from
the dilution air of the secondary tunnel, if the ambient air is
left unfiltered with the above conditions. This contribution to
particulate measurements may be enough to offset emission results
which are usually recorded to the nearest 0.01 gm/BHP-hr. Any
filter, of course, would reduce this effect dramatically.
While the "Draft Recommended Practice" only required filtra-
tion of the primary dilution air, it is apparent from this analysis
that the secondary dilution air should require filtering also,
especially when considering that approximately 75 percent of the
contribution of background particulate to the particulate measure-
-------
-17-
ment originates from the secondary dilution air. For the proposed
test procedure a filter should then be required for the dilution
air entrance of the secondary tunnel.
As manufacturers would be the prime beneficiary of good filter
design and they are the most knowledgeable about their individual
systems, the actual size and design of the filters has been left up
to the manufacturers. In this way, each manufacturer can choose a
filtration system that he believes is the most cost-effective.
The cyclonic separator, also diagrammed in Figure 83-4, was
based on the proposed test procedure for light-duty diesel vehi-
cles. A comment was submitted that addressed the need for the
cyclonic separator in the light-duty diesel proposed test proce-
dure, and it was concluded that cyclonic separators should be
necessary for catalyst vehicles only.l_/ The cyclonic separator was
made optional for light-duty diesel vehicles, and the same should
be done here in Figure 83-4 for heavy-duty diesel engines.
III. Temperature Residence Time Specifications
Primary-Dilution Air Temperature
A. Caterpillar
1. Comments (§86.1310-83(b)(5)
"Has the need to control primary-dilution air temperature
so closely been demonstrated? This will require special
conditioning of the dilution air, but in our case, a tempera-
ture range of 60°F to 100°F would not require special condi-
tioning."
2. Analysis
The recommended temperature range of 77+9°F (25^5°C) for the
dilution air entering the primary dilution tunnel was based on the
dilution air temperature required for the light-duty diesel test
procedure.^/ The use of a dilution air inlet temperature outside
of this range may affect particulate and gaseous hydrocarbon
interactions, which may affect particulate measurements. At the
present time no data are available showing the effects of the inlet
dilution air temperature outside of this range. If data are
submitted showing no effect, then the required temperature range
could be made larger. However, until data are submitted demon-
strating that there is no effect, the required dilution air inlet
temperature should remain at 77j^9°F as stated in the "Draft Recom-
mended Practice."
-------
-18-
Water Temperature
A. Cummins
1. Comments (§86.1310-83(b)(2)(b))
"Is there a requirement as to the maximum or minimum allowable
water temperatures used for heating or cooling the heat
exchanger? Can steam or 180°F water be used when the heat
exchanger is being used in the heating modes?"
2. Analysis
There is no specification for water temperature. Both steam or
180°F water can be used.
Residence Time
A. Caterpillar
1. Comments (§86.1310-83(b)(8)(b)(iii)(B))
"Shouldn't this read ' two seconds minimum '?"
2. Analysis
The original intent of the "Draft Recommended Practice" was to
read ". . . two seconds minimum. . . ." However, based on the
analysis of tunnel diameter in Section II, Equipment Specifica-
tions, the residence time should be revised to 0.25 seconds in the
secondary dilution tunnel. This change should be made in the
proposed test procedure.
IV. Engine-Related Requirements
Exhaust System Length and Diameter
A. Cummins
1. Comments (§86.1310-83(b)(3) and §86.1308-83(b)(3)(i)(A)
"Cummins in-use minimum exhaust system length is approximately
12 feet. In order to 'share' one CVS system between two
adjacent test cells, we would require an additional 25 feet
of insulated stainless steel tubing to reach the CVS system
inlet. We recommend that EPA specify a maximum length
from engine to CVS system of 35 feet and allow manufacturers
to choose the portion designated 'stock' exhaust system
and make up the remaining length with stainless steel in-
sulated tubing. We also recommend that the maximum inside
diameter of the stainless steel tubing be raised from 5" to 6"
because a 600 hp engine may not be able to meet typical in-use
-------
-19-
full load exhaust restriction requirements with a 5" diameter
tube.
B. Caterpillar
1. Comments (86.1308-83(b)(3)(i)(A)
"At original certification, the engine manufacturer may not
know all of the "in-use applications of the engine" in order to
establish minimum exhaust system length. The EPA should specify
range for acceptable exhaust system length such as 15 +_5 ft.
C. Analysis
The exhaust system length should be kept at a minimum to
prevent convective cooling and particle deposition, as this would
tend to artificially lower particulate measurements. Based on
Cummins' comment, a 12 foot exhaust system length would appear to
be a reasonable minimum length. Thus, since the minimum in-use
exhaust system length is difficult to determine for each engine, an
exhaust system length of 12 feet maximum should be recommended in
the proposed test procedure.
The length of the tubing from the exhaust system to the
entrance of the primary dilution tunnel must also be restricted
to prevent excessive particulate deposition. The maximum length
of 12 feet specified in the "Draft Recommended Practice" was
based on the light-duty diesel (LDD) proposed test procedure. In
the LDD Summary and Analysis of Comments, General Motors also
commented on the length of the tailpipe .J_/ Since that time, we
have run data justifying a longer pipe length of 20 feet, if it is
smooth and insulated. 17/ Otherwise a maximum length of 12 feet
still applies if the pipe is uninsulated. Thus, the tubing length
from the exhaust system to the dilution tunnel in the proposed
heavy-duty diesel test procedure should require a length of not
more than 12 feet (365 cm) if uninsulated, and of not more than 20
feet (610 cm) if insulated and composed of smooth stainless steel
tubing. Use of a longer length could be considered if further data
were presented.
To arrive at an appropriate pipe diameter for accomodating
the exhaust flows of heavy-duty diesel (HDD) vehicles, the HDD
flow rate will be compared to the LDD flow rate. The maximum
diameter of the tubing from exhaust to tunnel entrance is 4
inches in the LDD test procedure. An EPA study shows that a
larger diameter may cause an increase in measurement of partic-
ulate emissions .VTj However, a larger diameter for heavy-duty
testing is necessary to meet the full load exhaust requirements
of some heavy-duty diesel engines. The largest light-duty diesel
vehicle has an engine displacement that is about one-third to
one-half the size of engines used to power the heavier Class
VII and Class VIII vehicles. The rated engine speed for a light-
duty diesel vehicle is also approximately 50 percent higher
-------
-20-
than the engine speed for Class VII and Class VIII heavy-duty
diesel vehicles.J^/ Thus, on the basis of engine displacement
and engine speed, the exhaust flow rate should be roughly twice
that for the larger heavy-duty diesels than for light-duty die-
sels. In addition, the predominant use of turbochargers on
the larger heavy-duty diesel engines would increase the exhaust
flow to even more than twice that of light-duty diesel vehicles.
Cummins' request for a 6 inch exhaust pipe would be a 50 percent
increase over the light-duty pipe in terms of diameter and a 125
percent increase in terms of cross sectional area. Given that the
exhaust flow of heavy-duty diesels is more than twice that of
light-duty diesels, an increase to 6 inches would only take into
account the increased exhaust flow and no more. Thus, the request
should be accepted and the specification changed to a maximum of
6 inches. A diameter larger than 6 inches would be considered in
the future if sufficient data would show no effect on particulate
measurement.
Exhaust Back Pressure
A. Caterpillar (86.1308-83(b)(3)(ii)(B)
1. Comment
"Again, at original certification, the engine manufacturer may
not be able to identify "the maximum back pressure application
of the engine,1 but the manufacturer's recommended maximum
exhaust back pressure limit for the engine would be speci-
fied. The EPA should allow for this possibility."
2. Analysis
It is acknowledged that because of the many in-use applica-
tions of a heavy-duty diesel engine, it may be difficult to
identify the maximum back pressure application of the engine.
Therefore, the manufacturer's recommended maximum exhaust back
pressure limit for the engine should be acceptable for the partic-
ulate test procedure.
V. Filter Weighing Procedure
Post Test Filter Stabilization Period
A. Caterpillar
1. Comments (86.1339-83(a)(b) and (e)
"The upper limit of 56 hours is unreasonable. For any test
that is run late on any Friday, a person will have to work on
the weekend to weigh the filters. Also, new filters weighed
on a Friday could not be used for a Monday morning test. A
new filter weighed early on Friday should be usable anytime
(during normal working hours) on Monday. Also, a particulate
-------
-21-
sample obtained early on Friday should be allowed to be
weighed on Monday. An upper limit of 80 hours should be
adequate."
2. Analysis
The upper limit of 56 hours for the post test filter stabili-
zation period was based on the upper limit specified for the
light-duty diesel (LDD) test procedure. This conclusion from the
LDD test procedure came from allowing Saturday testing to be
weighed on Monday. Some facilities in the heavy-duty diesel engine
industry may not be able to have Friday test results weighed before
Monday morning. The upper limit of 56 hours should be increased to
80 hours to cover the additional time needed. This additional time
should not affect filter stabilization. It is also possible that
new filter weighings on Friday cannot be used until normal working
hours on Monday, and that particulate samples received on late
Friday cannot be weighed until Monday. Thus, an upper limit of 80
hours is recommended for these weighings also.
-------
-22-
References
_!_/ "Summary and Analysis of Comments on the Nature of Proposed
Rulemaking for the Control of Light-Duty Diesel Particulate
Emissions from 1981 and Later Model Year Vehicles," SDSB,
EPA, October 1979.
2j Southwest Research Institute Diesel Baseline Emissions Sum-
mary, EPA, June 1, 1980.
J3/ Telephone conversation with Brad Levin of Beckman Instruments,
Inc., Automotive Test Instruments Operations, August 20,
1980 and September 28, 1980.
4/ "Air Quality Criteria for Carbon Monoxide," U.S. Dept. of
Health, Education, and Welfare, No. AP-62, March 1970, p.
5-3.
5J "Standard for Emission of Particulate Regulation for Diesel-
Fueled Light-Duty Vehicles and Light-Duty Trucks," FR Vol.
45, No. 45, March 5, 1980, pp. 14496-14525.
6_/ Churchill, R.V., and Brown, J.W., Fourier Series and Boundary
~~ Value Problem, pp. 207-210, 3rd Ed., McGraw HTTl, 1978.
T_l Gray, A., and G.B. Mattews: A^ Treatise on Bessel Functions
and their Applications to Physics, 2nd Ed., Dover Publica-
tions, Inc., 1966.
B/ Black, Frank, "Comments on Recommended Practice for Measure-
ment of Gaseous and Particulate Emissions from Light-Duty
Diesel Vehicles," ORD, EPA, April 13, 1978.
_9_/ Holman, J.P., Experimental Methods for Engineers, McGraw-
Hill, Inc., Second Edition, 1971, pp. 37-38"!
10/ Zar, Jerold H., Biostatistical Analysis, Preatice-Hal1,
Inc., 1974, pp. 4.
ll/ Telephone conversation with Sherrill Martin of Southwest
Research on July 17, 1980.
12/ "Scientech, Inc., Electronic Top-Loading Balance," Scientech,
Inc., Series 222, March 1978.
13/ "Cahn Series-20 Automatic Electrobalance," Cahn Ventron
Corp., 1978.
-------
-23-
References (cont'd)
14/ "Sartorius Balances and Scales," Brinkman Instruments, Inc.
15/ "Mettler Balances," Mettler Instrument Corporation, 1978.
16/ Telephone conversation with Ray Giles of Mettler Instruments
Corporation, Inc., August 25, 1980.
17/ Penniga, T., "Evaluation of Exhaust Collection Configuration
on Diesel Particulate Measurement," TAEB, EPA, May, 1979.
18/ Automotive Industries, April 1978, pp. 79-85, 94-99.
------- |