United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S4-86/042 March 1987
v>EPA Project Summary
Investigation of Source Emission
PM10 Particulate Matter Field
Studies of Candidate Methods
William E. Farthing, Ashley D. Williamson, Sherry S. Dawes,
Randal S. Martin, and James W. Ragland
The full report describes the field
evaluation of two candidate methods
for source PM-10 measurement. The
two techniques are a new sampling
train design using emission gas recycle
(EGR) and a Simulated Method 5 (SIM-
SI approach using existing hardware
with a specific traversing protocol. Four
field tests were performed. At each test
site, the EGR and SIM-5 measurements
were compared with reference mea-
surements of PM10 and/or total par-
ticulate mass measurements. At two
sites, the EGR and SIM-5 measure-
ments were run simultaneously and
compared to each other. The test results
are presented, and the conclusions
derived from these results are discussed.
Also, recommendations are made for
procedural and hardware refinements
of each method.
Tn/s Pro/ecf Summary was developed
by EPA's Environmental Monitoring
Systems Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that Is fully docu-
mented In a separate report of the same
title (see Project Report ordering In-
formation at back).
Introduction
A size-specific PM,0 ambient-air par-
ticulate standard has been proposed and
promulgation is expected. The introduc-
tion of size into the definition of sus-
pended paniculate matter in the ambient
air suggests the need for measurement
of size-specific emissions from stationary
sources. Technology related to such mea-
surements has been developed in con-
nection with evaluation of control devices
on process streams. Inertia! impactors
and cyclones can separate or classify
aerosol particles in situ but are more
complex to operate than total paniculate
mass trains. Furthermore, operation of
these devices for typical engineering
evaluations does not require the docu-
mented accuracy and reproducibility
desired for established sampling
methodology.
The technical difficulties in size-specific
(i.e., PM,0) paniculate sampling are
greater than, but similar to, those of total
paniculate sampling by EPA Reference
Methods 5 or 17. Potential sampling
biases exist due to variations in the spatial
distribution of paniculate concentrations
across the sampling plane defined by the
duct cross-section. Likewise, temporal
variations in paniculate concentrations
due to process variations can cause in-
accurate or unrepresentative emissions
measurements. EPA Reference Methods
for paniculate sampling (Methods 5 and
17) deal with these problems by specifi-
cations on the sampling location to
minimize stratification (Method 1) and by
spatial and temporal averaging using a
traversing protocol which collects a
weighted average of the paniculate emis-
sions at an array of points spanning the
sample plane. Method 5 specifications
also require at least 3 separate mea-
surements, allowing further temporal
averaging. A second potential error in
paniculate measurements is duct/nozzle
sampling bias. Unless the gas velocity
entering the sampling nozzle equals the
local duct velocity, paniculate matter will
be selectively depleted or enriched in the
sample gas stream due to inertial separa-
-------
tion at the nozzle entrance. For total
paniculate sampling, this bias may be
restricted by specifying isokinetic sam-
pling (within a 10% tolerance) at each
point. This tolerance is easily attained in
sampling that is not size specific by vary-
ing the sample flowrate so that nozzle is
within 10% of stream velocity.
These potential problems are made
more severe in PM10 sampling by the fact
that inertia! size segregation must be
performed. Any inaccuracies in the in-
ertial cutoff diameter, which is determined
by sample flowrate, will lead to errors in
the PM,0 measurement by misclassifica-
tion of paniculate matter in the size range
near 10 jum. Inertial sizing devices (im-
pactors and cyclones) are available which
have a sufficiently sharp collection effici-
ency cut at 10/um, but only for specified
flowrates which are dependent on gas
temperature and composition. Without a
sampling nozzle of continuously variable
cross-sectional area, this fixed flowrate
requirement makes isokinetic sampling
in the manner of Method 5 impossible.
Since an isokinetic sampling bias can be
significant for particles near 10/im. this
effect cannot be ignored.
Previous work on this problem at
Southern Research Institute has led to
the development of two potential sampling
methods the Emission Gas Recycle
(EGR) sampling train, and the Simulated
Method 5 (SIM-5) traversing protocol.
The Emission Gas Recycle (EGR) Train in
principle eliminates the problem of
anisokinetic sampling bias by simulta-
neously allowing isokinetic sampling by
the nozzle and fixed flow operation at the
inertia! sizing device(s). The train design
allows the isokinetic flow of gas into the
sampling nozzle to be augmented by an
adjustable amount of filtered, recycled
stack gas upstream of the inertia! sizing
device.
The SIM-5 protocol is an alternate
candidate PM10 method in which existing
sampling equipment (cyclones or cascade
impactors without special gas recycle
adaptations) are used. The objective of
the protocol is to reduce anisokinetic
sampling errors to the approximate range
expected from spatial and temporal varia-
tion of emissions. Anisokinetic sampling
bias is kept in this range by synthesizing
a full duct traverse from partial traverses
at constant sample flowrate. Points for
each partial traverse are selected that
have duct velocities in the range to keep
anisokinetic sampling errors for 10 /im
particles below ±20% at each point.
Procedure
Emission Gas Recycle
(EGR) Train
A block diagram of the prototype field
EGR train is shown in Figure 1. Stack gas
is isokinetically extracted through the
sample portion of the EGR mixing nozzle
into the inertia! sizing component of the
sample train. After passing the inertia!
sizing device and instack sample filter,
the sample gas passes through the probe
and condenser or impinger train and into
the EGR flow control module. As in con-
ventional Method 5 control modules, the
gas flowrate entering the control module
is controlled by coarse and fine control
values (V, and V2, respectively) at the
entrance of the sealed pump. At the exit
of the pump and absolute filter, the total
flow is measured using a laminar flow
element (LFE). The gas stream is then
split into the recycle and sample flow
lines. The sample flow is monitored in
the normal manner using a dry gas meter
and a calibrated orifice. The recycle gas
flowrate is measured using a second
LFE. The partitioning between sample
and recycle gas is controlled by a valve
(V3) located downstream of the LFE. Valve
V4 was added to the system to extend the
range of control to higher recycle per-
centages by adding back pressure to the
sample flow line.
The recycle gas line, along with the
sample and pitot lines, passes through
the heated probe in which the recirculated
gas is reheated to the duct temperature.
Power to the heater is regulated by a
proportional temperature controller using
a thermocouple reference sensor located
in the gas stream downstream of the
heater.
In the course of this field measurement
program, size specific paniculate mea-
surements were made with two types of
inertia! sizing devices. Direct PM10 size
fractionation was performed using the
first cyclone of the SoRI/EPA Five Stage
Series Cyclone sampling train. This
cyclone was used for all EGR testing and
much of the SIM-5 testing. The remainder
of the SIM-5 testing and other reference
size distribution measurements were
performed using University of Washington
Mark V Cascade Impactors.
EGR Test 1
The first field test of the (EGR) system
at a stationary source took place at one of
two twin 56 MW coal-fired boilers at a
utility generating station. The test plan
consisted of two subtests. Subtest A in-
volved the comparison of traverses per-
formed with the EGR train and a standard
Method 17 train. To eliminate spatial
bias, a two-probe setup was configured.
One probe was configured with a cyclone
set using an EGR nozzle. Colocated with
the cyclone set was a Method 17 probe
with a 47mm filter. Three traverse points
were selected which represented the
maximum point-to-point velocity change
accessible to the EGR probe through one
of three six-inch ports. The recycle rate
was adjusted to achieve isokinetic sam-
pling at each point while maintaining the
chosen constant flowrate through the
cyclone set. The flowrate through the
colocated 47mm filter was adjusted ai
each point to sample isokinetically. Tota
paniculate mass concentrations mea
sured by the two trains were compared
In Subtest B, the EGR-Method 17 hard
ware previously described was usec
without modification. A third probe with
a cyclone using a non-recycle nozzle was
used for a "near-colocated" reference
Sampling for this subtest took place at i
single point. The nozzle for the non
recycle cyclone was chosen to provide t
10jum cut at the isokinetic sample flow
rate. The EGR cyclone was fitted with <
smaller nozzle. The recycle rate was ther
adjusted to provide the flowrate require*
through the cyclone for a 10/im size cu
as well. The gas flowrate in the Methoi
17 sampler was adjusted in the usua
fashion to maintain isokinetic conditions
Total mass concentrations measured b
all three trains and PM-10 concentration
obtained from the two cyclone trains wen
compared.
SIM-5 Test 1
The first field test of the SIM-5 samplin
protocol took place at the same 56 MV
coal-fired boiler used for the EGR shake
down. The test plan consisted of thre
subtests. In all subtests the SIM-5 sizin
devices were modified University c
Washington Mark V cascade impacton
In Subtest A, three probes were usec
Both a Method 17 and a SIM-5 impactc
were used in a twelve point traverse c
the duct area using the SIM-5 samplin
protocol. The impactor was operated at
constant flowrate while the flowrat
through the Method 17 was changed s
as to be isokinetic at each of the travers
points. A simultaneous 50 point travers
of the duct was made with the thir
probe operated according to EPA Refe
ence Method 17. The total particulal
mass determined by each device w£
compared. For Subtest B, three identic
-------
EGR Probe Assembly
Recycle
Line
Sample
Inlet
V3
Recycle Flow LFE
! !
Sealed Pump
Sample Orifice
Exhaust
Dry Gas Meter
Figure 1. Schematic of the emission gas recycle train.
impactors sampled simultaneously at the
same point. To quantify the magnitude of
error in PM,0 determinations resulting
from anisokinetic sampling error, one
impactor sampled isokinetically while the
remaining two operated at the upper and
lower limits of acceptability according to
the SIM-5 protocol. The PM-10 masses
obtained from each were then compared.
For Subtest C, three points in the sample
plane were chosen with widely differing
velocities. Identical impactors were op-
erated simultaneously, one at each point
while another identical sampler traversed
the same points in one sampling run
according to the SIM-5 traversing proce-
dure. The single point impactors sampled
isokinetically. Again, the PM10 masses
obtained by each impactor were
compared.
EGR/SIM-5 Test 2
The major purpose of the third field
test was to provide a direct comparison of
the SIM-5 and EGR techniques. Indepen-
dent reference measurements were ob-
tained with single point isokinetic
impactors. The site, a 500 MW coal-fired
power plant, was selected to access a
duct with velocities near 60 ft/sec, with
substantial velocity spread, and with an
aerosol mass median diameter in the 7 to
14/im range.
The test consisted of three subtests.
For Subtest A, a simultaneous three-
point traverse of the duct was performed
with both the SIM-5 and EGR samplers.
Both of these PM-10 sampling trains
used Cyclones I and IV of the SoRI/EPA
five stage cyclone set followed by a 47
mm filter. For Subtest B, three University
of Washington Mark V impactors were
run simultaneously as near in time as
practical to each set of Subtest A. Each
impactor was operated isokinetically at
one of the three traverse points used for
Subtest A. Subtest C consisted of a single
twelve point traverse of the duct with
both the EGR and the SIM-5 samplers.
The main purpose of this subtest was to
determine the problems associated with
a nozzle change during the SIM-5 run.
EGR/SIM-5 Test 3
The site for the fourth and final test
was a 221 MW pulverized coal-fired utility
boiler. The test plan was designed to
evaluate the EGR and SIM-5 protocols in
high and low particulate mass concentra-
tions. In the high concentration tests
each of the dual inlets of the control
device was sampled. Sampling was also
done at the outlet of the exhaust stack.
This test consisted of five subtests.
Subtest A involved a simultaneous twelve
point traverse of the duct with the EGR
and SIM-5 samplers. Subtest B deter-
mined reference size distributions using
University of Washington Mark V impac-
tors. The cascade impactors were op-
erated over the same twelve point traverse
used for Subtest A. The traverse was
subdivided into four subtraverses, each
subtraverse corresponding to a different
sample port. Each subtraverse was
sampled with a different impactor, using
SIM-5 protocol with the flowrate fixed so
that the nozzle velocity equaled the mean
measured velocity for the three points.
Subtest C consisted of Method 17 mass
train samples at the inlet to both units of
the control device. Subtest D involved
simultaneous impactor traverses in the
exhaust stack at the outlet of the control
device. For Subtest E, Method 17 mass
train samples were taken at the outlet
sampling site.
-------
Results and Discussion
As described previously, the EGR and
SIM-5 techniques were tested at four
sampling locations at three coal fired
utility boilers. The sites were selected to
provide a range of particulate concentra-
tions (15-4000 mg/dnm3) and duct veloc-
ities (15-100 ft/sec), significant fractions
of particulate larger than 10 micrometers
aerodynamic diameter (50-80%), and
substantial velocity non-uniformity. At
each site the EGR and SIM-5 measure-
ments were compared with reference
measurements of PM10 and/or total
particulate concentrations, and at two
sites the EGR and SIM-5 measurements
were run simultaneously and compared
to each other. Table 1 contains a summary
of these comparisons for the field tests in
the full report.
Several conclusions may be drawn from
the field data in Table 1. First,, in every
case, the average concentrations mea-
sured using different techniques agreed
within the combined 95% confidence
intervals. Since these intervals for some
tests reflect a substantial degree of vari-
ation presumably due to source fluctua-
tions, a more meaningful comparison can
be drawn from paired-run analysis of the
simultaneous measurements indicated in
Table 1. Since source fluctuations cancel
to first order in the comparisons, the
confidence intervals are smaller and some
observed differences are found to be
statistically significant. Thus, for example,
the EGR concentrations for both total
mass and PM10 are significantly lower
than the SIM-5 concentrations at Site 2.
Likewise, the EGR concentrations are
higher than the SIM-5 values at Site 3.
Such differences, even when significant
at the 95% confidence level, are not
consistent in direction from site to site,
and are typically on the order of 10%.
These differences may only reflect dif-
ferences in operating conditions. We
conclude that PM10 measurements using
each technique can be expected to com-
pare well with the other and with those
using other reference techniques. Like-
wise, EGR measurements of total mass
concentration may be expected to com-
pare well with Method 17 measurements
were particulate catches are sufficient to
allow good recovery from the EGR
sampler. Although SIM-5 was not de-
signed for total particulate mass concen-
trations, in the present test series SIM-5
total mass measurements fell within 15%
of values from the EGR and other iso-
kinetic samplers.
Table 1.
Percentage Differences and Confidence Intervals in Particulate Concentrations
Measured During Test Series"
Number of
Replications
PMn
Total
Concentration
EGR Initial Test Site 1
EGR Cyclone Isokinetic Cycloneb
EGR Method 17"
SIM-5 Initial Test Site 1
4
8
-8.3±27%
9.O±29%
-11.5±8.3%
SIM-5 Method 17±
SIM-5 Isokinetic Impactor^
EGR/ SIM-5 Comparison Test Site 2
EGR Cyclone - SIM-5 Cyclone*
EGR Isokinetic Impactors
SIM-5 Isokinetic Impactors
EGR/ SIM-5 Comparison Test Site 3
EGR-SIM-5*
EGR Impactor
SIM-5 Impactor
SIM-5 Impactor Method 1 7
3
4
5
5-6°
7-6f
Inlet
6
6-5=
6-5=
Outlet
6-7
-1.8±22%
-15.5±6.5%
-11±31%
3.8±25%
11 ±9.8%
27±16%
16±16%
-16±32%
-14.0±65%
-9.2±8.5%
1.3±38%
14±31%
1.7±21%
-9.8±16%
-11±14%
-7.4±23%
" All differences and confidence intervals expressed as percentages of the mean value. Confidence
intervals represent 95% significance level.
* These comparisons were analyzed as pairs since the measurements were simultaneous.
c Where two numbers of replications are given, the first number corresponds to the first listed
device and the second to the second device.
Recommendations
We have several recommendations for
further study. First, we feel that both
techniques are sufficiently advanced that
they should be documented in detail for
potential use as sampling methods if
source PM10 regulations are issued.
Second, they should be subjected to more
extensive validation and collaborative
testing than was possible during this
project in order to further define precision,
reproducibility, comparability with other
measurements, and possible sources of
interference of bias using each technique.
It is also recommended that further
development of both techniques continue
to refine procedural details and investigate
hardware improvements. These include
development of an EGR impactor train,
further characterization of impactor pre-
cutters and the existing PM10 Cyclone
(SoRI-l), investigation of an alternate
geometry Cyclone I, and investigation of
the optimum use of S-type pitot tubes
with instack PM10 samplers.
-------
William E. Farthing, Ashley D. Williamson. Sherry S. Dawes. Randal S. Martin,
and James W. Raglandare with the Southern Research Institute, Birmingham,
AL 35255-5305.
Alice C. Gagnon is the EPA Project Officer (see below).
The complete report, entitled "Investigation of Source Emission PM^P articulate
Matter Field Studies of Candidate Methods," (Order No. PB 87-132 841 /
AS; Cost: $18.95, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States Center for Environmental Research
Environmental Protection Information
Agency Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S4-86/042
0000329 PS
------- |