United States
Environmental Protection
Agency
National Risk Management
Research Laboratory
Cincinnati, OH 45268
Research and Development
EPA/600/SR-01/079 November 2001
Project Summary
Heavy Duty Diesel Fine
Particulate Matter Emissions:
Development and Application of
On-Road Measurement
Capabilities
J. Edward Brown
EPA's On-road Diesel Emissions Char-
acterization Facility, which has been
collecting real-world gaseous emissions
data for the past 6 years, has recently
undergone extensive modifications to
enhance the facility's participate mat-
ter (PM) measurement capabilities, with
a specific emphasis on fine PM or PM25
(particles less than 2.5 \un in aerody-
namic diameter). At present, the
facility's capabilities are focused on
continuous sampling and analysis, us-
ing fast-responding instruments such
as the Electrical Low-Pressure Impac-
tor (ELPI), the Tapered-Element Oscil-
lating Microbalance (TEOM), and a
particle-bound Polycyclic Aromatic Hy-
drocarbon (PAH) analyzer, all of which
require a dilute exhaust sample. This
dilute sample has been drawn directly
from the vehicle exhaust via a stack
dilution system, and sampled from the
ambient exhaust plume via probes in
the trailer. Dilute samples have also
been collected on filters for chemical
and gravimetric analysis. Experimental
results indicate that stack dilution sam-
pling does not adequately represent
real-world conditions as determined
from initial plume sampling. Therefore,
future efforts will be directed toward
improved plume characterization tech-
niques.
This Project Summary was developed
by the National Risk Management Re-
search Laboratory's Air Pollution Pre-
vention and Control Division, Research
Triangle Park, NC, to announce key find-
ings of the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
Because of the current level of interest
in fine particulate matter and its health
effects, EPA has refocused a substantial
amount of its research to study emissions
sources that produce fine PM. Diesel en-
gines, already under substantial EPA scru-
tiny for their nitrogen oxides (NOX)
emissions, are also known to emit large
quantities of small particles, although it is
not known how much of the fine PM in
ambient air actually comes from diesel
engines, especially those used in heavy
duty diesel vehicles (HDDVs).
A number of researchers have col-
lected diesel fine PM data, mostly using
engines mounted on dynamometers and,
to a lesser extent, chassis dynamometer
facilities. These facilities allow for the col-
lection of data under very controlled, re-
peatable conditions. Many of these
conditions are of a steady-state nature,
where the emissions are allowed to sta-
bilize before data or samples are col-
lected. Though steady-state tests prove
useful for producing repeatable results
and comparisons, there is no consensus
on how well steady-state tests represent
real-world emissions. More transient tests,
-------
on the other hand, typically suffer from
poor measurement repeatability. Fine PM
measurements in particular are problem-
atic because many of the most sensitive
instruments cannot follow such rapidly
changing conditions. Nonetheless, it is
likely that it will take a combination of
steady-state and transient tests to fully char-
acterize fine PM emissions from HDDVs.
The Air Pollution Prevention and Con-
trol Division of EPA's Office of Research
and Development has developed its on-
road approach as sort of a reality check
for HDDV emissions estimates. By replac-
ing assumptions with measurements, and
simulation parameters with real-world op-
erating conditions, the On-road Diesel
Emissions Characterization (ODEC) facil-
ity provides another dimension to the data
currently available for quantifying and
characterizing HDDV emissions. Fine PM
measurement capabilities have recently
been added as an extension of this origi-
nal purpose. The report describes the
facility's fine PM measurement capabili-
ties, and presents data from on-road test-
ing of two truck configurations.
Description of Heavy-Duty Test
Facility
The general capabilities of the ODEC
facility are shown in Figure 1. Its purpose
is to allow emissions testing of heavy-
duty diesel vehicles (HDDVs) in a man-
ner that represents the real world as
closely as possible. Fully integrated into
a Class 8b truck, the facility is designed
STACK MEASUREMENTS
— Opacity
to be completely self-contained, able to
collect several hours' worth of data while
traveling along public roadways. A ma-
jority of the data are collected in real time
by continuous analyzers which allows
comparisons between emissions and ve-
hicle operating modes. These data in-
clude vehicle parameters, engine
parameters, and emissions measure-
ments. Cumulatively, all of these mea-
surements form the core capabilities of
the facility, those that remain intact as
more task-specific capabilities are added
and removed.
Capabilities added for fine PM charac-
terization included a stack dilution sam-
pling system, plume sampling, and the
ability to operate a number of sophisti-
cated fine PM instruments. Of all the in-
strumentation that is available to measure
and characterize fine PM, none of it is
compatible with raw diesel exhaust. The
ODEC facility follows two approaches to
providing dilute samples to the PM instru-
ments: (1) a direct-dilution system that
draws samples from the raw exhaust and
dilutes them with clean air, and (2) probes
that draw naturally diluted samples di-
rectly from the truck's exhaust plume.
Each is described below.
The design of a direct-dilution system
is of considerable importance when sam-
pling fine PM. Several researchers have
demonstrated that particle size measure-
ments can be fundamentally altered by
changes in dilution ratios, residence times,
and physical characteristics of the sam-
Temperature
Velocity Head
T T T T f T I
ENGINE MEASUREMENTS
Intake, Exhaust, Coolant, --^
and Oil Temperatures •
- Speed, RPM —-*1
DRIVE SHAFT MEASUREMENTS
Torque
• Speed, RPM
OPERATIONAL
MEASUREMENTS
Speed, km/hr
Front-to-Rear G-Force
Computerized
Data Acquisition
System
02, %
CO2, %
CO, %
CO, ppm
NOX, ppm
THCs, ppm
EXHAUST SAMPLE
MEASUREMENTS
Figure 1. On-road Diesel Emissions Characterization Facility
pling system. Other considerations that
are specific to mobile test facilities in-
clude the size of the dilution system, its
power demands, and how the raw ex-
haust is delivered to its inlet.
Based on all of these concerns, the
facility currently utilizes the "ejector dilu-
tor" system used by David Kittelson at the
University of Minnesota in much of his
work. This system uses air-powered ejec-
tors to draw in samples and mix them
with a rapid flow of filtered air. Using a
combination of multiple stages and inlet
orifices, the system can deliver dilution
ratios as high as 1000:1 while using less
than 20 SCFM of compressed air. In de-
signing the system, intermediate stage
residence times and dilution ratios are
manipulated to create a system that at-
tempts to simulate the real-world dilution
of an ambient exhaust plume. The accu-
racy of this simulation, however, is cur-
rently limited by our knowledge of how
the plume itself dilutes.
Sampling directly from the plume is
possible because the high-dump (el-
evated) exhaust stack that is most com-
mon on Class 8 trucks typically creates a
plume that is seldom completely disrupted
as it passes along the length of the trailer.
Since the ODEC facility's laboratory is
already built into the trailer, plume sam-
pling is simply a matter of mounting
probes directly behind the truck's exhaust
stack, and locating the necessary instru-
mentation near these probes.
The fine PM instrumentation consists of:
• Electrical Low-Pressure Impactor
(ELPI) - uses a cascade impactor
and electrical particle detection to
provide real-time size classification
and quantification of particulate with
aerodynamic diameters from 0.03 to
10 urn
• Polycyclic Aromatic Hydrocarbon
(PAH) analyzer - produces real-time
measurements of surface-bound
PAHs using the principle of photo-
electric ionization.
• Tapered-Element Oscillating Mi-
crobalance (TEOM) - measures par-
ticulate mass in real time by
calculating the harmonic frequency
change of a vibrating element (where
the mass includes collected particu-
late).
• Aethalometer - uses an optical mea-
surement to calculate the amount of
black carbon deposited on a quartz
filter tape.
• Condensation Nucleus Counter
(CNC) - uses condensation particle
growth and optical detection to count
particles from 3 nm to 1 u,m in diam-
eter.
-------
Fine PM Emissions Data
Each technique used for fine PM sam-
pling (direct-dilution and plume sampling)
has its advantages and disadvantages.
For direct-dilution, the advantages relate
to the ability to control the dilution pro-
cess: a well-designed system should
maintain a steady dilution ratio and intro-
duce little or no background to the dilute
sample. This minimizes variability due to
factors unrelated to the source concen-
trations, and allows straightforward cal-
culation of those concentrations. The main
disadvantage is that this artificial dilution
may not accurately simulate real-world
plume dilution, and the fine PM measure-
ments may be biased as a result.
Plume sampling, however, uses some
of the most realistic dilution conditions
attainable, thus providing considerable
confidence in the representativeness of
the data, especially the size distribution
data (which is most sensitive to the dilu-
tion schedule). Unfortunately, because of
widely varying dilution ratios, the currently
available plume data can only provide
snapshots of the PM emissions.
Figure 2 shows some snapshot par-
ticle size distributions, as measured by
the ELPI, for a truck at a typical highway
speed (65 mph), as determined by stack
dilution sampling and plume sampling.
The bars represent the bin data recorded
by the ELPI software, where each curve
is a best-fit lognormal distribution profile
(or bimodal-lognormal, if it provides a
better fit). The -100 nm peak is typical of
the accumulation mode for diesel ex-
haust, where the leftward skewness (rep-
resented as a tail in the best-fit curve)
varies quite a bit with operating condi-
tions and dilution technique. The data
represented here would indicate that the
skewness may be a result of some bias
introduced by the direct-dilution sampling
system.
Of the fine PM instrumentation used in
this study, the ELPI gave the most useful
data for quantifying mass emissions. It
measures total PM mass concentration in
terms of equivalent aerodynamic diam-
eter (i.e., "unit density spheres"or u.d.s.),
based on its impactor cut points and stage
counts. In general, the highest mass emis-
sions are the spikes that correspond to
some identifiable event in the truck's op-
eration (e.g., pulling off from a stop, chang-
ing gears). The lowest emissions
correspond to low power conditions such
as low-speed/zero-grade tests, with emis-
sions generally increasing with increased
power demand. So, the on-road tests have
identified two primary contributors to PM
emissions, operating transients and power
demand.
Stack Dilution System
80
70 -
E
-y eo -
°- 50 -
o
"". 40-
Q.
§> 30-
T3
z 20-
10-
0 -
I
\
\
\ J
/
/
/
\
\
\
\
\
\
t
\.
n*"-- —
I
10 100 1000
Particle Diameter, nm
Plume Sampling
1 fin
140 -
| 120 -
ro
Q- 100 -
0
". 80 -
1 60 -
T3
z 40 -
T3
20 -
0 -
j
/
/
^
\ ^, — '
/
/^\
\
\
\
\
\
\
V
I
10 100
Particle Diameter, nm
""V^^ 1
1000
Figure 2. Particle Size Distributions for 105 kph (65 mph) operation.
Plume Dilution Characterization
One challenge of designing a dilution
system that will not bias fine PM mea-
surements is knowing the process that
the system is simulating. For on-road
trucks, the plume dilution process is poorly
understood, so any dilution system de-
sign is based on little more than an edu-
cated guess. Therefore, this project has
attempted to characterize the plume dilu-
tion process for Class-8 trucks.
Since dilution schedule is a function of
dilution ratios and residence times, the
project has measured both of these pa-
rameters at various points in the truck
plume. Dilution ratio is measured using
one or more gaseous tracer species
which is (are) not directly affected chemi-
cally by the dilution process. Concentrations
of this species are continuously measured
in the raw exhaust, in the plume, and
possibly in the background air away from
the plume. The dilution ratio is a measure
of how much background air has mixed
with the raw exhaust at the point where
the plume is sampled.
-------
The total residence time at any point in
the plume can also be measured using a
gaseous tracer. For these measurements,
the tracer (propane) is injected into the
exhaust at high concentration, so that it
causes an easily identified spike at the
plume sampling location. It is not neces-
sary to measure exhaust and background
concentrations, but it is necessary to char-
acterize the response time of the plume
sampling system. This response time is
subtracted from the delay time (from the
injection of the measurement spike) to
get a residence time measurement.
Figure 3 summarizes the plume dilu-
tion ratio measurements from 3 days' on-
road testing. The sampling locations are
identified at "2m" and "11m" to indicate
the distance from the exhaust stack in
meters. The data show little speed de-
pendency at speeds as low as 35 mph,
but the value and variability of the dilu-
tion ratio turn up sharply at lower speeds.
The data show reasonable run-to-run and
day-to-day repeatability at both the 2m
and 11m plume sampling locations. At
highway speeds (55-65 mph), the total
residence times are around 0.1 second
at the 2m location, and 1.7 seconds at
the 11m location.
Conclusions and
Recommendations
The following conclusions were
reached from the testing conducted thus
far in the research program:
• Although the stack dilution system
was designed and operated accord-
ing to currently accepted practice, the
character of the fine PM emissions
do not reflect real-world conditions
as found in the plume sampling.
• The overall process of cooling and
dilution of the exhaust plume in the
flow field of the moving tractor-trailer
is poorly understood, and should be
studied further.
• Although carbon dioxide (CO2), NOX,
and/or propane tracers have provided
1000
o
"ro
a:
100
10
o 11m-Oct 8
11m-Oct 7
* 11m-Oct 13
2m-Oct 7
2m-Oct 13
10 20 30 40
Truck Speed, mph
50
60
70
Figure 3. Plume Dilution Ratio Measurements.
all of the dilution schedule data so
far, a truly unique tracer is needed to
eliminate the problems with atmo-
spheric background and contributions
from other vehicles.
• A comparative evaluation of the cur-
rently utilized fine PM instrumenta-
tion is needed before proceeding with
future emissions testing.
The report has presented descriptive
details of the ODEC facility and fine PM
data from that facility. Since fine PM mea-
surement capabilities were added only
recently to the facility, these results repre-
sent work in progress. The following re-
finements should be considered
regarding future work:
• Per the third conclusion, above, it is
recommended that the use of a
unique tracer be implemented in con-
junction with current testing activities.
Plume delay times should be char-
acterized to a higher resolution, es-
pecially at the 2m location, where
the delay was measured as a single
sampling interval (0.1 second) using
the 10 Hz data rate. Measurements
should also include the 6m and 8m
locations, and possibly some plume
sampling behind the trailer.
It may be well worth the investment
to upgrade one or both ELPI units by
adding the filter stage option that is
currently offered by the manufacturer.
If the TEOM is to provide any useful
on-road data at all, the source of the
negative readings must be identified
and either eliminated or character-
ized (i.e., for data correction).
It is recommended that the various
options be explored to improve the
time resolution of the aethalometer.
-------
J. Edward Brown is with ARCADIS Geraghty & Miller, Inc., P.O. Box 13109,
Research Triangle Park, NC 27709.
John Kinsey is the EPA Project Officer (see below).
The complete report, entitled "Heavy Duty Diesel Fine Particulate Matter
Emissions: Development and Application of On-Road Measurement
Capabilities," will be available at http://www.epa.gov/ORD/NRMRL/Pubs
or as Order No. PB2002-100140; Cost: $36.00, subject to change, from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161-0001
Telephone: (703) 605-6000
(800) 553-6847 (U.S. only)
The EPA Project Officer can be contacted at:
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711-0001
United States
Environmental Protection Agency
CenterforEnvironmental Research Information
Cincinnati, OH 45268
PRESORTED STANDARD
POSTAGES FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use
$300
EPA/600/SR-01/079
------- |