United Stales EPA-600/9-89-Q85
Environmental Protection
September 1989
v>EPA Research and
Development
FIFTH SYMPOSIUM ON
FUGITIVE EMISSIONS:
MEASUREMENT AND CONTROL
(May 3-5, 1982, Charleston, South Carolina)
Prepared for
Office of Environmental Engineering
and Technology Demonstration
Prepared by
Air and Energy Engineering Research
Laboratory
Research Triangle Park NC 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the MISCELLANEOUS REPORTS series. This
series is reserved for reports whose content does not fit into one of the other specific
series. Conference proceedings, annual reports, and bibliographies are examples
of miscellaneous reports.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/9-89-085
September 1989
FIFTH SYMPOSIUM ON
FUGITIVE EMISSIONS: MEASUREMENT AND CONTROL
(May 3-5, 1982, Charleston, South Carolina)
EPA General Chairmen:
D. Bruce Harris and William B. Kuykendal
Air and Energy Engineering Research Laboratory
(Industrial Environmental Research Laboratory)
Research Triangle Park, North Carolina 27711
Prepared for:
U. S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
ABSTRACT
These proceedings document presentations at the Fifth Symposium
on Fugitive Emissions, held May 3-5, 1982, in Charleston, South Carolina.
The symposium was sponsored by the U. S. Environmental Protection
Agency's Industrial Environmental Research Laboratory (now known as
the Air and Energy Engineering Research Laboratory) in Research
Triangle Park, North Carolina, as part of the Agency's continuing effort
to develop methods for the measurement and control of airborne and water-
borne fugitive emissions from energy and industrial processes.
The objective of the symposium was to bring together people from
*
industrial, academic, research, and government organizations with exper-
ience or interest in fugitive emissions problems to exchange information of
mutual potential benefit.
The program included presentations by individuals from a variety of
organizations describing their experience and viewpoints regarding the impact,
measurement, and control of fugitive emissions. An international flavor
was provided by presentations by authors from Belgium, Canada, and Sweden.
11
-------
CONTENTS
SESSION I—Session Chairman Alfred B. Craig (EPA/IERL-RTP)
Results of Measurement Programs for Fugitive Hydrocarbons
Using a Downwind Crossectional Flux Analysis Method—
P. R. Harrison (Envirosol) 1-1
Test Protocol for Evaluating Fugitive Emissions from Active
Coal Storage Piles—D. Carnes, P. Catizone, T. Kincaid
(TRC Environmental Consultants); D. B. Harris (IERL-RTP) 2-1
Characterization of Fine Particulate Emission Factors for Paved
Roads—C. Cowherd, P. Englehart (Midwest Research Insti-
tute) 3-1
SESSION II--Session Chairman William Hague (Julius Koch,
USA, Inc.)
Micron Droplet Dust Suppression Proves Out in Variety of Fugi-
tive Dust Applications—W. Hartshorn (Sonic Development),
L. Strand (Andeze AB, Sweden) 4-1
The Optimization of Wind Screens for Fugitive Emission Control
Using Wind Tunnel Tests—C. J. Williams (MHTR)* 5-1
Evaluation of Field Test Results on Wind Screen Efficiency—
A. Larson (TRC Environmental Consultants)* 6-1
Effects of Street Sweeping of Fugitive Emissions from Urban
Roadways—D. F. Gatz (Illinois State Water Survey) 7-1
Evaluation of Road Carpets and Chemical Road Dust Suppressants—
A. Larson (TRC Environmental Consultants)* 8-1
Evaluation of Weathering Characteristics of Dust Suppressant
Chemical Additives—W.B. Kuykendal, D. C. Drehmel,
B. E. Daniel (IERL-RTP) 9-1
SESSION in--Session Chairman John E. Yocom (TRC Environ-
mental Consultants)
On the Use of SFg Tracer Releases for the Determination of Fugi-
tive Emissions—B. Vanderborght, J. G. Kretzschmar, T.
Rymen (SCK/CEN, Belgium); F. Candreva, R. Dams (INW-
RUG, Belgium) 10-1
(*) Alternate paper provided.
iii
-------
Contents (cont.)
An Atmospheric Tracer Investigation of Fugitive Emissions
Transport in the Colorado Oil Shale Region—G. A.
Sehmel (Pacific Northwest Laboratory) U-l
Laboratory Testing to Improve Rail Car Sealant Spray and
Loading Techniques for the Abatement of Fugitive Coal
Dust—C.J. Williams. W. F. Waechter (MHTR)** 12-1
Studies of Nontraditional Fugitive Particulate Control Tech-
nique s--B. M. Nicholson (EPA/OAQPS), M. Borcherding
(City of Portland), G. Ekhardt (State of Minnesota),
R. Mohr (State of Colorado)** 13-1
Evaluation of the Efficiency of a Charged Fog Generator in
Controlling Inhalable Particles at a Steel Plant (C. V.
Mathai, B. M. Muller (AeroVironment); W. B. Kuyken-
dal (EPA/IERL-RTP)* 14-1
SESSION IV--Session Chairman James A. Dorsey (EPA/
IERL-RTP)
Estimation of Ambient TSP Impacts of Coal Storage and Hand-
ling Facilities--R. C. Wells, D. C. Doll (Enviroplan),
J. Hattrup (Baltimore Gas and Electric) 15-1
A Determination of the Impact of Fugitive VOC Emissions
from a Municipal Hazardous Waste Incinerator on the
Surrounding Community--G. A. Holton (Oak Ridge National
Laboratory). L. J. Staley (EPA/IERL-Cin)* 16-1
Impact of Fugitive Emissions on PM-10 Concentrations—
T.G. Pace (EPA/OAQPS)** 17-1
Application of Dispersion Dictated Mass Balance for Calcula-
ting Fugitive Dust Emissions—C. F. Cole (TRC Environ-
mental Consultants), J. G. Moldovan (Anaconda Minerals),
P. B. Kunasz (Consultant) 18-1
Modeling the Emission of Aerosols in and Around a Metallur-
gical Plant--B. Vanderborght, I. Mertens, J. G. Kretz-
schmar (SCK/CEN, Belgium); F. Adams (UIA, Belgium);
R. Dams (INW, Belgium)** 19-1
SESSION V--Dennis J. Martin (TRC Environmental Consultants)
New York State Industrial Coal Pile Drainage Regulations and
Guidelines--C. Hornibrook (NYSERDA) 20-1
(*) Alternate paper provided.
(**) Abstract only.
IV
-------
Contents (cont.)
Calibration and Verification of a Coal Pile Drainage Model--
J. A. Ripp, G. T. Brookman. P. B. Katz (TRC Environ-
mental Consultants); J. G. Holsapple (New York Power
Pool) 21-1
Coal Pile Simulation Study--A. Schumacher, E.G. Hanson
(Environmental Science and Engineering) 22-1
Control of Acid Problems in Drainage from Coal Storage
Piles—H. Olem, T. L. Bell, J. J. Longaker (TVA) 23-1
ATTENDEES A-l
-------
RESULTS OF MEASUREMENT PROGRAMS FOR FUGITIVE
HYDROCARBONS USING A DOWNWIND CROSSECTIONAL
FLUX ANALYSIS METHOD
PAUL R. HARRISON
Envirosol
Environmental Solutions
1700 N. Fiske Avenue
Pasadena, California 91104
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
1-1
-------
ABSTRACT
A method to quantify fugitive volatiles organic compounds (hydro-
carbons) has been in the developmental process for the last seven years.
Recent activities have significantly improved the method and have quanti-
fied the precisions and accuracies under both controlled and field
conditions. This paper is a summary of various applications and their
results. Special attention is given to the precision (repeatability) of
the data in field conditions. Facilities, components, and other area
sources have been selected to demonstrate the flexibility of the method.
Actual data are given.
The major conclusions are that the technique is operational to most
applications with precisions of less than 30%, with 8% nominal. Ultimate
accuracies are not yet determined due to lack of an acceptable standard.
The method is the most flexible available and is providing emission rates
hitherto unavailable. The application to the "bubble rule" is encouraging,
1-2
-------
I. INTRODUCTION
This paper is a summary of a developmental program to further de-
velop a method to quantify fugitive emission rates of VOC's under all
reasonable meteorological and physical condition. The initial impetus
for the work was a problem with applying the U.S.EPA emission rates (AP-
42) for refineries to non-refinery operation, i.e. a small crude oil gas
treatment plant. In this initial program the measured leak rate for
individual components differed from the published factors by orders of
magnitude. By use of a downwind plume mapping technique in conjunction
with the first reported "Direct maintenance program", an understanding
of the differences was achieved.1 Since this initial program, performed
in 1976, the technique has been further perfected and applied to total
refineries and petrochemical facilities as well as their sub-units and
individual components. In addition, Natural Seeps, Impoundments, waste-
water treatment facilities, tankers and storage tank emission rates have
been quantified, usually to much better precision than achieved by more
passive measurements. Due to the response characteristics of the con-
tinuous VOC monitor preferred for these measurements, and due to the
question of quantitative determinations of precisions and accuracies
the method had not, until now, been advertised as operational. Since
very few data existed for fugitive emissions and due to the unusually high
variability of most fugitive sources, the task of objective evaluation
of the method was not at all trivial.
This paper will briefly describe the technique, summarize the past
applications and present data for some emission rates found. It will also
present the results of precision and accuracy tests performed on recent
data.
Results will be offered to show that the technique is quite flexible
and will provide adequate, defensible emission factors at a relatively
inexpensive cost (the more precision required, the more work required).
At this point the method is operational. Although it can be performed
by semi-technical personnel, tests should conducted under the direction
of an experienced person, especially if one desires a cost effective pro-
gram and a data set with optimum accuracy.
II. THE TECHNIQUE
The method makes very few assumptions. The most critical is that
the downwind signature has a Bionominal or "Gaussian" distribution,
especially in the horizontal. Since we know that on the average, the
turbulence structure will distribute itself to the Gaussian condition
in the micro and mesoscale meteorological volumes, the assumption is
relatively sound. The key is to choose the distance from the source
and the wind conditions so as to map the plume at the most likely isotropic
1-1
i \j
-------
turbulent condition. As a matter of practicability this condition
must be compromised, but one can recognize the condition and compensate
if necessary by early data reduction and a presurvey measurement.
The formula used is straight out .of Turner's handbook2 (or any
other pertinent resource for that matter) and is for a centerline con-
centration of a ground level source.
From diffusion theory and measurements, the following equation
applies for ground-level sources:2
Q = IT oy az AX max u c K
where
Q = Emission rate of hydrocarbons (gm/sec)
= Peak, net concentration of hydrocarbons in Gaussian fit
AY max . x
curve (ppm)
K - Conversion constant from ppm to yg/m3; 665 x 1CT6 gm/m3 for
methane at 20°C
av = Lateral extent of Gaussian plume (m)
az = Vertical extent of Gaussian plume (m)
C * Correction factor from methane equivalent to actual mass
emission rate due to instrument response
•7 - 3.141
u = Mean wind speed (m/sec)
All parameters are usually obtained from field measurements. (In some
instances, az is obtained theoretically from ay. Care must be taken
to prevent over or underestimation of this value in areas severely affected
by local,mechanically-induced turbulence.)
The transect method of sampling is used to actually map the dimen-
sions and concentrations of a plume. From this mapping, a reasonably
accurate emission rate can be calculated.
Two types of plume mapping have been used: single level transect
at an optimum distance and conditions from source, and a complete plume
crossectional mapping. The former method is less work (expense) but may
not be as accurate due to the necessity of calculating the vertical extent
of the plume (
-------
made except reasonable isotropic turbulence. For close-in measurements
of area sources this method is necessary since gz calculated from cry
may overestimate the emission rate. Please see Figure 3.
The actual physical measurement techniques will not be discussed at this
time since, although they are not particularly unique, they do require
specific protocols that must be followed in order to achieve optimal
accuracies.
III. EXAMPLES OF APPLICATIONS
This section will present a selection of the actual applications to
the quantification of fugitive VOC from various facilities and devices.
Please note that fugitive VOC emission rates vary greatly within and from
facility to facility. The following data will demonstrate the typical
precisions found in steady-state conditions as well as those in highly
varying conditions.
REFINERIES, TOTAL FACILITIES
REFINERY I
In this case the wind conditions were unusually ideal. A single level
set of transects were used in the calculations, but the vertical extent
of the plumes were verified by use of a slow aircraft. Five data sets
were obtained over a three month period. Each "signature" was unique
but,the data agreed very well, i.e.:
Test Q (tons/day)
1 29.0
2 24.3
3 27.1
4 24.3
5 28.1
AVERAGE 26.6 + 2.2, (8%)
REFINERY II
In this case the conditions were such that the transect line was very
close to the edge of the complex. Both vertical and a single level mapping
were taken. Figure 2 is a typical signature of the close-in facility fugi-
tive VOC emissions. Note that the single level transect would miss part
of the plumes. Comparison of the two types of data resulted in an emission
rate of 3.2 tons/day but with a precision of twenty three percent. These
data demonstrate the loss of precision due to the unfavorable transect or
mapping conditions as well as the possible effect of a non-uniform emission
rate (a turn-around was underway at Refinery II).
1-5
-------
0 PETROCHEMICAL (A sub-unit and components)
Most petrochemical data are proprietary; however, we can use one case
to demonstrate how a single unit can be isolated from the total complex-
In one case a sub-unit was quantified both upwind and downwind. At
the same time each individual emission point (component) or area was quan-
tified.
The sum of the most significant of the 19 leaking components was
706 + Ibs per day. The downwind minus the upwind value was 824 Ib per
day,~thus there was an agreement within fifteen percent between the
two methods. (Both single and multilevel transects were made but the
multilevel results were used preferentially in these calculations.) It
is interesting to note that the precision of these single day comparisons
were better than the comparison in results taken on separate days, i.e.
the variance of four separate emission rates using the plume mapping
technique was 21%. This, of course, was a result of process varia-
tions as well as the data sets themselves.
Finally, the small components were bagged and timed to find their
emission rates. A. tail gas compressor was two large to bag, thus both
a single level and multilevel transect was used to gain results of 88 and
81 Ibs/day respectively. As one can see, the agreement is quite good under
these conditions. Other individual areas quantified were loading racks,
a flare pipe leak, and a spill; each was identifiable within the upwind
transects. An intermittent process such as "Trap Blowing" was also quan-
tified (148 Ib/hr of operation). Please see Figure 4.
0 CRUDE OIL GAS TREATMENT PLANT
In this instance the fugitive VOC emission rate for the total facility
was initially 74 Ib/day. After the directed maintenance program the rate
was measured downwind as 7 Ib/day.
0 PROCESSED CRUDE OIL STORAGE TANK
This double seal storage tank contained warm crude that had under-
gone dewatering and used a heater/treater process to remove volitiles.
The emission rate during draw-down was 25 Ib/day. This measurement used
a single level transect and a plume centerline vertical profile for de-
termining the plume dimensions.
NATURAL OCEAN SEEPS
Over the last several years a large natural ocean seep has been
quantified by use of both the single and multilevel methods. The source
strength is known to vary from visual observations; thus precisions also
include natural variations:-^
1-6
-------
NATURAL SEEP EMISSION RATES
Emission Rate
Year T/day
1977 8.7
1978 10.7
1979 21.0
1980 16
25
20
24
22
AVERAGE 1980 21 + 3.2 (15%)
Since 32% was non-methane, it was determined to be equal to one-half the
county hydrocarbon inventory and that it was commercially feasible to trap
the gas for sale as product. Finally, a bonus was realized from some of
these measurements in that we were able to document revised dispersion
coefficients over the cold California oceans.
0 REFINERY WASTEWATER TREATMENT
Currently, under a U.S. EPA contract, an evaluation of the emission
rates of refinery wastewater treatment facilities is taking place. Sub-
units and components are also being quantified. These data will be
presented later after review. They will include preliminary information
such as effect of covers, operating characteristics, age and configura-
tion. Detailed comparisons of of purely statistical treatment of the
data versus the graphical treatment used are also being determined.
Since these devices are so dynamic in the material received it represents
a severe test for the method. For example, not only are the locations
of components usually not ideal, the input to them is highly variable
and represents a large mix of VOC constituents. Also, the liquid surface
conditions vary greatly and can control the fugitive emission rates as
much as the potential VOC in the bulk water. This study will also dis-
cuss the accuracy of the methods in greater detail.
IV. ACCURACIES AND/OR PRECISIONS OF THE METHOD
Although the "plume mapping" or "downwind crossectional analysis"
method of quantifying fugitive emissions has been used in even a more
primitive form for particulates as well as for other gases and tracer
studies, there has been some resistance to its acceptance. The primary
reason may be that, although it uses off-the-shelf equipment and is
relatively simple in theory, it is tedious and requires judgement as to
the best measurement location to optimize the turbulence conditions so
to approach the classical normal distributed plume. It marries the real
world with turbulence and statistical theory in a rather unique way that
crosses disciplines, thus making it suspect; i.e. it's a unique combina-
tion of established procedures.
1-7
-------
A recent obstacle that has been removed is the calibration of the
continuous monitor to the unique hydrocarbon mix encountered in most
areas (if only one VOC is present there is no problem). By using in-
field analyses for total, NMHC, and MHC, we can determine the calibration
factor C in equation (1) for each condition. Since the VOC are all con-
verted to methane, we can eliminate the necessity of calibration to each
constituent. Since the procedure is a step-wise progression of operations
to the data, the variations are easily exposed and not lost in the interven
ing calculations. Said another way, the process must not only "suit in a
reasonable final data point, each step or parameter must also make physical
sense.
As part of all these studies we have attempted to focus part of our
efforts toward objective determination of precisions and accuracies (few
of the data sets are complete)i
Table I is a summary of the precision of much data available to date.
Please note most of these cases are dynamic and contain process variations
as well as variations due to methodologies.
Recent data taken over several days verify these results but
also contain standard deviations varying as high as 69% and as low as
18%. (For data taken at different times during the same day, the preci-
sions improve to 15%.) Naturally we observe that the quality of the
data sets goes down as the variation among the data goes up. However,
the realistic test in these data is to compare data taken in close time
sequence on the same device. These comparisons are those shown in Table
I. These are a great improvement over the thousand percentage deviations
obtained in other studies on wastewater facilities.
Accuracy tests were also conducted by comparison of the results of
the downwind method to metered releases and to a temporarily ducted
source. The ducted sources were less accurate and more variable than
the transect method due to leaks an.* the effects of wind direction orien-
tation to the ducts. Local turbulence was also a major factor in the
poor showing of the ducted method. Standard deviations ranged from 220%
to 7%, much larger than the transects. Ratios of transect to ducted
resulted in an accuracy of 150% + 68%. Since we lost much of the VOC
from the ducts we are not surprised the ratios are usually in greater
than 100%. A similar test using a metered rate of release of propane
resulted in a ratio of transect to metered rate of 1.6 + 19%. Although
the precision was in the range seen before, the accuracy is not as good.
We have since discovered that as the propane was released, some of it
was nearly liquid at the rotometer. Thus, the reported metered release
rate is too low, especially since the metering apparatus is cooled by
the gas phase change. The resultant ratios confirm this finding. In
conclusion, we know the precision for these tests is within 30% and the
average accuracy is better than 60%. A definite test of the system's
ultimate accuracy has not yet been made due to these difficulties in the
accuracy of the calibration methodologies themselves. It is probably
equal or better than the precisions.
1-8
-------
TABLE I
PRECISION OF DATA SETS AVAILABLE TO DATE
Test a , Precision, % Number of Tests
Refinery I 85 days
Refinery II NA
Petrochemical 15 2 methods
Tail gas compressor 9 2 methods
Petrochemical 21 4 days
Ocean Seeps, 1980 15 5, same day
Refinery II Wastewater Treatment
0 Test in succession, 1 21 3
0 Test in succession, 2 31 3
0 Test in succession, 37 3
0 Test in succession, 4 28 3
0 All tests, same device 69* 13
*Conducted over several days and includes process variations.
1-9
-------
Finally, by using purely statistical manipulations on the raw data
sets, the agreements are also acceptable. Variations within the various'
statistical methods themselves vary between a standard deviation (a) of
7% to 13%. The purely statistical analysis and graphical methods agree
within a range of 60% to 120% (+ 40% to 20%) of each other with one of
three comparisons within 3%.
CONCLUSIONS
From the data presented we can make the following conclusions:
0 A method exists to quantify fugitive VOC with precisions typically
within 30% and to 8% for well behaved plumes.
0 Ultimate accuracies are not yet determined but are known to
be better than £ 60%.
0 The method uses a combination of classical theories in a
unique way.
0 The method has great utility and can be used on most sources
accessable downwind (one can also forecast wind shifts).
0 This method is providing data never before quantified with
acceptable precisions or accuracy.
0 The method has application to the bubble rule .and can verify
the effectiveness of the directed maintenance program.
REFERENCES
1. Harrison, P- R. , "Comparison of Component Emission Rates to AP-42 for
a Gas Treatment Plant at Ellwood, CA," AR.CO Production Co., Envirosol
Report No. 1461, 1976.
2. Turner, D. B. , "Workbook of Atmospheric Dispersion Estimates," U.S. EPA
(PHS) Publication 999-AP-26, 1970. (NTIS PB191482.)
3. Harrison, P. R. and Mass, S. T. , "Dispersion Over Water: A Case Study
of a Non-Bouyant Plume in the Santa Barbara Channel, California," AMS
Joint Conference, November 29, 1977.
1-10
-------
VIRTUAL SOURCE
OF FUGITIVE EMISSIONS
ELEVATED
POINT SOURCES
Figure 3» Schematic representation of vertical point profiles (1, 2. 3, 4, 5) and
elevated horizontal transect lines (A, B, C, D).
-------
(1)
SUMMIT
II HO
mxucT
UMf(J)
I
I—•
CO
Ficuin
(k)
COMFOMENI
(0
COUPONEM1
S1SIEH
IIINUCT LIMES
TMNSECI IINES
\
Figure 4: Schematic of cross-sectional measurement for subunits, components,
or component system.
-------
Questions
Greg Holt, Oak Ridge National Lab:
Q: How far downwind were you measuring for API unit?
A: Well, you know, when you are In structures, you have to be careful
because you have to go far downwind to try to get your turbulence
distributed; sometimes that 1s not practical. In this particular
case, we are approximately 5 meters downwind.
Q: What was your time frame for your measurements? For each transect or
total? Well, your averaging time, considering your air movement.
I'm thinking your Gaussian assumption 1s awfully close. I would think
perhaps a box model would be just as good.
A: Well, one of the things I can't argue with 1s when you take the data
you plotted up, 1t sure looks Gaussian and It also satisfied the
Gaussian tests within reason, there 1s no such thing as a perfect
Gaussian. The averaging times, typically, 1t takes a half hour to
get a data set, which Is good enough. Sometimes you get wind shifts,
that 1s why we've altered the technique, rather than take several at
one level, and then several at the next level, etc., to work your way
up, we take a couple at each level and move up and down - we keep going
that way as opposed to doing all the averaging at one level, because
a shear will come along sometimes and your plume will get bent over.
It's no real problem because you see that happening and you just
measure directly up rather than slantwise, so we've altered that.
We get kind of a data set within 15 minutes or so and just repeat it.
1-13
-------
George Sehmel, Pacific Northwest Laboratory
Q: What 1s downwind distance and how long 1s that wind effective? Do you
feel plume 1s contained 1n lower 2 weters? Not sure what you did there.
A: Well, these are more theoretical questions than practical questions.
The practical answer to that Is that you take sufficient height until you
look at your signature and you take sufficient points until you actually
see this coming out, so you see the data at Us real time and you map the
plume, and you don't care what the tall up there looks like that much
becasue the real Important part of 1t 1s the center line of the plume and
you take sufficient height and you clearly define that at least 7Q% of
the plume and the only problem you get Into 1s that you have to make sure
you take enough data points discrimination to get close to the maxima.
If you don't do enough of them within the maxima, then you have a problem
with what is the maxima. You can get fairly large error that way. Now,
1n some of these techniques you notice that we talk about selected inter-
perlated data you are actually searching and 1nterperlat1ng a maximum
concentration at the center line. And you will find that 1t helps a little
bit but 1t doesn't change 1t that much. The real question 1s that I don't
assume anything, I actually measure It. There are no assumptions here at
all, except that It's Gaussian distributed and then If you are really con-
cerned about that (1t being Gaussian) you can actually test 1t for its
Gaussian fit, which we do. The statistical techniques are purely statist-
ical techniques, using a logrlthmic Gaussian fit and those are your
results and you compare that to the graphical, which just assumes a
straightforward Gaussian. It is not a box model really, 1t certainly
looks like 1t doesn't have sharp wedges at all; if anything, 1t 1s a
Gaussian.
1-14
-------
Q: You were taking measurements at a horizontal, were you also taking
measurements at vertical as well?
A: We took several measurements at various levels. You will "find it is
easier to take the horizontal at various levels than try to go up and
down It at various horizontal positions.
Q: Then you were at more than one plane?
A: Absolutely, you map both the vertical. I showed one example of a single
plane which again was more or less a ground level source which assumed
certain vertical dispersion characteristics, but we progressed from that
a couple of years ago and by various means, either by long poles or actu-
ally a tethered sonde balloon system where we actually record the height
and tow the balloon back and forth 1f 1t 1s a very high type plume. Yes,
we get both vertical and horizontal dlsperson parameters.
Q: In one of your examples, you mentioned you were as close as 5 meters to a
unit source. What was your regular distance when doing entire refineries?
A: One-quarter of a mile.
Q: And how high did you go?
A: In a total refinery, remember we're looking for fugitives only, you'd
have to go maybe 500 feet. At this particular refinery, I don't have
that data here and It's somewhat proprietary, but we use an aircraft as
well as simultaneous surface measurements. The problem with the total
refineries 1s that you have cooling towers, you have flares, you have
stacks, etc., so the assumption was that 1f you took your transects close
enough In, and this happened to be the case, and flares were on the down-
wind side, so you get close enough so that the flares were not touching
1-15
-------
down 1n your transect line and took about one transect line, you can use
your assumption of vertical dispersion characteristics and come up with
a good number, using only one transect line of about 3 meters. On that
particular refinery, that was what we did and It worked very well. We
were looking only for fugitives, that was the purpose of that particular
task; although we did measure the total refinery. But to separate the
fugitives from the flares and other materials, we could not go too high.
Q: In a tethered balloon, what do you do, bring your sample down a small
diameter tube or how do you do that?
A: Well, 1t depends on the height. We can get up to about 500-600 feet
with tethered lines and bring the sample down. There's no real problem
with that, It's just a matter of lift and some of the other work we've
done on total chemistry of the detached plumes, we've been able to lift
quite a bit up Into the plume and get easily a few hundred feet of line
up there. If you get much higher than that, you just can't 11ft that
much line and then we go Into remote control samplers, but you don't
get a continuous sample, you get a point source sample. It's hard to
map a plume from one point or point data taken 1n sequence over a period
of time. We've towed the balloon several hundred feet high.
Q: What are your remote control samplers?
A: I've built these myself. I own the balloons, the telemetry and the
sampler.
Q: The best method?
A: It depends on what you're looking for. Most of the time, we've been
after sulfates, sulfur, etc., but you can either use absorbents, you
1-16
-------
can use a bubbler, whatever you want, 1t doesn't matter. It's just a
matter of turning the pump on and off at precise times, recording the
meteorological phenomena, time, location, depending on what your posi-
tion 1s, but It's nothing really sexy about It, you just turn It on and
off, apply 1t to whatever level you want 1n the plume, but once you get
above, 1f you want a continuous sample, say 500 feet, you almost have
to use a slow flying aircraft. But, for large facilities, that 1s not
so bad. It 1s that point source plume that 1s the problem with an air-
craft.
1-17
-------
TEST PROTOCOL FOR EVALUATING FUGITIVE EMISSIONS
FROM ACTIVE COAL STORAGE PILES
David Games
Pietro Catizone
Thomas Ciacaid
TSC Environmental Consul cants, Inc.
800 Connecticut Boulevard
East Hartford, Connecticut 06108
and
D. Bruce Harris
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved lor
presentation and publication.
2-1
-------
Introduction
A sampling program was developed to collect fugitive particulate matter
•missions data from an active coal storage pile servicing an electric
utility. The data were used to determine the relationship between program
costs and relative accuracy for variously sized monitoring networks.
The upwind-downwind sampling method was selected for the study because
it is the most appropriate method for assessing the source strength of
relatively large sources that cannot be sampled by traditional techniques.
This method entails measuring downwind concentrations, correcting this value
for ambient background concentrations (upwind contribution), and using
diffusion equations to calculate apparent source strength. These equations
assume that pollutant diffusion has a normal distribution about the plume
centerline in both the horizontal and vertical dimensions. The equations
also assume a uniform emission rate of pollutant and that total reflection
of the plume takes place at the Earth's surface. The area source is
represented as a virtual point source using the techniques of Turner.1
Because the position of the virtual point source varies with stability, the
downwind distance of the samplers changes slightly between tests.
Since emissions from fugitive dust sources are normally estimated from
measurements made at remote distances from the source, loss of material from
the plume between the source and the measurement point will affect the
source strength estimate. If the source being measured generates a
signficant number of large particles* many of the particles may be deposited
before the plume reaches the measurement point. To take the effect of plume
depletion into account, correction factors were developed to adjust the
apparent source strength. By collecting dustfall samples at downwind
distances where suspended particle concentration is also being sampled,
deposition velocity may be calculated by taking the ratio of the deposition
rate to the immediate ground level air concentration. Chamberlain and
others describe a method for computing a source depletion relationship by
modifying the equations published by Sutton.2 The depletion fraction
using this method is a function of downwind distance, Pasquill-Gifford
atmospheric stability class, source elevation and particle deposition
velocity. The estimates of source strength, corrected for dry deposition,
were statistically analyzed to determine the cost effectiveness of variously
sized monitoring networks for similar area sources.
Monitoring Program
Design of the sampling array to monitor source emissions is illustrated
in ?igure 1. The active area of the coal pile, which constitutes the major
source of fugitive coal dust, measured approximately 177 meters (580 feet)
long, 116 meters (380 feet) wide, and was oriented with the long axis in the
east-west direction. Activities within this area included loading onto the
active pile through a telescopic chute and spreading and reclaiming coal by
bulldozers and scrapers. The level of activity within the source area
varied from test to test and therefore the generation rate of the fugitive
emissions also varied.
The orientation of the sampling array with respect to the source was
based on 5-year average seasonal wind rose data collected at a nearby
2-2
-------
UGEHO
A HI6N-WUMC SAMPICI
vim sue sfuaivf
ANO MlfiH VOLUHI CASCADE INMCfM
O JO-FOOT TOJtt UIIH NICH WHUff SAMTLHIS
m 30-FOOf fttJEI UIIN MIGN VOLIMC SAMTIEU
" ANO NEIEOMN.OGICAL SfNSOM
O OUSI f All
* ME1EOMH08ICML SCiSMS Al I FCCT
Q STATION IB
H
MIT
'rtlCII VOIUMf .SANPtfR lOCATf* I401
APPBOI. 1000 Fill UPUINO OF SOIMCC
Figure ). Final monitoring network layout and site topography (elevations In feet).
-------
Rational Weather Service station. The upwind station was located
approximately 1,000 feet south of the aoucce area. Three downwind sampling
locations were established at approximately 150 feet, 1,000 feet and 1,400
feet, based on estimates and preliminary analyses of source strength and
deposition characterization. The crossvind dimension of the array spanned
approximately a 45° sector.
Initiation of a test occurred only when wind flow was at speeds greater
than 1 aps (2.2 mph). The length of each test was determined by the
activity level within the source area and the time required to collect a
measurable sample at all stations within the dust plume. Experience showed
that when filters became visibly dark in color, sufficient particulate
matter had been accumulated for accurate weighing. Site characteristics
during the test periods are described in Table I.
TABLE I. Site characteristics during test periods
Test
Number
Wind
Speed
Jm/s)
Wind Temp.
Direction+t (°C)
Pasquill-
Gifford
Stability
Category
Coal Pile
Activity
t
tt
2 bulldozers
2 scraper(s)
2 bulldozers
2 scraper(s)
3
4t
5t
6
7
8
9
2.3
3.6
1.5
3.6
3.1
3.6
2.9
146
220
60
147
146
145
143
23.9
23.9
27.8
24.4
24.4
21.1
23.3
B
B
B
B
B
1 bulldozer
1 scraper (s)
1 bulldozer
2 scraper (s)
1 bulldozer
1 scraper (s)
2 bulldozers
2 bulldozers
2 scraper (s)
2 bulldozers
2 scraper(s)
- ==^==
Preliminary tests to determine lateral and downwind configuration of
sampling array.
Tests that were terminated because of shifting wind flow out of the
desired sector.
Readings are in degrees relative to true north.
2-4
-------
Monitoring Equipment
The high-volume samplers used to collect total suspended particulate
matter (TSP) data were Misco Model 680 samplers. These samplers are equipped
with constant flow controller* to maintain the airflow through the filter
medium at a relatively constant volume rate of 40 cfm.* At the colinear
stations 13, 24 and 33, an additional high-volume sampler equipped with a
Sierra size selective inlet (SSI) and a Sierra five-stage cascade impactor
(CZ) was rua simultaneously at approximately 20 cfm. This lower flow rate
was used to reduce particle bounce within the impactor head. The design
cut-point of 15 um at 40 cfm for the SSI was altered to about 16 um by
the change in flow rate. Sierra glass-fiber filters. Models C-230-GF and
C-305-GF, were used as collection substrates for the cascade impactor and
standard high-volume measurements, respectively.
In addition to the high-volume data, particle deposition measurements
were made using 20.3 centimeter (8-inch) diameter dustfall buckets. The
buckets were mounted on 1.8-meter tripods to prevent contamination by
saltating particles near the ground. Grab samples of coal were collected
from the source area during each test and sealed in glass jars. These
samples were used to determine moisture content and grain size distribution
of the source material.
Meteorological data were collected routinely during each test at station
24 and included sky cover, temperature, wind speed and wind direction. Wind
speed and wind direction were also monitored near the source area. A
Climatronics Mark III Wind System was placed at the 10-meter level on a
telephone pole located at station 24. A similar wind system, Climatronics
Electronic Weather Station, was mounted on a tripod 1.8 meters (6 feet) high
and placed near station 12. wind speed and direction data were recorded on
analog charts at both locations. A standard mercury thermometer and a Bendix
aspirated psychrometer system were used to obtain ambient temperature,
dew-point temperature and relative humidity data.
Analytical Methods
Glass-fiber filters were inspected for defects, numbered and stored for
24 hours in a desiccator. Filters were weighed before and after testing in a
controlled atmosphere where the temperature was between 20°C and 25®C,
and the relative humidity below 50 percent. Basults were verified for 10
percent of the filters, randomly selected, according to the criteria
presented in the EPA Quality Assurance Handbook.3
Each dustfall sample was filtered to determine the total settleaole
particulate matter. The total water soluble content of the samples was
determined by the evaporation of an aliquot of the sample. Both values were
corrected by using a control sample for the fluid used (water). The sum of
the total suspended solid matter and total water soluble matter was defined
as the total dry settleable particulate matter.
The moisture content of the bulk samples of coal was determined following
the American Society for Testing and Materials (ASTM) procedure D3173-73,
"Moisture in the Analysis Sample of Coal and Coke." Grain size distribution
was determined following ASTM procedure D410-38, "Sieve Analysis of Coal."
(*) 1 cfm = 0.0005 m3/s. 2-5
-------
Hourly average meteorological data were determined from the data
collected during each t«at. Average wind «p««d and wind direction values
were obtained by reducing the analog charts and were converted to
appropriate engineering units. Atmospheric stability conditions were
characterized as Pasquili-Gifford stability categories and estimated as
suggested by Turner.^
••suits and Discussion
The calculated emission rates varied among sampler* because of a
combination of random and systematic errors. Random error in the data
results from factors influencing data collection, data reduction procedures
and from the contamination of samplers by downwind sources within the
sampling array. These factors were present in all five tests, and
theoretically are manifest as the deviation among individual estimates of
the source emission rate. Systematic error, introduced through the use of
the Gaussian equation, was not quantifiable in this study because actual
source emission rates were not known. The accuracy of the predictive
equations was considered to be a factor of 2.
Surface Sediments
Surface sediments within the source area were characterized in terms of
moisture content and grain size distribution. Both of these parameters are
important because of their influence on the emission rate. As the moisture
content increases, particles bind together and the erodability of the
surface material decreases. The grain size distribution provides an
indication of the availability of particles in the finer fractions and also
the surface roughness which influences wind flow in the vicinity of
individual grains. The moisture content on test days was between 3 percent
and 4 percent by weight. The surface sediment ranged from clay to pebble
size fractions. About 50 percent of the surface material was larger than 4
millimeters and the quantity of material finer than 62.5 vm ranged between
0.5 percent and 5 percent.
Suspended Sediment Characteristics
Within the source area, surface sediment is mechanically broken into
smaller size fractions by the activity of bulldozers and scrapers. This
activity occurs routinely and prevents the development of a surface crust by
cementing agents or an ablative surface where fines are selectively winnowed
and coarser materials stabilize the surface. The finer fractions may be
eroded solely by winds, or by a combination of winds and vehicles perturbing
the surface and ejecting particles upward into regions of higher wind
velocity. Airborne particles will remain in suspension if wind turbulence
can overcome gravity effects. The modes of transport include surface creep,
saltation and suspension.
TRC's study focused only on material leaving the source area in the
suspension mode. This material was analyzed to determine the distribution
of grain sizes. (Histogram plots of samplers located at downwind distances
of 219 meters, 481 meters and 612 meters are presented in Figure 2.) The
data indicated that there was a paucity of material in the 1.3-um to
4.2-wm size range. Also, a slight decrease was observed in the greater-
than-10-um size range with increasing downwind distance. This trend,
2-6
-------
100
3
i
I*
i
FIRST ROW
100
•0
§
i
5
r
SECOND ROW
100
i
i
I
*
l
THIRD ROW
STAGS 1
10.2 » 1C *
4.2
2
10.2 j*
3 I
2.1 - «.2 !• I
1.3 - 2.1 IN
BAOCUP
0.0 - l.J
HIW-WUJHE STASE MO O>U)£5POmiM GRAIN SIZE
Figure 2. Grain size distribution for center!ine samplers,
2-7
-------
coupled with a slight increase in the finer fractions with downwind
distance, is believed to reflect deposition of coarser materials.
Observed dustfall with downwind distance is illustrated in Figure 3.
The bulk of the material was deposited within 50 meters of the source where
the average deposition rate was 11 grana/m* /min. The deposition rate
decreased to 1.7 grams/rf /ain and 1.0 grams/a* /min at downwind distances
of 160 meters and 380 meters, respectively.
Source Emission Strength
Bach TSP measurement was used to back-calculate source emission strength
using Gaussian diffusion equations. To solve these equations for the source
strength term, the source area was represented by a virtual point source
which was located using Turner's techniques. The results for the complete
tests (nos. 3, 6, 7, 8 and 9) are presented in Table IX. TSP concentrations
were corrected for background contamination by subtracting the upwind sample
concentration value from the value observed downwind. Background
concentrations were low (approximately 30 ug/m* ) and did not change
significantly during the test period. The downwind TSP concentration
decreased dramatically between samplers located at line 1 and line 2. The
apparent source strength values were also corrected to account for the
depletion of the dust cloud as a result of the dry deposition of airborne
particulate matter. The calculated correction factors for the three
downwind distances are 1.24, 1.37, and 1.46 for lines 1, 2 and 3,
respectively.
Estimation of Program Cost
Program costs were estimated by grouping those tasks required to attain
a specific objective and by developing man-hour and other direct costs for
each group. Table III illustrates how tasks were grouped to estimate the
costs associated with using the upwind-downwind test method. The relative
importance of each component identified in Table III varies among programs.
For example, the proximity of the monitoring site to the organization
completing the tests affects travel expenses. However, basic assumptions
may be made and costs developed for variously sized programs. Figure 4
compares the total program cost with the number of high-volume samplers used
to monitor the fugitive emission source. The increase in cost is a step
function owing mainly to the addition of field personnel to handle the
increased work load.
2-8
-------
400
SO-
300
100-
M
lib-
40 80
160 200 240 280
OMMIBO OISTMCT (mtun)
Figure 3. Comparsion of total dustfall collected by sampler with
downwind distance from source.
M
40
m
i *
1"
10
0
i
•
•
"
•
•
i
j
i
1
•
1
•
•
0 1 2 34 S 6 7 8 9 10 II 12 13 U IS IS 17 18
HI-VOC SMPUXS
Figure 4. Comparison of total program cost and the number of high-
volume samplers using the upwind-downwind method to
determine fugitive source emissions.
2-9
-------
TABU II. BatlMtad Bmlaalon B«taa (or *ba Aiaa Bouroa
DO
I
T«it fl fast 1C Taat ff Taat IB »•« ft
Dlatanoa Bat IM tad Dlatanca Batlaatad Dlatanoa BatlMtad Dlatanoa Batlaatad Dlatanoa Bit luted
ttom Bnlaalon fro* Bnlaalon titm Bataalon fro* Eal>«lon '«>• Balaalon
Bowroa* R*t« Bouioa* IUt* Bouroa* B»k* Bowroa* Bat* Booroa* Bat*
at. t Ion (») |q/a| |.| (4/a| (•! la/a) !•! |9/«1 !•! <9/«l
11 111,
12 HO.
i> lot.
14 2)5.
IS 2i«.
21 171.
22 421.
2) 4S1.
24
2S 470.
It 47a.
17.
7.
7.
IB.
It.
27.
It.
IB.
—
14.
32.
27
31 9B7.4) IB.
)) CBB.B l».
14 «B7.t IB.
35 4««.l 34.
177.
1»4.
211.
23B.
2J».
—
4M.
41*.
471.
472.
477.
—
•»^
•04.
iO>.
iO«.
SB.
34.
32.
4S.
S3.
—
sc.
47.
33.
34.
SS.
—
—
3*.
44.
2».
1B4.
201.
21*.
24t.
2t7.
3B3.
432.
4a3.
411.
4*1.
St.
37.
24.
4B.
ii.
S3.
St.
St.
42.
S3.
1B3.
201.
21».
247.
2tl.
3B1.
431.
4(2.
401.
403.
4Bt.
4»S.
—
•11.7 7S.4 til.
•1B.S 43.2 t2B.
«l».
13B.
70.
S7.
•4.
1«2.
IDS.
<7.
Si.
3».
57.
Ot.
IBS.
—
St.
t3.
1B4.
1B7.
20t.
134.
253.
175.
303.
434.
4tS.
40t.
40*.
4*B.
sot.
—
~
•27.
• 3B.
**.
71.
40.
•3.
04.
4t.
4B.
31.
3t.
J7.
SO.
•1.
—
—
IS.
St.
•Olatanoa (roa »ouroa* !• tka 41«tanoa !• aiatais batwaam tk« •aaylaca and the victual point
batwaan taata baoauaa tha poaltlon of tha virtual point aoucoa Is a (unotlon o( ataoaphado atabllttv
TklB
v«(l««
-------
TABLE III. Components of program cost
o Management and program administration
o Development of test plan
o Calibration and preparation of equipment
o Travel to/from the monitoring site; shipment of equipment
o Data collection
o Checking and cleaning equipment
o Data reduction
o Data analysis
o fteport preparation
Determination of Optimum Sampling Protocol
To determine the optimum sampling protocol, estimates of source emission
rates were analyzed using standard statistical techniques. A summary of the
results for each test is contained in Table IV. The values for skewness
range from a low of 0.066 for test 7 to a high of 0.662 for test 8, which
indicates a slight to moderate bunching of the smaller values about the
mean. The values for kurtosis range from a low of -1.244 for test 6 to a
high of 1.485 for test 7. Hone of the values for kurtosis are indicative of
a significant departure from normality.
TABLE IV. Statistical analysis of emission rate estimates
Number Standard Coefficient
Test of Mean Deviation of
Number Observations (g/s) (g/s) Variation Skevness Kurtosis
3
6
7
8
9
14
12
12
15
14
29.9
57.4
68.8
113.0
77.1
13.6
12.6
19.8
41.8
28.1
45.6
21.9
28.7
37.0
36.4
0.382
0.102
0.066
0.662
0.639
-0.096
-1.244
1.485
-0.321
-1.047
The wide range of values exhibited by the coefficient of variation
indicates that the mean and standard deviation do not change together, and
for that reason the data from the five tests could not be combined to obtain
the desired solution. Instead each test was analyzed independently and the
results were compared to determine the variability of relative accuracy for
sampling arrays of different size. The data values from the five tests were
combined, however, to determine whether the sample population exhibited a
normal distribution and whether any significant differences exist among the
data values because of the techniques employed in estimating the source
emission rate.
2-11
-------
Test for normality. Because the mean of the emission rate values wert
different for each test, the values were standardized separately for each
test using the following expression:
Q ~ Q
Zt m-H - 1 i - 1, 2, 3r ... n U>
where:
Q. . represents the emission rate values
is the sample mean for the Q values for test j
S . is the sample standard deviation of Q values in test j
n. vas the Dumber of Q values in test j
The standardized values for the five tests were combined and grouped in nint
categories. A chi-square -goodness-of-fit' test indicated that the Zj/s
were values from a normal distribution.
Analysis of variance. A univariate analysis of variance was also
performed on the data values obtained for each test. The array of
high-volume samplers was divided into three lines and the samplers in each
line grouped into five positions. This division resulted in the layout shown
in Table V.
TABLE V. Layout of samplers by sampler -ID number
Line
1
2
3
1
11
21
31
2
12
22
32
Position
3
13
23,24,25
33
4
14
26
34
5
15
27
35
For each test, the following model for the Q values was employed:
QijK - 1 +ai +flj * WJij •»• eijk i - 1, 2, 3
j • 1, 2, 3, 4, 5 (2)
k • 1,
where:
7 is the general mean
ai is the effect due to line i
8 j is the effect due to position j
2-12
-------
(aft ) ij is the effect due to the interaction of line i with
position j
*ijk I* *&* error term
i* the number of observations in line i and position j
The e^ values are assumed to be independent random variables normally
distributed with a aean of zero and a variance of a1 . Also, all three
effects are assumed to be fixed effects.
An analysis of variance was performed for the five tests (3, 6, 7, 8 and
9). For all tests except test 9, there was no significant effect because of
the interaction of line and position. An analysis of variance was also
performed on the data for all five tests combined. For this case there was
not a significant effect because of the interaction of line and position.
There was also no difference among the lines, but there was a significant
difference among the positions. Duncan's multiple range test was performed
on the position means using an alpha level of 0.05 (se« Table 71). 4 This
test showed that position 5 was significantly different from positions 2 and
3. 'For all tests combined, the ranking of the position means from largest to
smallest was 5, 1, 4, 2 and 3. This ranking indicates a trend for the outer
positions in the rows to have a higher mean than the inner positions.
TABLE VI. Duncan's multiple range test for estimated emission rates
Sampler Mean Value of Estimated Emission Rates (g/s) All Tests
Grouping Test 3 Test 6 , Test 7 Test 8 Test 9 Combined
All samplers 29.9 57.4 68.8 113. 0 77.1 70.1
Line 1
Line 2
Line 3
Position 1
Position 2
Position 3
Position 4
Position 5
17.5
33.7
40.6
38.4
17.6
24.6
31.7
37.3
53.7
62.4
55.2
62.0
60.0
51.4
65.9
54.9
58.7
71.8
86.6
73.3
62.2
69.8
61.5
82.8
131.9
101.2
109.0
157.8
95.0
72.7
109.2
166.1
95.8
65.9
69.6
93.7
71.8
51.8
74.5
105.9
71.5
69.3
69.1
82.7
61.3
55.4
69.1
98.4
The difference between the positions within the sampling array suggests
that representing the area of fugitive emissions by a virtual point source
does introduce a bias into the estimates of emission strength. Samplers
located farthest from the plume centerline estimate higher source emission
rates than does the population mean. Conceivably, the source area might b«
better represented by a line source with segments weighted differently to
reflect the variability in emissions within the source area. An iterative
process may be used to determine the proper weighting factors for each line
segment to eliminate the bias in the data set. However, this technique would
require a different description for each test of each source with the result
2-13
-------
that determining the emission characteristics of the line source would be
difficult for programs that attempt to use a reduced number of samplers.
Relative Accuracy Os ing Reduced Humber of Samplers. It was desired to
determine the number of samplers that would be required to estimate the mean
within plus or minus (100 z r) %, where r is a number between 0 and 1. The
coefficient of variation was employed to obtain the desired estimates. The
coefficient of variation from each test was treated as a datum point. The
mean of five values was 0.339 and the standard error of this mean was 0.112.
A 95 percent confidence interval for the mean coefficient of variation may be
obtained by the following expression:
where:
x is the mean coefficient of variation
ta/2, v is the value of Students' t -distribution with a
probability level of a/2 and degrees of freedom, v
s/V"n"is the standard error of the mean
For this case a » 0.05
v • n-1 « 4, and the value of t is 2.776.
The lower bound of the 95 percent confidence interval on the mean is. 0.228
and the upper bound is 0.451.
•j.
For any number of samplers (n) the following equation may be solved for a
value of r:
m ^72, n-1 (c.v.)
/"n~" * '
where:
e.v. is a value of the coefficient of variation
Using successive values of n from 4 to 25 inclusively, values of r were
obtained for the mean coefficient of variation in addition to the upper and
lower confidence bounds for the coefficient of variation. To illustrate the
results in terms of relative accuracy, the number of high-volume samplers (n)
were tabulated with values of 1-r (see Table VII) . The total program cost
for incremental increases in program size were compared with relative
accuracy using the mean coefficient of variation. These results are
presented in Figure 5 and show that the most cost-effective program would use
10 downwind samplers.
Conclusions
A study was conducted to determine the effect of varying sampling network
size on the cost and relative accuracy of the information obtained. The
results of the study indicate that the most cost-effective program would use
10 downwind samplers and would provide, when deployed in accordance with the
guidelines of Kolnsberg, an estimate of emission strength accurate to within
2-14
-------
approximately 25 percent of the value possible using 30 or acre
samplers. These results, however, should only be assumed valid for
similar fugitive emission sources. Additional work is necessary to determine
the minimum size sampling network for more complicated source areas or
prevailing meteorological conditions.
TABLE VII. Values of 1-r for « of 0.05
n
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
3.182
2.776
2.571
2.447
2.365
2.306
2.262
2.228
2.201
2.179
2.160
2.145
2.131
2.120
2.110
2.101
2.093
2.086
2.080
2.074
2.064
2.056
Lower
Limit
C.V. - 0.228
.637
.717
.761
.789
.809
.825
.837
.847
.855
.862
.868
.874
.879
.883
.887
.890
.893
.896
.899
.901
.904
.906
1-r Values
Mean
C.V. • 0.339
.461
.579
.644
.689
.717
.739
.758
.772
.785
.795
.804
.812
.819
.826
.831
.837
.841
.846
.850
.853
.857
.861
Upper
Limit
C.V. • 0.451
.282
.440
.527
.583
.623
.653
.677
.697
.713
.727
.740
.750
.760
.768
i
.776
.783
.789
.794
.800
.805
.810
.815
2-15
-------
to
0>
Ul
»—I
t_l
•-•
s v
i
1.0
0.0
0.6
0.4
0
20
n
''
30 40
TOTAL PROGRAM COST X 1(T
• "
50
Figure 5. Comparison of relative accuracy (utilizing mean coefficient of variation)
to total program cost. Numbers denote the quantity of downwind samplers.
-------
References
1. O.B. Turner, Horkaook of Atmospheric Disunion Estimates. U.S.
Environmental Protection Agency AP-26. 1970. (NTIS PB191482.)
2. O.B. Slade, Ed., Meteorology and Atomic Energy. U.S. Atomic Energy
Cosniaaion, July 1363. 445 pp.
3. Environa«nt*l Protection Agency, Quality A««urance Handbook for Air
Pollution Measurement Systems. EPA-600/4-77-027A.1S77.(NTIS PB273518.)
4. O.B. Duncan, *t-tests and interrala for comparisons suggested by the
data.' Biometrics 31: 339-59 (1975).
2-17
-------
CHARACTERIZATION OF FINE PARTICULATE EMISSION FACTORS
FOR PAVED ROADS
by
Chatten Cowherd, Jr.
Phillip J. Englehart
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
3-1
-------
ABSTRACT
This paper presents the results of an expanded measurement program to
develop emission factors for particulate emissions generated by traffic
entrainment of paved road surface particulate matter. The emission sampling
procedure used in this program provided emission factors for the following
particle size ranges: < 15 ym, < 10 pm, and < 2.5 urn aerodynamic diameter.
Testing was performed at sites in the Kansas City and St. Louis areas. These
sites were representative of significant urban paved road emission sources
within the following land use categories: commercial/industrial, commercial/
residential, expressway, and rural town.
The measured inhalable particulate emission factors ranged from 0.06 to
8.8 g/VKT. Lowest emissions were measured for the "expressway" road category;
highest emissions were measured for the "rural town" road category.
Approximately 90% of the IP emissions (< 15 urn aerodynamic diameter) consisted
of particles smaller than 10 ym in aerodynamic diameter, and approximately 50%
of the IP emissions consisted of particles smaller than 2.5 ym in aerodynamic
diameter.
Correlation analysis of particulate emissions with parameters
characterizing the source conditions showed the existence of a relatively
strong positive relationship between intensity of emissions and roadway
surface silt loading. This relationship was used as the basis for derivation
of predictive emission factors for each particle size range. The equation for
IP emissions was found to represent measured IP emissions more accurately over
a much larger range of values than does the AP-42 single-valued factor.
To facilitate the use of these particle size specific equations in the
development of emission inventories, a classification system of mean or
typical silt loadings as a function of roadway category was derived. These
mean silt loadings were then inserted into the respective emission factor
equations to derive a matrix of emission factors for specific roadway
categories and particle size fractions.
3-2
-------
INTRODUCTION
Traffic-entrained particulate from paved roads has been identified as a
major cause of nonattainment of air quality standards for total suspended
particulates (TSP) in urban areas.1 Therefore, the quantification of this
source is necessary to the development of effective strategies for the
attainment and maintenance of the TSP standards, as well as the anticipated
standard for particles smaller than 10 micrometers (urn) in aerodynamic
diameter.
Few data on directly measured dust emissions from paved streets are
available in the literature. An isolated study of dust emissions from a paved
road in the Seattle area yielded an emission factor of 0.83 Ib/vehicle-mile at
20 mph.2»3 The test road was noticeably dusty and had no curbs or street
cleaning program; it was located adjacent to gravel roads and unpaved parking
lots from which dirt was tracked. Dust emissions generated by vehicular
traffic with average daily traffic exceeding 200 vehicles were estimated to
equal the amount removed by sweeping every 2 weeks.3
A single-valued emission factor of 3.7 g/vehicle-kilometer for dust
entrainment from paved roads was developed from another field study.1*
Emission measurements were obtained using the upwind-downwind technique with
high-volume samplers. Thirty-five successful tests were completed. It was
determined through microscopy that 78% (by weight) of the emissions consisted
of particulate less than 30 ym in size. Also through optical microscopy, it
was found that 59% of the particulate collected was mineral matter, while 40%
consisted of combustion products. It was also concluded in this study that
particulate emissions from a street are proportional to traffic volume but
independent of street surface dust loading.
In a third field study, quantitative emission factors for dust
entrainment from paved urban roads were developed using exposure profiling.s
Field testing was conducted at three representative sites in the Kansas City
area. At one location, controlled amounts of pulverized top soil and gravel
fines were applied to the road surface. Eight tests were performed at the
artificially loaded site, and five tests were made at a different site under
actual traffic conditions. Emissions were found to vary directly with traffic
volume and surface loading of silt (fines). The dust emission factor for
normally loaded urban streets ranged from 1 to 15 g/vehicle-kilometer,
depending upon land use. Approximately 90% of the emissions (by weight) were
found to be less than 30 ym in Stokes diameter and 50% less than 5 ym in
Stokes diameter, based on a particle density of 2.5 g/cm3. Measured emission
factors for street particulate reentrainment added to vehicle exhaust were
found to be an order of magnitude larger than the factors for vehicle exhaust
alone.6
This document presents the results of an expanded measurement program to
develop particulate emission factors for paved roads. The emission sampling
procedure used in this program provided emission factors for the following
particle size ranges.
3-3
-------
IP = Inhalable particulate matter consisting of particles smaller than
15 ym in aerodynamic diameter
PM-10 = Particulate matter consisting of particles smaller than 10 ym in
aerodynamic diameter
FP = Fine particulate matter consisting of particles smaller than
.2.5 ym in aerodynamic diameter
Results are presented for winter testing in the Kansas City, Missouri, area
and spring testing in areas of St. Louis, Missouri, and Granite City,
Illinois. These results are used as a basis for the derivation of a matrix of
emission factors for specific road categories and particle size ranges.
SAMPLING SITE SELECTION
Eight candidate sampling areas in Kansas, Missouri, and Illinois were
designated by the Environmental Protection Agency (EPA) as representative
sites for the field study. These areas represented a range of typical road,
traffic, geographical, and environmental conditions within residential,
commercial, and industrial land uses. Each sampling area contained a TSP
monitoring site providing historical air quality data.
A wide variety of road and traffic characteristics was found in the
presurveys of these areas. Equivalent hourly traffic volume ranged from
36 vehicles to 2,944 vehicles. Road width varied from 22 to 216 ft. Both
asphalt and concrete street surfaces, curbed and uncurbed, were included.
Street surface conditions ranged from smooth to rough, and surface particulate
loadings varied from light to heavy in comparison with typically observed
loadings.
Three major criteria were used to determine the suitability of each
candidate site for sampling of road dust emissions by the exposure profiling
technique.7
1. Adequate space for sampling equipment.
2. Sufficient traffic and/or surface dust loading so that adequate mass
would be captured on the lightest loaded collection substrate during a
reasonable sampling time period.
3. A wide range of acceptable wind directions, taking into account
(a) the street orientation relative to the predominant wind directions for the
locality, and (b) upwind obstacles (houses, buildings, or trees) to free wind
flow.
Although roads with light traffic were excluded from consideration, such roads
probably do not contribute substantially to total emissions of traffic
entrained" dust in urban areas.
3-4
-------
Based on the above criteria, nine sites were' selected for this testing
program:
Kansas City Area - three sites
7th Street in Kansas City, Kansas (commercial/industrial)
Volker Boulevard/Rockhill Road in Kansas City, Missouri (commercial/
residential)
4th Street in Tonganoxie, Kansas (rural town)
St. Louis, Missouri - two sites
1-44 (expressway)
Kingshighway (commercial/residential)
Granite City. Illinois - two sites
Madison Street (commercial/residential)
Benton Road (commercial/residential)
SAMPLING EQUIPMENT
A variety of sampling equipment was utilized in this study to measure
partfculate emissions, roadway surface particulate loadings, and traffic
characteristics.
The primary tool for quantification of emissions was the MRI exposure
profiler, which was developed under EPA Contract No. 68-02-0619.7 The
profiler consisted of a portable tower (4 m height) supporting an array of
four sampling heads. Each sampling head was operated as an isokinetic total
particulate matter exposure sampler directing passage of the flow stream
through a settling chamber (trapping particles larger than about 50 ym in
diameter) and then upward through a standard 8- by 10-in. glass fiber filter
positioned horizontally. Sampling intakes were pointed into the wind, and
sampling velocity of each intake was adjusted to match the local mean wind
speed, as determined prior to each test. Throughout each test, wind speed was
monitored by recording anemometers at two heights, and the vertical profile of
wind speed was determined by assuming a logarithmic distribution. Normally,
the exposure profiler was positioned at a distance of 5 m from the downwind
edge of the road.
The recently developed EPA version of the size selective inlet (SSI) for
the high volume air sampler was used to determine the IP concentrations. To
obtain the particle size distribution of IP, a high-volume parallel-slot
cascade impactor (CI) with greased substrates was positioned beneath the
SSI. This five stage cascade impactor has, at a flow rate of 40 SCFM, 50%
efficiency cutpoints at 7.2, 3.0, 1.5, 0.95, and 0.49 ym aerodynamic
diameter. SSIs fitted with high-volume cascade impactors were placed at 1-
and 3-m heights to determine the respective IP and FP mass fractions of the
total particulate emissions.
3-5
-------
Standard high-volume air samplers' were used to measure TSP matter
consisting of particles smaller than about 30 ym in aerodynamic diameter.
These samplers were operated at a height of 2 m.
The basic upwind equipment included SSIs and a standard high-volume air
sampler. In the Kansas City testing, two SSIs at heights of 2 and 4 m were
used to obtain the IP concentration of upwind particulate matter. In the
St. Louis testing, the primary upwind equipment included a high-volume air
sampler and an SSI/CI with greased substrates. For the secondary deployment
array, two SSIs were used to obtain the vertical distribution of IP.
Samples of the dust found on the roadway surface were collected during
the source tests. In order to collect this surface dust, it was necessary to
close each traffic lane for a period of approximately 15 min. Normally, an
area that was 3 m by the width of a lane was sampled. For each test,
collection of material from all travel lanes and curb areas (extending to
about 25 to 30 cm from the curbing) was attempted. A hand-held portable
vacuum cleaner was used to collect the roadway dust. The attached brush on
the collection inlet was used to abrade surface compacted dust and to remove
dust from the crevices of the road surface. Vacuuming was preceded by broom
sweeping if large aggregate was present.
The characteristics of the vehicular traffic during the source testing
were determined by both automatic and manual means. The vehicular
characteristics included: (a) total traffic count, (b) mean traffic speed,
and (c) vehicle mix.
Total vehicle count was determined by using pneumatic-tube counters. In
order to convert the axle counts to total vehicles, visual 1-min vehicle mix
summaries were tabulated every 15 min during the source testing. The vehicle
mix summaries recorded vehicle type, number of vehicle axles, and number of
vehicle wheels. From this information, the total axle counts were corrected
to the total number of vehicles by type.
The speed of the traveling vehicles was determined by noting the posted
speed limits of the roadway test section. As a check against this
determination method, speeds of the vehicles were determined through the
occasional use of a hand-held radar gun. The weights of the vehicle types
were estimated by consulting automobile literature and distributors of medium-
duty and semi-trailer type trucks.
SAMPLING AND ANALYSIS PROCEDURES
The sampling and analysis procedures employed in this study were subject
to the Quality Control guidelines which met or exceeded the requirements
specified by EPA.a.9 As part of the QC program for this study, routine audits
of sampling and analysis procedures were performed. The purpose of the audits
was to demonstrate that measurements were made within acceptable control
conditions for particulate source sampling and to assess the source testina
data for precision and accuracy. Examples of items audited include
3-6
-------
gravimetric analysis, flow rate calibration, data processing, and emission
factor calculation.
Participate samples were collected on type A slotted glass fiber impactor
substrates and on type AE (8- by 10-in.) glass fiber filters. To minimize the
problem of particle bounce, the glass fiber cascade impactor substrates were
greased. The grease solution was prepared by dissolving 140 g of stopcock
grease in one liter of reagent grade toluene. No grease was applied to the
borders and backs of the substrates. The substrates were handled,
transported, and stored in specially designed frames which protected the
greased surfaces.
Prior to the initial weighing, the greased substrates and filters were
equilibrated for 24 hr at constant temperature and humidity in a special
weighing room. During weighing, the balance was checked at frequent intervals
with standard weights to assure accuracy. The substrates and filters remained
in the same controlled environment for another 24 hr, after which a second
analyst reweighed them as a precision check. Substrates or filters that could
not pass audit limits were discarded. Ten percent of the substrates and
filters taken to the field were used as blanks. Paper bags for the vacuum
cleaner were conditioned and tared in a similar manner.
Prior to equipment deployment, a number of decisions were made as to the
potential for acceptable source testing conditions. These decisions were
based on forecast information obtained from the local U.S. Weather Service
office. A specific sampling location was identified based on the anticipated
wind direction. Sampling would be initiated only if the wind speed was
forecast between 4 and 20 mph. Sampling was not planned if there was a high
probability of measurable precipitation (normally > 20%) or if the road
surface was damp.
Sampling usually lasted 4 to 6 hr. Occasionally, sampling was
interrupted due to occurrence of unacceptable meteorological conditions and
then restarted when suitable conditions returned. The unacceptable
meteorological conditions most frequently encountered consisted of light winds
(below 4 mph) and insufficient angle (< 45 degrees) between mean (15-min
average) wind direction and road direction.
To prevent p'articulate losses, the exposed media were carefully
transferred at the end of each run to protective containers within the MRI
instrument van. Exposed filters and substrates were placed in individual
glassine envelopes and numbered file folders and then returned to the MRI
laboratory. Particulate that collected on the interior surfaces of each
exposure probe was rinsed with distilled water into separate glass jars.
When exposed substrates and filters (and the associated blanks) were
returned from the field, they were equilibrated under the same conditions as
the initial weighing. After reweighing, 10% were audited to check precision.
The vacuum bags were weighed to determine total net mass collected. Then
the dust was removed from the bags and was dry sieved. The screen sizes used
for the dry sieving process were the following: 3/8 in., 4, 10, 20, 40, 100,
3-7
-------
140, and 200 mesh. The material passing a 200 mesh screen is referred to as
silt content.
The vertical distributions of the product of plume concentration and mean
wind speed were numerically integrated to calculate emission factors. The
size selective inlet/cascade impactor sampler combinations provided reliable
point concentrations for IP and finer particle size fractions. Plume height
was determined by extrapolation of the vertical profile of total particulate
concentration as measured by the MRI exposure profiler.
TEST RESULTS
The winter testing was conducted during the months of February and March
1980 at three sites in the Kansas City area. The spring testing was conducted
during the month of May 1980, at two sites in St. Louis and at two closely
spaced sites in Granite City. Illinois.
The source tests were evaluated according to established QA criteria.
Seven of the nine Kansas City tests met all of the QA criteria, while only
three of the ten tests conducted in the St. Louis, Granite City area met the
QA criteria. The spring testing, in particular, was hampered by unseasonably
light wind conditions. Wind speed for four of the ten spring tests did not
meet the minimum wind speed criterion of 4 mph.
The results of the 10 runs which met the QA criteria were used as input
to Multiple Linear Regression (MLR) analysis (see below). These runs are
subsequently referred to as the "MLR" data set.
During each emissions sampling run and at other times when emissions
sampling was not being conducted, samples of street surface particulate were
collected to determine total particulate loadings and silt percentages. Silt
loadings on active travel lanes ranged from about 0.022 g/m2 on a freeway
(1-44) to more than 2.5 g/n)2 on a lightly traveled rural road in Tonganoxie.
As expected, loadings in curb areas substantially exceeded loadings in travel
lanes. The range of day-to-day variations in loadings at a given site was
generally within a factor of 2. Higher loadings tended to occur after a
precipitation event.
The upwind and downwind particulate mass concentrations for the various
particle size fractions measured during the field program were analyzed to
determine representative mass fraction ratios. The IP concentration measured
downwind of the test road segment was found to decrease with height. At a
sampling height of 2 m, the mean ratio of downwind IP to TSP concentration was
0.45 (a = 0.14), and the corresponding mean upwind ratio was 0 54 (0 =
0.18). This indicates that background TSP, although lower in concentration
contains a higher percentage of IP. Similar differences are also evident in
the mean upwind versus downwind PM-10 to TSP ratios and FP to TSP ratios.
The mean downwind ratio of FP to IP was 0.52 (o = 0.098) while the mean
upwind ratio was 0.53 (a = 0.085). This finding implies that there is no
3-8
-------
significant enrichment of fine particles attributable to the paved road
source.
Table 1 summarizes, by land use category and test series quality, the
emission factor data. As can be seen, the smallest emission factors were
measured in the freeway category which also had the lowest surface silt
loadings. The highest emission factor was measured in the rural town category
which showed a correspondingly high surface silt loading.
Intercomparison of emission factors by land-use category indicates that
relative to the mean expressway IP emissions: (a) mean commercial/residential
IP emissions were approximately 10 times larger; (b) commercial/industrial
emissions were approximately 20 times larger; and (c) the rural town roadway
produced IP emissions that were roughly 60 times larger. Relative to mean
expressway silt loading: (a) the silt loading for commercial/residential
roadways was approximately 25 times higher; (b) the silt loading for
commercial/industrial roadways was roughly 15 times higher; and (c) silt
loading on the rural town roadway was approximately 115 times higher.
MULTIPLE REGRESSION ANALYSIS
Stepwise MLR was the method used to evaluate independent variables for
possible use as correction factors in a predictive emission factor equation.
MLR is a statistical technique available in the Statistical Package for the
Social Sciences (SPSS).10 Because it was desirable to have multiplicative
rather than additive correction factors in the emission.factor equations, all
independent and dependent variable data were transformed to natural logarithms
before being entered in the MLR program.
The independent variables evaluated initially as possible correction
factors were silt loading (g/m2), total loading (g/m2), average vehicle speed
(Kph), and average vehicle weight (tonnes). The rationale for including
measures of roadway particulate loading stems from findings of an earlier MRI
programs which indicated that the magnitude of roadway emissions was directly
related to variations in surface loadings. The vehicle parameters—mean
weight and speed—were included largely by analogy to MRI's unpaved road
equation,11 although it was recognized that the dust generation mechanism for
paved roads may differ from that for unpaved roads. The moisture content of
the road surface particulate was not included as a correction parameter
because of the difficulty of collecting a sample without altering its moisture
content.
Preliminary MLR analysis of the entire data set indicated that all the
independent variables except vehicle weight were highly intercorrelated.
Although the stepwise algorithm would include vehicle speed first in a
predictive equation, silt loading and total loading showed essentially the
same correlation with IP emissions (r = 0.60). In other words, the variables
represent a common set of source conditions—either low vehicle speed, high
surface loading and emissions; or high vehicle speed, low loading and
emissions.
3-9
-------
TABLE 1. SUMMARY OF PAVED ROAD EMISSION FACTORS
TQCI. IP emission factor
Land use
category
Commercial/
industrial
Commercial/
residential
Expressway
Rural town
a RSD (relative
DH Iv*a1 at i wo
series
quality
All tests
MLR tests
All tests
MLR tests
All tests
MLR tests
All tests
MLR tests
standard
rlowi at-
-------
The decision was made to use silt loading rather than total loading or
vehicle speed in the development of the emission factor equation from the
"MLR" data set. This decision was based on the perception that (a) silt
loading is the most physically plausible indicator of the magnitude of IP
emissions, and (b) it will yield more reproducible results in independent
applications than total loading, a parameter which can be biased by the
presence of large particles (i.e., gravel).
Including silt loading as the primary predictor effectively precluded
total loading or vehicle speed from entering the equation for the "MLR" data
set. This follows from the high intercorrelations (multicollinearity)
mentioned above. Examination of the regression statistics indicated that
inclusion of vehicle weight as a second correction parameter could not be
justified.
The raw MLR equation for the "MLR" data set was as follows:
•IP
elp * 4.37 (sL)0'9 (1)
where:
eTP = IP emission factor expressed in grams per vehicle kilometer
traveled (g/VKT)
sL = Silt loading of road surface particulate matter expressed
in grams per square meter (g/m2)
This equation explained 73% of the variation in the emission factors. As
noted earlier, the "MLR" data set did contain data from all the land use
categories sampled during the field program.
The comparable predictive IP emission factor equation normalized to a
typical value for silt loading was as follows:
eiP = 2'54 < -07T )°'8 (2)
The emission factor equation was found to predict the "MLR" series test
data with a precision factor of 2.0. The precision factor (f) for an emission
factor is defined such that the 68% confidence interval for a predicted value
(P) extends from P/f to Pf. The precision factor is determined by
exponentiating the standard deviation of the differences (standard error of
the estimate) between the natural logarithms of the predicted and observed
emission factors.
The precision factor may be interpreted as a measure of "average" error
in predicting IP emissions from the regression equation. Assuming that the
actual IP emission factors are normally distributed about the regression line,
it can be stated that approximately 68% of the predictions are within a factor
of 2. The effective outer bounds of predictability are determined by
exponentiating twice the standard error of the estimate. The resultant
estimate of predictive accuracy, in this case 4.0, then encompasses
approximately 95% of the predictions.
3-11
-------
To put the precision factor of the IP predictive emission factor equation
emission factor into perspective, two comparisons were undertaken utilizing
the single-valued emission factor found in the current AP-42 manual.s
However, before valid comparisons could be made, it was necessary to convert
the AP-42 single-valued factor, which represents TSP emissions, to an
approximate IP emission factor. This was accomplished by multiplying the
AP-42 value by 0.4 which is the mean ratio of net IP (downwind minus upwind)
to net TSP concentrations as determined from the data collected in this study.
The first comparison involved the calculation of a precision factor for
the AP-42 data set. The resulting value of 2.1 is a measure of the ability of
the single-valued factor to represent the 40 pieces of data which were
averaged originally to produce the AP-42 factor. The second comparison
involved the calculation of a precision factor using the single-valued AP-42
factor to represent the "MLR" data set, as collected in this study. This
comparison yielded a precision factor of 4.4.
The precision factors and the range of the data values (emission factors)
upon which they are based are presented graphically in Figure 1. The ideal
model has a precision factor of 1.0, implying that each predicted value is
identical to the corresponding observed value, over an infinite range of
emission factors. The most important conclusion that can be drawn from
Figure 1 is that the emission factor equation, though far from ideal, does
predict IP emissions more accurately over a much greater range of values than
does the AP-42 single-valued factor over a considerably smaller range of data
values corresponding to the AP-42 data set. Furthermore, application of the
single-valued AP-42 factor to represent the wide range of IP emissions from
paved roads, as measured during this program, yields a precision factor which
is more than double (4.4 versus 2.0) that associated with the predictive
equation. This ability of the predictive equation to more accurately
represent variations in IP emissions is directly attributable to the
relatively strong relationship between roadway surface silt loading and IP
emissions.
Though not the primary focus of the program, it was possible to develop
predictive emission factor equations for the PM-10 and FP particle size
fractions using the same procedure as that applied in developing the equation
for IP. Derivation of TSP emission factors for use in developing a predictive
equation required different initial calculations, since only two TSP samplers
(one upwind, one downwind) were operated during the measurement phase of the
program. In essence, the initial calculation involved multiplication of the
IP emission factor for each run in the "MLR" series data set by the
corresponding net ratio of TSP to IP concentration as measured by appropriate
samplers. This procedure assumes that the TSP/IP ratio is constant over the
vertical extent of the plume.
The general form of the emission equation factors, applicable to all
particle size fractions, is as follows:
p
e = k (§75) (metric)
p
(3)
(nonmetric) M\
3-12
-------
- 3
o
1
2
Q_
2
AP-42 Applied to Present Data
\
Present Regression Equation
AP-42 Emission.Factor
Ideal Model
1 I I J I I III I I I I I I I ll I I 1 I I I I I
0.01 0.1 1 10
Emission Factor (g/VKT)
Figure 1. Comparison of emission factor precision
3-13
-------
The base emission factor coefficients (k, K), exponent (P), and precision
factor for each size fraction are listed in Table 2. For the metric equation,
silt loading is expressed as grams per square meter; for the nonmetric
equation, silt loading is expressed as grains per square foot.
TABLE 2. PAVED ROAD EMISSION FACTOR EQUATION PARAMETERS
(by particle size fraction)
Particle size fraction k (g/VKT) K (Ib/VMT) P Precision factor3
TSP
IP
< 10 um
FP
5.87
2.54
2.28
1.02
0.0208
0.0090
0.0081
0.0036
0.9
0.8
0.8
0.6
2.4
2.0
2.2
2.2
a Represents the interval encompassing 68% of the predicted values.
It should be noted that the tendency for the power term in the equation
to increase with larger particle size fraction is generally consistent with
MRI's previous paved road equation in which silt loading to the 1.0 power was
employed to account for variations in TSP emissions.
EMISSIONS INVENTORY APPLICATIONS
For the majority of emissions inventory applications involving urban
paved roads, actual measurements of silt loading will probably not be made.
Therefore, in order to facilitate the use of the previously described
equations, it is necessary to characterize silt loadings according to a
parameter(s) more readily available to persons developing emissions
inventories. After examination and analysis of silt loading and traffic data
collected during relevant MRI sampling programs, as well as surface loading
data gathered in connection with an extensive study of urban water pollution
the decision was made to characterize variations in silt loading based upon a
roadway classification system. This roadway classification system is
presented in Table 3. This system generally corresponds to the functional
classification systems employed by transportation agency personnel- and thus
the data necessary for an emissions inventory—number of road miles oer road
category and traffic counts—should be easily obtainable
3-14
-------
TABLE 3. PAVED ROADWAY CLASSIFICATION
Average daily traffic
Roadway type (ADT) No. of lanes
Freeways/expressways
Major streets/highways
Collector streets
Local streets
> 10,000
> 10,000
500-10,000
< 500
> 4
> 4
2a
2b
J Total roadway width > 32 ft.
D Total roadway width < 32 ft.
It should be recalled that traffic volume is not the only factor
affecting roadway silt loadings. For all roadways that provide access to
immediately adjacent areas, land use, particularly as it relates to the
potential for mud and dirt "tracking," is important. Silt loadings may also
be affected by street surface type and condition, the presence or absence of
curb, as well as public works practices and season of the year. However,
given the present data base, it is not possible to incorporate relationships
between these factors and silt loadings in a manner applicable to the majority
of emissions inventories.
The data base, made up of 44 samples collected and analyzed according to
the procedures outlined above, may be used to characterize the silt loadings
for each roadway category. These samples, obtained during MRI field sampling
programs over the past 3 years, represent a broad range of urban land use and
roadway conditions. Geometric means for this data set are broken out by
sampling location (i.e., city) and roadway category in Table 4.
The sampling locations can be considered representative of most large
urban areas in the United States with the possible exception of those located
in the Southwest. Except for the collector roadway category, the overall mean
silt loadings do not vary greatly from city to city, though the St. Louis mean
for major roads is somewhat lower than the other four cities. The substantial
variation within the collector roadway category is probably attributable to
the deposition effects of land use associated with the specific sample
locations. It should also be noted that an examination of data collected at
three cites in Montana during early spring indicates that winter road sanding
may produce loadings five to six times higher than the means of the loadings
given in Table 4 for the respective road categories.
3-15
-------
TABLE 4. SUMMARY OF SILT LOADINGS (g/raz) FOR URBAN PAVED ROADWAYS3 BY CITY
Roadway category
Local Collector Major Overall
City Xgb n Xg n Xg n Xg n
Baltimore0 1.42 2 0.72 4 0.39 3 0.68 9
Buffalod 1.41 5 0.29 2 0.24 4 0.56 11
Granite City (111.)6 - - - 0.82 3 0.82 3
Kansas City6 - - 2.11 4 0.41 13 0.60 17
St. Louis - - - 0.16 3 0.16 3
Overall 1.41 7 0.92 10 0.36 26 -
a Freeway/expressway data not included; only one value (0.022 g/m2) obtained.
Xg's are geometric means based on the corresponding n sample size.
~. Reference 12.
d Reference 13.
e From this study.
Table 5 presents the emission factors broken out by roadway category and
particle size. These were obtained by inserting the above mean silt loadings
into the emission factor equations with parameters defined in Table 2. These
emission factors can be utilized directly for emission inventory purposes. It
is important to note that the current AP-42 paved road emission factors* for
TSP agree quite well with those developed in this study. For example, those
cited in connection with MRI's previous testings were conducted at two roadway
sites in the major street and highway category. Those tests yielded a mean
TSP emission factor of 4.3 g/VKT versus 4.4 g/VKT as determined from the data
presented here.
SUMMARY AND CONCLUSIONS
The purpose of this study was to quantify inhalable particulate emissions
generated by traffic entrapment of paved road surface particulate matter
Paved road source testing was performed at sites representative of significant
emission sources within a broad range of urban land-use categories.
o -,-, Th«,,,;easured 1"halable Particulate emission factors ranged from 0.06 to
8.77 g/VKT. Lowest mean emissions were measured for the "expressway" road
category; highest mean emissions were measured for the "rural town"
category Approximately 90% of the IP emissions consisted of particles
smaller than 10 mn in aerodynamic diameter, and approximately 50% of the IP
emissions consisted of particles smaller than 2.5 ym in aerodynamic diameter.
3-16
-------
TABLE 5. RECOMMENDED EMISSION FACTORS FOR SPECIFIC ROADWAY CATEGORIES
AND PARTICLE SIZE FRACTIONS
Emission factor by particle size fraction
Roadway
category
Local
Collector
Major street
and highway
Expressway
TSP
g/VKT
15
10
4.4
0.35
Ib/VMT
0.053
0.035
0.016
0.0012
< 15 um
g/VKT
5.8
4.1
2.0
0.21
Ib/VMT
0.021
0.015
0.0071
0.00074
< 10 um
g/VKT
5.2
3.7
1.8
0.19
Ib/VMT
0.018
0.013
0.0064
0.00067
< 2.5 u
g/VKT
1.9
1.5
0.84
0.16
Ib/VMT
0.0067
0.0053
0.0030
0.00057
Correlation analysis of IP emissions with parameters characterizing the
source conditions showed the existence of a relatively strong positive
relationship between intensity of emissions and roadway surface silt
loading. This confirms the findings of earlier testing.5 Based on regression
analysis of a subset of acceptable ("MLR") test runs, the following predictive
IP emission factor equation was developed:
0.8
2.54
(5)
where:
elp = Inhalable particulate emission factor (g/VKT)
sL = Road surface silt loading (g/m2)
This predictive equation has an associated precision factor of 2.0 in
relation to the "MLR" data set. By way of comparison, the AP-42 single-valued
factor (corrected to represent IP emissions) has a precision factor of 2.1 for
its data set and a precision factor of 4.4 for the "MLR" data set, which spans
a much larger range of values than the AP-42 data set. Therefore, the
predictive equation, though far from ideal, does represent IP emissions more
accurately over a much larger range of values than does the AP-42 single-
valued factor. This fact is directly attributable to the relationship of IP
emissions to silt loading.
Extension of the regression analysis to include emission factor equations
for other particle size fractions—FP, PM-10, and TSP—yielded a set of
equations in which the power term for silt loading increased with larger
particle size fraction. This result is generally consistent with MRI's
previous paved road equation in which silt loading to the 1.0 power was
employed to account for variations in TSP emissions.
3-17
-------
To facilitate the use of these particle size specific equations in the
development of emission inventories, a classification system of mean or
typical silt loadings as a function of roadway category was derived. These
mean silt loadings were then inserted into the respective emission factor
equations. The resultant emission factors for specific roadway category and
particle size fractions can be utilized directly for emissions inventory
purposes. By accounting for variations in silt loading, these emission
factors are significantly more reliable than an overall average emission
factor in developing components of an urban paved road emission inventory.
ACKNOWLEDGMENT
The work upon which this paper is based was performed pursuant to EPA
Contract No. 68-02-2814, Assignment No. 32, and EPA Contract No. 68-02-3158,
Technical Directive No. 19. Dennis Drehmel and William Kuykendal
served as EPA project officer for the study.
REFERENCES
1. Lynn, D. L., 6. Deane, R. Galkiewicz, R. M. Bradway, and F. Record.
National Assessment of Urban Particulate Problem. Volume I - Summary of
National Assessment. U.S. Environmental Protection Agency. Publication
No. EPA 450/3-76-024, NTIS PB2636.65, July 1976.
2. Roberts, J. W., A. T. Rossano, P. T. Bosserman, G. C. Hofer, arid H. A.
Watters. The Measurement, Cost and Control of Traffic Dust and Gravel
Roads in Seattle's Duwamish Valley. Paper No. AP-72-5, Presented at the
Annual Meeting of the Pacific Northwest International Section of the Air
Pollution Control Association, Eugene, Oregon, November 1972.
3. Roberts, J. W., H. A. Watters, C. A. Margold, and A. T. Rossano. Cost
and Benefits of Road Dust Control in Seattle's Industrial Valley. Paper
No. 74-83, Presented at the 67th Annual Meeting of the Air Pollution
Control Association, Denver, Colorado, June 9 to 13, 1974.
4. Axetell, K., and J. Zell. Control of Reentrained Dust from Paved
Streets. EPA Publication No. EPA-907/9-77-007, NTIS PB288325, August 1977.
5. Cowherd, C., Jr., C. M. Maxwell, and D. W. Nelson. Quantification of
Dust Entrapment from Paved Roadways. U.S. Environmental Protection
Agency, Publication No. EPA-450/3-77-027, NTIS PB272613, July 1977.
6. Compilation of Air Pollutant Emission Factors, Third Edition, U.S.
EPA, Publication No. AP-42, NTIS PB275525, August 1977.
7. Cowherd, C., Jr., K. Axetell, Jr., C. M. Guenther, and G. Jutze.
Development of Emission Factors for Fugitive Dust Sources. Final Report
Midwest Research Institute for U.S. Environmental Protection Agency'
Publication No. EPA-450/3-74-037, NTIS PB238262, June 1974
3-18
-------
8. Quality Assurance Handbook for Air Pollution Measurement Systems.
Volume II - Ambient Air Specific Methods. U.S. Environmental Protection
Agency, Publication No. EPA 600/4-77-027a, NTIS PB273518, May 1977.
9. Ambient Monitoring Guidelines for Prevention of Significant Deteriora-
tion. U.S. Environmental Protection Agency, Publication No. EPA
450/2-78-019, NTIS PB283696, May 1978.
10. Nie, N. H., et al. Statistical Package for the Social Sciences, Second
Edition. McGraw-Hill, Inc., New York, 1975.
11. Cowherd, C., Jr., R. Bohn, and T. Cuscino, Jr. Iron and Steel Plant Open
Source Fugitive Emission Evaluation. Final Report, Midwest Research
Institute for U.S. Environmental Protection Agency, Publication
No. EPA-600/2-79-103, NTIS PB299385, May 1979.
12. Cuscino, T., Jr. Total Suspended Particulate Matter Analysis in
Baltimore, Maryland. State of Maryland, Baltimore, Maryland, October
1981.
13. Bohn, R. Evaluation of Open Dust Sources in the Vicinity of Buffalo,
New York. EPA Contract No. 68-02-2545, Assignment 1, Environmental
Protection Agency, New York, New York, March 1979.
3-19
-------
MICRON DROPLET DUST SUPPRESSION PROVES OUT
IN VARIETY OF FUGITIVE DUST APPLICATIONS
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
By: Wayne Hartshorn
Executive Vice President
Sonic Development Corporation
Mahwah, New Jersey 07430*
Lennart Strand
President
Andeze AB
Helsinborg, Sweden
* 305 Island Road
4-1
-------
ABSTRACT
In recent years, a little known dust suppression technology called
Dry Fog"* has proven successful in troublesome applications. Normally a
baghouse or conventional wet-spray system is assigned these problem
applications. This paper will detail this new technology and examine four
individual applications. Specifically, applications where it is successfully
controlling fugitive dust emissions in-situ — at 40 percent lower cost vs
baghouses.
Initially, the technology was applied almost exclusively to rock
product plants, which pose a dusting problem characterized by multiple,
widely dispersed dust generation points. The individual points are
located throughout the entire materials handling system. More recently
many quarries, chemical plants and coal handling operations have also
reported success. These will be discussed in specific terms.
In light of recent interest in coal conversions in many industries,
operating experience in coal handling operations at Dayton Power &
Light and Westvaco Pulp & Paper will be detailed fully.
Since Dry Fog Dust Suppression is a relative newcomer, the paper
will examine and discuss its basic principle of operation. Essentially,
small, micron-size water droplets — the size of airborne dust particles
themselves — are generated to blanket the dust and trigger in-situ
agglomeration. It is accomplished without discernible wetting and
with no chemical additives. It is the tiny droplet size that
differentiates this technology from any other.
The paper will also discuss the technological and economic factors
pertaining to matching the Dry Fog technology to other applications.
Included will be the pros and cons of Dry Fog vs other technologies
commonly considered for fugitive dust control. Feasibility, capital
costs, operating costs, space and ease of installation, additives and
wetting will be explored.
*Dry Fog™ is a registered trademark of Sonic Development Corporation,
305 Island Road, Mahwah, NJ 07430
4-2
-------
INTRODUCTION
As you know, wet spray dust suppression has been around for decades.
The Dry Fog concept itself is similar to the scrubbing process where a
spray of water droplets remove dust particles out of an air or gas
stream. However, in dust suppression, rather than pumping the mixture
from a scrubbing vessel, liquid drops agglomerate with dust particles
and knock them down in-situ.
Many of these wet spray dust suppression systems — commonly
termed chemical or conventional type systems — are saddled with a
major problem. While they do an adequate job of treating the material
being handled, often times the droplets sprayed do not evaporate quickly
or completely, thus contaminating the product. Without quick and
complete evaporation, product wetting is inevitable. In many materials
handling operations, such as cement, aggregate crushing, and coal
preparation operations, it is essential to keep the product dry. In
coal handling, for example, moisture directly impairs the heating
value of the coal. With wet coal, too much heat is dissipated in
evaporating that moisture. Therefore, less BTCJ's of heat per ton of
coal are available to generate steam.
DROPLET SIZE HINDERS EVAPORATION
The reason why these water droplets produced by conventional
wet-spray systems cannot evaporate quickly or completely is their
size. They are too large. The second problem with large droplets is
that when droplet size greatly exceeds that of the dust particles,
there is very little chance of particle-to-droplet contact to trigger
the desired agglomerating action. Instead, very little particle-to-
droplet contact actually occurs so the dust particles simply move
around the water droplets (Figure l).l
At Sonic, however, we have engineered a system that is capable of
producing a superfine atomization of water droplets that greatly
enhances particle-to-droplet contact. And it evaporates before wetting
anything but the dust. These atomized water droplets are best described
as fog. Since it doesn't wet product, call it Dry Fog.
The system operates on one underlying principle: By producing
water droplets of approximately the same size as the dust particles,
4-3
-------
Figure 1. Airflow around large water droplet (top) prevents coal dust
particles from contacting the droplet. The dust particle, however,
easily impacts a smaller droplet (bottom).
4-4
-------
the probability of collision between the two will be extremely high.
The superfine spray of water droplets rapidly agglomerates with the
individual dust particles and immediately knocks them out of the
atmosphere, causing them to fall down in-situ. Also, the droplets
equal the diameters of the dusts over a wide range of operating
conditions and particle sizes.
As mentioned, the tiny droplet size differentiates this technology
from any other wet-spray dust suppression system. The superfine fog
rapidly agglomerates with airborne particulate and immediately knocks
it down, in place. Knockdown is quick and it is accomplished without
product wetting.
The Dry Fog system is capable of consistently producing the uniformly
small droplets over all operating conditions. This is attributable to
the manner in which they are created. The Sonicore® spray nozzle is
the key to the Dry Fog system (Figure 2). It's actually an air-driven
acoustic oscillator that creates a sonic shock wave which shatters the
liquid, producing very fine droplets.
In addition to suppressing dust while insuring minimal moisture
addition to the product, the Dry Fog system has other advantages over
conventional methods of dust suppression — including baghouses. For
example, a Dry Fog system can be installed for as little as 40 percent
of the capital cost of installing a baghouse and in less than 20 percent
of the time. Since particle knockdown is achieved in-situ, there's no
need for long duct runs to convey the dust to a central collection
point. In most cases, it can also be installed while the plant operates
at 100 percent capacity. A typical baghouse installation can last up
to 3-6 weeks, and sometimes forces production to come to a standstill.
The system also offers these advantages over conventional wet-spray
dust suppression alternatives:
— low water consumption averaging only one to three gallons
per hour (per nozzle). Conventional hydraulic systems consume
anywhere from five to 15 gallons per hour to perform the
same task.
4-5
-------
RESONATOR CHAMBER
LIQUID
SONIC ENERGY
CORE
AIR OR GAS
Figure 2. Schematic of the Sonicore® spray nozzle.
4-6
-------
Low water pressure eliminates the need for costly pumping
systems. Usually 30-50 psi supply pressures are adequate.
The Sonicore nozzle will typically operate at 20 psi.
Air consumption is only seven scfm per nozzle at 65 psi.
Most plants have sufficient air available. Larger systems
will require supplemental air.
Almost no water is added to the process — less than 0.1
percent; this percentage runs as high as 1-10 percent
with wet spray systems.
Costly wetting agents and their associated controls are
eliminated with Dry Fog.
Dry Fog will not freeze. Independent research has proven
that water droplets less than 30 microns will not freeze
above minus (-) 40°F below zero.
— Self-cleaning nozzles. The Sonicore nozzle has no internal
screens or filters to plug. The liquid ports are large
enough to pass large particles and the sonic action of the
nozzle keeps it clean.
— Lowered maintenance costs. The system requires no handling
of chemicals and/or surface active agents.
— No major plant modifications required for installing the system.
Enclosures, skirting/ and shrouds, similar to baghouse
requirements, are all that are necessary.
In addition, many engineers overlook additional factors — such
as cost control, space management, and maintenance and operating costs —
when they search for a fugitive dust control system. The Dry Fog
system cuts capital costs over a baghouse and some conventional spray
4-7
-------
systems by as much as 20-60 percent. It also takes up less space and
can be installed in many applications where there isn't room for a
baghouse. For example, the entire system's space requirement is less
than 50 percent of conventional hydraulic systems, and less than
95 percent of the space of a baghouse. Only electricity and water
constitute annual operating costs. Also, because there are fewer
parts, equipment, and no chemical handling, maintenance is minijnal.
CONTROL STANDARDS
The principle objective of the Dry Fog dust suppression system
is to meet fugitive emission control regulations. Systems are designed
to control practically all types of dusts that measure one to 10
microns in diameter, as well as larger fugitive dusts that measure
up to 600 microns. Figure 3 shows average emission concentrations
at a typical aggregate crushing plant located in Sweden, with
and without the system in service.
IN-SITU, ON-SITE
As mentioned earlier, the main feature of this new technology
involves its ability to ensure that almost no moisture is added to
the product. Coal handling operations for example, require dry coal
at all times for maximum heating efficiency. Two very recent
applications, one a midwestern utility plant, the other a southern
pulp and paper mill, both solved major dust problems in their coal
prep operations. The systems have been successfully controlling
fugitive dust for close to two years and without a trace of wet or
frozen coal.
Dayton lower & Light Company's Longworth Station needed to comply
with tight EPA regulations. They had a problem suppressing fugitive
dust in their coal prep operation. Each time they ran the conveyor
that leads to its coal bunkers dust was generated.
Conventional wet-spray type dust suppression systems controlled
the dust, but left a residual moisture on the coal. This left them
with the dual problem of excessive energy costs for coal heating and
frozen coal during the winter months. With the wet-coal problem,
some of the heat generated inside the boiler was dissipated in
4-8
-------
Figure 3
AVERAGE RESULTS AT SWEDISH STONE CRUSHING PLANT
Points Treated Dust Level
System Off System On
Reclaim Tunnel 6.0 1.2
Crusher Discharge 190.0 16.0
Crusher Feed 73.0 2.6
Screen #1 2.7 0.4
Transfer Point 15. '0 1.0
Screen #2 11.0 0.9
4-9
-------
evaporating that moisture. Consequently, less BTU's of heat per ton
of coal were available to generate steam.
Dry Fog eliminated both problems at half the capital investment
compared to another alternate method — a baghouse.
The Dry Fog system was installed at several points in DP&L's
coal handling gallery. These points include a screen, a crusher,
three transfer points, and a 20' x 40' open bunker area. The
installation time required for the Dry Fog system was minimal and no
major modifications of the existing plant facilities were necessary.
Also, Dry Fog virtually eliminated maintenance costs.
For each gallon of water used in the process. Dry Fog yields a
much greater distribution of micron-sized droplets. This, of course,
means greater knockdown power since the size of the droplets is
equal or close to the size of the airborne dust particles.
The resulting dust suppressiqn satisfied the regulations
and brought a cleaner air environment for plant workers.
For DP&L's essentially intermittent operation, the Dry Fog system
functions only once or twice a day. The application of Dry Fog has
been so successful that they are considering applying it to other
facilities.
Westvaco Corporation of Charleston, SC holds the distinction of
being the first full scale coal prep installation of a Dry Fog system.
During its conversion from oil to coal, plant engineers found
a great amount of dust being generated over the crusher and transfer
towers. To control this problem, several wet-spray dust suppression
systems were examined. Finally Westvaco installed Dry Fog spray
bars at the inlets and outlets of its coal crusher and transfer points.
The system proved superior to a baghouse — a baghouse would
have required duct runs ranging from 150-200 feet between its coal
crusher and transfer tower. Also, lack of adequate space in its
coal prep room, and the high cost that accompanies a baghouse
installation, steered them toward Dry Fog.
4-10
-------
Dry Fog's tiny droplets triggered agglomeration so that the coal
dust was knocked down in-situ. This eliminated the need for long
duct runs to carry the dust to a central pickup point. And, since
Dry Fog's droplets evaporated quickly, Westvaco experienced no
moisture addition to the coal.
SCIENTIFICALLY PRCVEN
The emission concentrations at the stone crushing plant located
in Vilhelmina, referenced earlier, are fully supported by a series
of tests conducted by the Swedish National Board of Occupational
Safety and Health.
The test site was a fluidized stone crushing plant that operates
only six months out of the year due to harsh winter months.
Figure 4 gives a plot plan of the crushing operation showing
the location of the primary crusher, screens and hoppers.
In this particular crushing plant, raw, materials are crushed in
two stages — primary crushing and secondary crushing. These two
crushers are also the main dust generation points. After the raw
materials are crushed they are assorted into two fractional sizes —
0-8 mm and 8-16 mm. Daily output is about 200 to 250 cubic meters.
Crushing rate is close to 50 percent.
The Dry Fog system was tested over a three day period to determine
its efficiency as a dust control method. Samples were taken from the
test site with Dry Fog equipment in service for two days. The third
day of testing was conducted without the Dry Fog system in operation.
Temperature and wind velocity varied over the three days. Temperatures
ranged from 8°C to 13°C and the wind velocity varied between slight
westerly to northwesterly winds. Production output each day was
approximately 25-30m3/hr. Water usage was about 120 liters per
hour. The system consisted of a total of nine Sonicore nozzles.
Samples were taken at several points. The stone
crushing operation's dust levels were measured at the following
points: the reclaim tunnel, where coarse sand is loaded onto a
conveyor; at the jaw crusher; the main bell crusher; the primary and
secondary screens; and a transfer point.
4-11
-------
to
Finished
Product
8-16 mm
Finished
Product
0 - 8 mm
.^Transfer Point
Screen 2
Reclaim
Tunnel
Screen 1
IpTfop Crusher (Bell)
Bottom Crusher (Jaw)
Test Site
Raw Material In
Figure 4. Schematic of Swedish stone crushing plant.
-------
Figure 5 is an overall summary of test results with the Dry
Fog Dust Suppression system in service and without the system
in service. Dust concentrations were measured in mg/m3.
These government sponsored tests show that airborne particulate
levels are significantly reduced by the Dry Fog system. At the same
time, they also show that it accomplished the necessary reduction
without any discernible product wetting.
FURTHER TESTING
A second, more detailed experiment on the Dry Fog system was
performed at another Swedish crushing site in early 1981. This was
conducted at an underground mining operation. Figure 6 is a
plot plan of the mining operation. Over a four-day testing
period, the system's performance was measured for its ability to
suppress fugitive dusts on a section of this underground mine.
During normal mining operations, the Dry Fog system was activated.
At the same time,'six sampling units measured atmospheric particulate
levels. Figure 7 is an overall summary of test results with the
Dry Fog system in service.
Three tests were run during a four-day period. On the first
day, this particular section of the mine was shut down. However, the
fugitive dust samplers recorded atmospheric readings for particulate
content for comparison purposes vs particulate levels during normal
operation. The significance of the overall findings is uncovered in
the three days the mine was in actual operation vs the one day it
was not in service. The comparison shows that there was no appreciable
difference in airborne particulate levels when the section was
operating with the Dry Fog System in service and the day the entire
mining section along with the Dry Fog system was shut down.
In the final report, test engineers stated that "there was no
excess of applicable limit values for the total dust volume detected
during the testing period." This means they didn't exceed the total
dust volume limit allowable according to Swedish local and federal
regulations for the area. The allowable limit is 10 mg/m3.
REFERENCE
Schowengerdt, F.D., Brown, J.T., "Colorado School of
Mines Tackles Control of Respirable Coal Dust,"
COAL AGE, April 1976, Copyright 1976 McGraw Hill, Inc.
4-13
-------
Figure 5
INDIVIDUAL RESULTS AT SWEDISH STONE CRUSHING PLANT
with Dry Fog
Test: 1
2
3
Average
Reclaim
Tunnel
1.3
1.0
1.3
1.2
Crusher Crusher
Discharge Feed
(jaw) (bell)
9.3
16.1
22.9
16.0
1.5
3.0
3.2
2.6
Screen
(#1)
0.6
0.3
0.4
0.4
Transfer
Point
1.4
1.0
0.4
1.0
Screen
#2
1.3
0.6
0.7
0.9
without Dry Fog
Test: 1
2
3
Average
5.6
3.2
9.1
6.0
210
190
170
190
84
68
61
73
3.7
1.9
1.9
2.7
23
12
6.8
15
19
6.4
2.3
11
4-14
-------
Belt Conveyor
Screen
i
(—•
01
Sampling Unit
CT
Cone
Crusher
Belt Conveyor
Figure 6. Schematic of Swedish underground mine.
-------
Figure 7
TEST AT SWEDISH UNDERGROUND MINE
Points
Treated
System Off
Belt
Conveyor
System On
Screen fl
Screen #2
Transfer
Point 11
Transfer
Point #2
Transfer
Point #3
Total Dust Average
Measured mg/m3 Per Day
0.57
1.51
0.70
1.39
0.65
1.44
0.53
0.59
1.26
1.11
1.03
2.03
1.18
0.76
1.00
2.49
0.68
0.77
1.67
0.91
2.68
3.54
5.13
6.71
1.10
1.29
1.06
1.01
0.78
0.84
0.34
0.35
1.67
0.97
0.55
1.41
1.70
0.72
0.46
2.63
0.75
0.66
0.32
1.39
4.74
3.98
3.38
8.16
1.02
0.66
1.01
1.52
0.78
0.30
0.52
0.29
1.66
0.68
0.88
1.33
1.62
0.69
0.81
2.29
0.89
0.46
0.60
1.22
4.68
2.70
4.18
5.21
0.90
1.15
0.92
1.31
0.74
0.86
0.46
0.41
1.53
1.48
0.82
1.59
1.50
0.72
0.76
2.47
0.77
0.63
0.86
1.17
4.03
3.41
4.23
6.69
Total
Average
1.13
0.58
1.30
1.32
0.89
4.78
4-16
-------
,~ df CTibcd in this paper was not funded by the U.S. Environmental
!!!^ ™H SrS£ -^C c?ntents do «* necessarily reflect the views of the
Agency and no official endorsement should be inferred.
FUGITIVE DUST CONTROL STUDIES
USING SCALE MODELS AND
MASS LOSS ESTIMATES*
by: Michael F. Lepage, Anton E. Davies
Colin J. Williams
ROWAN WILLIAMS DAVIES & IRWIN Inc.
650 Woodlawn Road West
Guelph, Ontario, Canada
NIK IBS
(*) Although this is not the actual presentation made at the May 1S82
meeting, it closely resembles that presentation, which was
entitled, "The Optimization of Wind Screens for Fugitive Emission
Control Using Wind Tunnel Tests, " C. J. Williams (then of
MHTR. Ltd.).
5-1
-------
INTRODUCTION
Fluid model studies of bulk storage facilities provide an
inexpensive means of testing dust control strategies before
implementing them in the field. The potentially high cost of
implementing a successful dust control program requires careful
planning to ensure that the best result is obtained.
There has been growing pressure in recent years to improve
air quality around bulk storage sites (e.g. coal stockpiles,
sawdust piles and fly ash lagoons) by reducing dust emissions
(see Figure 1). Improved air quality, however, is only one of
the benefits of a dust control program, since dust losses often
represent losses of valuable inventory. In many cases, the cost
of implementing dust controls is more than offset in the long
run by the reduction in lost inventory.
There are two principal approaches to dust control at bulk
storage facilities. The first is to clean up the site, compact
loose surfaces and encrust exposed dust which would be available
for emission. The second approach is to reorganize the
stockpiles and erect wind breaks in order to reduce local wind
speeds below the threshold of dust emissions. The most
efficient dust control program is typically a combination of
both approaches.
Fluid model studies provide a means of optimizing dust
control while avoiding costly trial and error in the field. By
examining local wind conditions on a scale model in a wind
tunnel or water flume, one can identify the windiest and hence
most sensitive areas of the site and can evaluate the relative
merits of different schemes for localized wind speed reduction.
To give further meaning to the results, actual samples of
material from the site are tested in the wind tunnel to
determine the relationship between wind speed and the level of
dust emission (which can vary significantly from one site to
another). By feeding this information together with data from
the scale model tests into a computer simulation of dust
emission, wind reduction schemes can then be evaluated in terms
of their impact on overall emissions at the site.
The present paper describes the fluid modelling techniques
currently being practiced at Rowan Williams Davies & Irwin
Inc.'s fluid dynamics laboratory at Guelph, Ontario. The
techniques have undergone a number of -refinements since they
were first presented . The objective, which has remained
unchanged, is to determine the most effective configuration
5-2
-------
windbreaks and stockpile geometry for minimizing local wind
speeds. The techniques in question have been applied to
numerous sites, such as: the coal handling facilities of Kerr
McGee Inc. at Trona, California; the Pennsylvania Electric
Company at Homer City, Pennsylvania; the Dofasco steel mill at
Hamilton, Ontario; and Ontario Hydro's thermal generating
stations at Nanticoke and Atikokan, Ontario, Canada.
EXPERIMENTAL METHODS
Scale Model Tests
The fluid modelling techniques for dust control studies at
Rowan Williams Davies & Irwin Inc. include testing a scale model
of the study site in the simulated wind flow of both the
boundary layer wind tunnel and the open channel water flume.
Photographs of both pieces of equipment are shown in Figure 2.
The boundary layer wind tunnel is an open circulation type,
27m long, 2.4m wide and 1.8m high. The test section, located
near the downwind end of the tunnel, has a 2.4m diameter
automatic turntable built into the floor. Spires and roughness
elements can be placed on the floor upwind of the test section
to simulate the characteristics of the mean wind and turbulence
approaching the study site . Wind speed is controlled
automatically and can achieve values in excess of 25 m/s.
The open channel water flume is a 12m long, 1.2m wide and
0.5m deep apparatus that operates on the same principles as the
wind tunnel but uses flowing water to simulate wind. Flow
patterns on a scale model are visualized by injecting coloured
dye into the stream. Drifting of relatively large particles,
such as snow grains, is simulated using silica sand.
Scale models of bulk storage facilities are typically
constructed at scales between 1:400 and 1:600. The principal
modelling material is plexiglas, with stockpiles modelled in
clay to facilitate reshaping. The models are mounted on a 2.4m
diameter base that fits on the turntable of the wind tunnel test
section. Various wind directions are simulated by rotating the
turntable. The base of the model is divided into two sections:
an inner disk and an outer annulus. The 1. 2ra diameter inner
disk fits into the test section of the water flume.
Local wind speeds on the scale model are measured using an
array of surface-mounted omnidirectional sensors as shown in
5-3
-------
Figure 3. Speeds are measured at the equivalent height of 2.0m
above grade full scale and are sampled at a rate of 300 Hz. The
data are low-pass filtered at 200 Hz which is well above the
dominant frequency of atmospheric turbulence. The local speeds
are normalized by the free stream (gradient) speed, U ,
measured above the model using a pitot-static tube. The results
are recorded in the form U/U and s/U where U is the
one-hour mean speed and s is the RMS (standard deviation) of the
wind at each sensor location. Using previously measured
profiles, the ratios are then converted to the form U/UQ and
s/U where U is the mean speed at 2.0m above grade in an
unoSstructed area. The resulting ratios can be thought of as
local wind speed magnification or reduction factors due to the
presence of terrain features, buildings or windbreaks.
Dust Sample Tests
A schematic of the apparatus for testing dust samples in
the wind tunnel is shown in Figure 4. The quantity of dust
lifted into the atmosphere is evaluated by weighing portions of
the samples before and after exposure to the wind. The
apparatus is designed to distinguish between dust particles that
are sufficiently small to remain suspended in the atmosphere and
undergo turbulent diffusion (airborne particulates), and heavier
particulates that are subjected to the saltation process
(saltating particulates).
Three adjacent trays embedded in the floor of the wind
tunnel test section are filled with samples of the dust. The
middle tray is fixed to the floor while the upwind and downwind
trays are removable for the purpose of weighing. When the wind
is turned on, the saltation process quickly reaches equilibrium.
The middle tray is sufficiently long that the downwind tray
experiences a flux of saltating particles leaving the tray equal
to the flux entering from upwind. Hence, any measured weight
loss after a time t is attributable to airborne particulates.
The upwind tray, on the other hand, experiences no flux of
saltating material from upwind so that the measured weight loss
is a combination of both saltating and airborne particulates.
The difference in weight loss between the upwind and downwind
trays, therefore, gives the estimate of the outward flux of
saltating particulates.
In summary, if w is the upwind weight loss and W, is
the downwind weight loss after a time t, and if A^ and A,
are the surface areas of the upwind and downwind trays, \hen:
5-4
-------
W2/A2 = emissi°n of" airborne particulates per unit area
W,/A, - W0/A0 = outward flux of saltating particulates
.L J. £ £
Wind speed is monitored using a pitot-static tube, which
can be set at any desired height above the samples, and the
tests are repeated for a range of wind speeds. Spires and
roughness elements are used to generate turbulence. Although
full scale turbulence cannot be generated in the wind tunnel due
to the limitation of having walls and a ceiling, the turbulence
produced is thought to be sufficiently close to full scale for
the purpose of the dust sample tests. The impact of turbulence
intensity on these tests is the subject of ongoing
investigations.
Computer Simulation
The computer simulation combines the results of the dust
sample tests and the wind speed ratios described in the previous
sections with wind measurements from a meteorological station
near the study site in order to predict dust emissions for
individual wind events. Dust losses are calculated for the area
around each model sensor location and summed to obtain a global
dust loss for the entire site. The simulation has the capacity
of integrating the results for several years' worth of wind
events to predict average annual dust losses.
The dust loss from the area around each sensor location is
calculated hour by hour as follows. The reference hourly mean
wind speed, U , and wind direction are obtained from the
nearby meteorological station. The local mean speed at 2.0m
above grade is calculated by taking the product of U and the
local wind speed ratio, U/U , that was measured during the
scale model tests. The resufts of the dust sample tests are
then used to determine the local dust losses for the given hour.
The local RMS ratio, s/U , is not currently used in the
analysis but will be adde% in future studies as more data
becomes available on the relationship between turbulence
intensity and the level of dust emissions. For the present, it
is assumed that the level of turbulence throughout the study
site is uniformly the same as was generated in the wind tunnel
during the dust sample tests.
The dust sample results allow the local dust loss to be
divided into its airborne and saltating portions. When
calculating the saltating portion, consideration has to be given
to the fact that a partial balance exists between the flux of
material away from the sensor location and the flux into the
area from upwind. The net flux is determined using a
5-5
-------
two-dimensional centred-difference technique in which the
difference in dust loss between surrounding sensor locations is
examined.
Before the local dust losses can be calculated, initial
conditions have to be established - i.e. the amount of dust
available for erosion at each sensor location prior to the wind
event. The initial conditions are highly site-dependent and can
vary from one event to another. The calculated hourly dust
losses are subtracted from the initial amount until the
remaining amount reaches zero. When the remaining amount at a
particular sensor location has reached zero, no further dust
loss is permitted to occur at that location. If the area is
continually disturbed by vehicle traffic, then the material
available for erosion may be unlimited. To account for such
cases the initial condition is set sufficiently high so that the
remaining amount does not reach zero during the event.
Optimization and Evaluation of Dust Control Strategies
The fluid model studies are designed to evaluate the
following types of dust control strategies:
(i) development of an optimum stockpile geometry to achieve
maximum wind speed reductions in critical areas;
(ii) development of haul roads and operating procedures that
minimize disturbance of the surface of the storage piles
in critical areas;
(iii) development of an optimum arrangement of windscreens,
berms and other windbreaks upwind of critical areas.
Semi-permeable windbreaks may also be used downwind of
critical areas to trap saltating particles before they leave the
property, but are generally ineffective at trapping airborne
dust. Typically, it is not feasible to design a dust control
program that works for all wind directions but the program can
be designed for the directions most frequently responsible for
high wind events.
The following are some of the practical limitations by
which dust control programs tend to be constrained:
(a) the cost and availability of material for berms and other
windbreaks ;
5-6
-------
(b) the need for efficient haul routes, unobstructed by
windbreaks;
(c) the need for sufficiently large working areas uncluttered
by windbreaks;
(d) temporary windscreens must be practical to assemble,
disassemble or move about.
By and large, flow visualization on the scale model in the
water flume is adequate for determining which set of control
strategies works best . A qualitative judgement, however, is
less convincing than a numerical assessment, and the pros and
cons of various possible strategies are best weighed by using
predicted dust emissions from the computer simulation.
EXAMPLE OF RESULTS
Figure 5 shows a plan of the fly ash lagoon at Ontario
Hydro's Generating Station at Nanticoke, Ontario. During the
course of some thirteen years of ash storage operations the
level of deposited ash exceeded the waterline in some areas.
Eventually, dry ash became exposed resulting in significant dust
emissions during strong winds.
In order to continue using the lagoon, Ontario Hydro
drained off most of the water and compacted the ash surface
sufficiently to allow the traffic of heavy vehicles. Several
settling cells were created in the western half of the lagoon.
At present, the ash, which is a by-product of the coal-fired
boilers at the generating station, comes to the lagoon in a
liquid slurry poured from pipelines into the settling cells.
The ash is allowed to settle and is then removed by drag lines
and transported by truck to the permanent storage area. A fluid
model study was undertaken to determine the optimum method of
continuing this procedure over the next twenty years while
minimizing dust emissions at the site.
Since most observed incidents of high dust emissions
occurred during winds from WSW, the design of dust control
strategies focused on that direction. Figure 6 shows contours
of wind speed ratios derived from the scale model tests in the
wind tunnel. The highest values occur in the most elevated
areas which are exposed to the highest wind speeds.
5-7
-------
Ash samples were collected from the site and an example of
the data from wind tunnel tests on the samples is shown in
Figure 7. Weight losses of airborne particulates are plotted
for wind speeds of 6.0, 8.0 and 10.0 m/s (measured at 15 cm
height) and for various durations of wind exposure. Wind speeds
of 2.0 and 4.0 m/s were also tested but no weight loss was
observed, from which it was concluded that the threshold speed
of dust emissions is between 4.0 and 6.0 m/s. For durations
greater than 0.2 hours, the data shown in the figure fit to an
expression of the form W = at + b where W is the weight loss, t
is the duration of wind exposure, and a and b are constants
which vary with wind speed. The constants a and b were found to
be approximately related to the square of the wind speed.
A scheme for storing ash was developed, taking into account
all of the economical and practical limitations relating to the
site, and is shown in Figure 3. The plan shows the entire ash
lagoon as it would appear after approximately twenty years of
operation. The sections at the bottom show the progression of
the permanent storage area over the same period. Berms mounted
with windscreens are placed west-southwest of areas where ash is
being dumped. The dumping begins just east of the settling
cells (see Figure 5) and eventually progresses eastward until
the permanent storage area is filled to a height of 15m above
the level of the surrounding dyke. A 15m high porous windscreen
along the western perimeter of the lagoon protects the settling
cells area where much of the vehicle traffic occurs. Solidity
ratios between 0.5 and 0.8 were recommended for various
windscreens in order to obtain the optimum compromise between
maximum wind speed reduction and maximum extent of protection.
The scheme shown in Figure 8 was modelled and tested in the
wind tunnel. To assess the impact of the scheme on dust
emissions, the computer simulation was run using the wind
conditions that occurred during a particular dust event, on
January 13 and 14, 1985, which was chosen as a representative
case. Wind speed and direction during the event were recorded
at a tower located a few kilometres north of the lagoon. Wind
directions were observed to range between southwest and
northwest and the highest hourly mean wind speed recorded was 35
km/hr (at 10m above grade). As an initial condition for the
event, it was assumed that the entire ash storage area was
covered in a uniform thin layer of loose dust. Although this
was an oversimplification of the true situation, it was an
adequate assumption to assess the relative impact of the
recommended berms and wind screens.
5-8
-------
Figures 9 and 10 show contours of airborne dust emissions
as predicted by the computer simulation for the present ash
lagoon configuration (Figure 9) and for the ash lagoon after
twenty years of operation. The size of the shaded area,
indicating emissions of 0.2 kg/m or greater, is significantly
reduced in the future configuration despite the fact that the
ash is piled considerably higher than in the present case- In
the future configuration, the areas downwind of the protective
berms and windscreens experience minimal dust emissions. The
upwind slope of the westernmost berm in the permanent storage
area experiences accelerated winds and consequently high values
of dust emission, but in practice this area would be permanently
sealed and seeded with grass.
Integration of the dust losses over the entire area
indicated that the future configuration would experience about
half the total dust emissions of the present configuration for
the same initial conditions. If permanently sealed and grassed
areas are accounted for then the emissions would be further
reduced.
CONCLUSIONS
Fluid modelling studies are an' excellent visual and,
quantitative tool in the design of dust control programs for
bulk storage facilities. Areas where refinements to the
physical model approach are anticipated in the future include:
the establishment of more realistic initial conditions for the
computer dust simulation that account for the rate of
disturbance by vehicles, the moisture content of the surface
material and other factors; an indepth study of dust emissions
resulting directly from vehicle activity; a further study of the
relationship between airborne emissions and turbulence
intensity.
A further addition currently being implemented consists of
applying a numerical dispersion model to predict the rate of
dispersion of the dust emissions downwind of the study area, the
results of the numerical model can then be compared to field
measurements with high volume samplers.
5-2
-------
REFERENCES
1. A.E. Davies, "A physical model approach to the solution of
fugitive emission problems", Proceedings of the 73rd Annual
Meeting and Exhibition of the Air Pollution Control
Association, Montreal, June 22-27, 1980.
2. H.P.A. Irwin, "Design and use of spires for natural wind
simulation", National Research Council of Canada, N.A.E.
Report LTR-LA-233, 1979.
3. H.P.A. Irwin "A simple omnidirectional sensor for wind
tunnel studies of pedestrian level winds", Journal of Wind
Engineering and Industrial Aerodynamics, 7, pp. 219-239
(1981).
Figure 1. Dust emissions at a fly ash storage area.
5-10
-------
Figure 2. The boundary layer wind tunnel (top) and open channel water flume.
5-11
-------
Figure 3. Omnidirectional wind sensors on a model (top) and a side view of
an individual sensor.
5-12
-------
CBL1NG
N^y^s^^^^y^y^^^^
PITOT-STATIC TUBE
WIND
RAMPv*
AIRBORNE
DUST
xx''*
SALTAT1NG
1 ^DUST
^•~^
/I.3CM DEEP /
^7 DUST SAMPLE /
FLOOR/ 55
cm
240 cm
30cm
Figure 4. Schematic of the dust sample test apparatus,
5-IS
-------
12m HIGH
DYKE
SECONDARY
SETTLING
CELLS
AREA BETWEEN
CELLS IS LEVEL
WITH DYKE
TERTIARY
SETTLING POND
PERMANENT
STORAGE
AREA
PRIMARY
SETTLING
CELLS
ACCESS ROAD
PERIMETER^
ROAD
'ROADS CONSTRUCTED
-IN SOFT LOW LYING
AREA
NOTE- SHADED AREAS ARE IMMERSED IN WATER
OTHER AREAS CONSIST OF ASH
I
400
400
800m
Figure 5. Plan of the ash lagoon at Nanticoke, Ontario, in its present state.
5-14
-------
PREVAILING O.
WIND
Figure 6. Wind speed ratios, U/U , as measured in the wind tunnel for winds
from WSW.
5-15
-------
2800
2400 .
2000 _
cvj
*
CO
CO
3
1600 .
1200 -
O
u2
800 _
400 .
LOSS OF AIRBORNE PART1CULATES
DURATION (hrs.)
o
«
A
WIND SPEED: 6 m/s AT 15cm HEIGHT
8 m/s
10 m/s
Figure 7. Example of dust sample test results showing loss of airborne
material versus duration of wind exposure.
5-16
-------
PREVAILING
WIND
HAUL ROAD
15m HIGH
PERMANENT
80% SOLID
WINDSCREEN
SEE
BELOW
-PROTECTIVE BERMS
,3m HIGH WINDSCREEN
/80% SOLID
A—T
IOO
200
300
400 METERS
1st
.YEAR
10
»YEARS
20
a YEARS
Figure 8. Twenty-year storage plan.
5-17
-------
PREVAILING
WIND
2
Figure 9. Contours of airborne dust emission (kg/m ) during a typical dust
event for the ash lagoon in its present configuration.
5-18
-------
PREVAILING
WIND
20.2 kg /m
Figure 10. Same as Figure 9, but for the recommended scheme after 20 years,
5-19
-------
EVALUATION OF FIELD TEST RESULTS
ON WIND SCREEN EFFICIENCY*
Alan G. Larson**
TRC Environmental Consultants, Inc.
Englewood, Colorado
This paper has been reviewed in accordance with the U.5. Environmental
Protection Agency's peer and administrative review policies and approved lor
presentation and publication.
(*)Although this is not the actual presentation made at the
May 1982 meeting, it describes the same work and covers
the same general time period. It represents work funded
under EPA contract 68-02-3115, Technical Directive 117.
(**)Current address: The Kentwood Moore Co., 5690 Denver
Tech Center Blvd., Englewood, CO 80111.
6-1
-------
Abstract
A field measurement project was conducted to generate data which
would lead to a preliminary determination of the effectiveness of wind
screen material as an inhibitor of fugitive windblown dust emissions from
a stockpile. A stockpile comprised of fly ash was used as the test
material and a wind screen was erected upwind of it to comprise the test
configuration. Standard hivolume samplers were used to collect samples
of Total Suspended Particulates (TSP) and hivolume samplers fitted with
size selective heads were used to collect samples of Inhalable
Particulates (IP) (less than 15 /jn). Particulate concentration
measurements due to wind erosion from the stockpile both with and without
the wind screen were made. Low threshold anemometery was used to
document wind conditions. It was found that the wind screen did reduce
the fugitive emissions of TSP but the reduced emissions of IP were
insignificant.
Introduction
The emission of windblown fugitive dust from stockpiles can be a
significant problem at mines, power plants, steel mills, rock-handling
facilities, and many more industrial concerns which stockpile raw
materials, fuel, and waste materials. Stockpiles which are inactive (not
drawn upon) can be effectively treated for fugitive dust control by
several methods including chemical crusting agents, coverings, etc.
Active stockpiles, however, present a much more difficult dust control
problem because of the need for access and because of surface
disturbances caused during access. One approach which has been suggested
as a means of reducing windblown fugitive emissions from both active and
inactive stockpiles employs wind screen materials.
Wind screen material is comprised of a fabric-like construction of
high tenacity polyester fiber, woven to varying degrees of porosity and
prepared in different widths. The manufacturer holds that the material,
6-2
-------
when erected vertically like a fence, will reduce the wind speeds at
levels directly adjacent to the ground surface in the lee of the wind
screen. Further, the manufacturer contends, when the wind speed is
reduced due to the presence of the wind screen, windblown dust in the lee
of the screen will also be reduced. It is therefore expected that use of
the wind screen material as a wind break will result in reduced fugitive
windblown particulate matter in a limited region downwind from the wind
screen. This report describes a preliminary investigation which was
conceived to determine the actual effectiveness of one type of wind
screen material toward control of fugitive dust emission.
TRC Environmental Consultants, Inc. (TRC), under contract to the
EPA Industrial Environmental Research Lab (IERL) in Research Triangle
Park, North Carolina, has performed a preliminary field study which
investigated the effectiveness of wind screens to control fugitive
windblown dust from stockpiles. The study was conducted around a test
stockpile of fly ash material. The stockpile was located in a remote
portion of property owned by Colorado Public Service Company's Valoont
Power Plant in Boulder, Colorado. The site was chosen because of the
high incidence of consistent wind direction (westerly) in several wind
speed categories.
Fly ash material was chosen to comprise the test stockpile for this
field investigation because it is a common waste material where coal is
used for fuel, it is comprised of relatively small particulates, and thus
is susceptible to be windblown at moderate wind speeds, and it was
readily available at no cost. The fly ash being susceptible to wind
erosion simplified the experimental work because measurement of material
removed from the test stockpile was possible over a wide range of wind
conditions. All of this was judged as being optimum to lead to the most
expeditious determination of the effectiveness of the wind screen
material as an inhibitor of windblown particulate matter from stockpiles.
The study was designed to determine the total suspended particulate
matter (TSP) and the inhalable particulate matter (IP, less than
15 fjfn.) due to wind action. The emissions of particulate matter were
6-3
-------
measured from a simulated active stockpile before and after installation
of a wind screen. Hivolume samplers (hi-vols) and hi-vols fitted with
Sierra size selective heads were installed upwind and downwind of the
test stockpile to determine the TSP and IP concentrations under different
wind conditions. The upwind sampler measured background concentrations
and the downwind samplers measured the particulates emanating from the
test stockpile along with the background concentrations. Anemometers and
wind vanes were located upwind and downwind of the wind screen to record
wind speeds and directions that resulted in the windblown particulate
emissions.
Test Methodology
The test site was located on a native grass covered mesa. The mesa
is elevated above the surrounding terrain by about 50 feet* and is
oriented in an east-west direction which coincides with the direction of
the most frequent and highest winds. The test site was situated on the
eastern portion of the mesa in order to allow the west winds to stabilize
prior to reaching the test site. The test stockpile of fly ash material
was placed in a rectangular geometric shape. The fly ash material was
taken from the waste dump where the Valmont Plant waste material is
disposed. The elevated landform provided an isolated test site with an
unobstructed fetch of about 3/4 mile* (for west winds) and
was isolated from interfering particulate matter sources by about
3/4 mile to the west, about 1/2 mile to the northwest, and about 1/3 mile
to the east where the Valmont waste dump is located. Tests were not
conducted during easterly wind conditions, so the only interferences
experienced were those located to the northwest and west. It is believed
that during most conditions when interferences were experienced, those
interferences were adequately documented by upwind sampling. The
stockpile was shaped to a size of approximately 8 feet wide, 12 feet long
(N-S), and A feet high. The leading edge of the stockpile was located
approximately 12 feet downwind of the wind screen. The wind screen used
in the study was commercially available high tenacity polyester fabric
wind screen manufactured by Julius Koch USA, Inc. The screen was 6 feet
tall, 165 feet long, and had approximately 50 percent porosity. The
(*) For nonmetric units used in this paper, please use: 1 ft = 0. 3 m, and
1 mi = 1.61 km.
6-4
-------
screen material was mounted on steel poles set in the ground. Wind
screen material of 50 percent porosity was selected based on
recommendations of the manufacturer which were directed to configurations
most likely to result in maximum wind speed reduction at the levels
adjacent to the ground surface. Material six feet tall was chosen so as
to enable employment of a stockpile being a significant height above the
flat mesa test surface (A feet) and having a stockpile height to wind
screen height ratio of 2 to 3—also as recommended by the wind screen
manufacturer. A section of wind screen material 165 feet in length was
chosen so as to avoid end effects at the test stockpile (wind eddying
effects) and to allow for a reasonable wind direction variation of as
much as + 30°.
After the test stockpile was placed, it was allowed to dry before
tests were conducted. Following the drying period, the stockpile was
covered between tests with a tarpaulin to protect It from exposure to
precipitation. Samples of fly ash material were collected during each
field experimental day and were subsequently analyzed for moisture
content.
The instrument array used to generate test data was comprised of
three pairs of hi-vol samplers and three anemometer/wind vane sets. The
upwind hi-vol pair measuring TSP and IP were located 20 feet upwind of
the screen on the east-west centerline of the stockpile. Two
Climatronics Mark 3 anemometer/wind vane sets were located adjacent to
the upwind- hi-vol s at heights of A feet and 12 feet. Another
anemometer/wind vane set was located adjacent to the stockpile at a
height of A feet, level with the top of the stockpile. One pair of
hi-vols measuring TSP and IP were located on the E-W centerline 20 feet
downwind and another pair was located on the same centerline 50 feet
downwind of the stockpile. Two hi-vols measuring only TSP were located
50 feet downwind of the stockpile, one 30 feet north and one 30 feet
south of the east-west centerline. The wind screen was 165 feet in
length with approximately half of the screen on either side of the
east-west centerline. A diagram of the study site is shown in Figure 1.
6-5
-------
.20'
N
Wind Screen
165' Long
30'
O
A
121*4-8!*
t
-------
The plan for each test consisted of 6 separate steps:
1. Remove tarpaulin from stockpile.
2. Set up anemometer/wind vane equipment*
3. Shift hi-vols to be on the centerline of the wind and
stockpile.
4. Rake stockpile surface to loosen any packed ash.
5. Install filter paper in hi-vols.
6. Measure dust concentrations for approximately one hour
total sampling time.
Steps 1 and 2 were only required at the beginning of each test day
as usually two or three tests were conducted on a given day. During the
testing period, the hi-vols were moved, as necessary between individual
tests, to be on the centerline of the wind and the stockpile. Each test
period lasted approximately 60 minutes. At the end of the test day, the
equipment was removed to a storage area and the tarpaulin was replaced on
the stockpile. When more than one test was conducted during a single
day, Step 4 was repeated just prior to each additional test. This was
done to assure that each test dealt with the same particle size
distribution of fly ash as much as possible.
A total of 7 tests were run without the wind screen and 12 tests
were run with the wind screen installed.
Data Analyses
Assay Methods
The data collected during each hour-long test consisted of hi-vol
filters for both TSP and IP samplers, wind direction, wind speed, and
moisture content of the surface fly ash material. Particle size analysis
of the material on the stockpile surface was done from a single composite
sample obtained from all the faces of the stockpile.
6-7
-------
The hi-vol filters were assayed gravimetrically in accordance with
40 CFR 50 Appendix B Reference Method for Determination of Suspended
Particulates in the Atmosphere (Hivolume Method).
Wind data were compiled from visual strip chart recordings to
represent hourly average directions and speeds, peak gusts, and an index
consisting of a gust-duration product. The compilation of averages and
peak gusts is straightforward resulting in conventional values. The
gust-duration product index was devised in an attempt to derive a wind
representation that would directly relate to the physical processes of
the wind erosion phenomena. It was reasoned that, since wind erosion is
a direct function of both the wind speed and the duration of high wind
speeds, an index incorporating both wind speed gusts and the duration of
those gusts should relate directly to the measured sampler
concentrations. In practice it is difficult to determine the level at
which a wind gust becomes important. Therefore the gust-duration index
was derived simply by integrating the total area under the wind speed
curve for each experimental period. In this way a set of indices was
developed which relate to one another but not to other field experimental
work. The indices are in arbitrary units and can only be used as a
measure of the relative wind erosion potential for each of the tests
represented in this analysis.
Moisture content of the fly ash material was determined from
samples of the material collected during each experimental day- Each
sample of fly ash material consisted of about 60 grams of ash. Samples
were taken from each face of the stockpile, placed in a sealed bag, then
transported to the assay laboratory. Samples were placed in dried
crucibles, and promptly weighed to determine the mass of fly ash and
moisture prior to drying. Samples were baked at 105°C for one hour to
drive off all moisture, then placed in a dessicating chamber until
cooled. After cooling, the samples were weighed again to determine the
moisture loss. Moisture content was represented as percent reduction (by
mass) due to drying. The mean moisture content of the stockpile for an
experimental day was determined as a mean of the values from all faces of
the stockpile.
6-8
-------
Particle size analysis of a composite ash sample, which was
comprised of material from all the faces of the stockpile, was done
manually with eight sieves from number 10 to 325. The sieves are USA
Standard Testing Sieves which conform to ASTM specifications. The
sieving of the composite sample was done manually in order that the
mechanical action of the sieving process which might break up
agglomerated particles would be held to a minimum and the resulting
analysis would as nearly as possible represent the true surface particle
size distribution.
Selection of Data To Be Used for Analysis
A total of 19 tests were conducted to represent wind erosion both
with and without the presence of the wind screen. The data collected
were assayed as described above and a series of comparisons were made to
assist in selecting those tests that would best represent the effects of
the wind screen as an inhibitor of fugitive windblown dust.
For several reasons not all the test data collected were used in
the analysis. As a way of selecting the most representative TSP and IP
concentration data, all of the concentration and wind data were plotted
on scale diagrams so as to graphically illustrate the tests which
exhibited consistent wind and dust erosion conditions. It was found that
the wind direction variability (especially during peak gust conditions)
during some tests caused a portion of the eroded dust to pass by the
sampling array during some portions of some tests. In addition to this,
sampler failure resulted In inadequate representation of upwind or
background concentration levels on two tests. As a result, six of the
nineteen tests were only partially successful leaving eleven to be used
for analysis. The TSP and IP concentrations along with the supporting
wind data and measurements of other test conditions are listed In
Tables 1, 2, and 3.
6-9
-------
TABLE 1: TSP & IP CONCENTRATIONS
FOR TESTS WITHOUT WIND SCREEN
TEST
NUMBER
3
5
6
7
Means
AX 20 =
AX 50 =
TSP
UPWIND
225
118
70
82
124
X (at 20 feet) ~x
X (at 50 feet) -x
(yg/m3)
Ax 20
151
97
195
148
Upwind
Upwind
Ax 50
95
6
16
7
31
TSP & IP CONCENTRATIONS FOR
8
9
10
11
12
13
17
Means
300
72
80
88
—
44
201
131
138
53
16
59
—
4
189
76
72
86
7
—
—
24
88
55
IP (yg/o3)
UPWIND AY 20
158
35
57
70
80
TESTS WITH
126
35
72
53
20
6
86
57
26
40
22
—
29
WIND SCREEN
81
23
—
4
7
21
2
23
AY 50
34
21
19
—
25
•>
78
25
—
—
33
15
42
39
6-10
-------
TABLE 2: WIND DIRECTION, SPEED, PEAK GUST (nph)
AND GUST DURATION INDEX
DURING TESTS WITHOUT WIND SCREEN
TEST
NUMBER
3
5
6
7
Means
U-12
DIR*
285
225
240
220
U-12
SPEED
10.5
11.0
13.5
8.5
10.9
U-12
PK GUST
20.6
22.5
24.2
14.4
20.4
U-4
DIR
295
215
215
210
U-4
SPEED
8.0
10.0
11.5
7.0
9.1
U-4
PK GUST
15.5
• 19.8
19.5
16.6
17.8
D-4
DIR
290
205
210
220
D-4 D-4 GUST DURATION
SPEED PK GUST INDEX**
7.5
10.0
12.5
7.0
9.2
15.5
20.3
19.2
17.4
18.1
52
52
63
40
52
WIND DIRECTION, SPEED, PEAK GUST (mph) AND GUST DURATION INDEX
DURING TESTS WITH WIND SCREEN
8 290
9 280
10 250
11 230
12 285
13 300
17 210
Means
13.0
10.0
13.2
12.2
12.0
11.4
16.0
12.5
25.0
18.0
21.4
20.8
20.6
20.2
19.2
20.7
__
—
240
275
285
255
__
—
10.5
11.0
10.5
13.0
11.2
__
—
16.4
17.4
16.5
19.8
17.5
290
280
25.5
230
250
265
235
3.8
3.0
4.5
4.8
4.0
3.7
4.5
4.0
12.5
8.5
7.5
5.6
7.5
6.9
10.1
8.4
68
49
58
60
61
54
71
60
*U-12 » Upwind, 12 feet height. Direction (DIR) = bearing (true) in degrees.
**Gust Duration Index is in arbitrary units.
MOISTURE CONTENT OF SURFACE FLY ASH MATERIAL
DATA FROM FIVE FACES OF STOCKPILE ON DAYS HAVING WIDEST VARIATION OBSERVED
(PERCENT BY MASS)
NORTH FACE EAST FACE SOUTH FACE WEST FACE TOP AVERAGE
Day 1 2.7 1.6 1.1 1.9 1.7 1.8
Day 2 1.4 1.0 1.3 2.6 0.8 1.4
OVERALL AVERAGE hJOISTURE CONTENT 1.6
6-11
-------
TABLE 3: PARTICLE SIZE DISTRIBUTION
OF SURFACE FLY ASH MATERIAL
SIEVE
NUMBER
10
50
100
140
180
250
325
Pan
SIZE OF SIEVE
OPENINGS (ym)
2,000
300
150
106
83
58
45
RANGE OF PARTICLE
SIZES COLLECTED (ym)
> 2
300
151
107
84 -
59 -
46 -
< 45
,000
- 2,000
- 300
- 150
106
83
58
PERCENT OF
TOTAL SAMPLE
22.88
24.18
13.26
12.82
7.45
14.74
4.61
.07
100.01
Results
Wind Speed Reduction
A measure of the effectiveness of the wind screen can be shown
through an analysis of the reduction of wind speeds resulting frotn the
presence of the wind screen material. The data sample that can be
employed for this purpose is somewhat larger than is shown in Table 2
above since it does not matter about the success of the concentration
measurements. For this efficiency, the tests are used which were
conducted when the wind screen was in place and wind measurements were
made at the same heights above the ground both upwind and downwind of the
wind screen.
The average wind speeds and peak gust wind speeds as summarized in
Table 4 show that the average downwind wind speeds are reduced to
36 percent of the upwind value and the peak gusts are reduced to
41 percent of the upwind value. These reductions should be expected to
vary somewhat by location for both distance downwind of the wind screen
and for height above the ground. The values presented here should be
representative of the fugitive windblown dust measured in these tests
because the downwind wind speed was measured adjacent to the stockpile
and at the height of the stockpile.
6-12
-------
TABLE 4: WIND SPEEDS UPWIND AND DOWNWIND
OF WIND SCREEN AT 4 FEET HEIGHT (mph)
TEST
NUMBER
11
12
13
14
15
16
17
18
19
UPWIND
SPEED
10.5
11.0
10.5
10.0
12.5
12.7
13.0
11.5
10.5
DOWNWIND
SPEED
4.8
4.0
3.7
3.5
4.0
4.0
4.5
4.0
4.0
UPWIND
PEAK GUST
16.4
17.4
16.5
16.0
19.6
18.4
19.8
18.8
16.0
DOWNWIND
PEAK GUST
5.6
7.5
6.9
7.0
7.9
6.2
10.1
6.5
7.5
Means
% REDUCTION
WIND SPEED
54
64
65
65
68
69
65
65
62
64
Z REDUCTION
PEAK GUST
66
57
58
56
60
66
49
65
53
59
Airborne Concentration Reductions
The most graphic and direct measure of the effectiveness of the
wind screen as an inhibitor of windblown fugitive dust can be seen by
close examination of the TSP and IP concentration values as shown in
Table 1. Several relations of significance are described in the
following paragraphs.
Without the wind screen, an average of 148 y g/m increase in TSP
concentration is seen at 20 feet downwind. By the time the windblown
plume has traveled to 50 feet downwind the average increase in
concentration value has dropped to 31 yg/ffl above the upwind value.
With the wind screen in place, an average of 76 yg/nr increase in
concentration is seen at 20 feet downwind. However, an increase of
55 yg/m3 is still seen at 50 feet downwind. These differences are of
significance. Without the wind screen, the increase in TSP concentration
above background levels is high, but drops off rapidly probably due to
deposition. With the wind screen, the increase in TSP concentration
isn't as high above background levels as in the case without the screen
but the elevated concentrations remain elevated for a longer period.
This latter case is probably due to increased turbulent suspension of
particulates, where the increase in turbulence results from the air
flowing over the wind screen. fi ,_
-------
It can be seen from the wind speed averages and the wind gust
duration indices (as shown in Table 2) that the winds were higher during
the tests with the wind screen than during the tests without the screen.
Therefore, there was a greater potential for windblown fugitive emissions
during the tests with the wind screen. Some adjustment should be made so
as to allow a more direct comparison between the two sets of
concentration values. Probably the best manner in which such an
adjustment can be is to normalize the mean concentrations to the Gust
Duration Index (I), since this Index incorporates both wind speed and
duration. Such a normalization results in the following:
Normalized Concentration Without Screen - —=— - —rr • 2.85
and,
Normalized Concentration With Screen * -•;— « 777 • 1.27
I oU
then the ratio between concentrations with the wind screen and
concentrations without the wind screen is simply,
1.27
This says that when TSP concentration differences at 20 feet downwind are
normalized to a common base (wind conditions), the wind screen reduces
the windblown TSP dust concentration to approximately 45 percent of what
it is without the wind screen.
It is also seen from Table 1 (as noted above) that the
concentration differences stay higher with distance with the wind screen
than without it. Therefore, to obtain a more representative measure of
the total effectiveness of the wind screen, this difference with distance
should also be incorporated. This incorporation can be made as follows.
6-14
-------
(Ax 20 + Ax 50)
A 2 y .
T * 1 • .
Normalized Mean Concentration Without Wind Screen -^ 1 * « 1.73
and,
Normalized Mean Concentration With Wind Screen -\ * / « 1.10
The ratio between mean conditions with the wind screen and without it is
simply,
1.10
This then says that the total effectiveness of the wind screen is to
reduce the downwind fugitive windblown TSP emissions to 64 percent of
what they were without the wind screen.
The situation is considerably different for IP concentrations.
Also from Table 1 it is seen that the mean differences in IP
concentrations without the wind screen are 29 and 25 u g/m at 20 feet
and 50 feet, respectively. With the wind screen in place the mean
differences in IP concentrations is 23 and 29 ug/m . Considering the
accuracy and resolution of particulate sampling with the hi-vol sampler
and gravametric assay, these concentration differences are for all
intents and purposes the same. If, however, the same normalization
exercise as was used above is carried out, it is seen that,
x 20 + Ax50
\
J
>
Normalized IP Concentration Without Screen -^ ? f • .52
and,
(Ax 20
/A
Normalized IP Concentration With Screen -^ / - .52
then, the ratio between conditions with the wind screen and without it is,
.52 _
751 1
6-15
-------
To reiterate, the wind screen, although it reduces the wind speed in its
lee area where the windblown emissions originate, appears to make no
difference in the fugitive emission of windblown IP (small particle)
material.
It must be said that concentration data of the type collected
during this set of experiments are inherently variable. This variability
is due in large part to variations among the atmospheric processes which
result in the windblown activity. Also, various aspects of the sampling
system used to make the concentration measurements lead to data
variabilities. In this analysis, a great deal of care was taken to
screen the data collected in order to assure that as many of the
irregular occurrences as possible were eliminated from the data base.
This was done in an attempt to remove those factors that might obscure
the true behaviors. It is felt that this data screening was quite
successful and the resulting data base is of comparative high quality.
However, by screening the data base and eliminating those cases during
which atypical conditions existed, the data base was made to be quite
small. For this reason it is recommended that the results presented here
be respected as preliminary.
Conclusions
Wind screen material will reduce both the wind speed and fugitive
windblown dust emissions from stockpiled material. The amount of wind
speed reduction was found to be 64 percent of the mean wind and
59 percent of the peak gust values. That is the mean wind was reduced to
36 percent of the upwind value and the peak gusts were reduced to
41 percent of the upwind value in the conflguation tested. This
configuration was with a six feet high wind screen where the downwind
wind speed was measured at a point four feet above the ground and sixteen
feet downwind from the wind screen.
The amount of reduction of windblown fugitive TSP is 55 percent
lower at a location 20 feet downwind of the stockpile. That is, the TSP
was 45 percent of what it was without the wind screen. When considering
6-16
-------
both the data from 20 feet and 50 feet downwind of the stockpile, the TSP
reduction was 36-percent, or 64 percent of what it was without the wind
screen.
The analysis of reduction of IP (snail particulate emissions) did
reveal some small differences between the case with the wind screen and
without it. However, the differences are at an insignificant level. It
is concluded that the presence of the wind screen makes no significant
difference in the amount of windblown fugitive IP emissions.
The overall conclusion is that the wind screen both reduces the
wind speed in its aerodynamic wake and also reduces the particulate
emissions. There remains a question about the wind screen's
effectiveness at reducing the emissions of inhalable particulate matter
( < 15 urn) and additional measurements will be necessary to resolve that
issue.
6-17
-------
EFFECTS OF STREET SWEEPING ON FUGITIVE EMISSIONS
FROM URBAN ROADWAYS
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
Donald F. Gatz
Atmospheric Chemistry Section
Illinois State Water Survey
P.O. Box 5050, Station A
Champaign, Illinois 61820
ABSTRACT
To determine the effects of street sweeping on urban particle
concentrations, three standard high volume samplers were operated near a four
lane road in a commercial area, and another sampler was operated in a
residential area, in Champaign, Illinois. Results show that street-related
sources increased the total airborne particle mass about 20 yg/m3 at 7 m
downwind when the wind had a component perpendicular to the street. Analysis
of variance, comparing periods with and without sweeping in spring and summer
sampling periods, showed higher downwind-upwind concentration differences in
summer than in spring for east and west winds, apparently caused by higher
street dust loadings in summer, but no effect from street sweeping.
7-1
-------
1 INTRODUCTION
Provisions of the Clean Air Act enacted in 1977 required states to revise
their State Implementation Plans (SIPs) for all areas that had not attained
National Ambient Air Quality Standards (NAAQS). The Illinois SIP for Total
Suspended Particulates (TSP) was conditionally approved by the U.S.
Environmental Protection Agency (USEPA) with certain minor deficiencies. One
of the reasons for the conditional approval of the Illinois TSP SIP was
inadequate documentation of the impacts and effects of various controls on
non-traditional fugitive sources of TSP. These sources include paved and
unpaved roads, parking lots, and agricultural lands. This and similar studies
have been designed to correct those deficiencies and will be submitted to the
USEPA as part of the SIP for that purpose. The results of these studies will
be used by the Illinois Environmental Protection Agency (IEPA) to define
further the estimated impacts of non-traditional fugitive dust sources on TSP
non-attainment areas throughout the state. They will also be used to refine
the control strategies which may need to be applied to various non-traditional
fugitive dust sources.
The purpose of this study was to assess the effects of street sweeping on
air quality in Champaign, Illinois, by comparing aerosol concentrations
measured by high volume (hivol) samplers in the presence and absence of regular
street sweeping.
This study was planned so as to utilize the program of regular street
sweeping being conducted in Champaign by a research group from the Water
Survey's Surface Water Section and funded by USEPA as part of the Nationwide
Urban Runoff Program (NURP); the Principal Investigator of that study is
Michael L. Terstriep, Head of the Surface Water Section.
2 EXPERIMENTAL METHODS
2.1 AEROSOL SAMPLER LOCATIONS
A map of a portion of Champaign, showing the aerosol sampling sites within
the NURP study basins, is shown in Figure 1. Three hivol samplers were
operated in the commercial "Mattis South" basin along Mattis Avenue. As shown
in the figure, two samplers were located on the west side of Mattis Avenue, and
one on the east side. An additional hivol sampler was operated in a
residential Champaign neighborhood; this is designated the John Street site in
Figure 1.
7-2
-------
i
CO
JOHN
STREET
NORTH
BASIN
NO. 5
MATTIS
SOUTH
BASIN
NO. 2
MATTIS, EAST
BASIN ^
NO. 3
JOHN
STREET
SOUTH
- BASIN ^
NO. 4
MATTIS.
WEST (N)
MATTIS. WEST (S)
CHAMPAIGN. ILLINOIS STUDY AREA
IEPA-ISWS URBAN STORM RUNOFF STUDY
NATIONWIDE URBAN RUNOFF PROGRAM
Aerosol Sampler Site H Automatic Sampler Sites
SCALE 9 Wet and Dry Fallout Sampler
Raingage
Figure 1. Map of a portion of Champaign, Illinois, showing aerosol sampling sites,
drainage basins, and sampling sites of the Water Survey's project in the
Nationwide Urban Runoff Program (Modified from Terstriep et al, 1981).
-------
The hivol samplers near Mattis Avenue were all located 7.0 m ( t 0.4 m)
from the curb to minimize any effects of distance from the street on the
concentrations measured by the various samplers. At the John Street site the
hivol was 17.6 m north of the curb.
2.2 AEROSOL SAMPLING METHODS
All the aerosol samplers were installed so that their inlets were 2 m above
ground level. Thus, they conformed to the standard height, i.e., between 2 and
15 m, at which hivol inlets must be placed. The hivol samplers were all
positioned so that their inlets faced the street. It was important to be
consistent on this detail since there is evidence (Ortiz, 1978) that hivol
sampling effectiveness varies with sampler orientation to the air flow, at
least in strong winds.
The standard hivol samplers were provided and calibrated by the IEPA, which
also provided replacement pumps when airflow fell below minimum acceptable
rates. The IEPA also provided glass fiber filters for the hivol samplers,
weighed the filters before and after exposure, and calculated TSP
concentrations.
The hivol filters were changed at approximately 9:00 a.m. each Monday-
through Friday, from 16 April to 20 October, 1981. This schedule provided five
24-hr duration filters each week from each sampler- The first filter of the
week began Monday morning and the last ended Saturday morning.
2.3 STREET SWEEPING OPERATIONS
The street sweeper made single passes along both east and west curbs of the
Mattis South basin once per week, normally on Tuesday morning, between 15 April
and 26 May 1981. The sweeper used was a 1973 Elgin, Model Pelican "S," a
three-wheeled mechanical sweeper with right and left side gutter brooms and a
main rotary broom. Its sweeping path using one outside broom was 2.4 m (8 ft)
wide.
Measurements of collection efficiency by Bender (1982) indicated that the
sweeper collected about 67/& of the total dust mass on the street. This
approximate efficiency was maintained for all particle sizes down to 125 ym,
below which efficiency decreased markedly. Other measurements by Bender (1982)
indicated that 46$ of the total dust mass was in particles greater than
in diameter.
7-4
-------
In six sweepings of the Mattis South basin, the median mass collected was
28.8 g/curb m (102 Ib/curb mi). The maximum collection was 44.5 g/curb m (158
Ib/curb mi) and the minimum 13.2 g/curb m (47 Ib/curb mi).
2.4 SUPPORTING DATA
Several kinds of supporting data were also collected to aid in
interpretation of the air quality data. Precipitation was measured at the
Mattis Avenue and John Street raingage sites, as shown in Figure 1, and
provided to us, along with street sweeping data, by the Water Survey NURP
project mentioned earlier. Wind measurements were made at the Water Survey
Headquarters in Champaign, about 4 km east of the Mattis Avenue sites.
3 RESULTS AND DISCUSSION
An illustration of a small portion of the available data, plotted against
time, is given in Figure 2. The data include TSP measurements at four sites,
along with daily rainfall amount, dates of street sweeping, and winds. Wind
directions are shown by direction category. The westerly wind category (W in
Figure 2) includes days when the hourly mean wind direction was from the 202 to
338 degree sector for at least 75% of the hours between 6 a.m. and midnight.
Wind direction was not considered relevant in the remaining hours of the
sampling period because only a very small fraction of the traffic occurs then.
Similarly, the easterly wind category (E) includes days when the hourly mean
wind direction was from the 22 to 158 degree sector at least 75% of the hours
in the same time interval. The "other" wind direction category (0) includes
all other days. Mean wind speed during sample collection and daily maximum
1-minute gust speed are also shown.
Figure 2 shows that rainfall was frequent and relatively heavy during the
period shown, with seven days having 1.27 cm (0.5 in.) or more. Streets were
swept at approximately weekly intervals during the period. Figure 2 includes
only a small fraction of the data collected, but it illustrates several
relationships that occurred throughout the sampling period. One of these is
the generally high correlation between TSP observations from the several
samplers, including the one located in a residential area. Generally, on days
when TSP concentrations were unusually high at one site, they were unusually
high at all sites. There were still differences in concentration between
sites, of course. A very obvious difference in Figure 2 was that the
residential (John) site had consistently lower concentrations than the
commercial sites. Differences between collectors on opposite sides of the
commercial roadway are also apparent in Figure 2.
7-5
-------
280
200
100
20
1.5
1.0
•£0.5
TSP
MATTIS, EAST SIDE
MATTIS, WEST SIDE (N)
MATTIS, WEST SIDE (S)
JOHN STREET
RAINFALL
i
I
_ p
fin ,
I
n
n
i
n
I ' '
STREET SWEEPING
I I
I I
I
I
I
I
w
E
0
WIND DIRECTION CATEGORY
- O OO OO
O
-O OO O
II I 1
1 i
0 O
o
O OO O
o
O 0
OO OO OO O
1 1 1
-
18
15
- AVERAGE WIND SPEED
1 - 1
60
40
0
i 1 i 1 1 1 r 1 1
~ MAXIMUM WIND GUST
il 1 1 1 1 1 !
16 20 25 28 1 5 10 15 20
Apr May
STARTING DATE FOR FILTERS
Figure 2. Illustration of data collected in 1981 for evaluation of effects
of street sweeping on air quality.
7-6
-------
In assessing possible effects of street sweeping on air quality, it was
first necessary to establish whether the test roadway was indeed a source of
aerosols. To examine this question, I divided the data according to the wind
direction categories mentioned earlier- This was done to assess which samplers
were upwind, and which downwind, when the wind direction was either easterly or
westerly (i.e., having a large cross-wind component, relative to Mattis Avenue)
and to identify those days when the wind (1) was from either north or south, or
(2) blew from both east and west at different times during sample collection,
so that it was difficult to assign upwind and downwind. Results for the east
and west wind cases are shown in Figure 3, and those for the "other" winds in
Figure 4.
Figure 3 includes only measurements made in easterly or westerly winds, and
shows a clear tendency for the downwind samplers to have higher TSP values.
The upwind samplers experienced higher TSP concentrations with westerly winds.
For example, the upwind means were 71 ug/m3for 26 days with west winds and 60
ug/m^for 30 days with east winds, a difference of 18J, relative to the east
wind value. For both east and west winds, however, the downwind samplers had
higher TSP values, clearly showing the road to be a source of airborne
particles. As before, the solid line in the figure represents perfect
agreement between upwind and downwind samplers. The dashed line is the least
squares regression line for estimating the downwind concentration, D, from the
upwind value, U: D = 0.94(U) + 21.7. Since the slope of the regression line is
close to 1.0, the intercept, 21.7 ug/nr gives a good indication of the extra
airborne dust concentration that may be attributed to the roadway. The slope
near 1.0 shows that the added TSP concentration caused by the roadway is
approximately constant for all upwind concentrations.
To consider the possibility that the differences in mean upwind TSP
concentrations were the result of differences in mean wind speed, the following
wind speed data are presented. The mean value of the daily mean wind speed,
for 26 west wind samples, was 6.4 m/sec (S.D. =2.0 m/sec), while the
corresponding value for 30 east wind samples was 5.0 m/sec (S.D. = 2.1 m/sec).
Analogous means for the maximum gusts were 24.0 m/sec (S.D. = 8.6 m/sec) for
west wind samples and 18.2 m/sec (S.D. = 5.4 m/sec) for east wind samples.
Thus, the relative differences in mean wind speed (28$) and mean maximum gust
speed (32?) between easterly and westerly winds were somewhat greater than the
relative differences in TSP concentrations, but wind speed differences cannot
be ruled out as a possible cause. Another possible cause is that the site was
closer to agricultural sources with west winds, since it was located near the
western edge of the Champaign-Urbana metropolitan area.
Figure 4 shows the relationship between TSP concentrations on the east and
west sides of Mattis Avenue for the "other" wind category. There was no strong
tendency for either side of the street to have higher values, as would be
expected in the absence of a strong wind component perpendicular to the street.
The evidence presented thus far makes it clear that the street was indeed
the source of airborne particles, since the downwind sampler had higher
7-7
-------
° WEST WINDS
EAST WINDS
25
50
75 100 125 150 175 200 225 250
DOWNWIND TSP, M9 m'3
Figure 3« Comparison of TSP concentrations upwind and downwind of Mattis
Avenue on days when winds were predominantly from the east or
west.
7-8
-------
a.
CO
111
o
C/5
LU
S
180
162
144
126
108
90
72
54
36
18
i i
MATTIS AVENUE
18 36 54 72 90 108 126 144 162 180
EAST SIDE TSP,M9 m~3
Figure 4. Comparison of TSP concentrations on the east and west sides of
Mattis Avenue on days when winds were not predominantly from the
east or west.
7-9
-------
concentrations in both east and west winds, but neither side had- predominantly
higher concentrations with "other" winds. Next we address the question of
whether street sweeping had any effect on reducing the strength of this source.
This is done by comparing downwind-upwind differences during periods with and
without regular weekly sweeping. A day was considered to be in the "swept"
period if the street was swept on that day or any of the previous seven days,
and in the "unswept" period otherwise. A second comparison examines
commercial-residential differences with and without weekly sweeping (by the
same criteria) at the commercial site. (The residential site was not in one of
the experimental or control basins, and thus was swept irregularly, at
intervals of 2-4 weeks, during the sampling period.)
Table 1 presents results of a two-way analysis of variance (Brown, 1977) on
downwind-upwind differences ( TSP1). It examines both sweeping and time as
sources of variance. The upper part of the table gives mean downwind-upwind
differences in TSP concentrations for swept and unswept periods in both
"spring" (up to 30 June) and "summer" (after 30 June). There appear to be
differences between seasons (with higher concentrations in summer than in
spring), but not between periods of sweeping and not sweeping. The analysis of
variance table in the lower part of the table confirms that there were no
significant differences between periods of sweeping and no sweeping, but also
indicates that the differences between seasons were significant at the 0.10
level, but not at the 0.05 level. Table 2 presents results of a similar
analysis of variance on commercial-residential TSP differences ( TSP2). Again,
differences between seasons appear large (with concentration differences higher
in spring than in summer in this case), but differences between swept and
unswept periods look small. The analysis of variance table in the lower part
of Table 2 confirms that the differences between swept and unswept periods were
not significant, but also shows that seasonal differences were significant at
the 0.0001 level. Interactions between sweeping and time were not significant
in either comparison.
The main question addressed by this paper has now been answered. Although
the test roadway was shown to be a source of TSP, no evidence was found to
indicate that street sweeping had any effect, either beneficial or detrimental,
on air quality at 7 m downwind.
Table 3 summarizes TSP concentrations and the two TSP differences discussed
above. Mean TSP values are shown for three different data sets: (1) all
samples, (2) the samples used in the downwind-upwind comparison, and (3) the
samples used in the commercial-residential comparison. It is apparent that the
mean TSP concentrations in the commercial (Mattis ) and residential (John)
sites do not vary much between sample sets. For example, the Mattis mean shows
a maximum of 87 Pg/m3 and a minimum of 84 ug/m3 in the three data sets shown.
Similarly, at the John site the highest mean was 65ug/m3and the lowest 62
Pg/m3 Thus, regardless of the data set, mean TSP values were higher at both
sites in spring than in summer. The Mattis site had higher concentrations by
about 20ug/nH in the spring and about 8 pg/m3 in the summer -
7-10
-------
Table 1. Results of 2-way analysis of variance
on upwind-downwind TSP differences (ug/tn ) across
the test roadway
Spring
Summer
No. of Cases
Mean
S.E.M.
Source of
variance
Time
Sweeping
Interaction
Error
Sum of
squares
603.4
20.7
23. A
8369.3
Table 2.
Swept
12
12.9
4.2
Degrees
freedom
1
1
1
41
Unswept Swept
11 4
12.8 19.8
5.2 1.8
of Mean
square F
603.4 2.96
20.7 0.10
23.4 0.11
. 204.1
Unswept
18
23.0
3.1
Tail
probability
0.0931
0.7518
0.7368
Results of 2-way analysis of variance
of commercial-residential TSP differences
(ug/m )
Spring Summer
No. of cases
Mean
S.E.M.
Source of
variance
Time
Sweeping
Interaction
Error
Sum of
Squares
2251.4
8.1
159.1
12166.8
Swept
27
22.0
3.0
Degrees
freedom
1
1
1
86
Unswept Swept
19 8
19.6 7.2
2.0 3.5
of Mean
square F
2251.3 15.91
8.1 0.06
159.1 1.12
141.5
Unswep t
36
11.0
1.7
Tail
probability
0.0001
0.8116
0.2919
7-11
-------
Table 3. Summary of TSP concentrations and differences
by season for three different sample sets
Spring
Summer
Sample set
All samples
Site
*
Mattis
John
N Mean -
53 86 -
52 65 -
S.E.
6
5
N Mean
77 62
74 54 -
S.E.
2
2
A TSP1
Mattis 23 87 - 7
Downwind-upwind ,
difference
23 13
33 65
22 22
A TSP2
***
Mattis
John
Mattis-John
difference
46 84 - 5
46 62 - 4
46 21
68 63
68 55
44 10
*Includes days with at least one sample from both sides of Mattis Ave;
both sides weighted equally.
**Includes days with measurements on both sides of Mattis Ave., and east
or west winds (as defined in the text).
***Includes days with measurements on both sides of Mattis Ave. and at
John St., with no restriction on wind direction.
7-12
-------
TSP concentration differences, however, display different seasonal behavior
depending on the data set used. The mean commercial-residential difference
(&TSP2) was higher in spring than summer, as were TSP concentrations at both
sites. The mean downwind-upwind difference (ATSP1 ), on the other hand, was
larger in the summer, when the Mattis site had lower concentrations.
The higher TSP concentrations in spring are consistent with an agricultural
soil dust source, which would be expected to be stronger in the spring from
tilling operations, lack of ground cover by crops, and generally stronger
winds. It is reasonable for the Mattis site to have higher TSP concentrations
than the John St. site because it is closer to the sources with prevailing
winds, and because the John St. site is located in a somewhat older part of
town, where the more mature trees may help to remove particles from the
atmosphere.
The downwind-upwind differences were greatest in summer, when TSP
concentrations were lower at both the John and Mattis sites. Others (e.g.,
Cowherd and Englehart, this Symposium) have found that roadway emissions depend
on both traffic volume and roadway surface silt loading. We could find no
evidence for significantly greater traffic volumes in summer than in spring,
but according to data provided by Bender (1982) , the mean and standard error
of 11 spring measurements of total street loadings were 108 ± 5 g/curb-m, while
the same values for 10 summer measurements were 189 111 g/curb-m. Thus, the
spring/summer ratio of loadings, 1.75, is very close to that of ATSP1, 1.69,
and it appears that differences in street surface loadings between seasons
could well have caused the observed seasonal differences in ATSP1.
4 SUMMARY AND CONCLUSIONS
To determine the effects of street sweeping on urban TSP concentrations,
three standard hivol samplers were operated near a four lane commercial
roadway, and an additional sampler was operated in a residential area of
Champaign, Illinois, between April and October, 1981. When winds were blowing
across the test roadway in either direction, downwind concentrations at 7 m
from the curb exceeded upwind by an average of about 20yg/m^ . in a comparison
of downwind-upwind TSP concentration differences during a period of regular
weekly street sweeping with those from another period when sweeping was
discontinued, analysis of variance was unable to detect a significant effect of
sweeping on the amount of TSP produced by the roadway. Thus, for a commercial
urban roadway clearly shown to be a source of airborne particles, street
sweeping could not be shown to have any effect in reducing the amount of dust
produced.
7-13
-------
5 ACKNOWLEDGEMENTS
This research was carried out under support from the Illinois Department of
Energy and Natural Resources, Project 10.093. Mr. William P. Murphy was the
Project Manager.
The results were achieved only through the dedicated efforts of Susan
Wiley, who collected most of the samples and performed the data analysis, and
Bruce Komadina, who devised a convenient means of mounting the hivols and
supervised their installation. Thanks are also due Randall K. Stahlhut, for
programming assistance, Eberhard Brieschke, who helped with the installations,
and Mr- Kenneth Porter, Mr. Lawrence Boastick, and Dr. and Mrs. Glenn
Stout, who allowed us to operate samplers on their property- The Illinois EPA,
through the assistance of Mr. Arden Ahnell, Mr- Bob Button, and Mr- John
Shrock, provided, calibrated, and maintained the standard hivols, and provided
and weighed the filters. We also thank Mr. Michael Terstriep and Mr. Michael
Bender for their cooperation in supplying data.
6 REFERENCES
Bender, G. M., 1982. Personal communication. Mr. Bender is associated with
the Water Survey's Nationwide Urban Runoff Program.
Brown, M..B., Editor, 1977: BMDP-77 Biomedical Computer Programs, P-Series,
University of California Press, Berkeley, 880 pp.
Ortiz, C. A., 1978: Aerosol collection characteristics of ambient aerosol
samplers. M.S. Thesis, Texas A & M University, College Station, Texas.
Terstriep, M. L., G. M. Bender, and D. C. Noel, 1981: National Urban Runoff
Project, Champaign, Illinois. Evaluation of the Effectiveness of
Municipal Street Sweeping in the Control of Urban Storm Runoff
Pollution. First Annual Report, Prepared for the Illinois
Environmental Protection Agency and the U. S. Environmental Protection
Agency, Region V. Illinois State Water Survey, Champaign.
7-14
-------
EVALUATION OF THE EFFECTIVENESS
OF CIVIL ENGINEERING FABRICS AND
CHEMICAL STABILIZERS IN THE REDUCTION
OF FUGITIVE EMISSIONS FROM UNPAVED ROADS*
by
Alan G. Larson**
Donald L. Shearer
TRC ENVIRONMENTAL CONSULTANTS, INC.
Englewood, CO 80111
Dennis C. Drehmel
U.S. EPA, INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
RESEARCH TRIANGLE PARK, NC 27711
Gary W. Schanche
U.S. ARMY CORPS OF ENGINEERS,
CONSTRUCTION ENGINEERING RESEARCH LABORATORY
CHAMPAIGN, IL 61820
This paper has been reviewed in accordance with the U.5. Environmental
Protection Agency's peer and administrative review policies and approved ior
presentation and publication.
(*)Although this is not the actual presentation made at the May 1982
meeting, it describes the same work and covers the same general
time period. It represents work funded, in part, under EPA con-
tract 68-02-3115. The title of A.G. Larson's presentation was
"Evaluation of Road Carpets and Chemical Road Dust Suppressants.
(**) Current address: The Kentwood Moore Co., 5690 Denver Tech
Center Blvd., Englewood, CO 80111.
8-1
-------
Introduction
Emissions of fugitive dust from unpaved roads constitute a significant
portion of the total suspended particulate matter (TSP) emitted in the
United States. These roads may be managed by local governments for the
use of the public or managed privately by an industrial user (e.g.,
mining, lumber, oil and gas, manufacturing). Emissions of fugitive dust
from unpaved roads may exacerbate an air quality problem, and possibly
cause violations of the National Ambient Air Quality Standards (NAAQS).
Therefore, many state and local governments have required dust control
measures to be used on unpaved roads.
The U.S. EPA's Region 8, its Industrial Environmental Research
Laboratory-RTF (IERL), and the U.S. Army Construction Engineering Research
Laboratory agreed to conduct a field study at Ft. Carson, Colorado, to
evaluate the effectiveness of two potential control techniques—civil
engineering fabrics (road carpets) and chemical dust suppressants. TRC
was contracted by the EPA to design and manage a short-term study program
to evaluate these materials and to coordinate with the Army on a long-term
study- The Army was enthusiastic about conducting the tests at Fort Carson
because of the significant number of unpaved road miles on the facility.
The Army wanted to find an effective, but inexpensive, control technique
for fugitive road dust control. EPA Region 8 was interested in
participating in the study because an effective solution to the fugitive
road dust problem is of concern to Region 8. EPA-IERL was interested in
evaluating the overall effectiveness of these two specific dust control
techniques.
The short-term study was designed to measure the effectiveness of each
fugitive dust control application at two points in time—immediately
following application of the road carpet or chemical suppressants
(measuring maximum control efficiency) and after a few months of road use
by local vehicles (measuring average control efficiency). Also, by
limiting the measurements to a short time period (in this case,
approximately 1 hour), the influence of weather (i.e. precipitation) and
abnormally high winds was minimized and any bias in the results due to
these elements reduced. The long-term study was designed to measure the
overall efficiency of each dust control application. The effects of
weather, abnormal traffic patterns, and inconsistent wind patterns were
measured. By reclaiming the high-volume (hi-vol) filters every 3 days,
the decay in the effectiveness of the chemical test sections could also be
observed. In these studies, TSP and inhalable particulate matter
originating from each section of road carpet or chemical dust suppressant
were measured.
Several manufacturers of the road carpets and chemical dust
suppressants were contacted. It was the objective of the tests to
evaluate as many of the generically different products within each control
technology as possible. Space constraints limited the test to three road
8-2
-------
carpet and three chemical dust suppressant test sections. The road
carpets consisted of spunbonded woven polypropylene, needlepunched woven
polypropylene, and nonwoven polypropylene sections. A cross-sectional
drawing of the road carpet application is presented in Figure 1. The
chemical dust suppressants tested were an oil-latex emulsion, an
elastomeric. asphalt emulsion, and a calcium 1ignosulfonate. An untreated
road section was also maintained as a control.
Methodology
The test road is an existing gravel road located on Ft. Carson Army
Base, just south of Colorado Springs, Colorado. Traffic on the road is
generally light duty trucks and automobiles. Occasionally, a heavy duty
truck or troop transport uses the road. The test road was divided into
six test sections and one control section. Each section was approximately
200 yards (183 meters) long and 30 feet (9 meters) wide.
The three road carpets were each installed over the gravel road,
employing standard road constructor! methods with special techniques
applied when recommended by the manufacturers. The road carpets were
covered with approximately 6 inches (15 cm) or 2-inch (5 cm) or smaller
aggregate.
For application of the oil-latex emulsion and the elastomeric asphalt
emulsion, the gravel road was first scarified to a depth of 4 to 6
inches. The latex emulsion was applied as 1 part emulsion to 5 parts
water solution and blade-mixed by several passes with a road- grader. The
scarified section was then graded smooth and packed. After scarifying,
the asphalt emulsion section was windrowed with a road grader and the
emulsion applied over the scraped section. After several windrowing
passes with the grader to mix the road base and emulsion, and further
application of the emulsion, the section was smoothed and packed. The
calcium lignosulfonate section was graded smooth and the chemical applied
to the gravel road surface. No scarifying was recommended by the supplier.
Hi-vol samplers were installed on platforms along the test road. The
platforms each supported a pair of hi-vols—one for collecting TSP and one
for collecting particles in six size fractions smaller than 30 ym. The
hi-vol pairs were located 14 feet (4 meters) above ground downwind of each
test section at a distance of 50 yards (46 meters) from the downwind edge
of the section. One pair was located approximately 75 yards (69 meters)
upwind of the test road (Figure 2). The operation of the hi-vols was
regulated by wind direction controllers. The hi-vols operated only when
the wind was out of a direction ^60° either side of a line normal to the
road. This operating mode helped reduce cross contamination from one test
section to another. An additional hi-vol to measure background TSP was
located in an undisturbed area of the fort several miles downwind of the
experimental area.
8-3
-------
COARSE
AGGREGATE
« 1 •* * ••• • - • — •.•••••«.. ••»*•* ••• ••»• • • ^ • w • • •
-V:::v^;-v;:.Vr.:^^-:v:V.;.v://>::^^^^/:»;v>v>VV-y:^
• •: •-'• :.V'. v>''-/.-.'-."'-.:: • • ::-V-^; ;•".•:•: •>••: ••::: •:'•• •/•:-. -\V •/.;•:• :•:;.-: •• V: v; v.: f
:>:.« {•-•: .":••::;:.••;:.•;.• ;.\<.;::- >:-.v.v.^'.:-0': V >:' !>•:•: :VVA..^v:v;.; y.:j
••#vy:.;:^-;^
hTs*^l'laS^^^S^3a=22S^ "^*^
r»-Tra i-1'.*s>*^C*rSjy^j.iJLji^'^g^Vi^SirmrJS
ROAD
CARPET
EXISTING
GRAVEL
ROAD
FIGURE 1: ILLUSTRATION OF ROAD
CARPET APPLICATION
8-4
-------
DBACKGnOUND HI-VOL
RESIDENTIAL AREA
D
200YDSH
00
I
en
D
[
50 YDS
D
DHI-VOLS
D
D
a
DMETEOROLOGY
STATION
D
PAVED
ROAD
FABRIC
FABRIC FABRIC CONTROL
*2 *3 SEC
CHEMICAL
iM
>\ o^^ D UPWIND HI-V
ELEMENTARY
SCHOOL
IHAf-FlU -• »-«-w
COUNTER METER
CHEMICAL
*,2
CHEMICAL
OL V.
TRAFFIC
COUNTER
PAVED ROAD
NOT TO SCALE
N
FIGURE 2: FIELD TEST DESIGN
-------
Wind speed and direction were measured at a meteorological station
located approximately 150 yards (137 meters) downwind of Chemical Section
3. The test road was equipped with traffic counters at each end to
measure the traffic volume and patterns on the road and a speed meter in
the control section to measure vehicle speeds along the road.
Following the long-term tests, each chemical will be evaluated for the
following criteria:
• Effectiveness in reducing TSP and inhalable dust particles from the
road surface.
• Frequency of chemical retreatment of test road sections.
• Cost of chemicals for each application and costs over the course of
the 1-year test period.
From these criteria, the cost effectiveness of each chemical dust
suppressant will be determined.
Results
At the time the road carpets were installed, only 2-inch or smaller
sized aggregate was available. This aggregate had a relatively high
percentage (8 percent, by weight, less than 97 y m) of fine materials. At
the request of local residents, the road section was watered to reduce the
dust. Consequently, a hardpan road surface developed and any dust control
potential whi-ch the road carpets may have exhibited was negated. In
retrospect, a washed aggregate appears essential for testing the dust
control capabilities of road carpets. Plans are being made to treat the
road carpet sections with chemical stabilizers and to measure the
effectiveness of the chemicals in reducing fugitive road dust.
Raw data have been compiled for total suspended particulate matter and
for particles less than 30 ym in diameter for the three chemical test
sections, the control section, and the upwind and background stations.
The sub-30 ym particles were collected in six stages, with average
particles sizes of:
Stage 1 6.8 ym
Stage 2 3.3 ym
Stage 3 1.8 ym
Stage 4 0.9 ym
Stage 5 0.5 ym
Stage 6 Remainder of particles
8-6
-------
The respirable range of particles is a combination of Stages 2-6. The
total respirable particle concentration can be expressed as^':
X = (0.19 x S2) + (0.70 x S3) + S4 + S5 + S6
Where: X " respirable particle concentration (ug/m^)
S2 * concentration of particulate matter on Stage 2
S3 = concentration of particulate matter on Stage 3
S4 = concentration of particulate matter on Stage 4
S5 = concentration of particulate matter on Stage 5
S6 - concentration of particulate matter on Stage 6
Particulate matter concentrations have been tabulated for the first
nine long-term sampling periods. These concentrations are presented in
Table I. The particulate matter concentrations measured from the calcium
lignosulfonate, Chemical Section 1, are consistently higher than the
concentrations measured from Chemical Sections 2 and 3 and the control
section.
The manufacturers suggested application rate for the calcium
lignosulfonate was too low to provide sufficient covering of the road
surface or to allow for absorption of the chemical into the road and
therefore did not result in measurable dust control. The manufacturer has
since modified the recommended application procedures to include
scarifying' of the road surface prior to chemical application and increased
dosage of the chemical. The manufacturer has al.so required ambient
temperature consistently above 45-50°F before application of -the
chemicals can begin. Therefore, the collection of any reliable dust
concentrations will have to wait until the calcium lignosulfonate is
reapplied in spring 1981. One thing that appears to apply to the calcium
lignosulfonate test section is that grading the road prior to chemical
application loosened enough surface particles to cause particulate matter
concentrations which are greater than those measured from the control
section. The control section had been packed by the combined effects of
precipitation and traffic.
The concentrations from the road sections treated with the oil-latex
emulsion and the elastomeric asphalt emulsion are generally much lower
than the control section. When compared with the control section, control
efficiencies of the oil-latex emulsion range from approximately 6-70
percent for respirable particles and approximately 13-80 percent for TSP.
Control efficiencies for the elastomeric asphalt emulsion range from
approximately 30-96 percent for respirable particles and approximately
56-86 percent for TSP.
8-7
-------
iAim: i
HuiiHiirutl ttonuviit rut loiiu (im/1" ) "I inii|ilrnlilu mul lul«l mmi'didi'd p«rllciil«t* matttr
oo
oo
SAHI'LUIC
puiuon
12-14-00
Rttiplriiblf
TSP.
17-17-80
Re up 1 cob In
TSP
12-20-BO
RcMplmble
TSP
12-21-80
RtfNplrnble
TSP
13-26-BO
Rcupl ri'blu
TSP
11-29-80
Rniipl rub In
TSP
1 - 2-81
Rcxplrublc
TSI1
1 - 5-81
Hnn|il rnblu
TSP
1 - 8-81
Rciplmblf
TSP
CIICMICAL f)
OIL-UlliX
rnui.sioii
111
282
<>
91
m
54
15
51
27
5)
4 ft
102
}i
82
94
25)
tt
57)
it
909
cimiRoi.
SI:CTIOH
116
)26
20;
454
24
44
51
102
62
105
26
269
76
162
«
274
244
434
uruniD
STATIOII
«•
)65
90
it
26
116
76
66
)8
85
>t
149
)4
)9
1 1
167
<4
102
BACKCROUXO
STATION
9
24
> >
71
1
12
t
14
II
12
n
15
5
1)
* *
24
t«
• t
i The" conctrnt r«t loin irn niu dntii
" At tlm tlimi tlili r«purt UUM ptL-pinml, tliu ilnln fur lIu'Hi! pulnM lirnl IIVHII nuimurrd «nd
uflgliud but li«d not ln'ril prUCCIIIMM! llmnix') I'1" '|i»illly MIUIIIHIIII » |i(ci||r»i«ii.
-------
Because of the impending inclusion of three more chemical test
sections, the short-term tests have been postponed until all test sections
are operational. Two short-term tests have been completed for the
existing test sections. The limited data from these two tests show that
for the oil-latex emulsion, TSP has been reduced more than 50 percent.
For the elastomeric asphalt emulsion, TSP has been reduced by 90 percent.
Additional short-term tests are planned immediately after application of
the new chemicals and again in the summer. (Tables II and III.)
The influence of wind speed, wind direction, and vehicle use rates
have not been compiled for correlation -with the particulate matter
concentrations. These correlations will be made in spring and summer 1981.
Conclusions
The effectiveness of road carpets as a fugitive dust control technique
on unpaved roads is uncertain. The results of field tests to evaluate
this effectiveness are inconclusive. In order to adequately study the
emission control capability of the road carpets, a clean, washed aggregate
must be used over the carpet.
The application of chemical dust suppressants to unpaved roads is an
effective method of controlling fugitive dust. A wide range of control
efficiencies can be realized. For an oil-latex emulsion, up to 70 percent
control of respirable particles and up to 80 percent control of TSP has
been measured. For an elastomeric asphalt emulsion, up to approximately
96 percent control of res.pirable particles and up to 86 percent of TSP has
been measured. The dependence of the control efficiencies on wind speed,
vehicle speed, and weathering of the chemicals is still being evaluated.
The cost:benefit ratio for each chemical application is also being
evaluated. This evaluation will include the initial cost of chemical
application, costs of successive reapplication, and control efficiencies
achieved. From this costrbenefit ratio, chemical dust suppressants which
can be tailored to the economic and dust control requirements of the user
will be presented.
References
(i'V.A. Marble, "A Fundamental Study of Inertial Impactors" (Ph.D.
theses, University of Minnesota, 1970).
8-9
-------
TABLE II
LATEX EMULSION CONTROL EFFICIENCY
DATE
12-14-80
12-17
12-18
12-20
12-21
12-23
12-24
12-26
12-27
12-29
12-30
12-31
1-02-81
1-03
1-05
1-06
1-08
1-09
3-24
3-26
3-27
3-29
3-30
4-01
4-02
CONTROL
SECTION*
325.8
400.0
453.9
215.1
44 .4
80.4
102.0
103.6
105.0
105.0
214.6
269.3
206.9
162.4
204.1
273.6
333.7
433.8
118-7
68.0
37.5
42.3
45.7
58.7
68.0
LATZX
EMULSION*
282.4
146.6
90.8
67.8
54.0
51.8
50.6
52.1
53.1
71.6
102.3
102.3
88.8
82.1
103.6
133.6
137.3
143.5
64.9
60.0
54.5
43.5
37.9
54.4
64.3
CONTROL
/.
13.3
63.4
80.0
68.5
—
35.6
50.4
49.7
49.4
31. S
52.3
62.0
57.0
49.4
49.2
51.2
58.9
66.9
54.7
11.8
—
—
17.1
7.3
5.4
(**) Below this line represents respirable particles;
above this line represents TSP
8-10
-------
TABLE III
EMULSIFIED ASPHALT CONTROL EFFICIENCY
DATE
12-14-80
12-17
12-18
12-20
12-21
12-23
12-24
12-26
12-27
12-29
12-30
12-31
1-02-81
1-03
1-05
1-06
1-08
1-09
3-24
3-26
3-27
3-29
3-30
4-01
4-02
CONTROL
SECTION*
325-8
400.6
453.9
215.1
44.4
80.4
102.0
103.6
105.0
105.0
214.6
269.3
206.9
162.4
204.1
273.6
333.7
433.8
118.7
68.0
37.5
42.3
45.7
58.7
68.0
EMULSIFIED
ASPHALT*
64.1
64.3
64.3
36.1
19.2
25.5
29.2
24.5
20.4
31.9
51.1
51.1
40.1
33.5
42.7
58.1
70.3
90.6
30.6
40.1
44.8
26.1
12.8
21.5
26.8
REDUCTION
Z
80.3
83.9
85.8
83.2
56.8
68. 3
71.4
76.4
80.6
69.6
69.6
81. 0 **
80.6
79.4
79.1
78.8
78.9
79.1
74.2
41.0
—
38.3
72.0
63.3
60.5
i
(**) Below this line represents respirable particles;
above this line represents TSP
8-11
-------
EVALUATION OF WEATHERING CHARACTERISTICS OF DUST
SUPPRESSANT CHEMICAL ADDITIVES
by: William B. Kuykendal, Dennis C. Drehmel, Bobby E. Daniel
U. S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
ABSTRACT
This paper presents the results of an experimental program to evaluate
the effect of natural weathering on selected dust suppressant chemical addi-
tives. A previous study had evaluated the performance of eight commercially
_available additives. Four of these were selected for evaluation of their
'weathering characteristics. The four additives were: Coherex, Lignosulfonate,
SP 301, and Polyco 2151. Three concentrations of each additive were sprayed
on a test panel of pulverized coal and then exposed to the weather for periods
of 30, 60, and 90 days. The test panels were then evaluated in a wind tunnel
to determine the performance of the additives after weather exposure. The
results show that no significant performance degradation occurred for any of
the additives tested.
This paper has be«n reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
9-1
-------
INTRODUCTION
The Environmental Protection Agency has sponsored an active program
to develop technology to suppress fugitive particulate emissions. A
facet of this program has been aimed at the reduction of fugitive particulate
emissions through the application of dust suppressant chemical additives.
In addition to the work described by this paper, two other projects have
been conducted by EPA to evaluate the effectiveness of chemical dust
suppressants. In an earlier in-house study, Drehmel and Daniel1 evaluated
the relative effectiveness of 10 different dust suppressants using a
wind tunnel. Their results will be briefly reviewed in the following
section. Under a contract study, Larson^ evaluated four dust suppressants
in a field study conducted over several months on an unpaved road in
Colorado.
Before chemical dust suppressants can gain widespread use, they
must demonstrate the ability to effectively suppress fugitive emissions
in a manner that is cost effective when compared to other fugitive
control options. A major factor in determining cost effectiveness is
how often the dust suppressant needs to be reapplied. Two factors
determine the durability of a dust suppressant: wear and weathering.
In applications such as inactive storage piles where there is
little or no wear, the only factor affecting durability is weathering.
The purpose of this work is to evaluate the effectiveness of several
dust suppressants after exposure to natural weathering.
PRIOR WORK
The initial wind tunnel evaluation on the relative effectiveness of
dust suppressants by Drehmel and Daniel tested 10 commercially available
products. These products were evaluated in the same wind tunnel using
the same technique described in the experimental section of this paper.
Of the 10 dust suppressants evaluated, 8 gave consistent results and are
summarized in Table 1. Each dust suppressant was evaluated over a range
of application rates to determine the effectiveness of each as the
application rate varied. In order to compare the performance of each
dust suppressant on a common basis, the effectiveness for each, expressed
as entrainment velocity, was determined for an application rate calculated
at a cost of $300 per acre. Entrainment velocity is defined later in
the paper, but the higher the value, the more effective the additive.
It should be emphasized that the dilution ratios determined by the above
method were not necessarily the ones recommended by the manufacturer.
From these results four dust suppressants were selected for the
weathering study: Coherex, Lignosulfonate, SP-301 and Polyco 2151. The
selection was based on having a variety of chemical makeups and manufacturers
rather than on performance alone.
A brief literature review of weathering test procedures was con-
ducted to identify the key weather parameters that should be included in
the experimental plan. Since the purpose of the study was to evaluate
9-2
-------
TABLE 1. ADDITIVE PERFORMANCE RESULTS FROM DREHMEL AND DANIEL
Additive
Coal Dyne
Coherex
CPB-12
Lignosulfonate
Oil & Water
SP-301
Pentron DC-3
Polyco 2151
Chemical
Unknown
Petroleum Resin
in Water
Acrylic Latex
Lignosulfonate
Mineral Oil
Synthetic Polymer
Acrylic in Water
Synthetic '
Copolymer
Additive Supplier Performance
Aquadyne
Wit co
Wen Don
Wen Don
Wen Don
Johnson & March
Apollo
Borden
16.8
13.4
9.8
18.6
18.3
6.4
9.1
15.2
^ntrainment velocity (m/sec) for an application cost of $300/acre.
9-3
-------
weathering effects on dust suppressants, which are usually sprayed on coatings,
the available literature on paint and building materials was reviewed. Three
publications3*4*5 "were located along with one ASTM procedure.6 Each reference
identified the key weather related parameters as: solar radiation (particularly
ultraviolet), relative humidity, rainfall, and temperature. In addition,
it was decided to vary the weathering exposure time so that a degradation rate
could be determined.
EXPERIMENTAL
Since an evaluation of the weathering characteristics was the primary
objective of the study, the selection of the weathering method was important.
Natural weathering has the advantage of being "real world" testing and
requires a minimum amount of experimental apparatus. The disadvantage is, of
course, that the weather related variables cannot be controlled. Simulated
weathering tests conducted indoors do offer a controlled weather environment,
but there are no standard accepted test procedures (see References 3 and 4).
In addition, they are experimentally complex. For these reasons the natural
weathering method was selected.
Having selected the dust suppressants to be evaluated and identified the
key weather related parameters, it was then necessary to develop a test matrix.
The variables selected were dust suppressant, concentration of the dust
suppressant, and weather exposure time.
The selection of the dust suppressant concentrations required balancing
several factors. It was decided to select three concentrations for each of
the four dust suppressants. Furthermore, it was desirable from the stand-
point of the statistical design that the concentrations for each additive be
selected such that the medium concentration was half way between the high
and low concentrations. Care had to be taken in the selection of the high
and low concentrations to ensure that the experimental procedure would be
able to detect a result. If too high a concentration were used, the maximum
speed available in the wind tunnel (30 m/sec) would not erode any of the
test material. Conversely, too low a concentration would show no effect when
compared to an untreated test material. Based on Drehmel and Daniel's results,
it was possible to select dust suppressant concentrations that would yield
results within the range of the experimental capability. The values of the
dust suppressant concentrations selected are given in Table 2. It should be
emphasized that these concentrations are not the concentrations recommended
by the manufacturer. Rather they were selected to accommodate the experimental
procedure used so that weathering effects could be determined over a range of
concentrations.
From a practical standpoint it was decided to limit the outside exposure
time to a maximum of 90 days. It was felt that if a dust suppressant were
still effective after this length of time that 'it would be useful as a fugitive
control technique. Intermediate intervals of 30 days and 60 days were
selected in addition to zero days (no explosure). This variable of exposure
time posed a problem in that the weather occurring during any 30 day period
would not be the same as for any other 30 day period. The test matrix accommo-
dated this by including three replicates of the zero day tests, two replicates
9-4
-------
TABLE 2. ADDITIVE CONCENTRATIONS TESTED
Additive
Concentration, 1/m
Low Medium High
Coherex
Lignosulfonate
SP-301
Polyco 2151
0.022
0.005
0.005
0.0018
0.044
0.011
0.022
0.0034
0.066
0.016
0.038
0.0052
9-5
-------
of the 30 day tests, two replicates of the 60 day test, and one set of tests
for 90 days. All of the weathering tests were conducted between mid-August
and mid-November 1981 at EPA's Research Triangle Park, North Carolina, facility.
These variables were developed into a partial factorial design which
consisted of 96 tests. These tests are summarized in Table 3. One of the test
samples was destroyed during the course of the experiment so that 95 tests
were actually conducted.
The material selected for use in evaluating the dust suppressants was
pulverized coal. A sieve analysis gave an average size of about 200 pm.
The size distribution is shown in Figure 1. The coal was then placed in a
test panel made of plexiglass which had a test volume 305 mm x 305 mm x 19 mm
deep. The test panel was always loosely filled with the coal, even with the
top of the panel, and any excess coal was brushed off. The dust suppressant
was then applied with a spray bottle and allowed to dry overnight. After dry-
ing, the test panels were placed in outside racks and exposed to the weather.
Shortly after initiating the exposure tests a problem became apparent with
the planned exposure test procedure. A severe thunderstorm occurred with the
resulting large rain drops impacting on the treated coal causing the surface
to be destroyed. It was decided that the best solution would be to place
a cover over the test panels. The selection of the cover material became
quite important. Although the test panels would no longer be exposed directly
to rainfall it was necessary to maintain as many of the weather related par-
ameters as possible, particularly solar radiation. Several candidate cover
materials were selected and evaluated on an ultraviolet spectrophotometer.
Ultraviolet was selected since it had been identified as a major factor in
prior weathering studies. Two of the materials screened were judged to
be satisfactory. These were polyethylene (0.0762 mm thickness) and styrene
(0.1905 mm thickness), and the results are shown in Figure 2. Also shown in
Figure 2 is a plot of the relative energy of the sun as a function of wave
length. As can be seen from Figure 2 both polyethylene and styrene transmit
about 70% of the incident solar radiation. Polyethylene was selected as the
cover material because it was readily available. The polyethylene was
fashioned into an open tent and placed over racks each of which held eight
test panels as shown in Figure 3. This permitted free air circulation under
the cover which allowed the test panel to "see" ambient temperature and
relative humidity.
After exposure to the weather the test panels were brought inside for
evaluation in a wind tunnel. The wind tunnel had a 0.61 m square cross
section with an active length of 9.76 m. The first 4.88 m was used for
establishing the desired velocity profile prior to the location of the test
panel. The test panel was located in the wind tunnel such that the coal
surface was flush with the bottom of the wind tunnel. Past the test panel, the
remainder of the active section was used for disengagement. Beside the
active section of the tunnel is an air return loop. In the air return section
is a baghouse to clean recycled air to the active section of the tunnel.
Design of the active section was intended to produce a turbulent boundary
layer similar to that found for wind passing over open ground or a storage
pile. Details of this design are given by Viner et al.'
9-6
-------
TABLE. 3 TEST MATRIX
Additive
Coherex
Lignosulfonate
SP-301
Polyco 2151
Concentration
Levels
3
3
3
3
Exposure
Days
0
30
60
90
Replicates
3
2
2
1
TOTAL
Planned
Tests
36
24
24
J.2
96
Actual
Tests
36
24
23
ii
95b
aA total of 12 tests.
One test sample was destroyed during the experiment.
9-7
-------
1000
Figure 1. Sieve analysis of pulverized coal.
• AIR VS AIR
• AIR VS POLYETHYLENE
AAIR VSSTYRENE
• AVERAGE SUNLIGHT
200
225
250 275
WAVELENGTH, nm
Figure 2. Optical transmittance of candidate cover materials.
9-8
-------
Cover Frame
CO
i
CO
Polyethylene Cover
Test Panels
Figure 3. Exposure racks with test panels and polyethylene cover.
-------
Each test panel was initially weighed prior to being tested in the
wind tunnel. The wind tunnel was operated at several velocities, begin-
ning at the lowest and advancing monotonically. After maintenance of a
velocity for 3 minutes, the tunnel was shut down to remove and reweigh the
test panel. With the test panel put back, the next velocity was tested.
The process was repeated until the maximum test velocity was reached or until
the emission rate exceeded 13.04 g/min. The emission rate of 13.04 g/min
was selected arbitrarily as a very high emission rate which would indicate
that the dust suppressant had failed and the test should be terminated.
RESULTS AND DISCUSSION
In general each test conducted in the wind tunnel was similar in that the
dust suppressant nearly completely prevented erosion until a critical velocity
was reached. At this critical velocity the dust suppressant failed quickly
resulting in a large emission rate. This critical velocity, termed entrainment
velocity, was determined by measuring the velocity which produced an entrain-
ment rate of 4.89 g/min selected arbitrarily as indicating failure of the dust
suppressant. Figure 3 illustrates the rapid failure of the dust suppressant
as the velocity is increased to the entrainment velocity. For the case of
Figure 3 the entrainment velocity for 30 days outside is seen to be approxi-
mately 17 m/sec.
The results for each of the four dust suppressants evaluated are given in
Figures 4 through 7. These plots show entrainment velocity as a function of
weather exposure time for each of three concentrations. In general each plot
shows that the entrainment velocity increased with increased dust suppressant
concentration as expected. What was not expected was that there was no
detectable degradation in the performance of any dust suppressant over the
period tested.
Several factors should be kept In mind when examining these data. First,
as mentioned earlier, the concentrations tested were not intended to duplicate
concentrations suitable for field use. The selection of pulverized coal in a
shallow test panel is likewise a departure from a practical application. The
use of a polyethylene cover compromised the results by limiting the solar
radiation and eliminating direct contact with rain. The exposure procedure
used deviated from ASTM D 1435 in that the exposure racks were oriented flat
instead of at the prescribed 45° facing south. Finally, these were static
tests with no activity on the test panels. Nevertheless, these results are
believed to be valid for the intended purpose of the study; namely, to evalu-
ate the weathering characteristics of the four dust suppressants tested.
CONCLUSIONS
For the four dust suppressants evaluated it can be concluded that there
was no significant degradation in performance over the 90 day test period.
In addition, the experimental test procedures used are believed to be valid
and would have detected any significant degradation had it occurred. These
are welcome findings because they mean that dust suppressant chemical additives
do offer an effective means of suppressing fugitive particulate emissions.
9-10
-------
32.5
26.0
19.5
13.0
6.5
i i I r
DAYS OUTSJOC
0
— 30
mm. fiO
I I I I
J I
12 15 18 21 24 27 30
VELOCITY, m/sec
2
Figure 3. Wind tunnel data: Lignosulfonate (0.005 1/m )
o
0
CO
Z
30
27
24
21
IS
15
12
9
6
3
0
\
LOW CONCENTRATION
-- MEDIUM CONCENTRATION
-— HIGH CONCENTRATION
\ \ \ \ 1
\
\
10 20 30 40 50 60 70 80 90 100
TIME OUTSIDE, days
Figure 4. Weathering tests: Coherex.
y-n
-------
B
§
>
g
w
M
3
12'!_
9
6
3
C
_
, LOW CONCENTRATION
- _ ^^ MEDIUM CONCENTRATION
^. . HIGH CONCENTRATION
1 1 1 1 1 1 1
«•
—
_
1 1
Ed
10
20
80 90 100
30 40 50 60 70
TIME OUTSIDE, days
Figure 5. Weathering tests: Lignosulfonate,
u
-------
u
1
g
H
i
j • • * • ' ' •
24 fc_._ ._^ — — -11"*"*' *V»
J^« ^ •• ^* ^""^"^^••^^»^^. A ^J^
21 f ^"^
1
18
15
12
9
6
3
0
L * ^^"
•
-
_
_ » LOW CONCENTRATION
B MEDIUM CONCENTRATION
- ,-A.^ HIGH CONCENTRATION
1 1 1 1 1 1 1
1 1
^A.
r*-" "
^x
"^4 I
_
—
—
_
—
1 i
10 20 30 40 SO 60 70 80 90
TIME OUTSIDE, day*
Figure 7. Weathering tests: Polyco 2151
100
9-13
-------
REFERENCES
1. Drehmel, D. C., and Daniel, B. E. Relative Effectiveness of Chemical
Additives and Wind Screens for Fugitive Dust Control. In Proceedings:
Third Symposium on the Transfer and Utilization of Particulate Control
Technology: Vol. IV. Atypical Applications. EPA-600/9-82-005d,
July 1982.
2. Larson, A. Evaluation of Road Carpets and Chemical Road Dust
Suppressants. Presented at Fifth Symposium on Fugitive Emissions:
Measurement and Control, Charleston, S. C., May 1982.
3. Mitton, P.B., and Richards, D. P. A Model for Objective Development
of Accelerated Weathering Tests. Journal of Paint Technology, Vol. 43,
No. 563, December 1971.
4. Hoffmann, E. Weathering of Paint Films and Development of Accelerated
Tests. Journal of Paint Technology, Vol. 43, No. 563, December 1971.
5. Masters, L. W., Wolfe, W. C., Rossiter, W. J., Jr., and Shaver, J. R.
State of the Art on Durability Testing of Building Components and
Materials. NBS Report No. NBS1R73-132, March 1973.
6. ASTM D 1435, ASTM Book of Standards, Part. 27, 1971.
7. Viner, A. S., Ranade, M. B., Shaughnessy, E. J., Drehmel, D. C., and
Daniel, B. E. A Wind Tunnel for Dust Entrainment Studies. In
Proceedings: Third Symposium on the Transfer and Utilization of
Particulate Control Technology: Vol. IV. Atypical Applications.
EPA-600/9-82-005d, July 1982.
TABLE FOR CONVERSION TO ENGLISH UNITS
To Convert to Multiply By
m/sec ft/sec 3.33
gal./acre 1069.73
acre 4047
in 0.0394
ft 3.28
gr/min 15.34
9-14
-------
ON THE USE OF SF^-TRACER RELEASES FOR THE DETERMINATION OF FUGITIVE
EMISSIONS
B. VANDERBORGHT, J. KRETZSCHMAR and T. RYMEN
Nuclear Energy Research Center, B-2400 Mol, Belgium
F. CANDREVA, R. DAMS
Institute for Nuclear Sciences, B-9000 Gent, Belgium
Summary
The principle and applicability of tracer releases for the accurate
measurement of fugitive emissions is discussed and illustrated with
some practical examples. At the emission source SFg-tracer gas is
released at a well-known constant rate. With proper adaption of the
tracer emission device to the fugitive emission source thorough
mixing and homogeneous dispersion of tracer and pollutant can be
obtained. In this way the tracer to pollutant concentration ratio is
the same in the dispersed plume as in the emission and the pollutant
emission rate can be calculated by measuring the tracer to pollutant
ratio at any place in the plume.
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
1. INTRODUCTION
Although dispersion calculation for emissions from high stacks pre-
dict a maximum ambient air concentration at a certain distance from the
source, the highest concentrations are very often measured very close to
the factory. Control of point-source emissions has in certain problem
areas not produced the anticipated improvement in ambient air quality. The
reason appears.to be fugitive emissions at low- or ground-level altitude.
Fugitive emissions seem to be especially important for aerosol emissions
from metallurgical plants. For the complete understanding of air pollution
situations - through emission-inventories, mathematical modelling etc. -
it is necessary to be able to quantify the fugitive emissions.
Since the emission area is usually not well defined, and since volume
flow rates are usually unknown, the normal emission measurement techniques
cannot be applied. The measuring methods actually described in the litera-
ture can be classified into "quasistack" sampling or roof monitoring in
which the emission area is by some means physically restricted or into
"reversed modelling" where emission is calculated from immission measure-
ments and mathematical dispersion models.(1)
In this paper the principle and possibilities of a tracer technique for
the quantification of fugitive emissions from a metallurgical plant is
described.
A common feature of all mentioned procedures is the discontinuous charac-
ter preventing routine continuous emission measurements.
2. PRINCIPLE
The principle of the procedure is as follows :
A tracer component is discharged at a constant, well-known rate at the
emission site of the pollutant. When the tracer and the pollutant are well
mixed in the emission plume and when they are dispersed In the same way,
10-1
-------
then the concentration ratio of both components will remain constant all
over the plume and equal to the emission rate ratio :
or
where x * concentration [M.L~3
Q - emission rate [M.T'1
p - pollutant
t - tracer
Knowing the tracer emission rate and measuring the concentration of the
tracer and the pollutant in the plume give the possibility to calculate
the unknown pollutant emission rate Q . It is obvious that the quality
of the results will depend on the validity of equation (1) for which a
good homogenization and equal dispersion of tracer and pollutant is
necessary. The characteristics of the tracer release (place, time, flow-
rate, temperature, shape of nozzle) and the choice of the measuring points
for the simultaneous determination of pollutant and tracer concentrations
must be optimized as a function of this homogeneity.
When the fugitive emissions out of a factory workplace are to be
measured, different sampling strategies are possible. Suppose a building
in which two furnaces are the sources of uncontrolled emissions of metal
fumes (figure 1). Aerosol is evacuated from the inside atmosphere of the
workplace by means of natural draft through rooftop ventilation openings
and by means of a ventilator with the hood a few meters above one furnace
and the exhaust opening in the wall or the roof of the building. These
emissions are easily entrained in the wake on the lee of the building and
the source cannot be considered as a point source.
For the determination of these fugitive emissions - by the method of
tracer release and measuring tracer and pollutant concentration - three
different sampling locations are possible.
1) Tracer and pollutant are sampled outside at distances from a few meters
up to a few hundreds meters downwind the building. With this configura-
tion the total emission of the building is measured.
2) Sampling units are placed along the ventilation opening in the rooftop.
3) Isokinetic sampling is performed in the exhaust tube of the ventilator.
In the two latter cases homogenization of tracer and pollutant becomes
very critical. By proper adjustment of the tracer release and sampling
places the emissions of the different sources are discerned and the
emissions through the rooftop and through the ventilator can be calculated
separately. In this paper the properties of the three methods will be
compared.
3. EXPERIMENTAL
In an antimony (Sb) metallurgical plant the technique has been used
to quantify the fugitive emissions emanating from a 18m high 30mx20m long
workplace in which a converter and a refinery furnace were installed
10-2
-------
(figure 1). One of the experiments, in this workplace, will be described
here as an example. Uncontrolled Sb dust emissions originated mainly from
the filling and the slag removal from the convertor as well as from the
same actions on and the emptying of the refinery furnace. Minor leak
emissions on filter installations did also occur. Sb aerosol had a mean
aerodynamic diameter of 1 to 2 urn.
Sulfurhexafluoride (SFg) gas was used as a tracer. SFfi is an inert,
non-toxic gas, stable up to about 500°C, routinely detectable up to 50
ng.ra"3 and normal background concentrations are below the detection limit.
SF6 was discharged at the convertor (main emission place) at a rate of
about 27 g.min"*.
SFg and Sb were simultaneously sampled at three places in the roof of
the workplace (W^ to W3 in figure 1), at one place (WQ) in the exhaust
tube of the hood above the convertor and 8 places outside at distances
between 15 and 180m from the hall (Rj to Rg in figure 2).
Up to this distance sedimentation and deposition of Sb aerosol is neglec-
tible, and there is no basic difference between the dispersion of Sb-
aerosol and SFfi gas.
The sampling places outside were chosen in'such a way that interferences
from other parts of the factory were minimal. Four sampling periods of
each 30 minutes were performed. During sampling the Local wind direction
at the 30m-level was between 200° en 205° with a windspeed between 7 and
8,5 m.s"1 and neutral atmospheric stability. For the sampling points in
the roof (W1 to W3) the aerosol was sampled on the same filter for the 4
periods as the access to these places was difficult.
Aerosol samples were taken using LIB-type low volume samplers with
Whatman 41 filters. Sb was determined by neutron activation and X-ray
fluorescence. SF6 samples were collected by filling plastic bags and were
gas chroraatographically analysed. For further details on the sampling and
analysis procedures see reference (2).
4. Results and discussion
The results of the emission calculation are summarised in table I.
TABLE I : Sb emission from a workplace as determined by the concentration
ratio Sb/SF6 in : ambient air downwind of the building (R1 -> R7)
ventilator exhaust tubo (WQ)
rooftop of the building (Wj - W3)
sampling
period
1
2
3
4
Average
Sb emission in kg/h
based on sampling points
Rj -» R7 r.s.
1.6 ± 0.4 23 Z
1.1 ± 0.4 33 Z
2.6 ± 0.5 19 Z
1.0 ± 0.2 22 Z
1.6 24 Z
W0
2.5
1.3
2.7
0.7
1.8
Activities of
Convertor
8' filling
6' filling 4' slag
12' filling
9' filling
Refinery furnace
17' slag
short emissions
3' slag
r.s. relative standard deviation in procent on 7 measurements
10-3
-------
30 m
Ventilator
t
V/9////)
Con
tor
ver-
%.
/HO
'#,.
^Sb
% SF6 ^
00
'//
y/~ "
-f---j---T "
W, W2 W3
Refine-
ry
furnace
> _oE51Jn3
o
3
Sb203 aerosol
1-2
Fig. 1 - Layout of emission and sampling points in the workplace
Figure 2 - Layout of emission and sampling points
10-4
-------
sampling place
W0
Wl
w2
W3
average
Sb emission in kg/h
averaged over 4 periods
1.8
1.4
1.4
0.9
1.4 ± 0.37
The relative standard deviation on the emission determination for
each sampling period through the 7 sampling points (R1 + R7) outside the
building and on a distance larger than 30 m ranges between 19 and 33 %
with an average of 24 %, and no sampling place gave a systematic devi-
ation from the average. In sampling point Rg, 15 m downwind of the
roofcenter, on the other hand the Sb/SFg ratio was systematically higher
than in the other points. At this short distance, the dispersed plumes
from the two Sb sources (converter and furnace) were not yet well hom-
ogenized, resulting in a higher Sb/SFg ratio in the plume of the refinery
furnace than in the plume of the convertor, where the SFg emission was
performed. At larger distance the plumes of the two Installations seem to
be well mixed since the Sb/SFg remains from place to place constant
between acceptable limits. The considerable aerodynamic turbulence in the
wake at the leeside of the building, is responsible for this good re-
sult.
The calculated Sb emission for the consecutive sampling periods is
compatible with the emission causing activities on the plant apparatus,
except that the emission during period 4 is lower than expected in com-
parison with the three former periods.
The sampling points W1 to W3 and the hood of the ventilator, and
consequently also point WQ , are placed above the convertor. Visual obser-
vation of the emission plumes of the convertor and of the refinery fur-
nace indicates that emission results obtained by these points can in a
first approximation be equalized to the emission from the convertor and
that the contribution of the refinery furnace to the measurements in
those points is neglectlble.
With the measurements Inside the workplace - respectively in the
rooftop (W1 to W3) and in the exhaust tube (WQ) - two different methods
to calculate the emissions can be used. With the first method the global
pollutant emission is obtained, with the second it is possible to deter-
mine what fraction goes through the hood, respectively the roof.
a. By using the total 'SFg emission in equation (1) and the SFg and Sb
concentration respectively in the points WQ, W^, W2 and W3, the total
Sb-emission from the convertor is obtained. Averaged over the four
sampling periods, the emission results obtained by the points W« to W3
are respectively 1,8 - 1,4 - 1,4 and 0,9 kg Sb/h (table I) and aver-
aged over the four points : 1,38 ± 0,37 kg/h or 1,38 kg/h ± 27 %. If
SF6 and Sb were well homogenized in the plume above the convertor the
four results should be the same.
Considering the relatively long sampling time (4x30'), the 27 % rela-
tive standard deviation on the emission calculation indicates that the
homogenization of Sb and SFg right above the convertor is not com-
pletely attained.
10-5
-------
b. The measurement of the flow rate in the exhaust tube enables to cal-
culate the mass flow of SFg and Sb through the hood and the ventilator
(Qfl)- , T ,
Subtraction from the total SF6 emission (Qgp J gives the SFfi emission
through the roof (Qgp ) by draft ventilation. Substitution of this
value in equation (1) gives the Sb emission through the roof (Qst,)
instead of the total Sb emission, this of course with the condition of
sufficient homogenlzation of tracer and pollutant. The results of this
exercise are summarized in table II.
TABLE II : Sb and SF6 emission through the ventilator' and the roof of the
workplace
sampling
period
1
2
3
4
average
emission in kg/h
Q?F6 (i)
0.24
0.39
0.41
0.39
0.36
QSF6 (2>
1.48
1.18
1.20
1.23
1.27
Qsb CD
0.34
0.32
0.68
0.17
0.38
sampling
place
Wl
w2
W3
average
emission in
qR f'
kg/h
»)
1.08
1.07
0.70
0.95
(1) through the volume flow rate measurment in the exhaust tube
<3) Q?>, - Q* .
The sum of the Sb-emission through the roof and through the ventilator
should be the same as the total Sb emission from the convertor determined
by the previous procedure (sub a). Averaged over the four sampling
periods this gives :
h Qsb " °'95 "*" °'38 * **3 k8 Sb/h
which is only a 'few procents less than the 1.4 kg/Sb found by the other
procedure (table I).
It can easely been shown mathematically that both results should be
Identical with complete mixing of tracer and pollutant, I.e. when
- rxSb
The total Sb emission averaged over the four sampling periods, is
slightly higher when determined through places outside the building (Qsh
•1.6 kg/h) than through the sampling points in the roof and the venti-
lator (Qsb - 1.4 kg/h). This was to be expected since the places
outside are influenced by the emissions of the complete building while
the points inside were focussed on the emissions of the convertor only.
Moreover measurements outside can be positively interfered by the back-
ground caused by other parts of the factory. Measurements of the Sb con-
centration upwind the building showed that this was a minor problem.
10-6
-------
5. Conclusions and epiloge
Fugitive emissions can be quantified by the method of tracer release
and measuring the tracer to pollutant ratio in the dispersed plume. The
critical factor of the method is the homogenization of the tracer with
the pollutant. This can be optimized by proper adjustment of the tracer
emission device, the sampling time and the selection of the sampling
places•
With concentration measurements a few tens or hundreds meter away from
the source, in the example of the paper this is outside the building, the
tracer to pollutant concentration ratio remains constant from place to
place within acceptable limits. The total pollutant emission from the
building can be determined.
With concentration measurements close to the source, for example in the
rooftop and the ventilation of a workplace, sufficient homogeneity is
•ore difficult to obtain. Nevertheless can the emission be determined
with reasonable precision, even in presence of another source close to
the main source. The efficiency of an exhaust hood can be evaluated.
Within the sampling time, the emission of the pollutant can fluctuate. If
during this time the tracer emission and the dispersion conditions remain
constant, the time averaged pollutant emission can nevertheless be deter-
mined*
The procedure has been used to determine emission factors for
fugitive emissions. These data have subsequently been used for the
mathematical modelling of the Sb concentration in the environment around
the plant (2,3,4). The procedure has also been used to determine the
fugitive emissions of a metal refinery installation before and after
construction of a system for reduction of the fugitive emission. With
this information the reduction of the pollution level in the environment,
due to this investment, has been estimated using a bi-gaussian disperion
model.
This research was carried out in the "National R & D Programma,
Leefmilieu-Lucht" of the DPWB (Ministery of Science Policy).
References
(1) "Wanted : fugitive emissions"
S. Budiansky. Environmental Science and Technology V14, N8 (1980)
904-905.
(2) "Luchtverontreinlglng door een metallurgisch bedrijf. Eindrapport
Natlonaal R-D Programma Leefmilieu-Lucht
Diensten voor de Programmatie van het Wetenschapsbeleid, Brussel,
1981.
(3) "Influence of the meteorological input data on the comparison
between calculated and measured aerosol ground level concentrations
and depositions"
I. Mertens, J. Kretzschmar and B. Vanderborght
Proceedings of the "12th ITM on air pollution modelling and its
application"
NATO/CCMS Palo Alto August 1981
(4) "Depositie rondom een non ferro bedrijf"
B. Vanderborght, I. Mertens, J. Kretzschmar
Extern N4 (1981) in press
10-7
-------
AN ATMOSPHERIC TRACER INVESTIGATION OF FUGITIVE EMISSIONS
TRANSPORT IN THE COLORADO OIL SHALE REGION
George A. Sehmel
Pacific Northwest Laboratory
Richland, Washington 99352
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
This paper is based on work performed under the U.S. Department of Energy
Contract DE-AC06-76RLO 1830.. This work was done at the Pacific Northwest
Laboratory (PNL), operated for the Department of Energy by Battelle
Memorial Institute.
ABSTRACT
It is anticipated that spent -shale-disposal-areas could become prime
fugitive emissions sources. Atmospheric tracer experiments were conducted
to investigate transport and dispersion of fugitive emissions in the complex
terrain near a proposed spent shale disposal area. The atmospheric tracer
gas SFs was released from the Federal oil shale lease Tract C-a in Colorado
in July 1981. The SFs tracer gas release rate was 26 kg/hr at a tracer
release height of 1.7 m.
Airborne SF$ tracer gas concentrations were sampled at 32 radio-
controlled bag-sampling stations (at 15-min sampling times) and along roads
with syringe grab-samplers. Sampling sites were 0.6 to 13 km from the
tracer-gas release site.
There were four experimental time periods for releasing and sampling
tracer gas: two evening and two morning time periods. Results include
information on the following: a time-history of pollutant concentration
changes upwind and in gulches before, during, and after sunrise; identifi-
cation of major transport flow paths in the evening and morning; and the
behavior of pollutant mixing and channeling in valleys and downwind of valley
confluences.
11-1
-------
INTRODUCTION
Spatial variations in the airborne transport of pollutant plumes from
fugitive emissions from the developing shale oil industry are complex for
multiple local flows in the oil shale regions of Piceance Creek Basin of
northwestern Colorado. Significant local spatial variations in pollutant
plume transport are expected because of the complex terrain. For instance,
overall drainage flows affecting downwind airborne pollutant plume concen-
trations are a combination of multiplicity of drainage flows from gully trib-
utary systems. Each tributary has a drainage plume which may rapidly lose
its plume identity when mixing with other drainage flow plumes. The mixing
rates of these plumes cannot be adequately predicted. When the tributary is
large, i.e., a gulch, cross-gulch mixing of the two converging plume flows
may be slow. For instance, as was shown by Sehmel (1981) for August 1980,
cross-gulch mixing was usually incomplete at the confluence of the tracer
plume from Corral Gulch and the ambient air plume from Box Elder Gulch. A
mixing distance was defined, for nocturnal drainage flow, as the distance
downwind of the confluence for which the two plumes were uniformly mixed
across the gulch. The mixing distance is also compared to gulch widths
upwind of the confluence of the two gulches. For the site and times inves-
tigated, the mixing distance was from 7 to 20 times the gulch widths, 0.3
and 0.1 km widths upstream from the confluence of the two gulches.
In order to investigate downwind pollutant plume transport from poten-
tial fugitive emissions, atmospheric transport investigations were conducted
with tracers (Clements et al. 1981, Sehmel 1981, and Whiteman et al. 1981)
10 14 10 « M
—— *—f— «
VIlOHITCMS I
FIGURE 1. Map Showing Tract C-a Location in Piceance Basin
11-2
-------
in an area in an adjacent to Federal lease" Tract C-a in Colorado." The site
is shown on the map in Figure 1. The site is in the Piceance Creek basin,
-32 km southeast of Rangely, Colorado. Tract C-a is along Corral Gulch which
drains into Yellow Creek, a tributary of Whvte River. Previously, both non-
depositing SF5 tracer gas and depositing tracer particles were released
from the bottom of the Corral Gulch in August 1980 experiments (Sehmel 1981).
Air pollution control regulations are based partially upon 3-hr maximum
"concentr'atfons, 24-hr concentrations, and annual average concentrations. In
contrast, gas-bag sampling time periods were usually 15 min in this investi-
gation. This short time period, compared to longer time periods for compli-
ance with some air pollution regulations, was selected to investigate the
physics of air pollution transport. The rationale for using the 15-min time
period was that estimates of air pollution concentrations can be made for
longer time periods by integrating tracer concentration data as a function
of time, possibly a time period representative of a 3-hr air pollution regu-
lation. In contrast, transport and dispersion information is lost if time-
averaged tracer-gas concentrations are experimentally determined only for
the longer time periods specified in regulations for maximum air pollution
concentrations.
Even shorter sampling durations are important for investigating physics
required to develop either site-specific or more generic airborne pollutant
transport and air pollution budget models in complex terrain. The sampling
time requirement will be dependent upon the multiplicity and complexity of
pollutant plume flows and the required evaluated accuracy of predictive mod-
els for different distance scales. In this investigation, more spatial and
temporal details in identifying major and secondary plume transport flows and
interactions were obtained from syringe samples collected nearly instantane-
ously. An advantage of using syringe samplers was a dense sampling grid at
low cost. A disadvantage was that syringe samples were collected in a time
sequence, rather than simultaneously throughout the sampling grid. Neverthe-
less, concentrations investigated with syringe samplers were essential for
interpretation of concentrations investigated with the widely spaced time-
integrated gas-bag sampling sites and for identifying multiple plume flows
and interactions, both drainage and upslope flows.
The objective of the July 1981 experiments reported here were to simu-
late pollutant transport and dispersion from a proposed spent-shale disposal
site during both day and night time conditions. The tracer simulant was SF5
gas, a non-depositing and non-reactive gas. The July 1981 SF6 tracer-gas
experiments were conducted with the tracer release site at the proposed spent
shale disposal site on terrain just north of Corral Gulch. The terrain was
a sloping, undulating area between the hilltops along Corral Gulch and Dead
Horse Ridge.
EXPERIMENTAL PROCEDURES
Airborne tracer concentrations were investigated using one tracer
release site and multiple airborne concentration sampling sites. Details
follow which describe tracer release, site locations, gas-bag sampling,
syringe sampling, tracer analysis and meteorological data.
11-3
-------
Tracer release times raifged fro~m~3 to 9 hours. There were four experi-
mental time periods: two evening and two morning periods. During the
evening time periods, tracer release rates were constant from 1500 to
2100 MDT (July 15) and 1800 to 2100 MDT (July 17). During the morning time
periods, tracer release rates were constant from 0330 to 0930 (July 19) and
0300 to 1217 MDT (July 21). -
Tracer release times included airflow transition times around sunset and
sunrise. The evening experiments were conducted first. The results for eve-
ning experiments were less successful than those for morning releases. On
the first night, the radio control system malfunctioned for automatically
collecting gas-bag samples as a function of selected time periods at remote
sites. A power lead in the battery supply shorted. On the second night the
system was operational, but the experiment was stopped early due to potential
safety hazards from lightning in a severe local storm.
Tracer Release
During each experiment, the SF5 tracer gas was released at a constant
rate of 26 kg/hr at a height of 1.7 m. The tracer gas release rate was con-
trolled with rotameters. The reported release rate was calculated from pre-
and post-weighting tracer release cylinders. The S?s was released through
a 1.0 cm inside-diameter tube.
The SF5 was released from the gas phase in storage cylinders. Buoyance
effects on the airborne tracer plume were minimized because tracer-gas cool-
ing is minimized during this release procedure, as compared to the greater
cooling during tracer release and SFs plume subsidence that occurs if SFs
is released from the liquid phase in storage tanks.
Site Locations
Tracer release and downwind sampling locations are shown on the map in
Figure 2. Important locations are the tracer release site, the tracer-gas .
sampling sites and the meteorological measurement sites.
Airborne tracer-gas concentrations were investigated by examining gas
samples collected in either gas bags or syringes. Gas-bag sampling stations
were placed in four radio-controlled signal zones for controlling gas sampl-
ing pumps. The zones were chosen to selectively sample major plume flow
directions. Sampling pumps in each zone were controlled with tone-specific
radio-receivers placed in each zone.
Locations of gas-bag sampling stations 1 through 32 are shown by num-
bered locations in Figure 2. These thirty-two radio-controlled sampling sta-
tions encompassed the release site. Distances from the tracer release site
to these sampling stations range from 0.6 to 6 km. Sampling stations were
located at sites with road access.
The principal gas-bag sampling zones were zones 2 and 3, which encompass
regions of drainage flow from the tracer release site. Air pollution concen-
trations for upslope flows were also of interest. Consequently, gas-bag
sampling stations in all four zones were simultaneously activated (or deacti-
vated) for selected gas-bag sampling time periods.
11-4
-------
FIGURE 2. Topographic Map Showing Locations of the Tracer Release Site,
Sampling Sites 1 thru 32 in Four Radio-Control led Gas Bag
Sampling Zones and Roads Driven During Collection of
Grab Samples
Gas-Bag Sampling
Five electrical-control outlets at each gas-bag sampling station were
used for activating (or deactivating) air pumps for filling gas sample bags.
These 5 L, double-walled, sample bags* were prepared by encasing each sample
bag with an outer polyethylene bag as a diffusion barrier to prevent contami-
nation from SFs in both ambient and adjacent sample bags. An air pump, or
pumps, was attached to each electrical-control outlet, which was activated
(or deactivated) by a specific radio-signal tone. The pump sampling times
and non-operating times were controlled by operation of a tone-signal radio-
repeater, which was in a line-of-sight of antenna for each sampling station.
Each gas-bag sampling station had the following components. An antenna
was attached to a bamboo pole, 5.2 m in height, which was taped to a steel
fence post. The antenna was attached to a battery-powered radio-signal
decoder. The decoder was a radio receiver that responded to one of the four
transmitted tone signals. Each tone signal activated (or deactivated) sam-
pling pumps in only one sampling zone; the zones shown in Figure 2. Electri-
cal wires extended upward from the five electrical-control outlets on the
decoder. For most gas-bag sampling stations, each electrical-control outlet
was attached to only one sampling pump. In this case, sampling tube inlets
for all five pumps were located at a sampling height of 1.5 m. In other
cases, ten gas-sampling pumps were located at each gas-bag sampling station.
* Industry bag, 10 by 15 in., 2.5 mil, flow meter fully inserted, no tubes.
B Bar B, Suite 34, 121 West Whittier Blvd. Lahabra, CA 90631.
11-5
-------
Two sampling pumps were simultaneously operated at these stations by con-
necting each pump-signal outlet to two air sampling pumps. Sampling tube
inlets were located at heights of 0.3 and 5 m, respectively. Sample outlet
tubes from the sampling pumps extended into a double garbage bag within which
were five gas sampling bags. One dark garbage bag was inserted into another
dark garbage bag to reduce light transmission into_gas sampling bags as well
as to physically protect the bags.
Syringe Sampling
Relative locations of syringe-sampling sites are shown by dashed lines
in Figure 2. The 50 cm3 syringe* samples were collected at distances of 3
to 13 km from the tracer release site. Most syringe samples were collected
along the sides of a triangle formed by roads along Duck Creek, Yellow Creek,
and the road adjacent to gas-bag sampling stations 17, 18 and 20, hereafter
called road 18. Syringe samples were rapidly collected by driving along
roads, stopping, flushing a new syringe with ambient air by reaching out
through the driver's window, collecting a syringe sample, stoppering the
syringe inlet, marking the time and odometer reading on the syringe, and
immediatey driving to the next sampling location.
Tracer Analysis
Collected gas-bag and syringe samples were analyzed in a field labora-
tory for SFg tracer gas concentrations. The laboratory was in the area
labeled MOP in Figure 2. This laboratory was always outside the plume of
high SFg concentration. Neverthless to further minimize the possibility of
sample contamination, all samples were analyzed for SF5 concentrations
before tracer was released for the subsequent experiment.
The SFg concentrations were determined with an electron-capture gas
chromatograph** (GC). The GC was field-calibrated with SFg calibration gas
contained in cylinders at concentrations from 20 to 2000 pptv (parts per
trillion by volume/volume) . However, concentrations in some samples were
greater than the 2000 pptv upper limit for the calibration gas. In these
cases, a gas sample was diluted by using a 50 cm3 syringe. Dilutions were
made by partially expelling high-concentration gas from a syringe and sub-
sequently drawing ambient air into the syringe.
Meteorological Data
Meteorological data were recovered at the tracer release site and at
sites labeled MET. 1, MET. 2, and MET. 3 in Figure 2. Data are for 15-min.
averages. Meteorological data included wind speed, wind direction and ver-
tical temperature gradients (T60 m - T10 m). Winds coming from the north
are described as aoO wind direction while winds coming from the west are
described as a 270 wind direction. Meteorological data collected at the
MET. 1 site included wind speed and direction for both 10- and 60-m heights
* Single-use syringe, Plastipak, BD-5663, Becton-Dickinson, Rutherford,
NJ 07070.
** Field-Portable Tracer Gas Monitor, Model 215UP, Systems, Science and
Software, JaJolla, CA 92038.
11-6
-------
However at other sites, only wind speed and direction were collected at one
height. At the MET. 2 and MET. 3 sites, wind speed and direction were
measured at 10 m. At the tracer release site, wind direction and speed were
measured at 2.1 m.
-RESULTS "
Meteorological data and airborne SF§ concentrations are summarized:
one figure is used for each experiment. General features of these figures
are described here. Subsequently, data for each figure (experiment) are
discussed.
The order of presenting results for each experiment is the meteorologi-
cal data, the SFs concentrations determined from radio-controlled gas-bag
sampling stations, and SFs concentrations determined from syringe grab
samples. Concentrations for SFs tracer are reported as pptv, i.e., 1 pptv
equals 10-12 parts by volume. In the figures, airborne SFs concentrations
are superimposed on topographic maps. Each map is for successive time per-
iods for which radio-controlled gas-bag samples were collected or for which
a series of syringe grab-samples were collected or for which a series of
syringe grab-samples were collected. For both sample collection types, the
time period is in the figure title. In addition, to indicate the collection
sequence for syringe samples, selected sampling times are also printed
adjacent to selected concentrations.
Sulfur hexafluoride concentrations are shown in different graphical rep-
resentations. For gas-bag samp.les, concentrations are printed adjacent to
each sampling station location (station locations are indicated by solid cir-
cles). In some cases upper and lower concentrations are printed adjacent to
a solid circle. In these cases, the concentration represented by the upper
printing is for the 15-m height while the lower printing is for the 0.3-m
height. In most cases, however, only one concentration is printed adjacent
to a solid circle. These concentrations are for a 1.5-m sampling height,
i.e., near-respiration height.
Some SFs concentrations for gas-bag samples have special meanings. In
these cases, either a zero or blank is printed adjacent to the solid circle.
A zero indicates that a sample was analyzed, but the concentration was below
a detection limit of 1 pptv. If no concentration is printed, this means no
samples were collected because either an air pump malfunctioned or the
sampling station was not operated.
Concentrations from syringe grab-samples are shown graphically. The
graphical representations are oriented with the abscissa approximately par-
allel to roads used for collecting the respective syringe samples. In most
cases, the roads (along Duck Creek, Yellow Creek,, and road 18 adjacent to
gas-bag sampling station 18) approximate the sides of a triangle formed by
roads.
Concentrations shown are plotted at road-sampling locations projected
onto the abscissa. Road locations include cross-creek valley sampling
traverses for both main roads and two-wheel side trails. Main roads cross
11-7
-------
valleys at projected locations of about 2.6 and 10 km along Duck Creek and
2 5 and 6.5 km along Yellow Creek. Two-wheel side roads cross valleys at
projected locations of about 2.5, 3.7, 5.5 and 7.2 km along Duck Creek and
at 7.4 km along Yellow Creek. .... ...
Results, July 15
Meteorological data are shown as a function of time in Figure 3 for the
first evening tracer experiment. Since the frequency of temperature inver-
sions is relatively high, the occurrence of temperature inversions during
each tracer experiment is noted; i.e., T60 m - Tin m is Positive at
MET 1. A temperature inversion began about 1900 MDT. Other important
times noted are tracer release and tracer gas sampling periods. Tracer
release was from 1500 to 2100 MDT.
At present, these data are included for completeness since airborne
tracer concentration data reduction is incomplete. Analysis of the data col-
lection record will need to be completed before the airborne tracer concen-
trations can be reported as a function of sampling time. Record analysis is
required since this is the experiment in which an electrical short caused
transmission of spurious control signals to the radio-controlled gas-bag
sampling stations.
i
6O
40
20
360
340
320
300
280
260
240
220
200
14
12
10
8
6
4
2
0
+2
+1
0
-1
•2
MET-1.Ttom-TiOm
16 16 17 18 19 20 21 22 23 24
TIME. MOT
FIGURE 3. Wind Direction, Speed and Temperature Gradients
on July 15, 1981
11-8
-------
Results, July 17
Meteorological data are shown as a function of time in Figure 4a for
the second evening tracer experiment. Temperature inversions below 60 m
were intermittent and were only established at about 0945 and 2045 MDT.
Other important times are tracer release and the three tracer gas sampling
periods. Tracer release was from 1800 to 2100 MDT. Tracer release was
stopped due to a local storm. There were three tracer sampling time periods:
two gas-bag sampling periods from 1945 to 2000 and 2043 to 2053 MDT and one
syringe sampling time period from 1946 to 2235 MDT.
It is significant to note the first gas-bag sampling period was con-
ducted between 1945 and 2000 MDT. As shown in Figure 4a, this was a time
period of rapid wind direction change with a variable wind direction at the
tracer release site. Surface and upper wind directions were decoupled.
Upslope winds at the tracer release site were decoupled from upper wind
directions, even at 10 m. The 10- and 60-m wind directions at MET. 1,
MET. 2, and MET. 3 were significantly different (about 330 to 0° at
2000 MDT) from the 2.1-m surface wind direction (30 to 165° at 2000 MDT) at
the tracer release site.
Decoupled air flow directions are reflected in concentrations shown in
Figures 4b and 4c for both gas-bag sampling periods. As shown in Figure 4b,
tracer plume transport was to the northwest. Significant airborne tracer
in
Ul
HI
K
0
HI
a
2
g
u
Ul
K
5
Q
z
I
1
E
Q
HI
CL
CO
Q
Z
i
y_
(-"
80
360
280
200
120
40
0
-
_
, .. -
ff
- !
1 •
"" A r i ~
».X/ //'
- 3£] /t -
'1.1,1.1'
—
-
_
I
$
i/
K- \ ~
K * .
V
1- —
ll'l
S I-
w» ^* «'-
^> M K '
' \y p ~
V.
~ ' VARIABLE 2
/>_
r/v-J ,]
18 19
16
14
12
10
8
6
4
2
™ ^
"" »
- V //fVvy -
- ^h3r v •
_ ("\ _
- » » ~— ! -
r T T .sv', i
20 21
TIME. MDT
MET-1
— —60m
10m
+1
-1
- MET 1, Teom- Tt0m -
" I . I . i . I ~
SITES
MET-2 MET-3
——10m 10m
TRACER
RELEASE
1.5m
18 19 20 21
TIME, MDT
FIGURE 4a. Wind Direction, Speed and Temperature Gradients
on July 17, 1981
11-9
-------
l -v 1 -• .
1.6 km (1 MILE) GRID SPACING
(20 ft CONTOUR SPACING)
RADIO-CONTROLLED BAG
CONCENTRATIONS SHOWN
NEXT TO EACH SITE ARE
pptv
FIGURE 4b. Airborne Tracer Concentrations from 1945 to 2000 MDT,
Collected at the Radio-Controlled Sampling Sites
During the Evening of July 17, 1981
J--Z PROPOSED. 4 • - - .•.
£f SURFACE- SPENT ,^S ' " \
MILE) GRID SPACING
'2° *CONTOUR SPACING)
SF, SAMPLING
4 RADIO-CONTROLLED BAG
SAMPLING SITES I ,
CONCENTRATIONS SHOWN / i
NEXT TO EACH SITE ARE
pptv
FIGURE 4c. Airborne Tracer Concentrations from 2043 to 2053 MDT,
Collected at the Radio-Control led Sampling Sites
During the Evening of July 17, 1981
11-10
-------
concentrations were measured at gas-bag sampling sites 1, 11, and 10 (refer
back to Figure 2 for site identification). The maximum concentration was
2400 pptv at site 1.
Indices of both horizontal and vertical diffusion were measured during
this experiment. Since tracer concentrations at site 11 were 58 pptv at
0.3 m and 42 pptv at 5 m, concentrations decreased over one order of magni-
tude in the 1.3 km horizontal distance between site 1 (2400 pptv) and site 11
(58/2400 = 0.02). The tracer plume tended to remain adjacent to the ground
during upslope flow. The concentration decrease with height at site 11 was
0.7; i.e., the ratio of 42 pptv at 5 m compared to 58 pptv at 0.3 m.
The direction of the maximum tracer concentration shifted and concen-
trations decreased in the subsequent sampling time period, from 2043 to
2053 MDT. The tracer plume flowed counter-clockwise from the tracer release
site. Tracer concentrations were essentially constant at 2 pptv west of the
MDP area.
For this time period, upslope plume flow from the tracer release site
spilled into Corral Gulch. Vertical concentration profiles at sites 25 and
27 reflect plume spillage into Corral Gulch. A maximum concentration of
680 pptv was measured at site 25. In contrast to the previously discussed
vertical concentration profile decrease for upslope flow at site 11, the
relative concentration increase with height at site 25 was a factor of 1.3;
i.e., the ratio of 680 pptv at 5 m compared to 510 pptv at 0.3 m. At this
site, the tracer plume was overriding drainage flow in Corral Gulch. How-
ever at down-gulch site 27, vertical mixing was complete for the two heights
investigated.
Plume flow inferences are made since the tracer-plume path was almost
certainly between sampling sites 26 and 31. Concentrations were 0 pptv at
site 26 and 1 pptv at site 31. Although the tracer plume rotated counter-
clockwise, the main plume width was either relatively narrow (the 1.5 km
between sites 26 and 31), or the main plume was elevated at sampling sites
with 0 or 1 pptv concentrations.
Concentrations from the more spacially frequent syringe samples shown
in Figure 4d confirm, in more detail, plume-location and plume-width results
from gas-bag samples shown in/igure 4b. Syringe samples were rollected
along Ridge Road, which is adjacent to sampling sites 1 and 11. The main
plume width was narrow; concentrations decreased over two orders of magni-
tude from the maximum along a plume width of about 0.7 km.
Concentrations from syringe samples reflect tracer plume segments
extending across Ridge Road, especially between 4 and 7 km. These rela-
tively high concentration plume segments were of short time duration since
corresponding 15 min gas-bag sample concentrations were less, e.g., 1 and
2 pptv at site 10 in Figure 4b.
Tracer plume flow extended into the bottom of Big Duck Creek Valley, a
valley immediately north of Ridge Road. Concentrations are shown in Fig-
ure 4d; data for a time period between 2145 and 2235 MDT. Samples were col-
lected during heavy rain after tracer release was stopped at 2100 MDT.
11-11
-------
FIGURE 4/1. Airborne SFs Concentrations Along Roads for Sampling
"~ Times from 1946 to 2235 MOT on July 17, 1981
Results. July 19
Meteorological data are shown as a function of time in Figure 5a for the
first morning tracer experiment. Establishment of a temperature inversion is
unknown since the 60-m height wind and temperature instruments at MET. 1
were not operated. Important times are tracer release and the tracer
gas-sampling periods. Tracer release was from 0330 to 0930 MDT. There were
seven tracer sampling time periods: five gas-bag sampling periods were 0533
to 0548, 0700 to 0715, 0745 to 0800, 0930 to 0845 and 0937 to 0952 MDT; and
two syringe sampling time periods were 0537 to 0652 and 0704 to 0946 MDT.
Concentration data reflect little tracer-plume spread during drainage
flows before sunrise, plume spread and dilution after sunrise, and subsequent
plume lifting. A coherent plume is shown in Figure 5b with relatively high
centerline concentrations, 2200 to 2300 pptv, along sampling sites 18, 19 and
20. As shown in Figure 5c between 0700 to 0715 MDT (an hour after sunrise at
0615 MDT, sunrise was delayed by a cloud), plume spread is reflected in con-
centrations of 14 and 19 pptv at sampling site 10. Plume spread may reflect
increased wind direction variability at the tracer release site. Even with
increased plume spread, the concentration remained 2200 pptv at site 18.
This high concentration was caused by interactions of a main drainage flow
plume with drainage flow containing tracer. (More detailed evidence for
plume interaction was obtained in the next experiment on July 21). Further
plume spread and also plume lifting are shown 45 min later in Figure 5d.
Concentrations at site 18 decreased an order of magnitude to 171 pptv.
Concentrations continued to reflect increased surface heating effects
with increasing time. As shown in Figure 5e for 0830 to 0845 MDT, continued
11-12
-------
6 7 8 9 10
TIME. MOT
FIGURE 5a. Wind Direction and Speed on Ouly 19, 1981
a PROPOSED: .57' -- •*.
SURFACE • SPENT . «Z300 " \
RETORT . SHALE ". ,«2
'*""' "
\ .M»W." ». Wi.S -^
.6 km (1 MILE) GRID SPACING U
tn *• nnhiTnuR GDAPIURI i A
(20 ft CONTOUR SPACING)
SF. SAMPLING
• RADIO-CONTROLLED BAG /
cASCtf
FIGURE 5b. Airborne Tracer Concentrations from 0533 to 0548 MDT,
Collected at the Radio-Control led Sampling Sites
During the Morning of July 19, 1981
11-13
-------
^i'ki'.w?^
9 ,:
S"5.FA«- •«"! 2so " *
1.6 km (1 MILE) GRID SPACING
120 ft CONTOUR SPACING)
SF. SAMPLING
RADIO-CONTROLLED BAG
SAMPLING SITES
CONCENTRATIONS SHOWN
NEXT TO EACH SITE ARE
pptv
i -J-* '
FIGURE 5c.
Airborne Tracer Concentrations from 0700 to 0715 MDT,
Collected at the Radio-Controlled Sampling Sites
During the Morning of July 19, 1981
'. "- " .4j^i^^}jU^=«gSi$g,
.k^^rP^ff-': >'
: ' T" ^^^—
-'L v.j^s5L*r* ;..' ,<-,
^^_J,9,,.;
"'y^x'^ff/V'i^' -' \\ •••''
-0
©»!;••••;' .,
^t:LV-;r^
^»aK*
; -LEASE TRACT C-.i
l! '»."•'>-^!.. •••.•!•>
\ .•-_•!• .»»>1«>'.-' v ^-.rs^
1.6 km (1 MILE) GRID SPACING Vj
(20 ft CONTOUR SPACING) ^
~-j~ Y~'~* i SF»SAMPL|NG *-
>• ' I ' f • RADIO-CONTROLLED BAG •/
"&' ' '*•' V -"- SAMPLING SITES ;;
•'"''l1 ;
-------
<3*^
j-|-3S**-,)>-Tr,
' {- ^K^"~'
• .^f^^
'.<*?*&*.. ff/X-' - ::\ -,
-•]*?• ."X^Uu..:_.[-i™e..
W^- . I-1 -"yJ-^SR.'*. 3oi"°' -A
SURFACE- SPENT ,,in' " V
' 910 DUCK CHEEK
^^••'&M.['t''^ -x
_> •./ > JF'i* -.- A .' : • -»•*. i
<^:^fffe
-r^?Wl^!f--
, T-7'-Y*$Hf'-VV;"f
v. • • •-.IjfcT*^*«™"v. .-.;•/. •
)-K - 5-^>i'A--^/'.••!.- ••-'
1.6km(1 MILE) GRID SPACING
(20 ft CONTOUR SPACING)
SF. SAMPLING
RADIO-CONTROLLED BAG
SAMPLING SITES
CONCENTRATIONS SHOWN
NEXT TO EACH SITE ARE
pptv
l'\
FIGURE 5e. Airborne Tracer Concentrations from 0830 to 0845 MDT,
Collected at the Radio-Control!ed Sampling Sites
During the Morning of July 19, 1981
plume spread included site 1. However by 0937 to 0952 MDT, plume lifting
resulted in low-surface concentrations shown in Figure 5f.
Concentrations determined from syringe samples confirm that tracer-plume
spread was limited during drainage flows before sunrise. As shown in Fig-
ure 5g near site 18, syringe-sample concentrations were similar to 15-min
gas-bag concentrations shown in Figure 5b. The main-plume width was narrow;
concentrations decreased over two orders of magnitude from the maximum along
a plume width of about 1.2 km.
Except for concentrations along Ridge Road, syringe data reflect that
the tracer plume flowed beyond the terrain investigated with gas-bag sam-
plers. Syringe-sample concentration data in Figure 5g and 5h show the
following:
• a time-history of pollutant concentration changes in gulches
before, during and after sunrise
• identification of major drainage transport flow paths at greater
distances than investigated with gas-bag sampling sites
• pollutant mixing at the confluence of two valleys.
11-15
-------
BIG BUCK CBEEK^
-j^/-,. -•.:; • > 1* PROPOSED ^ 2,
/ '/ " x-S'SURFACE- SPENT ,
- RETORT SHALE
CORRAL GULCH
' 'SV1--' Vv '"•* :' i W':; A^V*-;
It'-l' j -vi i . i „! liCi' .Jj 'vJ w J>-
^^CTsasraS^SftN ^\
•.^vIr;1/i--?'-/f^/..tV'»-
2 39 ,-^S ' • ^T/tveaowcBEEK?-,- - '•.' ,
rr; '/s,'-.. ^^.Si!?'.'"":,''
'J7x' 'f '•' 1.6 km (1 MILE) GRID SPAC
//•;• I" '^ (20 ft CONTOUR SPACING)
FV1 '
'^'
/
-r I 2
!-J * '
.f -H.
."•' • ,«;Jl I
_: ,. 'i ''So 1
;'-v*il
SF. SAMPLING
RADIO-CONTROLLED BAG ,'
SAMPLING SITES
1 ^ CONCENTRATIONS SHOWN
NEXT TO EACH SITE ARE
pptv
FIGURE 5f. Airborne Tracer Concentrations from 0937 to 0952 MDT,
Collected at the Radio-Control!ed Sampling Sites
During the Morning of July 19, 1981
FIGURE 5g. Airborne SF6 Concentrations Along Roads for Sampling
Times from 0537 to 0652 MDT on July 19, 1981
11-16
-------
FIGURE 5h. Airborne SF6 Concentrations Along Roads for Sampling
- Times from 0704 to 0946 MDT on July 19, 1981
In addition Figure 5g shows that the effects of surface heating imme-
diately- alter sunrise9 (06ll MDT) directed a surface ff%£%^*£«
Horse Road. Upslope flow was present between 0629 and 0639 MDT, the tracer
o?ume divided along Ridge Road into two identifiable plumes. The high con-
?en?rat on near site 17 at 0629 MDT continued to reflect the tracer following
drainage flSwT In contrast, concentrations near 3.5 km reflect surface
% o I flow Iron, surface heating. Surface wind velocities for upj slope
flows are significant since tracer was measurable (0639 MDT) even at the
r ad junction9 between sites 1 and 2. Upslope flows al so transports the
tracer near to site 15; concentration is indicated by the dashed line.
Concentrations along Duck Creek also reflect plume splitting j^o major
and secondary plumes. The major plume passed across the creek near 9 km
while the secondary plume was near 3 km. The secondary plume followed a
ma or tributary Into Duck Creek. In contrast, the major plume passed over
ridges along Duck Creek. The major plume was uniformly mixed across Duck
Creek Valley at the confluence of Duck and Yellow Creeks.
Results, July 21
Meteorological data are shown as a function of time in Figure 6a for
the second morning tracer experiment. A temperature Aversion dissipated at
about 0645 MDT. Important times are tracer release and the tracer gas-
am in eHods. Tracer release was from 0300 to 1217 MDT There were en
tracer sampling periods: seven gas-bag sampling periods from 0500 to 051b,
0700 to OTIS] 0745 to 0800, 0830 to 0845, 0937 to 0952, 1100 to 1115; and
11-17
-------
360
340
320
300
280
260
240
220
210
140
120
- — — 60m
10m
TRACER
, RELEASE
10m
flu
34567
9 10 11 12
TIME. MOT
.FIGURE 6a. Wind Direction, Speed and Temperature Gradients
on July 21, 1981
1200 to 1215 MDT and three syringe sampling time periods from 0500 to 0647,
0548 to 0748, and 0800 to 0938 MDT.
Tracer-plume flow trends were similar to those during the first morning
experiment. Concentration data reflect little tracer-plume spread during
drainage flows before sunrise, plume spread and dilution after sunrise and
subsequent plume lifting. A coherent plume during 0500 to 0515 MDT is
ononnCted ^ n Fi9ure 6b by 1 ar§6 centerline concentrations, 1800 and
20800 pptv, along sampling sites 22 and 20, respectively. Since tracer
concentration was an order of magnitude greater at downwind site 20, the
tracer-plume axis passed either to the side or over site 20.
10 oS SCe]aol° is Su99ested to explain relative concentrations at sites 19,
18, 20, and 22, concentrations of 130, 210, 20800, and 1800 pptv, respec-
tively in Figure 6b. In the region of sites 19 and 18, there are two
significant drainage plumes with concentrations reflecting partial mixing of
tw° ^n^396 Plumes. The smallest of the four 1 istid concentrations!
-TJ3?,PptV-at,site 19» reflects little mixing of the tracer drainage
J 9 J6 mT drainage plume' The main drainage plume is directed 9
to f H \ re^10ns near site 19 b* the deePest local gully, a dis-
tance for developing drainage flow depths which is almost twice as long as
the distance from the tracer release site. Drainage flow from the tracer
11-18
-------
w^-'^ym-i-:-
!;:--• ^t''?i
H. '- /,• V^^TTF^V^ - - • -
v^>^- ••.,,' • .1^;PROPOSED. 7. — '- -v
f-^' / '.- " ! jfftZ SURfACE SPBNT , . • ' V
.' / / .,/* RSTORT SHALE '. ,.1800 •'
y•': -x: • li 7 r'r/^^••••-'• -,«.«
FIGURE 6b. Airborne Tracer Concentrations from 0500 to 0515 MDT,
Collected at the Radio-Control led Sampling Sites
During the Morning of July 21, 1981
release site may override the main drainage flow until the two plume mix
near site 18, a concentration of 210 pptv. The main plume is deflected
northward near site 22, a concentration of 1800 pptv, by drainage flow from
the tracer release site. The "undiluted" tracer-plume axis passes near
site 20, a concentration of 20800 pptv.
There is subsequent supporting evidence for this two-plume scenario
from the succeeding gas-bag sampling period between 0700 and 0715 MDT (an
hour after sunrise at 0605 MDT); see Figure 6c. The tracer plume axis was
deflected and lifted by interaction with a main drainage flow of less rela-
tive strength than earlier because of surface heating. All concentrations
were significantly reduced. Nevertheless, concentrations at downwind site 20
remained on an order of magnitude greater than at site 22; i.e., 2070 pptv
compared to 235 pptv. Tracer plume spread to the north included sites 10,
9, and 17. The conclusions are that the "undiluted" tracer plume axis was
near site 20 and plume interactions resulted in significant tracer
concentrations at sites 10, 9, and 17.
Concentrations continued to reflect increased surface heating effects
with increasing time. As shown in Figure 6d for 0745 to 0800 MDT and in Fig-
ure 6e for 0830 to 0845 MDT, plume spreading and interaction continued.
Differential plume heating effects became important. The "major" drainage
flow became relatively less significant because of valley side-wall
heating. In contrast, the "secondary" drainage plume persisted from the
tracer release site. This persistence directed tracer plume flow into the
gully between sites 17 and 18. After drainage flows dissipated, plume
11-19
-------
300' "*2070
V. PROPOSED. .570
-'* SURFACE • SPENT . * '
. HAOIO-CONTROLLED B*O -"'=- S
SAMPUNG SITES t l"'
J? • j >
CONCENTRATIONS SHOWN
HEX! TO EACH SITE ARE
F^jOWvmi^t
FIGURE 6c. Airborne Tracer Concentrations from 0700 to 0715 MDT,
Collected at the Radio-Control!ed Sampling Sites
During the Morning of July 21, 1981
/•V.'"'' V HB>1«.J ^ •fV»'
-/Y ' ' !.' ,' V« Urn (1 mll.| GRID SPACING >9
t- . f S 120 ft CONTOUR SPACINGI ^
FIGURE 6d. Airborne Tracer Concentrations from 0745 to 0800 MDT,
Collected at the Radio-Control led Sampling Sites
During the Morning of July 21, 1981
11-20
-------
• LITTLE DUCK CREEK
.'•.PROPOSED .80
•*" SURFACE SPENT , *260 " '
RETORT SHALE '. «7
. ' MO
".-,-fpK": pS;
>-""'•. KV 'm
k^::sM*Mi
. .. „ W.^...Sp^,Jffi\,^Jg£-
_,^y. - |) ^/^flrOwy -
FIGURE 6e. Airborne Tracer Concentrations from 0830 to 0845 MDT,
Collected at the Radio-Control!ed Sampling Sites
During the Morning of July 21, 1981
lifting occurred. Plume lifting is shown by low concentrations in Figure 6f
at times from 0937 to 0952 MDT.
Gas-bag samplers were subsequently collected in zone 1; see zone loca-
tion in Figure 2 to further investigate pollutant transport during upslope
flows. As shown in Figure 6b, downslope winds ceased at about 1000 MDT at
the tracer release site, with subsequent variable wind directions. As a
result of this variation, upslope surface winds transported the tracer west-
ward towards sampling zone 1. Concentrations for two time periods are shown
from 1100 to 1115 MDT in Figure 6g and from 1200 to 1215 MDT in Figure 6h.
Concentrations above 1 pptv were not detected in zone 1.
Syringe samples were collected to investigate in detail tracer plume
segmentation, effects of tracer plume flow through side tributaries along
Yellow Creek, and cross-valley tracer plume flows into Duck Creek. The
syringe-sample concentration data in Figures 6i through 6k show the
following:
• a time-history of pollutant concentration changes in gulches
before, during and after sunrise
• further identification of major drainage transport flow paths at
distances greater than those investigated with gas-bag sampling
sites
• cross-valley concentration profiles along Duck and Yellow Creeks
11-21
-------
•vi .1 x\ ''••" ; ^;C }:^v*:<» •
o^^^iS£Si*§ v^ •
^Tl^J^^{T;:;
0:lV';^Li(:W: -:
FIGURE 6f. Airborne Tracer Concentrations from 0937 to 0952 MDT,
Collected at the Radio-Controlled Sampling Sites
During the Morning of July 21, 1981
1
-------
'W1-- VI '"•* •" My- 'i'S#? 'v<>
•~^'^jD^EEij^^j^fr;-1^& V ^;:-
- ->5*- • . ,.•' • yfepnoposeD: .
' " Z SUKFACE-SMNT . .» "
. RETOIIT SHALE -
-.-'!
• - ,•>• '/r- ;r-
"* --' - --•' •-
f'jf. '( {.' 1.6 km (1 MILE) GRID SPAC
//. I" ''1 (20 ft CONTOUR SPACING)
y"'I. ;.Jr'..'<:..' i S
•*"• ' \' ' -. i •
SF« SAMPLING
RADIO-CONTROLLED BAG
SAMPLING SITES
CONCENTRATIONS SHOWN
NEXT TO EACH SITE ARE
pptv
'...ft .!»._»£
FIGURE 6h. Airborne Trace Concentrations from 1200 to 1215 MDT,
Collected at the Radio-Control!ed Sampling Sites
During the Morning of July 21, 1981
FIGURE 6i. Airborne SFe Concentrations Along Roads for Sampling
Times from 0500 to 0647 MDT on July 21, 1981
11-23
-------
O——O ALONG MAIN ROADS
O—O ALONG CftOSS-V ALLEY T« AILS
FIGURE 6j. Airborne SF§ Concentrations Along Roads from Sampling
Times from 0548 to 0748 MDT on July 21, 1981
FIGURE 6k. Airborne SF5 Concentrations Along Roads for Sampling
Times from 0800 to 0938 MDT on July 21, 1981
11-24
-------
• pollutant flow from side tributaries leading into Yellow Creek and
• pollutant mixing and channeling in valleys at valley confluences, and
downwind of valley confluences.
Concentrations determined from syringe samples confirm that tracer-plume
spread was limited during drainage flows before sunrise. As shown in Fig-
ure 6i near site 18, the maximum syringe-sample concentration is comparable
to the maximum 15-min gas-bag concentration of 20800 pptv shown in Figure 6b.
The main-plume width was narrow, and concentrations decreased rapidly as a
function of cross-wind distance. Relative concentration decrease rates fur-
ther support the scenario concept: a major drainage flow interacting with a
smaller drainage flow containing tracer; Concentrations decreased over two
orders of magnitude in a distance of about 0.1 km at a projected road dis-
tance of 2.7 km. Hence, the "edge" of the main drainage flow is near 2.7 km.
In contrast, the gradual concentration decrease from 3 to 6 km reflects the
diverted prinicipal plume flow from the tracer release site.
Tracer plume flow segmented into multiple tracer plume flows as plumes
followed along terrain depressions toward Duck and Yellow Creeks. Terrain
effects were investigated by collecting some samples either near side-
tributaries along main roads or along cross-valley traverses. Cross-valley
concentrations were investigated by using side trails from the main roads
along Duck and Yellow Creeks.
A more detailed description of main road and side-trail locations is
needed to emphasize concentration measurement locations, more detailed than
can be read from the figures. Main roads were usually located along either
side of each valley. In Duck Creek Valley, the main road was along the val-
ley's northern.edge. Side trails extended approximately half-way across the
valley towards the tracer release site and terminated at the creek. Beyond
the confluence of Duck and Yellow Creeks, the main road continued along the
northern edge of the valley to 7 km along Yellow Creek. Beyond 7 km, the
main road crossed to the southern side of the valley. Several syringe sam-
ples were usually collected along this valley crossing. For the remainder
of Yellow Creek Valley used in these experiments, the main road was located
along the southern edge of Yellow Creek Valley between 0 and 2.4 km and along
the northern edge between 2.4 and 7 km. At 7.5 km, a side trail extended
completely across the valley.
Concentrations along Duck Creek shown in Figure 6i reflect tracer plume
impingement on a valley side wall at 2.5 km; a major tracer plume with a cen-
ter! ine near 5.5 km, and a secondary plume near 8.8 km. Impingement was
along the northern side of a major gulch leading in to Duck Creek, near the
juncture with Duck Creek. Impingement occurred prior to the cross-valley
traverse: concentrations are indicated by the dashed line. At the traverse,
the plume axis was near the center of this major gulch. The traverse was
near the confluence with Duck Creek Valley. Beyond the confluence, the
tracer plume impacted drainage flow from upper Duck Creek. Consequently, the
tracer plume flowed along the southern edge of Duck Creek. At 4 km, concen-
trations along the center of Duck Creek were an order of magnitude greater
than they were along the northern edge of Duck Creek. Nevertheless, plume
mixing continued down valley. However, mixing was incomplete since concen-
trations were always greater near the center of Duck Creek. Also, these
11-25
-------
greater center!ine concentrations probably reflect contributions from seg-
mented tracer plumes flowing over the southern edge of Duck Creek.
The tracer plume segmented in the region between Duck and Yellow Creeks.
Segmentation is confirmed by concentrations along Yellow Creek. Along Yellow
Creek, the maximum concentration was at 4.5 km, a location adjacent to a
side tributary which directed a tracer plume segment into Yellow Creek.
Beyond the confluence of Duck and Yellow Creeks, the tracer plume flowed
along the northern valley edge. At 7.5 km along Yellow Creek, concentrations
were nearly an order of magnitude less than along the southern edge. It is
noted for subsequent referral in Figure 6j that although sunrise was 0605 MDT
at the tracer release site, this southern edge was still shaded at 0618 MDT.
Concentrations were nearly uniform across the valley further down creek,
about 10.2 km along Duck Creek.
Tracer plume segmentation and interaction with ambient air plumes are
further confirmed with concentrations shown in Figure 6j. First discussed
are concentrations near site 18. Between about 2 and 2.5 km, concentrations
cycled about an order of magnitude between succeeding sampling locations.
This cycling may have occurred because samples were collected near the undu-
lating interface between the tracer plume and the major drainage plume from
near site 2, and the interface width was greater than reflected in Figure 6i.
Although plume-interfacing characteristics were changing, a portion of the
tracer plume continued to flow as a secondary plume near 1.2 km.
The main concentration trends (cross valley concentration differences
and segmented tracer plume flows along side tributaries) continued along
Duck and Yellow Creeks. However, significant cross-valley differences also
occurred along Yellow Creek. At 2.5 km, concentration near the northern
edge was about four times greater than near the southern edge of the valley.
At 7.5 km, where the entire valley was not being heated by the sun, there
was no longer a cross-valley concentration difference.
Concentrations in Figure 6k are more influenced by surface heating
effects since samples were collected two or three hours after sunrise. Sur-
face heating directed a segmented tracer-plume flow between 1 to 3 km along
Ridge Road. Near site 18, the tracer plume was also directed between 3 km
(the road along site 18) and 6 km along Ridge Road. This last plume direc-
tion confirms in more detail the main tracer-plume direction concluded from
gas-bag concentrations shown in Figure 6e for times between 0830 and
0845 MDT.
CONCLUSIONS
Fugitive emissions transport and dispersion were simulated in tracer
experiments during conditions of relatively low wind speeds; i.e., wind
speeds less than the threshold wind speed causing soil suspension. Although
spent^shale will probably suspend only for greater wind speeds, fugitive
emission rates will be nearly independent of wind speed if suspension were
mainly caused by mechanical stresses. For instance, constant vehicular
traffic on spent shale can produce a relatively constant fugitive emission
11-26
-------
flux which is independent of wind speed. If this proposed emission scenario
occurs, the maximum downwind airborne concentrations will similarly occur
during low wind speed conditions. Airborne concentrations will remain rela-
tively large since airborne pollutant concentrations will not be diluted by
high wind speeds, u, i.e., downwind airborne concentration decreases are
proportional to 1/u.
A multiplicity of plume flows and plume interactions were identified
for fugitive emissions transport from the proposed spent shale disposal
site. The tracer plume segmented. The greatest concentrations occurred
during drainage flows. Plume interactions were between traced plumes and
ambient air plumes. Since multiple plume flows and plume interactions were
identified, model predictions of downwind transport of fugitive emission
plumes on a local scale must include predictions of plume segmentation and
interaction.
Predictions on a local scale are essential if maximum "fence-line" con-
centrations are considered. Two parameters in model predictions are fugitive
emission rates, Q, and airborne concentrations, x. The only controllable
parameters in these models are pollutant release heights, release times, and
release rate, Q. Predictions of X/Q ratios both at the fence-line and
downwind are used by industry and in regulations. Although further analysis
is planned for evaluations of predictive models, the x/Q ratios can
nevertheless be estimated from the reported tracer concentrations; x with
units of pptv, and-the constant tracer release rate, Q = 26 kg/hr.
REFERENCES
Clements, W. E., S. Goff, J. A. Archuleta and S. Barr. 1981. Experimental
Design and Data of the August 1980 Corral Gulch Nocturnal Wind Experiment,
Piceance Basin, Northwestern Colorado.LA-8895, Los Alamos National Labora-
tory, Los Alamos, New Mexico.
Sehmel, G. A. 1981. "A Dual-Tracer Experiment to Investigate Pollutant
Transport, Dispersion, and Particle Dry Deposition at the Rio Blanco Oil
Shale Site in Colorado." PNL-SA-9327. In Proceedings of the Second
Conference on Mountain Meteorology, November 10-31, 1981, Steamboat Springs,
Colorado.Published by the American Meteorology Society, Beacon Hill,
Massachusetts, pp. 137-146,
Whiteman, D. C., N. S. Laulainen, G. A. Sehmel and J. M. Thorp. 1981. Mete-
orological Data Report for Field Studies Conducted in August 1980 in the Oil
Shale Region of Northwestern Colorado.PNL-3734, Pacific Northwest Labora-
tory, Richland, Washington.
11-27
-------
Laboratory Testing To Improve Rail Car Sealant Spray and Loading
Techniques for the Abatement of Fugitive Coal Dust
Colin J. Williams and William F. Waechter, MHTR Ltd. *
ABSTRACT
The technique of spraying coal rail cars with latex or similar sealants to con-
trol fugitive emissions has been used for several years. There are serious
deficiencies in the systems used to load coal cars and also in the application
techniques of the sealants. This paper reviews these problems and gives prac-
tical advice, resulting from laboratory tests and field visits to four coal load-
ing facilities, on how to improve the loading and spraying methods currently
used.
A series of laboratory tests has been conducted which subjected over 100 samp-
les of coal to simulated train journeys. These tests, unlike previous tests on
sealants, simulated the important parameters of the motion induced vibration
of the coal surface and also wind effects on the erosion of the sealed surface.
Measurements of the actual vibration levels in a full scale coal rail car were
used to design the vibration table and MHTR's boundary layer wind tunnel was
used for the wind erosion tests.
The following variables affecting sealant performance were investigated:
sloped surfaces; centre humps; settling during transit; concentration of sealant
solution; alternative sealants; various application techniques; effect of rainfall;
preferential spraying and ambient air temperatures of 15°C and 30°C.
(This is an abstract of a presentation for which a paper is not available.)
(*) Current address: Rowan, Williams, Davies, and Irwin, Ltd., 650 Wood-
lawn Road, West, Guelph, Ontario, Canada NIK IB 8.
12-1
-------
Studies of Nontraditional Fugitive Particulate Control Techniques
Brock M. Nicholson, EPA/OAQPS-RTP
Maxine Borcherding, City of Portland (Oregon)
Gary Ekhardt, State of Minnesota
Ray Mohr. State of Colorado
ABSTRACT
Demonstration studies have been conducted with funding assistance from EPA in
three U. S. cities to determine the effects of controlling nontraditional sources of
fugitive particulate on ambient air quality. These studies involve application of
control measures which have not been traditionally required as part of the State
Implementation Plans (SIPs) for attaining ambient air standards. It is now gener-
ally recognized that many urban areas of the country will be unable to attain the
ambient standards solely with application of traditional or stack emissions con-
trol. Because of the magnitide of this fugitive problem and the difficulty in defi-
ning its character, EPA has allowed states to perform demonstration studies of
nontraditional fugitive control techniques as a logical first step in a control stra-
tegy for attainment of ambient standards.
The study design, execution, and results to date of these studies are described.
Attention is given to the technical and institutional aspects that were followed in
an effort to ensure results more reliable than previously obtained in such studies.
The SIP development process, as envisioned by EPA in allowing studies, has not
resulted in the initiation of a large number of studies. Therefore, it is hoped
that the knowledge gained by a few major studies, on the least understood of the
nontraditional control measures, will be applicable in many areas throughout the
country.
In Denver, the study involves modified winter snow and ice control practices,
such as sand cleanup and use of salt as an alternative. In Portland, vacuum sweep-
ing of heavily soiled paved road surfaces in industrial and commercial areas is
be.ing studied. In Minneapolis, the study focuses on the control of unpaved road-
ways at construction sites and the cleanup of mud and dirt carried onto paved
roads around construction sites.
(This is an abstract of verbal presentations for which papers are not available.)
13-1
-------
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
A New Charged Fog Generator
for Inhalable Particle Control *
by
C. V. Mathai
Arizona Public Service Company
P. O. Box 53999, Station 5680
Phoenix, Arizona 85072-3999
(*) Although this is not the actual presentation made at the May 1982
meeting, it closely resembles that presentation, which was en-
titled, "Evaluation of the Efficiency of a Charged Fog Generator
in Controlling Inhalable Particles at a Steel Plant, " by C. V.
Mathai and Bradley M. Muller (then of AeroVironment, Inc.,
Pasadena, CA 91107) and William B. Kuykendal (then of EPA/
IERL-RTP, but now of EPA/OAQPS/AQMD). The work was
supported by EPA contract 68-02-3145.
14-1
-------
A review of the literature shows that
control efficiency of inhalable particles
using water droplets can be improved
significantly if the droplets are electrically
charged. A spinning cup fog thrower
was developed initially to generate
electrically charged water droplets. The
poor performance of this device in wind
tunnel tests was attributed to the short
lifetime of the fine droplets generated.
the ineffective ionizer ring method of
charging the droplets, and the nonuni-
form charge distribution observed
along different regions of the fog spray
pattern.
A new charged fog generator (CFG)
was then developed by modifying a
commercial rotary atomizer. In this
device, the droplets generated are
contact-charged to provide a high
charge-to-mass ratio of 1.2 fiC/g. The
droplets have a number concentration
median diameter of about 100 fjm and a
mass median diameter of about 200
/an. The water flow rate is variable (4-
70 l/h), and the fog spray pattern can
be easily changed from long and narrow
to broad and short, with a typical spray
coverage of 16-24 m3. The device uses
about 1 kW power (110 VAC) and is
portable.
The CFG (at a bentonite ore loading
operation) was extensively field-tested
to determine the dependence of its
inhalable particle control efficiency
(PCE) on various instrument settings
and field conditions. These tests show
that the overall mean PCE is 78% higher
than the corresponding value for
uncharged fog. Individual PCEs as high
as 88% were achieved. The lifetime of
the droplets seems to be the dominant
factor determining the PCE; and PCE
values were higher for higher applied
voltages and higher water flow rates.
The data suggest that, under optimum
instrument settings. PCE of water
droplets could be doubled by charging
the droplets.
Introduction
Although the total particle mass
loading of anthropogenic aerosols is only
about 10% of that from natural sources,
their effects are significant and largely
detrimental. Recent reports indicate
that inhalable particles (15 i/m and
smaller in aerodynamic diameter), in
general, and fine particles (2-3 //m and
smaller), in particular, may be a human
health hazard, and degrade atmospheric
visibility. Devices such as electrostatic
precipitators and wet and dry scrubbers
have performed exceedingly well in
controlling pollutants from conventional
industrial stack sources. However, effec-
tive and economically feasible methods of
controlling inhalable particles from non-
stack sources in open areas are lacking.
Recognizing this situation, the U.S. EPA
has been encouraging the development
of new methods, such as charged fog
technology, to control fugitive emissions.
Removing fine particles from a gas
stream in an open area is difficult
because of the particles' low mobility and
unfavorable inertia! properties, as well as
uncontrollable external factors (mete-
orological parameters). The most com-
monly used dust control (ordinary water
sprays) in mining and other material
handling areas is only 30%-40% efficient
in controlling inhalable particles. Only
during the last few years has electrostatics
been used to augment particle collection
efficiency of water droplets. Numerous
studies have shown that most industrial
pollutants and naturally occurring dust
particles acquire electric charges as they
are dispersed into the air. Studies have
also shown that the polarity and magni-
tude of the charges on these particles
depend on their size and origin (coal, soil,
mineral, etc.). Therefore, the particle
collection efficiency (PCE) of water
droplets can be significantly enhanced
via electrostatic forces of attraction if the
droplets are charged to the opposite
polarity.
Commercially available charged fog
devices for fugitive emission control have
certain disadvantages: the need for high
pressure air or water to properly atomize
the water; the possibility of spray nozzle
clogging if the water supply contains high
concentrations of 'dissolved salts and
suspended solids; and poor charge-to-
mass ratio of the droplets to provide a
high degree of fine particle control. Under
the sponsorship of the U.S. EPA, Aero-
Vironment, Inc., has developed a new
charged fog generator (CFG) which
overcomes these difficulties. This research
and development project consisted of two
phases: Phase I involved examining the
theoretical aspects of charged fog
technology, developing a bench-scale
prototype instrument and evaluating its
inhalable PCE in a controlled laboratory
experiment, and an economic analysis of
the feasibility of the technique to control
fugitive emissions; and Phase II involved
construction and field-testing of a
prototype to evaluate its PCE. This report
summarizes both phases.
14-2
-------
Theoretical Background
When an aerosol particle approaches a
water droplet with a relative velocity, it
may collide directly with the droplet
(impaction), barely touch the droplet
(interception), or miss the droplet entirely.
The collection of an aerosol particle by a
charged droplet is the result of a number
of simultaneous mechanisms of interac-
tion between them; e.g., inertial impaction,
direct interception, Brownian diffusion,
and electrostatic, diffusiophoretic, and
thermophoretic forces. The relative effect
of the mechanisms of interaction between
the droplet and the particle depends on
the size of the particle. For large particles
(aerodynamic diameter greater than 2-3
the dominant mechanisms of
10°
particle collection by droplets are impac-
tion and interception. For particles
smaller than 0.1 fjm, Brownian diffusion
becomes very important. For particles
between these two ranges, electrostatic
forces are the dominant interaction
mechanisms.
The PCE of uncharged water sprays
(where inertial impaction is the major
collection mechanism) is given by,
1-exp -.- -2k-.!=.•
2 QG D
(D
where QL and QG are the volumetric flow
rates of the water (liquid) and air (gas)
components of the dust cloud, respec-
tively; L is a characteristic length for the
total capture process; 0 is the mean
droplet diameter; and rj is the single
droplet collection efficiency. Figure 1 is a
plot of calculated E versus particle radius
for a droplet radius, R, of 106 fjm and RH
of 75% (curve C). This curve has a
minimum (whose magnitude depends on
the droplet size) for particles with radius,
r, near 1 fjm. This minimum is caused by
the ineffectiveness of inertial and diffusive
interaction mechanisms for particles in
that size range. However, if the droplet
and the particle are oppositely charged,
the minimum in E is eliminated (Curves A
and B) as electrostatic forces become the
dominant mechanism of particle collection
in this size range. This effect is the
fundamental principle on which charged
fog technology is based. When the
droplets and the particles are charged, rj
is given by,
77 = - 4CQC Qp/24 77* c0r R2 pU0, (2)
10'
Iff
0.01
0.1 1.0
Particle Radius, tun
10.0
Figure 1.
Calculated single droplet collec-
tion efficiency in air of 10aCand
900 mb as a function of panicle
radius for 106fjm radius droplet
at 75% relative humidity for
droplet and particle charges of
(A) ±20 esu cm'2. (8> ±2 esu
cm'*, and (Cl zero charge.
where C is the Cunningham slip correction
factor; Qe and Qp are the charges on the
droplet and particle, respectively; E0isthe
dielectric constant; p is the viscosity; and
U0 is the free-stream velocity. Thus, for a
given particle size distribution, the
electrostatic forces are proportional to
the magnitude of the charges and
inversely proportional to the droplet's size
and its free-stream velocity.
Although Figure 1 shows that the
addition of electric charges on particles
and droplets eliminates the minimum in
PCE (near particle radius, r = 1 um) and
yields values of E which are 5 to 10 times
higher in certain size ranges than for
uncharged sprays, the overall PCE of an
operating system may not be that high.
Increases in PCEs of about 15% for 1 -//m
particles to over 45% for 0.3-//m particles
(compared to uncharged droplets) have
been reported. Values of 50%-80%, with
charged droplets under controlled experi-
mental conditions, have also been
reported.
When sprayed into the air, the charged
droplets will evaporate unless the air is
saturated with water vapor. The droplet's
lifetime determines the effective contact
time between the droplet and particles
and thus strongly influences the overall
14-3
-------
PCE of the system. The lifetime of a
water droplet depends on the temperature
and relative humidity of the medium into
which it is introduced. To obtain the best
PCE. the droplets must be small enough
to provide both an adequate spray rate per
volume of gas treated and sufficient
contact time, yet large enough so as not to
evaporate too quickly.
The maximum charge a droplet can
carry before it disintegrates is reached
when the outward pressure produced by
the electric field at the surface of the drop
is equal to the inward pressure produced
by the surface tension. This limiting
charge is called the Rayleigh limit, given
by,
0,.y = 87r [CoffR3]1'2 O)
where Qp.y is the limiting charge on the
droplet (coulombs); c0 is the permittivity of
the medium in which the droplet is
located; a is the surface tension of the
liquid; and R is the droplet radius, in
micrometers.
Charged Fog Generator
Development
Water droplets may be generated by a
spray nozzle or rotating cup. Droplets are
generally charged by electrostatic induc-
tion, ionized field, or contact charging.
Commercially available charged fog
devices use water spray nozzles and
induction charging.
In this study, a prototype unit, called the
spinning cup fog thrower (SCFT), was
initially developed in association with the
University of Arizona in Tucson. The SCFT
consisted of a rotary atomizer, an ionizer,
and a vane-axial blower. The rotary
atomizer consists of a small hollow-shaft
motor and a spinning cup. Water from a
low-pressure source is introduced into
the hollow shaft and flows toward the
other end where the spinning cup is
mounted. Entering the rear of the cup, the
water stream strikes a rotating "spider"
which deflects the water to the sides. A
sheet of water then flows toward the lip of
the cup where droplets are formed by
centrifugal force and the air stream's
striking the thin water layer. The droplets
were then charged by a stream of positive
ions produced by the ionizer containing
numerous small discharge needles. It
was expected that the ions produced in
the region of the ionizer ring would follow
the airflow from the vane-axial fan, mix
with the droplets, and charge them. The
charged droplets were then deflected and
projected forward by a stream of air
supplied by the vane-axial blower. These
droplets had a median diameter of about
20 fjm and a charge-to-mass ratio of 1 x
"
To evaluate the PCE of the SCFT, tests
were conducted in a wind tunnel at the
University of Arizona. These tests showed
only about 50% inhalable PCE. This poor
performance was attributed to the very
short lifetime of the fine droplets gene-
rated, the ionizer ring method of droplet
charging (which proved ineffective), and
the nonuniform charge distribution
observed along different regions of -the
fog pattern. The SCFT was therefore
abandoned, and a new charged fog
generator (CFG) was developed by
modifying a commercial rotary atomizer.
Figure 2 schematically represents the
new CFG. Water is introduced through
the water tube into the 3600-rpm rotating
cup, whose inside is fabricated to a
gradual smooth taper. A small deflecting
baffle is attached to the open end of the
water tube so that the water will be
deflected 90° and strike the rear end of the
rotating cup. Because of the centrifugal
forces, the water becomes a thin film and
moves forward into a high velocity
airstream from the vane-axial fan. The
impact of the high velocity air on the thin
water film breaks the water film into fine
water droplets.
The water tube, air cone, and rotating
cup are made of nonconductive materials.
The water tube is attached to the rotating
cup, thus rotating with the cup. The other
end of the water tube is attached to the
water supply through a rotating seal. The
water for atomization is stored in an
electrically isolated reservoir (120-1
capacity), and a small pump is used to
pump it to the inlet of the rotating seal.
The water flow rate can be varied from
about 4 to 70 l/h. The airstream from the
fan, controlled by an air butterfly setting,
projects the fog forward. By adjusting this
setting, the airflow speed can be controlled,
thereby controlling the shape of the fog
spray pattern to better conform to the
shape and size of the source of dust upon
which the charged fog is to be applied.
The spray pattern covers a volume of 16-
24m3.
To achieve a high charge-to-mass ratio
for the droplets, contact charging by
directly connecting a high voltage source
to the inflowing water was found to be the
most efficient method. However, this
method requires that the entire water
supply (reservoir) and associated tubing
be electrically isolated so that there is no
current leakage.
The size distribution of the droplets was
determined using a cloud optical array
probe and a precipitation optical array
probe. These measurements gave a
14-4
-------
number concentration median diameter
of about 100 fjm and a mass median
droplet diameter of about 200 fjm.
Collecting the droplets on greased glass
slides and observing them under a
microscope yielded values consistent
with the above results.
The charge-to-mass ratio of the droplets
was determined using a special sample
train (developed by AeroVironment),
consisting of an insulated stainless steel
probe tip mounted on a standard glass
midget impinger. The probe was connect-
ed electrically to copper wool, packing
inside the midget impinger, which was
also connected by a shielded cable to an
electrometer. The impinger was immersed
in a Dewar flask containing dry ice. An
isokinetic sample of droplets was then
extracted from the fog spray. As the
droplets moved through the impinger
they were condensed, transferring their
charge to the copper wool packing. The
charge transferred to the copper wool
was then measured by the electrometer.
Knowing the mass of droplets collected
and the current produced by them, the
charge-to-mass ratio was calculated.
This method gave a typical value of 1.2 x
10~* C/g with an applied voltage of 15 kV,
about 25% of the maximum allowed value
(Rayleigh limit) for 200-//m droplets.
The charge-to-mass ratio was also
estimated using another experimental
setup, in which droplets were allowed'to
transfer their charges to a 125 um mesh
size standard sieve placed in the path of
the fog spray and the current generated in
the mesh was measured. Knowing the
size of the droplets and the number of
droplets generated by the CFG per
second, the charge-to-mass ratio was
estimated. This method gave values
consistent with those reported for the
gravimetric method.
In summary, the CFG is fairly small,
portable, and mounted on a moveable
platform. The total power requirement is
about 1 kW (110 V AC power supply). It
has no nozzle-clogging problem and does
not require compressed air. Using a small
inverter, the unit can be operated by an
automobile battery. This feature may be
important because the system has high
potential application in locations where
commercial electric power is not available.
The physical characteristics of the
charged droplets from the CFG are such
that high innalable PCE is expected.
Evaluation of the Inhalable
PCE of the CFG
The CFG was field tested at the Kaycee
Bentonite Corporation's bentonite pro-
cessing plant in Worland, WY, during the
first half of 1981. Bentonite is a water-
absorbing material used mainly to seal
water leaks in oil wells. Bentonite ore is
unloaded from front-end loaders onto the
Air Fan
DC Power Supply
Nonconductive Air Cone
Water-Deflecting
Baffle
Nonconductive
Spinning Cup
Nonconductive
Water Tube \
DC Water Pump
Figure 2. Schematic of the charged fog generator.
14-5
-------
grill of a hopper which is attached to the
west wall of the plant building. From the
bottom of the hopper (approximately 4 m
below the hopper grill level, inside the
plant), the ore is carried by conveyor belts
to processing areas. The hopper is
completely enclosed except on the side
through which the front-end loaders
unload. The hopper is 6.7 m wide, about 2
m deep and, from the grill level, about 3 m
high. The grill is inclined about 30° to the
horizontal.
Bentonite ore from two large piles,
approximately 100 m to the northwest
and southwest of the hopper, are carried
to the hopper by front-end loaders and
dropped on the grill. The bucket of the
front-end loader is about 2.4 m wide and
successive loads are unloaded uniformly
over the 6.7 m wide hopper. It takes about
10 front-end loader dumps to fill the
hopper to the grill level. Those 10 dumps
are accomplished in about 25 minutes.
The bucket is removed from inside the
hopperjirea in 20-25 seconds. One full
hopper of ore will be carried away by the
conveyor belt in about an hour.
The land around the plant is fairly flat. A
railroad track is east of the plant and a
paved road (very little traffic) beyond that.
Bottom-dump trucks filled with bentonite
ore arrive near the storage piles from the
south and, therefore, the dust in the
hopper area due to transport of ore to the
plant area is negligible compared to the
dust generated in the hopper itself. In
other words, the hopper can be considered
as a fairly isolated source.
The CFG and the particle sampling
instruments were mounted on the
outside of the south wall of the hopper,
about 4 m above ground level. A platform
was built to mount these instruments.
Ideally, the particle sampler inlet should
have been mounted on the east (rear) wall
of the hopper, but for practical reasons
could not be. The CFG sprayed water
droplets across the hopper above the grill.
The total volume to be treated by the
charged fog was about (6.7 m x 2 m x 3 m)
40 m3, somewhat larger than the max-
imum coverage of the CFG fog. Unfortu-
nately, only one prototype CFG was
available for these tests. To have had a
second unit, mounted on the north wall of
the hopper and operated concurrently,
would have been ideal.
Particle samples were collected using a
Sierra Model 230 CP cyclone preseparator
followed by a Sierra two-stage cascade
impactor. The cyclone's air inlet protruded
0.3 m into the hopper. The particle
sampling system was operated at a flow
rate of 0.85 mVmin (30 cfm), which was
calibrated at regular intervals during the
entire test program. At this flow rate, the
cyclone has a particle cut-point of about
7.3 //m. The impaction plates of the
cascade impactor were chosen so that
the upper filter would collect particles
larger than 1.8 fan. The lower backup
filter collects all particles smaller than 1.8
fjm. For this report, particles collected on
the backup filter are characterized as the
fine particle fraction, and those on the
upper filter as the coarse fraction. The
mass of coarse fraction collected was
often an order of magnitude smaller than
the fine fraction (possibly indicating a
particle bounce problem with particles
larger than the cut-point of 1.8 /urn
reaching the backup filter).-Subdividing
the samples into more size ranges would
have required either much longer sam-
pling time or much higher flow rates to
obtain acceptable filter loadings. Neither
alternative was desirable.
Field Test Design
Three test scenarios were designed to
determine the PCE of theCFG. In the first,
no attempt was made to control the dust
inside the hopper, except that the CFG's
fan ran continuously (this ensured that
the mixing of the dust cloud was nearly
identical among various runs in order to
make direct comparison of tests with and
without charged fog). In the second,
uncharged fog was applied on the dust
cloud. In the third, charged fog was
applied. Other governing parameters
were varied under each scenario to yield
a statistically acceptable set of test data.
The parameters varied were water flow
rate, fog pattern (short or long), applied
high voltage, wind conditions, and
relative humidity. In addition, tests were
performed with the polarity of charges on
the droplets reversed. This test protocol
called for 32 test runs. However, in actual
practice, data had to be collected under
prevailing field conditions; hence, 96
runs were performed.
Particle samples were collected on
preconditioned and preweighed glass
fiber filters. Each particle sample was
collected during 12-15 front-end loader
dumps (roughly 30-40 minutes). Wind
speed and direction and relative humidity
were also recorded simultaneously. After
each sample was collected, the filters
were transferred to special envelopes
and brought to AeroVironment's labora-
tory for analysis.
14-6
-------
Data Processing
The mass of the fine and coarse
fraction and sum of the two were
obtained from the final and initial weights
of the filters. Using known values of the
sample time, number of dumps, and flow
rate, particle mass concentration for each
dump was calculated.
The amount of dust generated in the
hopper fluctuated from dump to dump. To
overcome this variation, each sample
was collected during 12-15 dumps and
the mean value for each dump was
calculated. The amount of dust generated
also varied from day to day, caused by
factors such as changes in ambient
conditions and the moisture content of
the bentonite ore. The effect of the latter
is not too significant since the ore stored
in each pile comes from the same mine
and a pile is completely processed before
a new pile is started. To overcome the
day-to-day dust level fluctuations, sample
concentrations were normalized for each
day with respect to the background value
(the fan-only value), and a percentage
particle collection (control) efficiency was
calculated. Percentage PCE is obtained
from:
where C0 and C are the particle concen-
trations corresponding to "fan only" and
"fog test" scenarios, respectively.
Values of E are calculated for the fine
fraction, the coarse fraction, and the sum
of fine and coarse. The report gives these
percentage PCEs and the corresponding
field condition data, water flow rate, and
applied high voltage for the whole field
test program.
Comparison Between Charged
and Uncharged Fog
Figure 3 shows the mean values of the
measured percentage fine PCE and total
PCE for charged (striped bars) and
uncharged (solid bars) fog. For this
comparison, all test runs under all
instrument settings and field conditions
are included. The mean and standard
deviations of the fine PCE are 48.1% and
23.0%, respectively, for the charged fog
and 27.8% and 25.3%, respectively, for
the uncharged fog. The corresponding
values for all the particles (fine and
coarse together) are 44.5%, 21.8%,
25.0%, and 24.4%, respectively. These
numbers show that, even under average
field conditions and instrument settings,
inhalable PCE can be almost doubled by
electrically charging the water droplets.
However, under optimum instrument
settings and favorable field conditions,
50
Control Efficiency. %
8 S
Particle
K>
O
10
0
, • Uncharged Fog
0 Charged (+ & -) Fog
-
-
I
'/
I
I
\
'/>
-
-
Fine Panicles Only Fine & Coarse Particles
Figure 3. Mean inflatable PCE of all test
runs (for various CFG settings
and field conditions/ for charged
and uncharged fog.
the improvement in inhalable PCE can be
expected to be higher. Note that the
volume of the dust cloud treated was
somewhat larger than the maximum
coverage of the CFG.
The size range of particles collected in
the coarse mode was fairly narrow, and the
mass collected was often an order of
magnitude smaller than the fine fraction.
Another problem which may have inhibit-
ed coarse particle collection is the particle
bounce effect, by which some of the
larger particles pass the upper impaction
plate and settle on the backup filter with
the fine particles. Consequently, most of
the ensuing discussion concentrates on
the sum of fine and coarse fractions of the
particles collected. Therefore, this partic-
ular experiment could not demonstrate
the full effect of charged fog on the coarse
particles.
Figure 4 shows total PCE with E plotted
as a function of ambient RH for two sets of
instrument settings for an applied high
voltage of 4 kV (positive charges) and a
water flow rate of 60 l/h. The circles
represent a broad spray pattern, and
squares represent a narrow spray pattern.
The method of least squares was used to
fit a straight line to the data sets, shown
in the figure, yielding a correlation
coefficient of 0.93 for the broad spray and
-0.19 for the narrow spray. The corre-
sponding slopes are 0.99 and -0.09,
respectively. Although the wind conditions
were not identical for all data points, the
14-7
-------
I as
u
s
•575
[65
o
•45
!•
CFG Settings
Water flow: 60 Iph
Voltage: 4 kV
Spray Type
Broad o
Narrow n
'Slope: 0.395
Corr.
Coeff.: 0.927
° Slope: -0.087
Corr.
Coeff.: -0.189
35 45 55 65 75 85 95
Ambient Relative Humidity. %
Figure 4. PCS of the CFG platted as a
function of ambient relative
humidity for broad to) and
narrow (at spray patterns.
fine PCE increases with increases in
ambient relative humidity for a broad
spray, but is fairly independent of RH for a
narrow spray. The generally lower values
of E for the narrow spray can be easily
explained: a board spray covers most of
the dust cloud in the hopper, while the
narrow spray covers a smaller portion,
resulting in lower E values.
The difference in the dependence of E
on RH for the broad and narrow sprays is
also explainable. For a narrow spray,
most droplets occupy a volume away from
the open side of the hopper. When fog is
applied continuously, this area becomes
more humid than outside the hopper or
near the hopper opening, assuming the
wind is not too strong. Thus, there is
minimal droplet evaporation and. con-
sequently, fairly steady PCE. However, for
a broad spray, the droplets are distributed
from the rear wall of the hopper to outside
the open side of the hopper. Thus, when
the ambient RH is high, fewer droplets
will evaporate, leaving more droplets to
collect dust particles; on the other hand,
when the ambient RH is low, more
droplets are lost due to evaporation near
the open side of the hopper and outside the
hopper.
The effect of the longer droplet lifetime
in a higher RH atmosphere increases the
PCE. These test observations are thus
consistent with the conclusion that the
droplets should be small enough to
provide high PCE, yet large enough not to
evaporate too quickly.
Figure 5 shows PCEs for negatively and
positively charged fog for the fine fraction
separately, and for both fine and coarse
particles combined, with the same water
40
SS
i;
1 30
,£*
l£
1
1 20
U
a
•«• 10
0
S Charged H fog
0 Charged (+1 Fog
•
•
.
'
K
^
•^
\
\
V////////77/77777S
\
^
V
x
O
Y/////////////777/,
m
'
!
-
Fine Particles Only Fine & Coarse Particles
Figure 5. Comparison of PCE of the CFG
for positively and negatively
charged fog. with all other
parameters nearly identical.
flow rate, the same applied voltage, and
nearly identical field conditions. The
higher value of E for negatively charged
fog suggests that inhalable bentonite
particles carry a net excess positive charge.
Figure 6 is a good example of the effect of
water flow rate in the CFG on its ability to
control particles. The PCE in this case
decreased by 42% when the water flow
rate was reduced from 60 l/h to 30 l/h,
with all other instrument settings and
field conditions nearly identical. This
effect is expected: increased water flow
rate increases the number of charged
droplets available for particle collection.
Water flow rate was particularly signifi-
cant in this experiment because the CFG
was being applied to control dust in a vol-
ume larger than its maximum coverage;
therefore, any decrease in the water
flow rate further decreased the coverage.
Under a controlled experimental setup
one would expect the PCE of charged
droplets to decrease if the wind speed in
the volume being treated were increased.
The increased wind would reduce the
time available for the droplet and particle
to interact. However, in this experimental
situation, because of the location and
method of particle sampling, this effect is
not evident.
Conclusions
After the initial setbacks with a
prototype spinning cup fog thrower, a
new charged fog generator was devel-
oped. Water droplets are produced in this
14-8
-------
as 40
1
•$
u
§30
•5
o
^ 20
1
2
-------
industrial hygienist. Control of
germs and possible decontamina-
tion of an area using proper chemi-
cal additives in the water spray is
also of interest to our defense
forces. Preliminary investigation of
this application could be coupled
with (1), above.
(4) The potential of the CFG to control
dust from a mobile source is worth
examining. No alternative method
of control is available, other than
ordinary water sprays, which are
not efficient in controlling inhalable
particles.
(5) Possible application of charged fog
in recovering airborne precious
metals should be of great interest to
the gold and other precious metal
and mineral mining industries.
14-10
-------
ESTIMATION OF AMBIENT TSP IMPACTS OF COAL
STORAGE AND HANDLING FACILITIES
Robert C. Wells
Enviroplan, Inc.
Dennis C. Doll
Enviroplan, Inc.
John Hattrup
Baltimore Gas & Electric Co.
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
ABSTRACT
This paper is directed to those concerned with evaluating the air
quality impact of fugitive particulate emissions from coal storage and
handling. Site-specific analyses were conducted for the coal handling system
planned for Baltimore Gas & Electric's Brandon Shores Power Plant in Anne
Arundel County, Maryland. Fugitive particulate emissions were considered
significant due to the presence of a designated nonattainment area for TSP
near the plant. The analyses employed dispersion modeling with a modified
version of the Industrial Source Complex (ISC) Dispersion Model. Emission
factors and control efficiencies were taken from available published
information. Available emission factors were critically evaluated and
specific factors were required to be consistent with available measurements
and site-specific conditions. Significant problems developed with the
meteorological data base using the standard U.S. EPA preprocessor. Since
these problems produce unrealistic prediction under some meteorological
conditions, approximations were employed to eliminate the unreasonable
predictions from the analysis. Site-specific analyses have aided in
defining the appropriate dust control equipment for the coal handling
system.
15-1
-------
1.0 INTRODUCTION
National Ambient Air Quality Standards (NAAQS) have been
established for Total Suspended Particulates (TSP) which define
concentrations of suspended particulates associated with health problems and
with secondary impact such as property damage. TSF measurements indicate
that many areas of the United States do not attain NAAQS for TSF. Further,
emissions from "traditional" sources of pollutants (e.g. smoke stacks) often
do not account for the observed TSF concentrations. Fugitive particulate
emissions, including both natural dust and anthropogenic sources, contribute
to measured concentrations. As such, quantification and control of fugitive
dust is an important part of achieving and maintaining the NAAQS for TSP.
This paper describes a case study of the application of available
estimates of uncontrolled fugitive emissions, control efficiency, and
atmospheric dispersion for environmental analysis of a coal-fired power
plant. Regulatory criteria are stated in terms of maximum permissible
ambient concentrations. Thus, dispersion modeling based on emission
estimates is the only suitable technique to adapt available information on
fugitive emissions to the specific situation of a proposed facility.
A wide range of control measures is available to suppress or
contain fugitive emissions resulting from material storage and handling.
Costs associated with different control measures for fugitive emissions vary
greatly. In addition, some control measures can affect efficiency and
reliability of operations. Therefore, one cannot assume that it is
automatically desirable to control fugitive emissions to the maximum degree
possible. Ambient impacts of fugitive emissions under various levels of
control are important in determining cost-effective measures of attaining
ambient air quality goals as defined by the NAAQS.
Additionally, the analyses described here for evaluation of
fugitive emissions of a new facility are also useful in identifying emission
reduction credits (ERC) for use in emissions trades. Preliminary
investigations based on dispersion modeling are useful in defining the most
cost-effective controls associated with emission trade-offs, as well as the
spatial effects of trades.«involving both fugitive and stack sources.
Current U.S. EPA policy requires that proposed emission trades which
control fugitive emissions rather than stacks demonstrate the equivalence of
emission reductions using dispersion modeling. These trades will typically
require post-approval monitoring to confirm modeling results; however,
pre-approval analyses are an important first step in demonstrating the
appropriateness of ERC.
The facility evaluated in this study is the Brandon Shores Power
Plant being constructed in Anne Arundel County by the Baltimore Gas &
Electric Company. Figure 1 details the location of the plant and air
quality in the area. The plant was originally designed to burn either oil
or coal. Plans initially called for operation with oil and the plant
received environmental permits on that basis. During construction, however,
it was determined that it would be economically advantageous to burn coal.
15-2
-------
Therefore, the plant has been treated for regulatory purposes as an existing
facility undergoing coal conversion even though the plant is currently
under construction.
The Brandon Shores Power Plant consists of tvo units each with 620
MW generating capacity. Units I and II are projected to come on-line in
1984 and 1988, respectively. The lay-out of the pover plant and on-site
coal handling system is detailed in Figure 2.
2.0 STUDY DESIGN CONSIDERATIONS
2.1 REGULATORY APPROACH
There are tvo basic approaches to regulating fugitive emissions
associated with coal handling and storage. First, specific controls can be
required as a condition of the appropriate permit. With some restrictions,
this approach can apply to federal Prevention of Significant Deterioration
(PSD) permits as veil as to construction and operating permits issued by a
state government. This is common practice in some states. Controls can be
either specified by the regulator or proposed by the operator and approved
by the regulatory agency.
Second, a facility may be limited to specific maximum ambient
impacts. These impacts are defined either as maximum total concentrations
including background (e.g., NAAQS) or as maximum incremental concentrations
due to the facility (e.g., PSD increments).
The Brandon Shores Pover Plant is subject to the latter form of
regulation. While a minimum level of control is not dictated by the
regulatory agency, the facility must account for ambient concentrations from
fugitive emissions. This approach potentially provides for the most
cost-effective controls, in that the specific control required vill be
directly related to the ambient impact of a new facility and to the existing
air quality. In practice, the analysis is complicated by limitations of
available emission factors and dispersion models.
Three ambient criteria potentially constrain fugitive emissions at
Brandon Shores. First, the NAAQS apply in any area of public access. This
is generally taken to apply at any point beyond the facility's property
line.
Second, in areas that are nov attaining the NAAQS (vhich includes
the Brandon Shores Plant site), incremental increases in TSF concentrations
due to new sources are limited. As with the NAAQS, these PSD increments
apply in all areas of public access. They include impacts from all sources
constructed after a specific date. In this case, no other sources are
projected to consume PSD increments of TSP in the vicinity of the Brandon
Shores facility; therefore, PSD increments apply to concentrations from the
Brandon Shores facility alone.
15-3
-------
The third case,- involving areas that do not attain the NAAQS, is
more complicated. The Brandon Shores Plant site is located near a
(secondary) nonattainment area for TSP. If the facility does not contribute
significantly to ambient concentrations in the nonattainment area, then the
facility is permissible. If the facility does contribute significantly to
the nonattainment area, it is then subject to two types of control. First,
the facility must employ extensive controls to achieve the lowest achievable
emission rate (LAER) for particulates. Second, the facility must provide
for emission reductions in the nonattainment area to offset the increased
emissions at the facility. While it is possible to satisfy these two
requirements in some cases, they represent a severe limitation on the siting
of new sources. It is often preferable to voluntarily control sources to
the point where they exert no significant impact in a nonattainment area.
Criteria for ambient concentrations and incremental impacts are
based on both long-term (annual) and short-term (24-hour) averages.
Specific concentrations applicable here are presented in Table 1. Criteria
for short-term NAAQS and PSD compliance limit the largest second-highest
concentration predicted based on approved dispersion modeling procedures.
The significance criteria for nonattainment area impact (de minimis
increments) are subject to interpretation depending upon the regulatory
context and the discretion of the regulatory agency.
The increments were proposed by U.S. EPA to relate to the
incremental contribution to a violation of the NAAQS. Hence they should
apply only at the time and place of an actual violation of the NAAQS. For
long-term averages this includes the entire nonattainment area by
definition. For short-term averages, prediction impacts from the proposed
facility may or may.not coincide with short-term concentrations in excess of
the NAAQS.
In this instance, the State of Maryland is free to set its own
criteria for significant impact in a nonattainment area. The State has
chosen to employ the numerical criteria suggested in federal regulations as
applicable to the overall highest concentration predicted from the proposed
facility in the designated nonattainment area. This determination ignores
the coincidence of the facility's impacts with high background
concentrations.
Projection of TSP concentrations in the Baltimore area developed
by the State of Maryland for the years 1982 and 1986 indicate that this
boundary of the nonattainment area is projected to change drastically in the
1980's (Figure 2). The location of this boundary is an important part of
the definition of significant impact. Concentrations at the boundary of
the designated nonattainment area ultimately prove to be the limiting
concentrations, which determine the level of control required. As such,
projecting air quality to the time of plant start-up would significantly
alter the conclusions of this analysis.
15-4
-------
3.0 METHODOLOGY
Under the regulatory criteria outlined in Section 2, ambient
impacts from the operation of coal storage and handling equipment at the
Brandon Shores Power Plant and resulting total ambient concentrations must
be determined. Projection of ambient impacts for proposed facilities
requires the independent estimate of emission rates from the facility and
dispersion from the source. In addition, projections of total
concentrations resulting from proposed facilities require some consideration
of other sources of a pollutant. These three issues are addressed in
separate subsections below.
3.1 EMISSION RATE ESTIMATES
Available procedures for estimating emission rates from sources of
fugitive particulate emissions require (1) selection of an emission factor
which describes uncontrolled emissions from the source, (2) identification
of a control efficiency associated with the specified controls, and (3)
specification of source extent (area or operating rate).
Factors
Published emission factors quantify emissions from all potential
sources of fugitive particulates which are of concern here. The quality of
these emission factors varies greatly, however. Quantitative estimates
available for some sources are merely assumptions. Other sources have been
subjected to empirical field studies and emission factors exist which
account for a wide range of potential conditions. In some cases, different
field studies employing different methodologies have been conducted to
quantify similar sources. It is disturbing to note that field studies
conducted with reasonable methodologies arrive at different results.
Appropriate emission factors for a particular facility must be
selected from the wide range of available estimates. These emission factors
should be based on two criteria. First, the emission factors should be
determined from a specific field study or estimation procedure of high
quality. Second, the field study or estimation procedure should be as
closely related as possible to the specific sources of concern. A third
criterion often raised in a regulatory context is conservatism within a
range of uncertainty.'
An overall quality index has been suggested in past work conducted
by U.S. EPA contractors. The proposed rating scale is presented in
Table 2. Essentially, the quality index asserts that (1) any data are
better than no data; (2) representative data are better than extrapolated
data; (3) accurate data are better than inaccurate data; and (4) large
quantities of accurate data are better than small quantities of accurate
data. The qualitative scale is intuitively reasonable, and will provide a
useful measure of the relative quality of different emission factors from
different sources. Unfortunately, selection of alternate emission factors
is often a choice between extrapolation of high quality data from different
15-5
-------
processes or the use of lower quality data for more representative
processes.
In practice, the authors know of no emission factors for fugitive
particulate emissions which deserve an A rating. Some emission factors are
available which would be rated B. Most available emission factors would be
rated C or D, and a few potentially significant sources must be quantified
using E-rated emission factors (assumptions). Generally, D-rated or E-rated
data should be disregarded if better information is available. B and C
ratings are actually a continuum. Use of any emission factor requires
"extrapolation from similar processes." The relative degrees of
extrapolation and inherent data accuracy must be balanced in selecting an
emission factor.
The regulatory concern for conservatism is often not practical
with fugitive emissions estimates. The range of emission factors for some
sources can span more than three orders of magnitude. It seems entirely
unreasonable to rely on a questionable assumption simply because it is more
conservative than the results are a measurement program.
Three sources of published information potentially provide
emission factors useful in this study. First, work has been conducted by
Midwest Research Institute (MRI) under U.S. EPA contract. Their work began
in 1924-x with additional useful information published in 1978 and
1979. ' Emission factors potentially applicable to coal have been
developed based on limited sampling of fugitive emissions from coal
handling, and based on extrapolation from more extensive sampling of
different materials.
Second, FEDCo Environmental Inc. conducted a limited field study
of fugitive particulate emissions from coal mining in the western United
States. While coal mines are not identical to power plants, the study
represents a moderately large data base specific to coal. An extrapolation
of these emission factors may be preferable to use of emission factors
developed for other materials and for a specific process.
Climatic conditions must also be accounted for in using emission
factors developed for coal mines in the arid west for operations at a power
plant in the eastern United States. The effect of climatic conditions on
fugitive particulate emissions are poorly understood. However, both
empirical studies and theoretical considerations ' suggest that climate
is important.
Third, a wide range of emission factors from a variety of sources
was summarized in 1977 under U.S. EPA contract. This summary included
earlier MRI work, as well as numerous limited field studies, and emission
factors which had been used in regulatory determinations in the past.
Generally, emission factors cited in this document which are not included in
work discussed above would be rated D or E. The compilation provides a
useful summary of emission factors which may be considered based on
conservatism of regulatory precedent. However, field studies subsequent to
15-6
-------
its publication generally have found lover emission rates than those
previously employed in regulatory determinations.
The uncertainty noted in estimating fugitive particulate emission
factors, combined with regulatory concerns of precedent and conservatism,
precluded the selection of a single "correct" set of emission factors.
First, a set of emission factors was formulated based on adherence to recent
regulatory precedent. It was intended that these emission factors provide
an upper bound on ambient impacts associated with coal handling equipment at
the Brandon Shores Power Plant. A determination that the plant would be
environmentally acceptable based on these emission factors would provide
strong support for issuance of permits.
A second set of emission factors was formulated based on what
seemed to be the most representative factors available. An analysis based
on these emission factors provides the best assessment of ambient impacts
from fugitive particulate emissions. The substantial margin of conservatism
typical of the former set of factors is not the case here, however.
Emission factors judged to be best are frequently at the low end of the
range of factors quoted.
At the Brandon Shores Power Plant there are three important
processes with the potential to produce fugitive particulate emissions.
First, continuous transfer equipment removes coal from barges, transfers
coal between conveyors, places coal in and out of storage, and supplies
bunkers. Second, wind erosion from storage piles on-site is likely to
produce emissions. Third, coal must be crushed prior to combustion. In
addition, some emissions would be expected from conveyors.
Conservative emission factors for continuous handling equipment
were based on a number of emission factors that had been used in a
regulatory context. The emission factors for barge unloading and transfer
houses are based on commonly used factors with unclear origins. '
Emission factors for load-in and loadr-out of storage piles were taken from
MRI's earlier work with stone quarries. The available emission factor
judged the best to quantify emissions from continuous handling operations
was developed for fugitive emissions from the iron and steel industry.
The emission factor is based on similar activities with different materials.
Site-specific correction factors are suggested to minimize the uncertainty
associated with extrapolation.
A conservative estimate of wind erosion from coal storage piles
was taken from emission factors developed for stone quarries. While the
factor is not particularly applicable to coal, it has frequently been used
in a regulatory context. The wind erosion emission factor developed for
coal mines was judged to be more appropriate for an accurate emission
estimate. This factor is based on data directly relevant to coal. The data
were collected in an arid region and corrections for climatic differences
and storage pile activity were assumed based on past work. The emission
factor is stated in terms of a functional relationship with wind speed,
which is.intuitively reasonable though different from theoretical functional
forms. ' This functional relationship is particularly important in
15-7
-------
avoiding severe overpredictions of ambient impacts from coal storage and
handling under low wind speed conditions.
Conservative emission factors for.coal crushing were based on U.S.
EPA emission factors for stone crushing. Alternative emission^xfactors
have been proposed based on limited data specific to coal. While
neither factor could be considered ideal, the choice of the more
source-specific factor seems preferable to the assumed applicability of
stone crushing emission factors based on regulatory precedent.
An available estimate for emissions from conveyors was based on
the assumption that 251 of the emissions from "transfer and conyeving" or
332 of the emissions from transfers would be due to conveyors. This
estimate is based on visual observations and is employed here both for
conservative and best estimates.
Emission factors employed in this study are presented in Table 3.
Comparison of emission factors is difficult because of different assumptions
for the two sets of factors. The conservative emission factors often lump
sources together (e.g., all transfers and conveyors), whereas the best
emission factors are specific to single sources (e.g., single transfer
points). Conservative wind erosion estimates are based on material
throughout; best estimates are based on surface area of storage piles and on
the wind speed. Table 4 summarizes the alternative emission factors in
terms of annual mass emissions. This summary illustrates the assumptions
inherent in selection of one factor over another.
Control Estimates
The uncertainties noted above for estimates of uncontrolled
fugitive particulate emission rates are compounded in estimating control
efficiencies. With the exception of baghouses, where control efficiencies
can be readily demonstrated, available control techniques for fugitive
particulate emissions are not quantified at all. Available.estimates of
control efficiency are based entirely on subjective estimates. While use
of these subjective estimates is not desirable, there is little choice. The
added confusion associated with developing alternative "conservative" and
"accurate" estimates of control efficiencies is not warranted. Neither
conservatism nor accuracy can be demonstrated, and extensive regulatory
precedent does not exist with respect to assumed control efficiencies.
It is important to specify control which will mitigate
unacceptable environmental impacts without excessive costs or operational
limitation on the facility. An initial set of controls was suggested based
on engineering judgement. Selective controls were added as potential
problems were identified in dispersion modeling. Table 5 summarizes
controls considered for the Brandon Shores Power Plant, together with
estimates of control efficiencies employed in dispersion modeling.
15-8
-------
3.2 DISPERSION MODELING
In order to accurately represent dispersion of particulate
emissions from coal handling, a dispersion model should be able to represent
(1) isolated area sources, (2) local turbulence associated with buildings
and obstructions, (3) dispersed emissions associated with enclosures and
conveyors, and (4) settling and deposition associated with the large
particles typically emitted as fugitive dust. Regulatory considerations
dictate that the dispersion model also correspond to established D.S. EPA
guidelines on dispersion modeling for environmental permitting.
Based on these criteria, the U.S. EPA's ISC Model appears to be a
clear choice. The model is flexible enough to account for a variety of
source geometries (points, lines, volumes, areas) and specifically accounts
for settling and deposition associated with particulates. The ISC Model
includes features to account for dispersion due to the mechanical turbulence
(downwash) in the wakes of buildings and similar large structures. The
downwash algorithm has been extended here to include dispersion associated
with area sources which are, in themselves, obstructions to air flow.
The flexibility in specifying source geometry available in ISC was
fully utilized in this application. Storage piles were represented as area
sources. Sources controlled by baghouses were represented as point sources
(without buoyant plume rise) which are subject to building downwash.
Sources with controls which do not duct emissions to a specific point were
modeled as volume sources, where emissions are assumed to be emitted over
the effective cross section of the source.
The particle data needed to determine settling and deposition were
based OB,sphysical diameters determined from microscopic analysis of hi-vol
filters. This corresponds to the implicit definition of TSP as particles
subject to hi-vol capture. The continuous particle size distribution
observed was divided into three categories: (1) aerosols, (2) slightly
settling particles, and (3) rapidly settling particles based on the
assumption of stokes settling. Table 6 details the particle size ranges and
particle characteristics employed to represent each group.
Receptor Selection
Dispersion modeling* requires identification of locations where
concentrations are expected to be maximized. Ground level sources will show
maximum ground level concentrations near the source. Elevated releases may
show maximum ground level concentrations further downwind. Preliminary
modeling of the significant sources of fugitive particulates at the Brandon
Shores Power Plant yielded the intuitively reasonable result that the
release heights of fugitive emissions at Brandon Shores were low enough so
that maximum ambient concentrations would be expected at the property line.
This corresponds to the nearest source-receptor distance where the NAAQS
would apply. Thirty-four receptors were therefore located along the
property line. Eight additional receptors were located near the barge
unloader at the minimum source-receptor distance recommended for accurate
15-9
-------
prediction modeling with ISC (100 meters). A total of 42 receptors were
located to quantify overall maximum impacts.
Maximum ambient impacts in the designated nonattaimnent area to
the vest and north of the plant are anticipated at the nonattainment area
boundary. Thirteen receptors were thus located on the boundary. Six
receptors were also located at nearby monitoring locations identified in
Figure 1. A total of 63 receptors were employed in this analysis.
Meteorological Data
Dispersion modeling employed five years of meteorological data
(1964-1968). Surface observations from the Baltimore-Washington
International Airport were combined with upper air data from Dulles Airport
using the standard U.S. EPA meteorological preprocessor.
Two important anomalies were noted which required corrections.
First, extremely low interpolated mixing depths were eliminated because they
are not reasonable and because these unreasonable mixing heights affect
predictions for low-level releases. Mixing depths from the U.S. EPA
meteorological preprocessor were uniformly truncated to a minimum of 75
meters. Second, periods of calm winds recorded in airport data presented
both a theoretical problem and a practical problem. In theory, a
steady-state Gaussian Dispersion Model does not apply where an established
mean flow (wind) does not exist. This problem is typically dealt with by
assuming a minimum speed of 1.0 meter per second.
The practical problem is created by the fact that typical airport
meteorological instruments cannot measure wind directions accurately at low
wind speeds associated with calm conditions. Wind directions are typically
assumed to persist over the period of observed calms. This creates the
unrealistic situation of multiple hours of persistent wind direction
accompanied by low wind speeds.
In any case where prediction of maximum concentrations resulted
from this unrealistic representation of calm wind conditions, predictions
were reevaluated without the periods of calm winds. The deletion of calm
winds provides an. important perspective on the limitations of the current
state of the art of dispersion modeling.
3.3 BACKGROUND CONCENTRATIONS
Estimating total ambient concentrations associated with a given
source requires the inclusion of background concentrations from known and
unknown sources which also emit the same pollutants. In the case of
particulates, it is impossible to include all other potential sources in
dispersion modeling. Ambient concentrations may be dominated by an
industrial facility, a highway, a farm, or a wide range of sources. In this
situation it is often expedient to make the conservative assumption that
high measured background concentrations— which do not include the proposed
facility —added to high predicted concentrations from the proposed facility
represent likely maximum total impacts.
15-10
-------
Background concentrations representative of the Brandon Shores
Plant site were taken from the nearest continuous monitoring station, which
is located in the attainment area identified in Figure 1. Measurement data
from monitoring stations located in nonattainment areas would obviously show
violations of the NAAQS which are assumed not to ezist based on the
designated attainment status of the Brandon Shores Plant site. In this
study, monitoring data for the calendar year 1979 was employed.
Differences in background particulate concentrations should be
expected as a function of meteorology. In the case of a pollutant with
continuous monitoring data available (e.g., SO.), background concentrations
can be related to specific meteorological conditions. In the case of TSP,
which is measured as a daily 24-hour average, this relationship is difficult
to define.
The most obvious relationship expected is high ambient
concentration associated with a limited range of wind directions. Days of
high measured concentrations were examined for persistent meteorological
conditions which identify background concentrations as a function of wind
direction.
The highest measured concentration (114 ug/m ) was associated with
persistent winds from the south, therefore, this concentration was taken as
representative background for receptors to the north of the Brandon Shores
Plant site. The second-highest measured concentration (106 ug/m ) could not
be characterized with respect to wind direction. The second-highest
measured concentration was thus used to represent background concentrations
for all other receptors, as none could be specifically excluded based on
meteorological considerations.
4.0 RESULTS
Results of dispersion modeling are evaluated in terms of three
regulatory criteria identified above. Prediction modeling generally
proceeded from conservative assumptions, which would have unambiguously
supported permitting, to better assumptions, which indicate best estimates
of projected impacts (but with less conservatism).
Conservative emission factors combined with initial control
specifications indicated the potential for exceedence of all three
regulatory criteria noted above. While annual average impacts, were
generally acceptable, appropriate 24-hour impacts were 230 ug/m , 124
ug/m , and 46 ug/m for the NAAQS, PSD increments, and nonattainment
impact, respectively. Examination of source contributions to the predicted
concentrations indicated that two sources (inactive storage pile and main
yard conveyor) were primarily responsible for the high predicted
concentrations. At this point, it was elected to identify more effective
controls for these sources based on prediction results.
Additional controls (Phase Two in Table 5) were applied. These
controls suggested compliance with the secondary NAAQS; however, applicable
PSD increments at the plant boundary and de minimis thresholds for the
15-11
-------
nonattainment area were both exceeded. Relevant maximum concentrations
over five years were predicted to be 43 ug/m and 16 ug/m for PSD
increments and nonattaimnent impact, respectively. It was noted that calm
wind conditions contributed significantly to the predicted exceedence of
the PSD increment. The predicted PSD increment consumption without calm
winds was 40 ug/m . Analysis of source contributions to predicted
concentrations indicated that the active storage pile area and main yard
belt were primarily responsible for predicted high concentrations.
Additional controls were specified for the main yard conveyor
(Phase Three of Table- 5). Predicted concentrations with additional
controls were 40 ug/m and 13 ug/m for PSD and nonattainment impact,
respectively.
At this point, it was clear that the conservative assumptions
implicit in emission factors employed thus far would not support the
permitting of a coal handling facility with conventional fugitive
particulate controls. The identical analysis was then conducted using the
best estimates of fugitive emission rates identified above.
The control efficiencies for some of the sources were revised for
this analysis. These revisions were necessary because, in a few cases, the
emission factor itself implicitly accounts for some control. Also,
additional information regarding the controls considered for the Brandon
Shores Plant became available which necessitated some revision. These
revisions are identified in Phase Four of Table 5.
The predicted largest second-highest 24-hour average
concentrations at the property line were now 2 ug/m . The maximum first
highest 24-hour average concentration predicted at the boundary of the
2
nonattainment area was 2 ug/m . Both of these concentrations are
significantly below the appropriate regulatory criteria. Addition of
appropriate background concentrations suggests that ambient air quality
standards will also be maintained.
Prediction modeling based on conservative emission factors would
have strongly supported environmental permitting. Prediction modeling
based on the best estimates does not present as convincing a demonstration
of the acceptability of environmental impacts. The analysis does, however,
suggest that requirements of extremely stringent controls (e.g. , silos for
active storage) are not consistent with the best available information
concerning fugitive dust emissions.
15-12
-------
5.0 DISCUSSION
5.1 RESOLUTION OF PERMITTING FOR THE BRANDON SHORES PLANT
The analysis based on best emission estimates described above
demonstrates that the environmental impact from fugitive particulate
emissions from coal handling at the Brandon Shores Plant is likely to be
veil below established regulatory criteria for ambient impacts and low
enough to maintain the NAAQS. Analysis based on conservative emission
estimates, which are more consistent with regulatory precedent than with
scientific evidence, suggests the possibility of significant short term
impacts under some conditions. Unfortunately, the conditions associated
with maximum impacts are also conditions not treated well by either the
conservative emission factors or Gaussian dispersion modeling.
While the possibility of ezceedence of applicable regulatory
criteria exists, it would seem unreasonable to require extensive (and
expensive) controls on the basis of this analysis. The more prudent
approach would be to issue environmental permits based on best estimates of
environmental impacts. A monitoring program could be established to
confirm that environmental impacts are, indeed, acceptable. If problems
were identified due to fugitive emissions from coal handling, mitigative
measures (e.g. modification of management practices, retrofitting of
control equipment, obtaining emission offsets) could be implemented. While
there is a risk that more expensive controls would be ultimately required
if problems were identified, this risk is more than compensated by the
likelihood that environmental impacts will be acceptable.
This course of action was, in fact, pursued at the Brandon Shores
Power Plant. The facility received environmental permits with the
stipulation that TSP monitors be sited to identify potential maximum
impacts from the coal handling facilities.
5.2 LIMITATIONS OF THE METHODOLOGY EMPLOYED
The methodology discussed above did not conclusively predict the
environmental impact of fugitive emissions from coal handling. All
predictive analyses will be similarly limited by the quality of available
emission data and dispersion models.
The requirement of specific controls is a viable alternative.
The problem with this approach is that the possibility of minimal impact
with little control is not explored and the cost effectiveness of
alternative controls for specific problems is not considered.
The regulation of coal handling based on ambient criteria for TSP
is difficult. Analytical tools for the analysis must be improved for
reasonable and timely permitting based on predicted impacts. Several areas
can be readily identified where available analytical tools need
improvement. These can be generally classified as problems of estimating
emission factors, of control effectiveness, and of dispersion modeling.
15-13
-------
Emission Factors
Emission factors specifically for coal handling and storage need
to be improved. Demonstrated measurement techniques are probably adequate
but the amount of data is not. Material handling emission factors should
be based on the full range of typical operating parameters (e.g., drop
height). Parameters related to surface moisture and silt content need to
be extensively measured in varying climates and/or related to readily
available parameters. Some potentially significant sources (e.g., train
unloading, crushing) need much more study.
Emission factors for vind erosion need to be related to climate,
pile configurations and cumulative erosion. Furthermore, currently
available emission factors, which are linearly dependent on wind speed, do
not agree with either theoretical relationships or empirical data from wind
tunnel testing. At a minimum, more extensive data to quantify the range of
variation expected are needed.
Control Effectiveness
Quantification of control effectiveness is extremely important
and urgently needed. With the possible exception of baghouses, no controls
are quantified. Even baghouses may not be accurately characterized by a
simple percentage reduction. Control effectiveness needs to be specified
over the range of options available for generic types of control. For
example, spray systems should be distinguished on the basis of droplet size
(fog vs. spray) and wetting agents.
Dispersion Modeling
Dispersion models employed in regulation development have not
been extensively validated for impacts from low-level, non-buoyant releases
of particulates near flow obstructions. The area source algorithm
(critical to this analysis) has not been validated at all.
Several problems with conventional dispersion modeling are
magnified by the situation here. Plume dimensions based on 3-minute
average data produce extremes (high and low) of impact at short
source-receptor distances when assumed representative of one-hour averages.
Near field dispersion from mechanical turbulence near ground level (typical
of industrial facilities) is completely neglected except in the context of
building downwash. Treatment of extremely low wind speed conditions is
neither physically reasonable nor empirically accurate. These problems are
of lesser importance for stack-type emissions than for fugitive
particulates. It is not surprising that new applications of dispersion
models lead to new problems in their application.
In summary, the current state of the art of fugitive dust
analysis has provided an adequate quantitative description of fugitive
emmisions and resulting ambient impacts from a proposed coal handling
tacility. An informed regultory decision which accounted for inevitable
uncertainties resulted. Much productive work remains, however, which can
narrow the range of uncertainty and expedite the consideration of fugitive
dust in environmental permitting. s^^vc
15-14
-------
CJ1
I—1
01
Designated TSP Nonattainment Area
Projected 1982 Nonattainment Area
———— Projected 1986 Nonattainment Area
A TSP Monitor Site
}|C Brandon Shores Plant
Figure 1: Location of the Brandon Shores Plant and TSP Nonattainment Area
-------
01
i
^
PROPERTY LINE
DUST SILO.
•
•
CRUSHER
BUILDING
INACTIVE
STOCKPILE
COAL BUNKER UNIT +2
COAL BUNKER UNIT + 1
CONVEYOR
INACTIVE STOCKPILE
ACTIVE STOCKPILE
STACKER/RECLAIMER
TRANSFER HOUSE +2
TRANSFER
HOUSE +1
"""«««„
Figure 2: Coal Handling System at the Brandon Shores Plant
-------
TABLE 1: REGULATORY CRITERIA FOR EVALUATING
AMBIENT IKS ACT
Criterion Critical Concentration (ug/m )
24-Hour Basis Annual Basis
Clean Areas
NAAQS (Secondary) 150* 60
PSD Increment 37* 19
Nonattainment Areas
Impact Threshold 5** 1
Second-Highest
** Maximum First-Highest
TABLE 2; CRITERIA FOR FUGITIVE PARTICULATE EMISSION FACTORS
Quality Index Criteria
A Based on a statistically representative number of
accurate field measurements of a specific process
vhich span expected ranges of parameters
B Based on a limited number of accurate field
measurements of a specific process
C Based on a limited number of field measurements of
a specific process with undetermined accuracy — or
on — extrapolation of B-rated data from a similar
process
D Estimate based on professional judgment
E Assumption
15-17
-------
TABLE 3; SUMMARY OF EMISSION FACTORS FOR COAL HANKLING
Source
AND ASH HANDLING EMISSIONS
Conservative
Factor
Ub/ton)
iader 0.4
0.15(2)
0.05(2)
0.2
0.15
.rage .028(4)
Estimated
Quality
Index
D
D
E
E
E
D
Best
Factor
Ub/ton;
2.56 x 10-3(1)
1.31. x 10-3(1)
4.32 x 10"4(3)
0.06
3.60 x 10-3(1)
2.18 x 10"4(1)
Estimated
Qual ity
Index
C
B
E
D
B
B
Transfers
Conveyors
Crusher
Bunkers
Active Sti
Pile Load-In
Active Storage .035
Pile Reclaim
Active Storage .320
Pile Wind
Erosion
Inactive Storage .320
Pile Wind
Erosion
Fly Ash Unloading 0.30
Fly Ash Storage 0.30
(5)
(6)
(6)
(8)
(8)
D
D
2.18 x 10~4(1)
.091u(7)
(Ib/acre-hr)
.091u(7)
(Ib/acre-hr)
5.08 x 10
.288(1)
-4(2)
B
,(9)
Notes to Table 3:
(1) Emission factor determined from;
(0.0018(5/5)(D/5)(H/10))/(M/2)2
15-18
-------
TABLE 3; SUMMARY OF EMISSION FACTORS FOR COAL HANDLING
AND ASH HANDLING EMISSIONS
(Continued)
vbere:
S • Silt content of material in percent (varies)
U » Average wind speed (9.5 mph)
h - Drop height in feet (varies)
M - Surface moisture content in percent (varies)
(2) Based on 0.2 Ib/ton for all transfers and conveyors with the assumption
that 25Z of the total emissions are due to conveyors.
(3) Assumed to be 33Z of transfer emissions detemined vith the emission
factor equation in Note 1. This corresponds to 25Z of total, emissions
from transfer and conveying, as assumed for conservative emission
factors.
(4) Emission factor determined from:
0.04/(PE/100)2 (Ib/ton)
where
PE * Thornthwaite's precipitation - evaporation index (118 for
Baltimore area)
(5) Emission factor determined from:
0.05/(PE/100)2 (Ib/ton)
(6) Emission factor determined from:
(0.11/(PE/100)2) (D/90) (Ib/ton)
vhere D - duration of material in storage (365 days)
(7) Emission factor determined from:
A (1.6u) (repEDCo/PEBALTIMORE)2 db/acre-hr)
vhere:
A « Activity factor • .33
U - Hourly mean wind speed (m/s)
PEPEDCo * Mean PE index for PEDCo field studies - 49
PEBALTIMORE " Mean PE index for Baltimore area - 118
(8) Personal communication with Mr. William Wagner of U.S. EPA Region IV.
(9) Substantial extrapolation from another material involved.
15-19
-------
TABLE 4: ESTIMATES OF ANNUAL EMISSIONS WITH ALTERNATIVE EMISSION
FACTORS BASED ON FINAL CONTROL
Source Conservative Estimate Best Estimate
tons/year tons/year
Barge Unloader 6.0 0.2
Transfers 1.1 0.3
Conveyors 6.4 0.2
Crusher 1.5 0.4
Bunkers 0.9 <.l
Active Storage Pile 5.4 <.l
Load-in
Active Storage File 5.4 <.l
Reclaim
Active Storage Pile 1.0 0.7
Wind Erosion
Inactive Storage Pile 4.3 1.2
Wind Erosion
Fly Ash Unloading 1.5 <.l
Fly Ash Storage 0.2 .2
Total 33.7 3.6
Notes to Table 4:
(1) Based on an annual average wind speed of 9.5 miles per hour.
15-20
-------
TABLE 5; SPECIFICATION OF CONTROLS FOR THE
BRANDON SHORES PLANT
Source
Phase One
Barge Unloader
Conveyor (Barge
Unloader to
Transfer House #1)
Conveyor (Transfer
House £l to Transfer
House #2)
Transfer Houses fl
and #2 and Crusher
Building
Main Yard Conveyor
Active Storage File
Load-in
Active Storage File
Reclaim
Conveyor (Crusher to
Unit No.l Bunker
House)
Coal Bunkers
Fly Ash Storage
Fly Ash Handling
Control Method
Enclosed with baghouse
dust collection system
and vater sprays
Enclosed conveyor and
gallery
Hooded conveyor with
wind break return
Totally enclosed with
baghouse dust collection
Uncontrolled
Stacker spray system
Reclaimer spray system
Enclosed
Enclosed with baghouse
dust collection
Enclosed in silo with
baghouse dust collection
Wet suppression from
enclosed rotary unloader
Efficiency (Z)
99
99
90
99
75
80
70
99
99
95
15-21
-------
Inactive and Active
Storage Pile Wind
Erosion
TABLE 5: SPECIFICATION OF CONTROLS FOR THE
BRANDON SHORES PLANT
(Continued)
Uncontrolled
Phase Tvo
Inactive Storage Pile
Wind Erosion
Main Yard Conveyor
Crusting Agent
Water spray
95
50
Phase Three
Main Yard Conveyor
Foam suppressant
75
Phase Four
Barge Unloader
Conveyors (except
main yard conveyor
and the conveyor
from Transfer House
#1 to Transfer House
#2)
Main Yard Conveyor
Transfer Houses,
Coal Bunkers and
Crusher
Active Storage Pile
Load-in
Enclosure with vetting
agents
Enclosed conveyors and
galleries
Foaming agent, vind
shields and hooded
conveyor/wind break return
Enclosed with baghouse
dust collection
Stacker spray system
(1)
95
99.5
90/87.5
99.5
(2)
67
(3)
15-22
-------
TABLE 5: SPECIFICATION OF CONTROLS FOR THE
BRANDON SHORES PLANT
(Continued)
Active Storage Pile Reclaimer spray system 67
Reclaim
Fly Ash Storage Enclosed in silo with 99.5
vith baghouse dust
collection
(4)
Fly Ash Unloading Vet suppression from 90
enclosed rotary unloader
Active Storage Pile Wind activated spray system 90
Wind Erosion vith vetting agents
Notes to Table 5;
(1) Based on the average control for the two transfers, conveyor and the
scoop which are part of the barge unloader.
(2) First figure refers to load-in. Second figure refers to reclaim.
Differences arise because different fractions of the conveying system
are enclosed during each operation.
(3) These figures are less than those used for the conservative emission
estimates (75Z load-in, 802 reclaim) because part of the control
mechanism is explicitly accounted for in the "uncontrolled emissions"
determined from the best emission estimates.
(4) Control is for enclosure only. Control from uniform moisture addition
is accounted for in estimating "uncontrolled emissions".
TABLE 6: REPRESENTATIVE PARTICLE PROPERTIES USED
IN DISPERSION MODELING
Representative
Mass Particle Reflection Settling
Fraction (Z) Diameter Coefficient Velocity (m/sec)
16 7.07 urn .97 2.0075 x 10~3
54 19.31 urn .7375 1.498 x 10"2
30 40.33 urn .5812 6.533 x 10~2
15-23
-------
REFERENCES
(1) U.S. EPA, Emission Trading Policy Statement, 47 FR 15076, April 7,
1982.
(2) Maryland State Department of Health and Mental Hygiene, Air Quality
Program. Plan for Implementation of the National Ambient Air
Quality Standards for Total Suspended Particulate Matter.
Photochemical Oxidants and Carbon Monoxide for the Metropolitan
Baltimore Intrastate Air Quality Control Region. December 15, 1978.
(3) Bohn, R., Cusci.no, T. and Cowherd, C. "Fugitive Emissions from
Integrated Iron and Steel Plants," U.S. EPA Contract #68-02-2120
(Midwest Research Institute), Research Triangle Park, N.C., March
1978, EPA-600/2-78-505.
(4) Cowherd C., Jr. et al. Development of Mission Factors for Fugitive
Dust Sources. U.S. EPA, OAQPS, EPA-450/3-74-037, June 1974.
(5) Cowherd, C., Jr. et al. Iron and Steel Plant Open Source Fugitive
Emission Evaluation. U.S. EPA, IERL, EPA-600/2-79-103, May 1979.
(6) PEDCo-Environmental Inc. Survey of Fugitive Dust from Coal Mines.
U.S. EPA, Region VIII, Office of Energy Activities,
EPA-908/1-78-003, February 1978.
(7) Chepil, W.S. "Influence of Moisture on Erodability of Soil by Wind,"
Soil Science Society Proceedings, pp. 289-292, 1956.
(8) PEDCo-Environmental, Inc. Technical Guidance for Control of
Industrial Process Fugitive Particulate Emissions. U.S. EPA Contract
#68-02-1375, Research Triangle Park, N.C. March 1977,
EPA-450/3-77-010.
(9) United Nations. Air Pollution bv Coking Plants. United Nations
Report SECE/COAL/26, 1968.
(10) Johns Hopkins University, Applied Physics Laboratory. Fugitive Dust
Emissions from the Proposed Vienna Unit No. 9. Prepared for the
Maryland Power Plant Siting Program, JHU/PPSE 8-14, February 1981.
(11) Gillette, D. "Tests with a Portable Wind Tunnel for Determining Wind
Erosion Threshold Velocities," Atmospheric Environment. Vol. 12, No.
12, pp. 2309-2313, December 1978.
(12) Chepil, W.S. "Dynamics of Wind Erosion: II. Initiation of Soil
Movement." Soil Science. Vol. 60, No. 5, P.397-411, 1945.
15-24
-------
(13) U.S. EPA. Compilation of Air Pollutant Emission Factors. Third
Edition (including Supplements 1-10), U.S. EFA, OAQPS, February
1980.
(14) Jutze, G.A. et al. Evaluation of Fugitive Dust Emissions from
Mining. U.S. EPA, IERL, EPA-600/9-76-001 (Contract Nos. 68-02-1321,
Task No. 36), June 1976.
(15) W.P. Reefe. Emission Factors For Mining Operations. Colorado
Department of Health, Air Pollution Control Division, Denver,
Colorado , Unpublished, March 1978, Cited by Currier, E.L., and B.D.
Neal. "Fugitive Emission from Coal-Fired Power Plants," Air
Pollution Control Association, Proceedings of the 72nd Annual
Meeting (Cincinnati, Ohio, 1979), Publication 79-11.4.
(16) U.S. EPA. Regional Workshops on Air Quality Modeling; A Summary
Report. U.S. EPA, OAQPS, April 1981.
(17) Bowers, J.F., et al. Industrial Source Complex (ISC) Dispersion
Model User"s Guide. U.S. EPA, Source Receptors Analysis Branch,
EPA-450/4-79-030, December 1979.
15-25
-------
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
INHALATION PATHWAY RISK ASSESSMENT
OF HAZARDOUS WASTE INCINERATION FACILITIES*
GREGORY A. HOLTON**
THE MAXIMA CORPORATION
OAK RIDGE, TENNESSEE
CURTIS C. TRAVIS
ELIZABETH L. ETNIER
FRANCIS R. O'DONNELL
OAK RIDGE NATIONAL LABORATORY
OAK RIDGE, TENNESSEE
(*) Although this is not the actual presentation made at the May 1982
meeting, it closely resembles that presentation, which was
entitled. "A Determination of the Impact of Fugitive VOC Emis-
sions from a Municipal Hazardous Waste Incinerator on the Sur-
rounding Community, " G. Holton (Oak Ridge National Laboratory)
and L.Staley (EPA/IERL-Cin).
(**) Current address: First Environment. 314 W. Broadway. Lenoir
City. TN 37771.
16-1
-------
I. INTRODUCTION
One objective of quantitative risk assessment is to arrive at a value or, more
realistically, a range of values that describe the possible adverse effects on human health
associated with low, chronic exposure to a known or suspected toxic substance. There is
increasing impetus to use risk assessment to prioritize research needs in health and
environmental areas. Quantitative risk assessment typically consists of two components,
the measurement or estimation of the concentrations of pollutants to which a population
will be exposed and an estimation of resultant health effects based on available dose-
response relationships. These two components are called exposure assessment and health
effects assessment, respectively.
Exposure assessment is defined as the determination of the concentration of toxic
materials in space and time at the interface of target populations. This description
should include identification of all major pathways (air, water, and soil) for movement
and transformation of a toxic material in a selected environmental setting. The present
assessment focuses only on atmospheric concentrations and exposure via the inhalation
pathway.
Health effects assessment consists of relating population exposure to such potential
human health impacts as carcinpgenicity, mutogenicity, teratogenicity, or
neurotoxicity. Unfortunately, adequate data for assessing excess human risk resulting
from exposure to most lexicological substances do not exist.
The U. S. Environmental Protection Agency is conducting, research to determine air
concentrations, public inhalation exposure, and health risk resulting from chemicals
emitted during incineration of hazardous wastes. Incineration facilities produce stock
and fugitive (non-stock) atmospheric emissions. The magnitude of public exposure to
such emissions is dependent on both plant design and operation as well as the
physicochemical properties of the waste. Engineering variables include facility design,
size, destruction and removal efficiency (ORE), and emission source strength (stack vs
fugitive). Physiocochemical waste variables include incinerability, volatility, and
toxicity. The present assessment was performed to determine the relative importance of
these variables to human inhalation exposure and health risk.
It should be emphasized that both the exposure assessment and health effects assessment
methodologies presented in this paper are very generalized, and caution should be
exercised in interpreting the results. Site-specific application of our results would
require careful evaluation of the extent to which the models and parameter values used
in this paper are representative of conditions prevailing at the specific site.
2. INCINERATION FACILITY DESIGN
In the United States, there are at least 219 facilities operating 284 hazardous waste
incinerators of varying design and capacity. The liquid injection (LI) design is the most
common type of hazardous waste incinerator (51%) employed in the United States.
Liquid injection incinerators can be applied to virtually all liquid wastes (liquid, slurries,
sludges). Liquid injection systems employ a waste nozzle (burner) which atomizes the
16-2
-------
waste and mixes it with air to form a suspension that promotes complete combustion.
Atomization is achieved mechanically or by pressure atomization systems using high-
pressure air or steam.
The rotary kiln (KK) design is employed at 7% of the incineration facilities in the United
States.' A rotary kiln is a cylindrical refractory-lined shell mounted at a slight incline
from the horizontal plane. Rotation of the shell provides for movement of the waste
through the kiln as wejl as for enhanced mixing. Rotary kilns can be applied to both
liquid and solid wastes.
Three incinerator sizes (I, 10, ISO x 10^ Btu/h) were chosen for analysis to provide a
profile of existing incinerator capacities. Two facility designs (liquid injection and
rotary kiln) for each of the three sizes were selected.
Facility design may be important in determining fugitive emmisions, since such emissions
are dependent on the number of pumps, tanks, valves and flanges in operation. Each
incineration facility was designed following a review of existing-practice incinerators.
For example, the hypothetical ISO x 10° Btu/h incinerator used in this study was assumed
to possess one receiving tank for each of four categories of waste: clean or dirty high-
Bru waste and clean or dirty low-Btu waste (dirty waste is any waste that requires
pretreatment to enhance viscosity). Two additional tanks were included to provide extra
storage capacity for irregular shipments. The storage area of the facility was designed
with sufficient capacity for 14 days continuous operation. The feed area was assumed to
have sufficient storage for three days of operation. Other design factors such as piping
or numbers of pumps were also considered.
A compilation of important facility design parameters is presented in Table I. Equipment
quantities and operating assumptions for the receiving area are identical for both
incinerator types, but vary slightly with size. The number of storage tanks in the storage
and feed areas remains the same for all sizes of incinerators, but the holding capacity of
the tanks increases. The number of pumps, valves, and flanges are identical in the
storage and feed areas of the 10 and ISO x 10° Btu/h incinerators. Since these areas are
the major contributors to fugitive emissions, one should not expect a large change in
fugitive emissions as a function of incinerator size.
3. WASTE CHARACTERIZATION AND EMISSION RATES
Operation of a commercial incinerator is characterized by receipt of waste of widely
varying composition. Except in settings where an incinerator is dedicated to a particular
chemical process waste stream, a detailed qualitative or quantitative makeup of the
waste being burned is usually unknown. The EPA has funded several surveys to determine
the composition of hazardous waste streams currently being incinerated. A total of 237
different constituents have been identified as present in one or more of the 413
hazardous waste streams reviewed. Table II lists the ten most prevalent constituents of
hazardous waste streams currently being incinerated, according to the most recent
survey.3 Unfortunately, data from this survey were not available when this assessment
was performed.
To deal with the uncertainties inherent in the characterization of incineration waste
streams, it was decided to examine the behavior of representative chemicals from three
generic waste classes. To choose these waste classes, data from an earlier survey of
industrial wastes4 were analyzed and grouped by incinerability characteristics.
Incinerability in this report is defined as the inherent heating value (Btu/lb) of the waste
16-3
-------
before fuel oil is added. The three waste groups selected for study, in order of increasing
incinerability were: (I) pesticide-related chemicals (3,021 Btu/lb); (2) API separator
sludge chemicals (4,049 Btu/lb); and (3) phenol/acetone distillation chemicals (15,850
Btu/lb). Four chemicals from each waste group were selected for analysis. Criteria for
choosing specific chemicals were high and low volatility, high and low toxicity, and
prevalence of the chemical in U. S. industrial waste streams. The chemicals chosen for
the three waste groups were: (I) chloroform, ethylene bichloride, hexachlorabutadiene,
and 1,1,2,2-tetrachloroethane; (2) chromium, lead, arsenic, and phenol; and (3) toluene,
pyridine, phthalic anhydride, and methyl styrene.
The rate of release (mass per unit time) of specific chemicals in stack emissions is
controlled by three facility variables: waste throughput, chemical concentration in the
waste stream, and ORE. Waste throughput in an incineration facility is determined by
the percent contribution of the waste to the total waste stream after supplementary
addition of No. 2 fuel oil to insure combustibility. If the waste has a high enough Btu
content (10,000 Btu/lb) to bum without the addition of supplementary fuel (as is the case
with phenol/acetone distillation waste), no additional fuel oil was assumed.
To determine the effect of ORE on human exposure and risk estimates, three scenarios
were chosen. These scenarios were 99.99, 99.9, and 99.0%. An additional analysis of
95.0% ORE for arsenic, chromium, and lead was performed. Predicted annual
incineration rates (MT/y), and annual stock emission rates (g/y) for the 99.99% ORE
scenario are given in Table III.
Fugitive emissions consist of tank, pump, in-line valve, open-ended valve, flange,
instrument connection, and sampling connection emissions from the receiving, storage,
and feed areas of the incinerator. Fugitive .emissions were calculated using Monte Carlo
procedures reported elsewhere. Expected total fugitive emission rates for each
chemical are given in Table III. It should be emphasized that these chemical-specific
fugitive emission rates are only first approximations. An exact methodology for
estimating volatilization rates of specific chemicals in complex mixtures does not exist.
4. EXPOSURE ASSESSMENT
4.1 SITE DESCRIPTION
The location chosen for this assessment was a hypothetical northern Midwest site (S-l)
located in Marathon County, Wisconsin at 44° 55' latitude and 115° 30* longitude. This
site is located in a rural area adjacent to several large cities. Thus, even though the
total population within 100 km of the site is 0.45 x 10° people, there ore no people
residing within 875 m of the incinerator site (see Table IV). This location is neither a
current nor a planned incineration facility site.
A circular area of 100-km radius around the incinerator facility was assumed for the
assessment. Exposure and risk due to long-ronge transport (greater than 100 km) was not
performed in this assessment. Although long-range transport can significantly contribute
to total population exposure by incorporating large population centers, the average
individual exposure to such transported materials is very low (orders of magnitude less
than background in some cases).
16-4
-------
4.2 ATMOSPHERIC TRANSPORT MODELS
Annual-overage ground-level air concentrations of representative chemical pollutants
were estimated using IEM, an automated inhalation exposure methodology. This
methodology employs a slightly modified version of the Industrial .Source Complex Long
Term Model (ISCLTM), a Gaussian plume model developed for EPA.7
A circular area of 100-km radius around the incinerator facility was assumed for the
assessment. Gaussian plume models are generally applied for distances of 20-50 km
around a site. However, this type of dispersion model has been validated out to ISO km
over flat terrain at one location," predicting annual-average air concentrations within a
factor of three of those measured.
The assessment area was divided into sector segments consisting of 20 concentric circles
about the origin and sixteen radial direction vectors. The circles had radii of 0.15, 0.2S,
0.50, 0.75, 1.0, 2.0, 3.0, 4.0, 5.0, 7.5, 10, 20, 30, 40, 50, 60, 70, 80, 90, and 100 km. Each
radial vector was separated by 22.5° intervals with the first sector being centered on due
north (0°), and proceeding clockwise. ISCLTM was then employed to calculate
concentration estimates at segment centroids.
The total exposure in a sector segment is calculated by multiplying the annual-average
ground level air concentration at the segment centroid by the number of persons located
in the segment, as obtained from I960 Census data tapes using an adaptation of the
APORT computer code.'
4.3 MODEL INPUT PARAMETERS
IEM input parameters include model plant descriptors, pollutant behavior variables, and
region-specific meteorological data. Each of the six hypothetical incineration facilities
(two designs in three sizes) was assumed to contain four separate emission sources: a
stack and three area sources (the receiving, storage, and feed areas). The stack was
located at the center of the circular grid (site origin), and the area sources were
represented as square areas centered on the origin. This potentially artificial area
source configuration of squares was necessitated by IEM modeling constraints and will
not affect pollutant concentration estimates unduly. Each area was centered on the
stack to simplify computation and to prevent possible bias of atmospheric dispersion
estimates. Stack and area source quantities are summarized in Table I. Inspection of
these values shows that stack and receiving parameters are identical for both designs.
However, differences in area, width, and generic gas emission rates for the storage (10
and ISO x 10° Btu/h sizes) and feed (ISO x 10° Btu/h size) areas do occur between designs
because of size and number of storage tanks. The specific parameters employed are
summarized in Table V.
Region-specific meteorological data were obtained from Stability Array (STAR) data
tapes"' and from a compendium of weather statistics''. The STAR data were organized
into six Pasquill stability categories (A through F) and six wind speed classes with
average wind speeds of 0.75, 2.5, 4.3, 6.8, 9.5, and 12.5 m/s. The remaining
meteorological parameters were obtained or derived from Ruffner (1978). For the
principal site they include an average air temperature of 280.2° K and mixing layer
heights of 1819.5, 1213.0, 1213.0, 1024.3, 10000.0, and 10000.0 m for stability categories
A through F, respectively. Mixing layer heights were assumed equal for all wind speed
classes within each stability category.
16-5
-------
Site-specific population data were obtained from 1980 Census data tapes that hod been
reformatted into coarse and fine rectangular latitude/longitude grids. Approximately
0.45 x 10° people reside within c 100-km radius around the generic site.
5. HEALTH EFFECTS ASSESSMENT
A key component of health effects assessment is the determination of human health
response to a given exposure. Quantification of the exposure-response relationship can
be made in an number of ways, depending on the availability of data and the health
impact under consideration. There are many potential health endpoints in humans,
including carcinogenicity, mutagenicity, and teratogenicity. Unfortunately, adequate
human data for assessing risk to most toxicological endpoints do not exist, although a
wealth of animal data are available for selected endpoints.
Toxicity data employed in this assessment are taken from a health effects assessment
summary for 300 hazardous organic jaxistituents prepared by EPA's Environmental
Criteria Assessment Office (ECAO). Chemicals studied were classified as either
carcinogens or non-carcinogens, and separate measures of toxicity were developed for
these two classes. Carcinogenic toxicities are etimated in terms of lifetime excess
cancer risk factors, while noncarcinogenic toxicity is estimated by acceptable daily
intake (ADD, and threshold limit values (TLV).
Health effects data for compounds considered in this assessment are summarized in
Table VI. It is apparent that not all health risk estimators are available for each
chemical studied. The estimate of excess human lifetime risk to cancer was used for all
compounds in the pesticide-related waste. In the absence of cancer risk estimates, the
available ADI value was used for all compounds in the API separator sludge waste except
for arsenic; the TLV1J was converted to an ADI for arsenic and for compounds in the
phenol/acetone distillation waste for which there is no reported AOI.
6. RESULTS ANO CONCLUSIONS
The purpose of this assessment was to determine the relative importance of plant design
and waste physicochemical variables on human inhalation exposure and health risk
resulting from hazardous waste incineration. A hypothetical waste incineration site in
the northern Midwest was chosen for analysis. This site has a population of 0.45 x 10°
persons, with the closest individuals residing 1500 m from the incineration site.
To account for engineering variables, two facility designs (liquid injector and rotary kiln)
of three sizes (I, 10, and ISO x 10° Btu/h), each burning three generic wastes, were
selected. These are designated LI-1, LI-IO, LI-150, and RK-I, RK-IO, and RK-150,
respectively. Three levels of destruction and removal efficiency (ORE) were considered
(99.0, 99.9 and 99.99% ORE). The three waste groups selected for study, in order of
increasing incinerability were: (I) pesticide-related chemicals (chloroform, ethylene
dichloride, hexachlorobutodiene, and 1,1,2,2-tetrachloroethane); (2) API separator sludge
chemicals (arsenic, chromium, lead, and phenol) and (3) phenol/acetone distillation
chemicals (toluene, pyridine, phthalic anhydride, and methyl styrene).
Annual-average ground level air concentrations of representative chemical pollutants
were estimated using IEM, an automated inhalation exposure methodology.5 Air
concentrations were estimated for both stack and fugitive emissions using region-specific
meteorological and climatological data.
16-6
-------
Estimates of individual and total population exposure resulting from incineration of
hazardous materials were calculated by multiplying ground-level air concentrations by
site-specific population data obtained from the I960 Census. Estimates of risk
associated with incineration of hazardous materials were obtained by multiplying the
population exposure estimates by health risk estimators supplied by the U. S.
Environmental Protection Agency (EPA) Environmental Criteria Assessment Office
(ECAO). Chemicals studied were classified as either carcinogens or noncarcinogens, and
separate measures of toxicity were used for those two classes. Carcinogenic toxicities-
are measured in terms of lifetime excess cancer risk, while noncarcinogenic toxicity is
measured in terms of acceptable daily intake (ADI). Estimates of life-time excess
cancer risk were used for chloroform, ethylene dichloride, hexachlorobutadiene, 1,1,2,2-
tetrachloroethane, and arsenic. The AOI value was used far chromium, lead, phenol,
toluene, pyridine, phthalic anhydride, and methyl styrene.
Major conclusions to be drawn from this report are:
I. Fugitive emissions may be an important contributor to total po!Mutant emission
rates. The largest contribution occurs when stock emission rates are lowest, that is
for small incinerators (I x I06 Btu/h) with OKE's ranging from 99.9 to 99.99%. For
large incinerators, fugitive emissions are a relatively unimportant contributor to
total emissions at all ORE'S studied (Table VII). Caution should be used in
interpreting these results as it appears that the fugitive emission predictions in this
report may be higher than those actually experienced at incinerator sites.
2. Total population exposure is relatively insensitive to changes in ORE for small
incinerators (I x 10° Btu/h), increasing by only a factor of 7 from 99.99 to 99.0%.
Total population exposure to volatile compounds is sensitive to incinerator size for
low ORE'S (99.0%) and is relatively insensitive to incinerator size for high ORE'S
(99.99%) (see Table VIII).
3. For all chemicals studied, human health risk from incineration was small. For
incineration of pesticide-related waste at a site with a population of 0.45 x 10°,
expected number of cancers over' 70 years are less than 1.6 x IO"3 for a ORE of
99.99% (see Table IX). The excess cancer risk over 70 years to the maximally
exposed individual at site S-l from incineration of pesticide-related waste is less
than 1.2 x IO'7 for a ORE of 99.99% (see Table X).
4. None of the four selected constituents of the phenol/acetone distillation wastes are
known carcinogens. Therefore, using the AOI as a measure of noncarcinogenic risk,
estimates of average daily intake as a fraction of the AOI were mode. These risk
estimates, which account for both stack and fugitive emissions, are highest for
pyridine releases from ISO x 10° BTU/lb plants, but even these risk values are less
than lO'7 of the AOI (see Table XI).
5. Risk to exposed individuals (as a fraction of AOI) from incineration of heavy metals
is small (see Table XII). Even for the maximally exposed individual, all inhalation
exposures to heavy metals result in average daily intakes which are at least 6
orders of magnitude less than the acceptable daily intake.
16-7
-------
7. REFERENCES
'Keitz, E. L., L. J. Boberschmidt, and C. C. Lee, "A Profile of Existing Hazardous
Waste Incineration Facilities," MITRE Corporation, McLean, Virginia (1982 — in
preparation).
2Bonner, T., 8. Oesai, J. Fullenkamp, T. hughes, E. Kennedy, R. McCormick, J. Peters,
and O. Zanders, Engineering Handbook for Hazardous Waste Incineration, EPA
Contract No. 68-03-2550, Monsanto Research Corporation, Dayton, Ohio (1980).
3MITRE Corporation, "Composition of Hazardous Waste Streams Currently
Incinerated." Working paper (1983).
*U. S. Environmental Protection Agency, Background Document, Resource and
Recovery Act. Subtitle C. Identification and Listing of Hazardous Waste, Office
of Solid Waste, Washington, D.C. (1980).
Vtolton, G. A. and C. C. Travis, "A Methodology for Predicting Fugitive Emissions for
Incinerator Facilities," Proceedings, 1983 Notional Meeting of the American
Institute of Chemical Engineers, Washington, U.C. (November, 1983).
6ODonnell, F. R., and G. A. Hoi ton, "An Automated Methodology OEM) for Assessing
Inhalation Exposure to Hazardous Waste Incineration Emissions," Incineration and
Treatment of Hazardous Waste; Proceedings of the 9th Annual Research
Symposium at Fort Mitchell, Kentucky, May 2-4, 1983, U. S. Environmental
Protection Agency, Cincinnati, Ohio (in press).
7Bowers, J. F., J. R. Bjorklund, and C. S. Cheney, Industrial Source Complex (ISC)
Dispersion Model User's Guide. (Volume I), • EPA-450/4-79-030, U§7
Environmental Protection Agency, Research Triangle Park, North Carolina (1979).
^uckner, M. R. (ed.), Proceedings of the 1st SRL Model Validation Workshop. Nov. 19-
20, 1980, Hilton Head, S. C., Savannah River Laboratory, DP-1597 (1981).
'Fields, O. C. and C. A. Little, APORT — A Program for the Area-Bosed
Apportionment of County Variables to Cells of a Polar Grid. ORNL/TM-6418. Oak
Ridge National Laboratory, Oak Ridge, Tennessee (I97T5E
'"National Oceanic and Atmospheric Administration, Seasonal ond Annual Distribution
by Posquill Stability Classes STAR Program. National Climatic Center, U. S.
Department of Commerce, Asheville, North Carolina, 1974.
"Ruffner, J. A., Climates of the States. Vols. I and II, Gayle Research Company, Book
Tower, Detroit, Michigan (1978).
I2'VI. S. Environmental Protection Agency, Environmental Criteria Assessment Office,
Health Effects Assessment Summary for 300 Hazardous Organic Constituents in
Support of Regulatory Impact Analysis of the Land Disposal Branch (LOB) and
Interim Final Incinerator Regulations of the Technical Branch (TB) of the Office
of Solid Wastes," OSWER/EPA (1982).
13American Conference of Governmental Industrial Hygienists, TLVs - Threshold Limit
Values for Chemical Substances and Physical Agents in the Workroom
Environment with Intended Changes for 1974, Cincinnati. Ohio (1974).
16-8
-------
Table I. Expected equipment quantities and operating assumptions
for each incineration facility design
Rotary
kila
liqvid iaj*etor
Six*. 10* Bttt/n
•BCEIVDW AREA:
Capacity, track*
Capacity, taak car*
Pwpa (pain)
la-lia* fair**
Op*a-*ad*d ralr**
Flaagea
laatr. eoaacctioaa
1
1
0
1
13
2
24
4
Saatpliag ooaneotioas 0
Storage taaka
S1DRA6E AREA:
Poapa (pain)
Pvapa (aiagl*)
Ia-lia* talrea
Op*a-*ad*d >*lr**
Flaagea
laatr. coaaectioaa
0
1
0
13
.2
24
4
Saatpliag coaaectioaa 4
Storage taaka
FEED AREA:
Pnpa (pairs)
la— liae fair**
Opea-eaded fair* a
Flaagea
laatr. eoaaectioa*
4'
1
13
2
24
4
Saatpliag aoaaectioaa 1
Storage taak*
•1.0 x 104 gal
b2.0 x 104 gal
e1.5 x 10S gal
"1.25 x 104 gal
•2.5 x 104 gal
ll
capacity.
capacity.
capacity.
capacity.
capacity.
10
2
0
3
39
6
72
12
2
2*
1
2
39
<
72
12
4
4b
1
13
2
24
4
1
1*
150
2
1
4
52
1
96
16
6
'*
1
2
39
6
72
12
4
4«
1
13
2
24
4
1
1
*2.0 103
15.0 103
•1 .2 103
i(.Q JO3
Jl.5 103
1
1
0
1
13
2
24
4
0
0
1
0
13
2
24
4
4
4d
1
13
2
24
4
2
2
«•!
gal
gal
gal
gal
10
2
0
3
39
6
72
12
2
2«
1
2
39
<
72
12
4
4«
1
13
2
24
4
2
2d
capacity.
capacity.
capacity.
capacity.
capacity.
150
2
1
4
52
8
96
16
6
<*
1
2
39
6
72
12
4.
4*
1
13
2
24
4
2
2-i
16-9
-------
Table II. Ten most prevalent constituents of hazardous
waste streams
Cam«titB»t iBoamt Incinerated (JfT/y)
lUtkuol 133.168
Ao«tomitril» St.646
TO!M«« it.620
Btianol 55.090
4«rl «a«t*t* 54.926
Ae«tom« 51.535
XTl»n« 49.453
lUtiyl «thyl t«too« 42.520
Adlple told 36.135
Etiyl toctatt 32.576
16-10
-------
Table III. Consumption, concentration, and emission rates of
incinerated chemicals (99.99% ORE)
lM Ml«, 0/7
•allBIIBt
I FiMlloB
|BB!B- •••la
•r«t«4* • tfB»B
mitt it)
UU
UjMlor
10
ISO
10
BUck P«|lll» ll««l FBflllt* lock P«(IU>B link PB|lllf* BlBBk FB|ltl» luck F<|lllt<
OJ
i
PESnCIDE-BELATBb BASTE
Cklorof or»
ElkfltBi «Ukloil4B
B«»eklorokBU4l»*
I.I1E*4
7.1JE.4
4. If 1.4
l.l.l.l-l«tr>cklo(0*tkBB«l.lOE*4
API IEPABATOI
SLUDGE BASTE
Cklo.lue
Ui4c
AI»B!CC
Pk.B.I
PBENOL/ ACETONE
DISTILLATION BASIC
TalB»*
PycHlBB
PklktIU (BkytfrKt
lUlkjrl |.) 0 1.441.4 0 1.101.) 0 l.(*l»l 0 1.4*1.4 0 1.10B*! O
1.101*1 0 1.101*4 0 I.lOBtl 0 1.101*1 0 1.101*4 0 1.101*1 0
4.001*1 0 4.001*1 0 (.001*4 0 4.001*1 0 4.001*1 0 4.001*4 0
(.111-1 1.4)1-1 (.171*0 *.(BB-1 1.0)1*1 I.IOB-I 4.I7B-1 1.I1B-1 (.111*0 ».IIB-1 l.OJf.l 1.I1B-1
1.»1B«1 4.4)B») I.*1B«4 1.1(1*4 1.171* ) 1.111*4 l.UB.l (.111*1 l.»ll*4 1.101*4 1.171*1 1.40E*4
2.41E.1 1.411*1 l.(BB*l ». 411*1 4.011*4 1.111*1 l.(IB*l 1.441*1 1.411.1 ».(»•*! 4.011*4 1.111*1
l.Olt-1 1.141-1 1.011*0 1.4(1-1 1.0*1*1 1.44B-1 1.01B-1 1.401-1 1.01B*0 1.4*E-1 1.0*1*1 1.101-1
1.121*1 *.»11*1 1.711*4 1.111*1 1.111*1 4.1*1.) 1.111*1 1.011*1 1.111*4 1.1*1*1 1.1JB.J 3.1JE.J
•fiom oo.pll>llo> •( ml* l>l«n>llo> (DIETA. 1*M«).
^Calcvlklcd ky •*«r«|l«| f«ill«l4*-v*lft*«4 •••!• Bka«l
•i ME af »1.M ll B1IBM4 101 Ik* k«fj MUU.
-------
Table IV. Cumulative population by distance from the incinerator site
Distance to Ring Center (m)
200
375
£25
875
1,500
2,500
3,500
4,500
6,250
8,750
15,000
35,000
55,000
75,000
95,000
Population
0
0
0
0
203
949
2,378
4,053
6,039
7,437
76,576
126,614
263,097
349,969
448,187
Table V. Stack and area source parameters employed in atmospheric
dispersion modeling
Rotary kiln
Liquid iajcetoz
Sis*. 10* BtWh
Six*, 10* Btu/h
10
150
10
150
STACI:
H*i|At. • 15.24
Exit «a* t*»p«rator*. I 355
Exit fa* f.locity, •/* 1.10
Di*B*t*z. • 0.41
8»*rlc (a* taiui OB rat*. (/• 1.0
RECEIVING AREA:
H.iffct, • 2
Ar*a (iqvar*), •
Width of ar*a. •
Q*a*xic |t« Minion rat*, f/t
STOKAQE AREA:
B*i(Jit. . 2
Area (iqnar*), •
lidth of az*a. •
0«n«rie |*« ••ii*ion rate, f/i
FEED AREA:
Araa (tqnaz*). •
Widti of u*a. •
G«n«rie (*• aaiitiioa rat*, 1/1
4.1
16.2
8.73
76.2
30.49
355
11.0
0.61
1.0
6.1
130
11.4
130
30.41
355
41.1
1.22
1.0
15.24
355
1.10
0.41
1.0
30.48
355
11.0
0.61
1.0
f.l
S»9
2«.3
689
S.I
76.2
8.73
76.2
f.l
130
11.4
130
30. 4«
355
41.1
1.22
1.0
6.1
74.3
8.6
74.3
6.1
130
11.4
130
6.1
390
19.8
390
6.1
74.3
8.6
74.3
6.1
130
11.4
130
6.1
390
19.8
390
6.1
325.2
18.0
325.2
6.1
345
18.6
345
6.1
2165
46.5
2165
6.1
325.2
18.0
325.2
6.1
423
20.6
423
6.1
2737
52.3
2737
S.I
S66
29.4
866
6-12
-------
Table VI. Health risk estimators for chemicals in three generic wastes
Be«lth li«k E«ti««tor
Pollutant
Exce** tiik* ADIb
TLV«
PE3TXCIDE-IBLATED WASTE
Cklorefoa
Btarlea* dicaloride
BexsoAlorobatadlene
1.1.2.2-tetr»ckloroet««ae
API SHPAIAIOK SUOBB WASTE
BB 71
III
Uad
ArMaie
Pkcaol
FUMOL/ACETDNE DISTILLATION IASTE
TelMM
Ptridlm.
Pkti«He tniydrid*
.tyr.n.
1.8E-1
3.7B-2
7.7SB-2
2.0B-1
NA
NA
NA
1.4E+1
NA
NA
NA
NA
NA
I. SEX)
NA*
1.4B-1
NA
1.4E-1
1.3E+J
l.OE-1
7.0B-2
6.IB+0
1.3E+2
5.4B+0
2.1B+0
1.7E+1
5.08+1
4.0B+1
NA
3.JB+1
5. OB- 2
5.0B-1
1.3E-1
2.0B-1
1.9E>1
3.75E+2
l.JEfl
6.0E+0
4.JOE+2
c»ae«r tf t«r » lifetia* «rpo»r(
i«t»k« (OSEPA. 1983).
•Ese««* ri»k of
(USEPA. 1982).
*Aee«pt«bl« dail
°Thz*«hold li»it t*la« 1* U« ti««-»«i»ht»d conc*ntr«tion for »n
S-h verkd>7 »d 40- h workweek to be ued t* i |«ide in tie control of
kascrda (ACOIH, 1980).
'Not t»«iUbl».
16-13
-------
Table VII. Percent contribution of fugitive emissions to total
emissions of chloroform at a liquid injection incinerator
« ME
99. 99
99.90
99.00
LI-1
93.5
51. •
12.5
CamtribatiOB (%>
U-10
71. <
37.1
3.6
U-150
24.1
3.2
0.3
Table VIM. The effect of ORE on total population exposure
(person-ug/m-') to chloroform emissions from
a liquid injection incinerator
« MB
99.99
99.90
99.00
LI-1
2.5B+0
3.7B+0
1.7E+1*
U-10
7.1E+0
1.7E+1
1.1B+2
U-150
2.08+1
1.2S+2
1.2B+3
•l»«d >i 1.7 z 101
Table IX. Expected number of excess cancers over 70 years
from incineration of pesticide-related waste at a liquid
injection incinerator (for a population of 0.45 x 10*)
* ME U-l U-10 U-150
99.99
99.90
99.00
1.3B-4
2.4B-4
1.6B-3
4.0B-4
1.4B-3
l.lfi-2
1.6B-3
1.2B-2
1.2E-1
16-14
-------
Table X. Excess cancer risk over 70 years to the maximally
exposed individual from incineration of pesticide-related
waste at liquid injection incinerators (99.99% ORE)
U-l
LI-10
U-150
2.SB-8
7.8E-8
1.2E-7
Table XI. Average daily intake of selected phenol/acetone distillation
waste constituents released by hazardous waste incinerators
(presented as a fraction of the ADI)
Tolnca*
Pyridin.
Piti. tafcyd.
Methyl > tyrant
AOI
134
5.4
2,1
171
U-l
2.7E-10
3.9E-10
3.2E-13
9.1E-11
Inoin*
U-10
1.2E-9
1.8E-9
1.4E-12
5.8B-10
r*tio« Facility (99.
U-150
8.4E-9
1.4E-8
2.SE-11
5.7E-9
LI-1
2.8E-10
4.1E-10
3 .2E-13
9.3E-11
99% ORE)
U-10
1.2E-9
1.9E-9
2.4E-12
5.8E-10
Ll-150
8.5E-9
1.4E-8
2.5E-11
S.7E-9
Table XII. Average daily intake of heavy metals (as a fraction
of AOI) for the maximally exposed individual from incineration
at a liquid injection incinerator (95% ORE)
Chraaia'
Ari«nic
Utd
U-l
7.4E-8
3.8E-8
1.4E-7
U-10
3.0E-7
1.5E-7
5 .3E-7
U-l 50
7.4B-7
3.8E-7
1.4E-4
•Tb« AOI f»la» for ehroaini IV *•* u«d.
16-15
-------
Impact of Fugitive Emissions on PM-10 Concentrations
Thompson G. Pace, EPA/OAQPS-RTP
ABSTRACT
The National Ambient Air Quality Standard for Particulate Matter is cur-
rently being reviewed. Indications are that a revised standard for inhal-
able particles (IP) will be proposed. This will consider only those par-
ticles smaller than 10 jion aerodynamic diameter. Knowledge of the bi-
modal distribution of particles in ambient air suggests that, while fugi-
tive sources were a major component of TSP, they will be much less sig-
nificant under an IP standard.
Ambient chemical and size distribution data and assessment of source
contributions based on a number of ambient studies will be consolidated
into an appraisal of the impact of fugitive emissions on the inhalable par-
ticle fraction of Particulate Matter. Empasis will be on size segregated
data from dichotomous sampler studies where the coarse mode contribu-
tion to ambient IP can be isolated.
»
(This is an abstract of a verbal presentation for which a paper is not available.)
17-1
-------
APPLICATION OF DISPERSION DICTATED
MASS BALANCE FOR CALCULATING
FUGITIVE DUST EMISSIONS
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views oi the
Agency and no official endorsement should be inferred.
CLIFFORD F. COLE, TRC ENVIRONMENTAL CONSULTANTS, INC.
JOHN G. MOLDOVAN, ANACONDA MINERALS COMPANY
PAUL B. KUNASZ, CONSULTANT
ABSTRACT
This paper describes the development and application of a new
receptor-source model which provides estimates of emission rates from
fugitive dust sources by employing a dispersion dictated mass balance
(DDMB) technique. Unlike other chemical or morphological receptor
models, the DDMB method considers only one constituent, which in most
practical applications will be total suspended particulate matter. One
equation is written for each receptor, such that the total concentration
of particulate collected at each receptor is set equal to the sum of a
contributing source's emission rate times a dispersion factor. The
dispersion factor must be calculated from a dispersion model, and in this
sense, the DDMB method relies on meteorological data. Given measured
particulate concentrations, values of emission rates can be computed by
solving a set of simultaneous linear equations.
The DDMB method was applied to compute the emission rates and the
contribution of TSP levels of a major mining operation in Butte,
Montana. The DDMB results provided an estimate of the mine's "effective
emissions" which differed greatly from emissions determined by the
traditional emission factor technique. The Butte study is used to
demonstrate the DDMB methodology, and findings of the study are discussed.
18-1
-------
1.0 BACKGROUND
In airsheds characterized by high particulate concentrations, the
quantification of each major source's contribution to ambient air
particulate levels is critically important. Only when it is known where
particulates come from, and how much an individual source adds to the
particulate burden, can controls be applied in a cost-effective way. The
large number of areas in the U.S. in which concentrations still exceed
particulate ambient air standards, despite expensive control efforts,
attests to the difficulty of identifying and quantifying dust
sources(l).
Source apportionment—the determination of the mass of particulate
matter at a receptor that emanates from various sources—has
traditionally been accomplished with source modeling. Emission rates of
sources are estimated, and input along with meteorological data into a
dispersion model that analytically simulates the transport, diffusion,
and deposition of particulate matter. In the vicinity of discrete,
well-defined point sources, this modeling method appears to work well.
However, in regions where a large portion of the particulate matter is
fugitive dust the source modeling method has proven inadequate^). The
failure of traditional modeling to properly account for observed
particulate concentrations is due in part to the difficulty in
quantifying fugitive dust emission rates. Emission rates from dust
sources are hard to measure, and change dramatically with meteorological
conditions, type and extent of source activity, and other parameters.
In the last few years receptor modeling has been touted as a remedy
to the fugitive dust source apportionment problem(3). Proponents of
receptor modeling correctly point out that chemical mass balance (CMB),
enrichment factor, time series, and multivariate receptor models
eliminate some of the uncertainty associated with source modeling.
However; there are two distinct drawbacks that limit most receptor
models' use in regions where fugitive dust is suspected of inducing large
hi-vol concentrations. First, receptor models are unable to
differentiate physically or chemically similar constituents that may be
entrained from different sources; and second, receptor models assume that
the relative mass fractions of all chemical species are conserved, and
that no selective deposition occurs(^). To illustrate that these
drawbacks can be severely limiting, consider a TSP nonattainment area
where road dust is found to be a significant contributor to
concentrations. If the road dust is chemically and physically
homogeneous, and if deposition and resuspension of the dust occur, then
the receptor model cannot pinpoint which are the offending roadways(2).
This paper discusses a new receptor model method termed Dispersion
Dictated Mass Balance (DDMB). The DDMB method utilizes features from
both source and receptor modeling technology, thereby overcoming some of
the deficiencies in either modeling approach used alone. Because DDMB
considers only one constituent, usually total suspended particulate, and
because the effects of deposition and reentrainment can be taken into
account, the method is especially well suited to fugitive dust source
apportionment. The DDMB method is simple and inexpensive to apply, and
yields statistically sound results.
18-2
-------
2.0 ANALYTICAL METHOD
2.1 GOVERNING EQUATIONS
Like all receptor models, the basis for the DDMB method is a simple
model in which the pollutant contribution from a given source at a
receptor is set equal to the emission rate of that source times a
dispersion factor:
Xj - (X/Q)j Qj (1)
where X j is the concentration at the receptor due to source j
Qj is the emission rate from source j
(X /Q)j is the dispersion coefficient from source j to the
receptor
Equation (1) can be summed over a number of sources, P, so that
P
X - , *r i 0(/Q) . Q. (2)
The only restrictions imposed by equation (2) are that concentrations are
additive, and directly proportional to emission rates. If there are no
more than P sources in the vicinity of the receptor that contribute
pollutant, then equation (2) will account for all of the pollutant
concentration detected. More likely, it will be impossible to identify
all of the sources that impact the receptor, but the contribution of
unknown sources can be accounted for by an assumed background
concentration, B, which is added to the right-hand side of equation (2).
Obviously, equation (2) can be written for any number of independent
receptors to produce a set of linear equations. Generalizing for n
receptors,
Xn
(x/Q)
nj
+ B
(3)
If the x/Q values and the background concentration are known, if the
measured concentrations at each receptor are available, and if the number
of receptors equals or exceeds the number of sources, then equation (3)
yields a fully determined (or overdetermined) system that can be solved
to find source emission rates.
While the above list of necessary data conditions appears
restrictive, in fact for many regions these data may already be available
or can be readily calculated. Consider an area deemed nonattainment for
TSP. Annual average hi-vol concentrations at several locations will
undoubtedly be available, and an annual average background TSP
concentration can easily be assumed or can be determined from these
measured concentrations using a variety of methods^). The X/Q values
may already be known if adequate modeling studies have been performed in
18-3
-------
the area. If not, model simulations of TSP concentration at each of the
hi-vol receptors must be performed. The X/Q values are determined by
dividing modeled concentrations by the source emission rates assumed in
the model.
It is important to note that the emission rates of each source or
source grouping ultimately determined by the DDMB method depend upon the
X/Q dispersion coefficients that are introduced in equation (3). For
this reason every attempt should be made to accurately simulate
dispersion in the initial modeling phase. If deposition is thought to be
an important phenomenon in the area being studied, then a model which
accurately refects deposition should be used. Similarly, meteorological
data input to the dispersion model should be screened carefully. If
sources are combined, then the resultant x/Q values depend upon the
intial choice of emission rates assumed in the dispersion model, and
special care should be taken to estimate these initial emission rates
accurately.
2.2 METHODS OF SOLUTION
The set of fully determined or overdetermined DDMB linear equations
can be solved with almost any matrix solution software package, although
.software packages that provide statistical analysis of the data are
preferable. For overdetermined systems the standard deviation of the
predicted emission rates, and the matrix of correlations between emission
rates, are extremely valuable statistics. The standard deviation
provides a measure of how "good" a computed fugitive dust emission rate
is, and allows one to compute statistical confidence intervals associated
with each emission rate. Given the uncertainty in fugitive dust emission
factors, confidence intervals should be demanded by every investigator.
The correlation matrix between emission rates provides guidance in
combining sources, a technique which greatly enhances the DDMB method.
2.3 SOURCE COMBINATION
In order for equation (3) to be solvable the linear system must be
fully determined or overdetermined, that is, there must be as many or
more hi-vol receptors than sources. One way to achieve this is to
combine sources by grouping them together for the purpose of the DDMB
computations. For example, two adjacent unpaved parking lots could be
considered a single fugitive dust source, thereby reducing the order of
the linear system of governing equations by one. Of course, source
combination like this reduces the resolution of unknown emission rates,
but it may be required to make the matrix fully determined. Even more
important, careful source combination can dramatically improve the
accuracy of the computed emission rates, as will be demonstrated later in
this paper.
18-4
-------
In deciding which sources to combine, common sense dictates that
sources which have similar impacts on the hi-vol receptors be combined.
The hi-vol data cannot discriminate one of these sources from another, so
it is not reasonable to ask the DDMB method to make a distinction.
Frequently such sources will be located near one another, as in the
adjacent parking lot example above. More rigorously, one can determine
which sources are linked by examining the correlation between predicted
sources' emission rates. When the correlation between two sources
approaches -1.0 this indicates that the two emission rates are linked in
such a way that if one rate is varied, the other must react oppositely in
order to minimize the sum of squares of the residuals. Two strongly
negatively correlated sources cannot be resolved by the data at hand, and
attempts to do so would yield arbitrary results which would be extremely
sensitive to any errors in the data.
3.0 DDMB APPLICATION
The DDMB method was used to determine source apportionment and TSP
emission rates from various sources in Butte, Montana for the year 1978.
Butte provides an especially attractive study area for the DDMB method
because there are a number of fugitive dust sources in and near Butte
whose emission rates have proven very difficult to quantify.
Conventional fugitive dust inventories have yielded drastically different
emission rates for the same sources^). This is illustrated in
Table 1., which shows a summary of emission rate estimates determined by
three independent studies.
TABLE 1.
1978 ANNUAL ESTIMATED EMISSION RATE (tons/year)
FROM ANACONDA SOURCES AND BUTTE SOURCES
EMISSION SOURCE
ANACONDA
BUTTE
STUDY 1
15,000
11,000
STUDY 2
5,612
3,402
STUDY 3
26,142
11,564
In the following sections the application of the DDMB method at
Butte is illustrated. The major sources in Butte are described, hi-vol
and x/Q input data are tabulated, and the sequential steps in the DDMB
process are discussed.
3.1 SOURCE AND RECEPTOR DESCRIPTION
Figure 1. shows the location of major particulate matter sources
and hi-vol samplers in Butte. The hi-vol samplers, indicated by stars in
Figure 1., provide every-sixth-day measures of total suspended
particulate. Of the nine hi-vols displayed in Figure 1., three are
operated by the Montana Air Quality Bureau, and the remainder are
operated by Anaconda Minerals Company.
18-5
-------
1. ALPINE 6. KAW
2. HILLCREST 7 GREELEY
3. MONTANA POWER 8. KEBGEN
4. MT. CON 9 DR. CANTY
5. YATES
9 J 234 5 6 Km.
SCALE
FIGURE 1
SOURCE AND HI-VOL
LOCATIONS
18-6
-------
Anaconda operations are located-from the northeast to the east of
the city of Butte, and involve the mining and concentrating of copper
ores from the Berkeley Pit. Drilling, blasting, and removal of
overburden and copper ore by shovels are dust producing activities that
are confined to the 550 meter deep Berkeley Pit. The dumping and storage
of ore, as well as crushing and concentrating activities, occur at the
southern end of the pit perimeter. Overburden is temporarily stored in
primary dumps to the northwest, northeast, and southeast of the pit
before being returned in a backfilling operation. Haul roads and access
roads link the Berkeley Pit with the crusher, concentrator, and dumps.
The city of Butte, adjacent to the pit, is itself a source of
fugitive dust resulting from many different activities. Previous studies
have identified paved and unpaved roads, fuel combustion, motor vehicle
exhaust, construction, burning, and wind erosion from cleared areas as
major particulate sources.
3.2 INPUT DATA
Two separate types of input data are required by the DDMB model:
measured TSP concentrations at hi-vol receptors, and the x /Q values for
each source-receptor pair.
The 1978 measured TSP concentrations at nine hi-vols in the Butte
area are shown in Table 2. A uniform annual average background
concentration of 20..0 y g/m^ is assumed to characterize the airshed, so
that the contribution of local sources to measured concentrations is
determined by .subtracting 20.0 from the measured concentrations.
TABLE 2.
1978 BUTTE AMBIENT TSP CONCENTRATIONS
MEASURED MEASURED MINUS
CONCENTRATION BACKGROUND BACKGROUND
HI-VOL
GREELEY
HILLCREST
YATES
ALPINE
RAW
MT. CON
HEBGEN
DR. CANTY
MONTANA POWER
(Ug/m3)
79
37
73
89
54
34
70
56
47
(y g/m )
20
20
20
20
20
20
20
20
20
(yg/m )
59
17
53
69
34
14
50
36
27
The x/Q values used in this study were determined from previously
run ISCLT dispersion model simulations of the Butte area. Using 1978
18-7
-------
00
00
TABLE 3
X/Q VALUES FOR SOURCE-RECEPTOR COMBINATIONS
HI -VOL
GREELEY
HILLCREST
YATES
ALPINE
KAW
MT. CON
HEBGEN
DR. CANTY
MONTANA POWER
INPIT
0.4426
0.1190
0.4077
0.5391
0.1097
0.2098
0.2636
0.0792
0.0738
_ T<
CRUSHER
0.7797
0.1161
0.8189
0.4287
0.2464
0.1972
0.7001
0.1415
0.1298
?P FMT^TON 9
STORAGE
0.8525
0.1195
0.8092
0.4919
0.2083
0.2264
0.5294
0.1338
0.1145
OIIRrFS —
HAUL ROADS
0.4896
0.1390
0.4544
0.6426
0.0925
0.1412
0.2132
0.0668
0.0585
BACKFILL
0.1566
0.0523
0.1603
0.1204
0.1583
0.9977
0.5135
0.3044
0.1262
DUMPS
0.1351
0.1574
0.1283
0.1823
0.0352
0.0702
0.0603
0.0296
0.0263
CITY
0.6002
0.1850
0.5490
0.6781
0.6817
0.2560
0.6457
0.5517
0.4408
NOTE: x/Q values in sec/m3 x 106
-------
meteorological data, the ISCLT model computed the impact of seven
distinct sources:
1. Berkeley Pit
2. Crusher dump
3. Active storage
4. Anaconda haul roads and access roads
5. Backfill
6. Primary dumps
7. Butte (unpaved roads, fuel combustion, wind erosion, etc.)
The first six sources are controlled by Anaconda, while the seventh
source represents the city of Butte. The source grouping option
available in the ISCLT model was used to compute the impact of combined
sources where appropriate. The city of Butte, for example, was modeled
as 134 discrete area sources.
Values of x /Q, determined by dividing the model-predicted
concentrations by the model-assumed emission rates, appear in Table 3.
Each x/Q value in Table 3. represents a measure of the TSP dispersion
coefficient from a source to a hi-vol receptor.
3.3 DDMB SOLUTIONS
Initially, attempts were made to compute TSP emission rates for
each of the individual seven major sources in Butte (pit, haul roads,
crusher, storage, backfill, dump, and city of Butte). The effective
emission rates and confidence intervals resulting from this effort are
shown in Table 4.
TABLE 4.
TSP EMISSION RATES AND CONFIDENCE
INTERVALS FOR UNCOMBINED SOURCES
EMISSION RATE 95% CONFIDENCE INTERVAL
SOURCE (gm/sec) (gm/sec)
INPIT
HAUL ROADS
CRUSHER
STORAGE
BACKFILL
DUMP
TOWN
-176
47
-35
196
11
-6
52
+ 1,540
+ 237
+ 224
+ 1,230
-1- 139
+ 230
+ 41
Clearly, the DDMB method is unable to resolve each of the seven sources,
as indicated by the physically unrealistic emission rates, and the very
wide confidence intervals which suggest that there is a great deal of
uncertainty associated with each predicted emission rate. An explanation
for such large error bars is found in the matrix of correlations between
emission rates, displayed in Table 5.
18-9
-------
TABLE 5.
COFRELATION MATRIX
CRUSHER
INPIT -.68
CRUSHER
STORAGE
HAUL ROADS
BACKFILL
DUMPS
STORAGE
.40
-.92
HAUL ROADS
-.995
.71
-.45
BACKFILL
-.978
.66
-.40
.98
DUMPS
.11
-.06
-.08
-.17
-.15
CITY
-.30
-.10
.16
.27
.23
-.13
The correlation between the inpit and haul road sources is very close to
-1.0, meaning that these two sources cannot be distinguished by the
available data. In essence, the hi-vols "see" these two sources as the
same source. This is not surprising since the haul roads are located
within and surround the Berkeley Pit. Similarly,'the crusher and storage
sources are correlated, as are the inpit and backfill sources. These
strong negative correlations suggest that emission rate accuracy could be
greatly improved by combining sources.
The seven sources were systematically combined and a matrix
solution achieved for each combination, partly in an attempt to find the
most accurate emission rates, and partly to experiment with the source
combination method. Table 6. illustrates the improvement in DDMB
performance as the souces are combined. In Case "B" the pit and haul
roads are combined as if they constituted a single, indistinguishable
source, and the confidence interval associated with the combined source
is smaller than for either individual source. Case "C" in Table 6.
illustrates the effect of combining the crusher and storage sources,
while Case "D" further combines the backfill and primary dump sources.
In Case "D" all but one of the physically unrealistic negative emission
rates have been eliminated, but confidence intervals still remain large.
This condition is not remedied until all of the Anaconda sources are
combined (Case "E"). When they are, confidence intervals shrink
dramatically, demonstrating the resolution sought between Anaconda and
the city of Butte, but at the expense of lack of resolution among the
Anaconda sources.
Using the emission rates shown for Case "E" of Table 6., the TSP
contribution from the Anaconda sources and from the town can be computed
at each hi-vol receptor. The results of these computations appear in
Table 7. which lists the concentration contribution in percent for
Anaconda sources, for the city of Butte, and for background. The
Anaconda contribution is generally lower than that predicted by
conventional source modeling techniques, and is in good agreement with
recent Anaconda estimates of source contribution^5).
18-10
-------
TABLE 6
TSP EMISSION RATES AND CONFIDENCE INTERVALS FOR VARIOUS
CASES OF COMBINED SOURCES
oo
i
SOURCE
IN. PIT
HAUL ROADS
CRUSHER
STORAGE
BACKFILL
DUMP
CITY
CASE "A"
NO SOURCES
COMBINED
-176 ±1540
47 ± 237
- 35 ± 224
CASE "B"
CASE "C"
CASE "D" CASE "E"
PIT & HAUL ROADS PIT & HAUL ROADS PIT & HAUL ROADS, ALL ANACONDA
COMBINED
61.8 t 64
25.2 - 82
196 t 1230 23.1 ± 97
11 t 139 -7.4 ± 14
CRUSHER & STORAGE CRUSHER & STORAGE, SOURCES COMBINED
COMBINED
53.0 ± 63
5.5 ± 31
-7.0 ± 15
- 6 ± 230 -2.1 ± 110 -3.4 - 121
52 ± 41 50.0 * 19 52.4 * 19
BACKFILL & DUMPS
COMBINED
57.4 ± 37
4.7 t 25
-21.1 ± 38
59.1 - 2?.. 9
52.4 t 16 48.3 ± 13.7
Note: All emission rates expressed In gm/sec
All confidence Intervals at 95% level
-------
TABLE 7-
1978 TSP CONTRIBUTIONS AT SELECTED HI-VOLS
ANACONDA CONTRIBUTION BUTTE CONTRIBUTION BACKGROUND CONTRIBUTION
HI-VOL (%) (%) (%)
GREELEY
HILLCREST
YATES
ALPINE
RAW
MT. CON
HEBGEN
DR. CANTY
MONTANA POWER
37.7
20.8
37.4
38.7
11.2
25.3
24.0
9.5
9.6
36.8
24.6
35.8
38.1
55.3
28.6
46.4
51.7
44.9
25.4
54.6
26.9
23.2
33.5
46.1
29.6
38.8
45.6
4.0 DISCUSSION
The DDMB method provides a means of estimating emission rates and
source apportionment from a number of simultaneously emitting fugitive
dust sources. Analytically the DDMB method finds that combination of
source emission rates that yields the best fit (in the "least squares"
sense) between measured and modeled concentrations. The method enjoys
several advantages over current fugitive dust source apportionment
techniques:
o DDMB does not require the detailed laboratory work
associated with many receptor models such as CMS, SEM,
X-ray diffraction, or optical microscopy.
o DDMB can discriminate among sources that emit chemically
and physically identical particulate matter.
o Unlike many spatial receptor models such as trajectory
analysis, cluster analysis, and pollution wind roses,
DDMB provides quantitative measures of source
contribution.
o Depending upon the software package used to solve the
linear system of equation, the DDMB method can "police
itself" by providing confidence intervals for each
computed emission rate.
o The DDMB method is an inexpensive, cost-effective
analysis.
18-12
-------
Naturally there are disadvantages, too. The number of distinct sources
that can be resolved is constrained by the number of hi-vols available.
Furthermore, the improvement in confidence intervals achieved by
combining sources is made at the expense of the number of sources that
can be resolved. It may be that the DDMB method will be most useful when
investigators are willing to partition all sources into just a few source
categories, as in the example application presented above.
Will the DDMB method always work? Application of the procedures
outlined in this paper will always yield emission rate estimates as long
as the governing equations are fully determined or overdetermined. But
there is no guarantee that the answers will be "correct," or that the
confidence intervals will not be so wide as to make the answers
meaningless. Based on this study, and other ' applications of the DDMB
method, it seems that there are some criteria that make success of the
DDMB method more likely: first, there must be no large systematic errors
in the dispersion model used to compute X/Q values. Random errors in the
dispersion model are tolerable, and their effect can be minimized by
combining sources. Second, if the measured hi-vol concentrations are
appreciably greater than the background concentration, then the
unavoidable errors inherent in the hi-vol measurements will have the
least impact on computed emission rates. In other words, the DDMB method
probably works best in areas where large fugitive dust sources are
present. Finally, if particulate sources are spatially separated, and if
hi-vol data are available from many different locations, then the
method's ability to discriminate sources will be enhanced.
REFERENCES
1. Cooper, J. A. and J. G. Watson, Jr., "Receptor Oriented Methods of
Air Particulate Source Apportionment," JAPCA 30; 1,116 (1980).
2. Watson, J. G. and J. C. Chow, "An Overview of Source-Receptor Source
Apportionment," presented at APCA meeting, Philadelphia,
Pennsylvania, June 21-26, 1981.
3. Core, J. E. and T. G. Pace, "Receptor Models—How Great Thou
ArtJ(?)," presented at APCA meeting, Philadelphia, Pennsylvania,
June 21-26, 1981.
A. Yocom, J. E., E. T. Brookman, R. C. Westman, and 0. P. Ambardar,
"Determining the Contributions of Traditional and Nontraditional
Sources of Particulate Matter," JAPCA 31: 17 (1981).
5. Moldovan, G. J. and P- A. Doughty, "The Butte, Montana Particulate
Apportionment Analysis," presented at APCA meeting, New Orleans,
Louisiana, June 20-25, 1982.
18-13
-------
*
**
Modeling the Emission of Aerosols in and Around a Metallurgical
Plant
B. Vanderborght, I. Mertens, and J. G. Kretzschmar, SCK/CEN,
Belgium; F. Adams, UIA, Belgium; and R. Dams, INW, Belgium
ABSTRACT
Dust emissions from a metallurgical plant can cause considerable nuisance and
health hazard to the environment. The cost to benefit analysis of pollution
abatement investments can only be performed by proper modeling of the emis-
sion-immission situation. The problem was thoroughly examined in a case
study by daily antimony immission measurements around a metallurgical plant
over a period of one year. It soon appeared that low level fugitive emissions
and not the high stacks were the major sources of the high immission levels.
These sources cannot be monitored continuously so that direct emission data for
dispersion model calculations are not available. For every process in the plant
an emission factor was obtained by means of a combination of tracer releases,
immission measurements, reversed modeling for the fugitive emissions and in
stack measurements for chimney emissions. Emission factors, inventory of
the plants' production and meteorological observations formed the input data
for bi-Gaussian dispersion modeling. In the calculations, the contributions
from the fugitive and point source emission can be discerned in the immission.
Calculated values fitted the measurements of suspended Sb-aerosols very well.
Validation is done by means of scatter diagrams, the comparison of measured
and calculated cumulative frequency distributions, time series and pollution
roses. Not only the global statistics but also individual measuring periods show
very good agreement. Although the global Sb-dust deposition can be well des-
cribed by wet and dry deposition parameters, and source depletion, some indi-
vidual measuring sites show systematic deviations from the calculations. It
seems that local turbulence conditions were responsible.
(This is an abstract of a presentation for which a paper is not available.)
(*) Nuclear Energy Research Center, B-2400 Mol, Belgium.
(**) Institute for Nuclear Sciences, B-9000 Gent, Belgium.
19-1
-------
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
NEW YORK STATE INDUSTRIAL COAL PILE DRAINAGE REGULATIONS AND GUIDELINES
By Carol Hornibrook, Senior Project Manager, NYSERDA
2 Rockefeller Plaza, Albany, NY 12223
REGULATIONS
Existing waste water discharge standards applicable for New York Coal Pile
Drainage are found in the State Pollution Discharge Elimination System Permit
(SPDES). At this time regulations applicable to the industrial sector control
only the total dissolved solids at .5 mg/liter/day. This unlike utility coal
pile drainage which, in addition to total dissolved solids, requires the
regulation of total suspended solids and pH.
GUIDELINES
In 1980 the New York State Department of Environmental Conservation (NYSDEC)
published a document entitled "Coal Pile Guidance for SPDES (State Pollution
Discharge Elimination System)" which lists a number of requirements for new and
existing coal piles. These requirements call for the monitoring of possible
toxic pollutants emitted from coal piles and the treatment facilities required
to prevent contamination of ground water by leachate and the contamination of
surface waters by runoff. Specifically, the New York State guidelines require
natural and/or manmade liners under new and some existing coal piles, ground
water monitoring systems, storm collection ponds to contain runoff for pH
neutralization and heavy metal removal.
At this point I would like to clarify that a 10 year-24 hour storm is a
statistical term referring to a certain size storm event (expressed in inches
of rain) occurring over a 24 hour period, which has a recurrence interval of 10
years.
The specific limitations for each of the requirements previously mentioned are
listed in Table 1. The reader should note that these guidelines are being
implemented in the same fashion as regulations and specific discharge
limitations are being set in new and renewed SPDES permits.
20-1
-------
TABLE 1
COAL PILE GUIDANCE FOR SPDES
To comply with surface and ground water classifications, standards and
limitations, (Title 6, Official Compilation of Codes, Rules and Regulations,
Part 701 and 703) for the protection of the State's water resources and
compliance with promulgated federal best management practices (BMPs)
regulations (40 CFR, Part 125) designed for the protection of said resources,
the following requirements will apply.
I. New Coal Piles
A. Site Considerations
The control of wastewater containing runoff and leachate from new
coal piles shall be as follows:
1) The site shall be designed to prevent runoff from entering the
pile.
2) The coal pile shall be placed on an impervious surface. This
surface may be a liner or any in-place material which meets
Department criteria for impermeability.
3) Liner material shall be selected so as not to deteriorate after
contact with coal pile wastewater and shall be protected from
puncture from operating equipment. Liner protection and
repairability shall be evaluated in the liner design.
4) Wastewater shall be collected and treated.
5) Collection ponds, if used, shall be lined with an imperivous
material, and designed to contain the runoff from the ten-year,
24-hour storm.
6) A ground water monitoring system may be required to evaluate
and/or monitor the integrity of the facility's impervious lining.
20-2
-------
7) Discharge from treatment facilities to surface waters are not to
contravene Part 701, (see Table 2) 'Classifications and
Standards of Quality and Purity" and discharge to the ground are
to meet Part 703, (see Table 3) "Ground Water Classifications
Quality Standards and Effluent Standards and/or Limitation".
8) Coal pile wastewater may be:
a) Treated separately;
b) Recycled for other uses at the facility;
c) Combined with other waste streams for treatment except for
the purposed of dilution.
B. Permit Conditions (SPDES)
1) The permittee shall prepare an engineering report and final
plans for construction of facilitites designed to prevent,
control, and treat runoff and leachate from coal piles for the
purposes of maintaining and protecting water quality standards
and compliance with any applicable federal requirements.
2) Organic and inorganic substances will be limited and require
treatment, depending on the coal pile characteristics. Most
heavy metals may be found in coal pile wastewater. Organic
substances which have been detected in coal pile wastewater
include chlorobenzene, dichlorethylene, chloroform, pthalates,
and methylene chloride. Monitoring and treatment requirements
for these compounds will be evaluated on a case and site
specific basis because wastewater characteristics are dependent
on the characteristics of the coal used and discharge objectives
relating to the receiving water.
II. Existing Coal Piles
A. Site Considerations
Existing coal piles will be evaluated on a case-by-case basis
utilizing new coal pile considerations. The permittee may be
required to undertake a ground and/or surface water monitoring
program to determine the extent of possible contamination. As a
minimum, ground water monitoring and treatment of contaminated runoff
will be required. Ground water monitoring results will be evaluated
to determine applicable new coal pile considerations.
20-3
-------
B. Permit Conditions (SPDES)
Specific permit conditions may include the following:
1) Permittee to submit a proposed plan of study designed to assess
coal pile operations at the applicant's facility, including a
study program to evaluate the operation's effect on surface and
ground water quality.
The study program would be subject to Department approval and
shall consist of a ground and surface water monitoring program
to determine compliance with applicable water quality standards,
and monitoring of coal pile runoff and leachate to determine and
assess wastewater characteristics.
2) Based upon NYSDEC review of the study program, the permittee may
be required to prepare an engineering report and final plans foe
construction of facilities designed to prevent, control, and
treat runoff and leachate from coal piles for the purposes of
maintaining and protecting water quality standards and
compliance with any applicable federal requirements.
III. Imposition of limitations and monitoring requirements and/or any partic-
ular policy guidelines if deemed appropriate for the protection of the
environment may precede any of the above considerations for either new oc
existing coal piles.
Taken from: Bureau of Industrial Program
Division of Water
Dates: April 14, 1980
It becomes immediately evident that compliance with these regulations, from the
engineering design study to placement of the liner material and monitoring of
its effectiveness, can be fairly costly. NYSERDA sponsored the R&D project
presented today with specific intent of better understanding the hydrology of
coal piles through extensive data collection and verifying and calibrating a
coal pile drainage model, in order to potentially decrease industries cost of
compliance with these guidelines.
I hope that the information presented today will prove useful in future efforts
to comply with these guidelines and to obtain and renew SPDES permits.
20-4
-------
Calibration and Verification of a Coal Pile Drainage Model
John G. Holsapple, New York Power Pool
Gordon T. Brookman, John A. Rlpp and
Pamela B. Katz, TRC-Environmental
Consultants, Inc.
Abstract
In 1980 the New York State Department of Environmental Conservation (NYSDEC)
published a document entitled "Coal Pile Guidance for SPDES" which lists a number
of requirements for new and existing coal piles. These requirements emphasize the
monitoring of possible toxic pollutants emitted from coal piles and the treatment
facilities required to control leachate contamination of ground water and runoff
contamination of surface waters. Specifically, the New York State regulations
require natural and/or manmade liners under new and some existing coal piles, ground
water monitoring, storm collection ponds to contain runoff from the ten-year 24-
hour storm, and treatment facilities for pH and most heavy metals.
Because of these new requirements, New York State utilities must plan for the
most efficient coal pile drainage treatment and avoid the unneeded expense of
constructing new coal pile runoff storage capacity to collect drainage that does
not require treatment.
This paper will present a summary of a research study being performed by
TRC-Environmental Consultants for ESEERCO (Empire State Electric Energy Research
Corporation) which has the following objectives:
1) The calibration and verification of a mathematical
model to simulate coal pile drainage flows and pollutant
loadings under historical and design storm conditions.
The model will be used to size and design treatment
systems to meet NYSDEC specifications. The model will
also simulate the changes in coal pile runoff due to
changes in coal source and coal characteristics.
2) The development of a data bank on coal pile drainage
from two plants which will be useful in characterizing
which pollutant parameters appear in significant
quantities.
The presentation will include a detailed description of the sampling programs
at the two plants and an overview of the mathematical model Including its uses and
its limitations.
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
21-1
-------
INTRODUCTION
The United States Environmental Protection Agency (EPA) presently requires
that area runoff from coal storage piles at steam electric plants be controlled
to prevent the discharge of acidic water and total suspended solids (TSS).
Accepted practice dictates that storage/treatment be designed to accommodate
all of the runoff from a 10-year, 24-hour storm. Also, TSS must not exceed
50 mg/1 and the discharge pH must fall within the range 6.0 to 9.0.
Most state water pollution regulatory agencies have adopted the EPA limi-
tations for permitting purposes. However, some states, including New York, are
requiring more stringent controls for coal pile runoff. The New York State
Department of Environmental Conservation's (NYSDEC) "Coal Pile Guidance for
SPDES" requires natural and/or man-made liners under new and some existing coal
piles, as well as groundwater monitoring, and 10-year, 24-hour storm runoff collection
ponds and treatment facilities for pH, solids and several trace metals. New York
utilities must submit an engineering report with their SPDES permit applications.
The report must contain sufficient field data to show what pollutants are being
emitted in existing coal pile drainage and the means to predict quantitative and
qualitative characteristics for new coal piles. Surprisingly, there is little
information on the characteristics of coal pile runoff. There is an incentive
and a need, then, in New York State to develop a technique to accurately charac-
terize coal pile drainage.
In June of 1981, TRC Environmental Consultants, Inc. initiated a research
project for the Empire State Electric Energy Research Corporation (ESEERCO)
involving the calibration and verification of a coal pile drainage model. The
21-2
-------
project is also partially funded by the New York State Energy Research and
Development Authority (NYSERDA). The objectives of this study include 1)
field testing and verification of a coal pile drainage model and 2) the develop-
ment of a large data base on coal pile runoff and coal pile characteristics.
The scope of work, involves sampling at two sites. Sampling at Site 1 was com-
pleted in the fall of 1981 at the Greenidge Station of New York State Electric
and Gas Corporation. Site 2 sampling is presently being performed at the
Russell Station of Rochester Gas and Electric Corporation.
Upon completion of this project in late summer/early fall of 1982, TB.C
will provide ESEERCO with a calibrated and verified model to simulate coal
pile drainage flows, pollutant loadings under historical and design storm
conditions, and a model user's manual. The model will be a useful tool for
sizing and designing coal pile drainage treatment systems. It will also be
useful for simulating changes in coal pile runoff due to changes in the coal
source and in coal characteristics.
This paper presents an overall of the project to date, including the site
selection process, a detailed discussion on the sampling program, a description
of the model and the calibration and verification process.
SITE SELECTION
In 1981, there were nine major utility plants burning coal in New York
State. As part of the site selection process, TRC visited each plant and
developed a subjective ranking system, which included the following factors:
- coal pile size
- percentage of sulfur in coal
- coal pile containment
- ease of sampling on pile
- variability of pile shape and size
- climate
- availability of data 21-3
-------
The Russell Station (Site 2) was the highest ranked plant for the following
reasons:
- The plant has an average size of 96,000 tons, which is near the average
size of the piles evaluated;
- The average sulfur content is 2.4 percent, which was one of the highest
of all plants evaluated;
- The plant has a partial concrete-lined gutter around the pile which
collects most of the runoff;
- Approximately 25 percent of the pile is reserve, which facilitates the
installation of in-pile piezometers;
- The pile size and shape remains fairly constant throughout the year.
The Greenidge Station (Site 1) ranked second for the following reasons:
- The plant has an average size of 110,000 tons, which approximates the
survey average of 112,000 tons.
- Runoff and leachate are completely contained by an impervious plastic
liner under the pile. All runoff and leachate are directed to a single
discharge point.
- Approximately 50 percent of the pile is reserve and, therefore, facili-
tates the placing of in-pile monitoring equipment.
- Size and shape of the pile does not vary significantly during the year.
•These two sites were then recommended to ESEERCO as test sites and permission
was received from the member utilities to proceed.
21-4
-------
FIELD MONITORING PROGRAM
The field monitoring program was designed co provide input data to calibrate
and verify the drainage model and to collect an extensive data base on pile
hydrology and chemistry along with site meteorology and runoff quantity and
quality. The scope of work for the field program was based on a TRC report to
EPA and EEI entitled "Planning Study to Model and
Monitor Coal Pile Runoff."1
DATA REQUIREMENTS
Figure 1 illustrates the categories of data necessary to the modeling
effort and the data base. Site description data includes general information
on pile size and shape, proximate analysis of delivered coal, methods of stacking
and reclaiming and pertinent existing data on site geohydrology. Meteorological-
data being collected includes historic information from local U.S. Weather Service
stations and continuous monitoring of precipitation, air temperature, relative
i
humidity, solar radiation and pan evaporation during the survey period. Runoff
data being measured includes continuous recording of all drainage flows from the
coal pile and when flow does exist continuous measurements of pH and conductivity.
For each storm event samples are collected on a 30-minute interval a't the start
of each storm and after 4 hours of runoff on a 60-minute interval for the fol-
lowing analysis:
- Acidity/alkalinity
- Total suspended solids
- Total dissolved solids
- Total organic carbon
G. T. Brookman, at al., "Planning Study to Model and Monitor Coal Pile
Runoff," EFA-600/7-81-016 (NTIS PB81-152530), February 1981.
21-5
-------
SITE DESCRIPTION UAIA
GFNERAL PI ANT
INFORMATION
COAL PILE(S)
DATA (PHYSICAL)
COAL(S)
DATA (CHEMICAL)
GROUND MATER
DATA
II
METEOIIOLOGICAL DATA
to
i—j
0>
1
PRECIPIIATION
HISTORICAL DATA
(3 MONTHS PRIOR
TO FIEIO PROGRAM)
1 II
AIR
flMI'ERAIURE
SOIAR
RADIATION
EVAPORATION
1
PRECIPITATION
FIELD DATA
(DURING FIELD
PROGRAM)
1
I
IEMPU'MUR.
III 1 RUNOFF DAI A
1
1 1
SOl Alt 1
j i.nn f WAD All AT 1 fUJ 1
HAD 1 AT 1 ON Vnr UHA 1 1 Uff 1
*
DRY DAYS (INCLUDE STORM
EVENTS WITH
-------
- Sulfate
- Hardness
- Total and dissolved forms of 20 metals
Not illustrated in Figure 1 is the geohydrological data being collected at each
site to define pile moisture conditions. At the start of each site monitoring,
a drilling program is conducted on the pile. During that program, hollow stem
auger borings are made with split spoon sampling through the pile at 5-foot
intervals. Coal samples are then analyzed for moisture content. Where pos-
sible, in-pile piezometers are installed to observe water level fluctuations at
the pile base. Simultaneous with the drilling program a resistivity survey is
conducted on the pile to map the saturated areas of the pile. Throughout the
10 weeks of runoff monitoring, piezometers are observed for water levels, pH,
conductivity and periodically sampled for the runoff parameters. Pile perme-
ability and infiltration rates are also measured throughout each monitoring
season to correlate with runoff percentages.
LOGISTICAL CONSIDERATIONS
As mentioned above, 2 sites are being monitored for 10 weeks each. The
field program is built around a portable laboratory in which solids and acidity
analysis are performed on all runoff samples. Also this laboratory serves as
of base of operations for calibrating instruments, preserving and storing samples
and cleaning equipment and glassware. An instrument shelter is also located on
each site to house recorders for flow, pfl and conductivity, along with the auto-
matic sampling equipment. The samplers are programmed to start collecting
samples at the start of each storm through an actuator switch which senses an
increase in runoff flow. This insures the complete sampling of each storm event.
A technical support person is kept on-site during the entire monitoring period
to maintain all equipment and perform on-site analysis.
21-7
-------
RESULTS TO DATE
The monitoring program at Site 1 has been completed and a total of 4
runoff events were measured. Storms ranging in size from 0.19 inches to 1.8
inches of rain were measured. A series of hyetograph/hydrographs is plotted
for the 4 storms and drainage coefficients were determined. As illustrated in
Figures 2 to 6, drainage coefficients never went higher than 39 percent. This
indicates that at least 60 percent or more of the rain hitting this pile was
absorbed by the coal and lost to either evaporation, pyrite oxidation or base
flow seepage. Two observations made during the monitoring at this site lend
to the fact that the pile is acting like a large sponge. First, the pile was
continuously growing with new coal from 38,000 tons to 120,000 tons and second,
the moisture of the pre-fired coal was significantly greater than that of the
delivered coal. These measured drainage coefficients are substantially lower than the
reported in literature, i.e., 70 to 80 percent.
Most of the chemical analysis data has not been finalized to date, but a
number of obvious observations can be made concerning the quality of the runoff.
Briefly they include:
1. The pH of both storm runoff and base flow seepage did not vary. All
pH's were in the 2.0 to 2.4 range.
2. The conductivity of the runoff was depressed during storm events, as
low as 4,000 micromhos per centimeter (umho/cm) and during base flow
seepage remained at a higher concentration, 8,000to 9,000 umho/cm.
3. Conductivities correlated well with dissolved solids.
4. Because of the low pH, most of the metals found in the runoff were in
a dissolved form rather than suspended.
5. Analysis of metals data during storm events can indicate the interflow
hydrograph.
The completed results of the data collected at Site 1 will be included in the
final report to ESEERCO.
21-8
-------
to
I—1
CO
o
P 0.1
£L Z 0.2
is
10
a
6
4
2
TOTAL PRECIP: 0.19 IN.
DRAINAGE COEFFICIENT: 0.06
SITE: G STATION
DATE: SEPT. 8, 1981
i—i—r
10 12
NOON
HYETOGRAPH/HYDROGRAPH
STORM EVENT 1
FIGURE 2
-------
Z
O
0.1
t ~i
Q. Z 0.2
O ~
g 0.3
Q.
60
40
30
20
10 -
I
TOTAL PRECIP: 0.61 IN.
DRAINAGE COEFFICIENT: 0.06
SITE: Q STATION
DATE: SEPT. 14, 1082
10 12
NOON
HYETOQRAPH/HYDROGRAPH
STORM EVENT 2
FIGURE 3
10
-------
DO
t—i
z
o
t
D.
O
UJ
oc
o.
*•: o.i
g
70
60
60
40-
30
20 -
10 -
TOTAL PRECIP: O.08IN.
DRAINAGE COEFFICIENT: 0.12
SITE: Q STATION
DATE: SEPT. 21, 1081
6
0
i i
10
i i i i i
4 6
10
MIDNIGHT
HYETOGRAPH/HYDROGRAPH
STORM EVENT 3
i i
12
NOON
T
FIGURIL
-------
N3
I—i
I
z
o
0.1
CL Z 0.2
O ~
UJ o o
tE O.3
O.
30O
g£
n!2
200
100
TOTAL PRECIP: 1.6 IN.
DRAINAGE COEFFICIENT: 0.29
SITE: G STATION
DATE: OCT. 27, 1982
1 4
2
• 1 i 1 * .1 i i
6 a 10 12
NOON
HYETOGRAPH/ HYDROGRAPH
STORM EVENT 4 - 1st PHASE
FIGURE 5
-------
o
UJ
DC
Q.
to
(—i
t—i
CO
6OO
50O
£$400
^& VM
"• ^-
3OO
2OO
10O
12
T I I
24
6
8
I I I I I I
1O MID- 2
NITE
HYETOGRAPH/HYDROGRAPH
STORM 4 2nd PHASE
FIGURE 6
-------
MODEL DEVELOPMENT
A 1976 T5.C study of nonpoint source pollution at a utility site in Penn-
sylvania attempted to address sheet washoff from coal storage piles using the
SWMM/RECEIV II model. This model was eventually thought deficient in the sim-
ulation of coal pile runoff because it did not take into account:
1. Erosion of coal from the pile surface,
2. Percolation of storm-water through the pile,
3. Pyrite oxidation/acid production in the coal pile.
In 1979, a follow-up study was conducted by TRC to determine what physical
and chemical phenomena associated with coal pile runoff needed to be addressed
so that the model could be used as a prediction tool. A number of phenomena
were identified as characteristic of runoff and should, therefore, be included
in the model. They are:
1. Precipitation in the form of rain and snow,
2. Surface runoff and.infiltration through the pile,
3. Moisture content of the pile,
4. Snowmelt from piles subject to heavy winter accumulations,
5. Groundwater infiltration through the base of the pile,
6. Pyrite oxidation and acid production,
7- Freeze/thaw cycles,
8. Gully erosion along the sides of the pile,
9. Washoff of coal from the pile surface.
A comprehensive search of existing models found that the Ohio State University
(OSU) version of the Stanford Watershed Model was a good basis for a coal pile
runoff model. With extensive modification of the OSU model, the new model
treats the coal storage pile as a small unvegetated watershed with many features
similar to larger watersheds. A TRC coal pile runoff model was then developed
and consists of two major components: the hydraulic and the qualitative.
21-14
-------
BASIC FUNCTION OF HYDRAULIC COMPONENT
A diagram of the hydrologic c-ycle of a coal pile is shown, in Figure 7.
The model reads hourly precipitation data in the form of rain or snow input on
a meteorological data tape or cards. A continuous water balance simulation in
the model uses precipitation, pile moisture content and pile water storage as
additions and surface runoff, interflow and base seepage as reductions. The
hydraulic component of the model also predicts pile erosion caused by the gullying
of the pile slopes during intense storms.
Hydraulic data are presently being measured in New York State at a utility
coal pile for the calibration of the hydraulic model.
HIGHLIGHTS OF A WORKING HYDRAULIC MODEL
Once the hydraulic model has been calibrated and verified using data from
several coal pile monitoring programs, it will become a working design tool.
Input ^111 consists of two data sets: 1) site specific data such as volume and
acreage of the coal, average moisture content of the delivered coal and pile
compaction data and 2) meteorological data either in the form of historical
weather data vailable from the National Climatic Center (NCC) or local on-site
measurements taken by the owner of the coal pile. Site-specific data are
entered into the model by card deck while the historical weather data are
entered on magnetic tape.
The working model lists daily, monthly and annual runoff flows plus the
optional output of selected precipitation events with hourly or subhourly data.
Optional output from the hydraulic model are also available. The output
are described below.
BASIC OUTPUT OPTIONAL OUTPUT
1. Table of average daily rainfall. 1. All input data.
21-15
-------
UPPER ZONE STORAGE -
DEPRESSION STORAGE
to
RUNOFF ^
STREAM
DEPRESSION
STORAGE
LOWER ZONE
STORAGE
-S-^WATER TARIE
• .
GROUND WATER
RUNOFF
STREAM
TO DEEP STORAGE
HYDROLOGICAL CYCLE OF COAL STORAGE PILE
FIGURE 7
-------
BASIC OUTPUT
2. Summation of daily runoff rates
for each month.
3. Monthly and annual totals of
runoff, seepage and total flow.
4. Monthly and annual total of
frozen precipitation.
5. End-of-month values of pile,
surface and ground-water moisture.
6. End-of-month values for snowpack.
7. Annual balance of unaccounted
moisture.
OPTIONAL OUTPUT
2. Echo of recorded runoff (when
measured) as a comparison with
simulated flows.
3. Daily values of snowpack, snow-
melt and snowfall.
4. Average daily temperatures.
5. Number of freeze/thaw cycles
6. Simulated gully erosion
7. Plots the hyetograph and hydro-
graph for selected storms.
BASIC FUNCTION OF THE QUALITATIVE COMPONENT OF THE MODEL
Stored coal exposed to the atmosphere undergoes the physical and chemical
process of pyrite oxidation. Products of these reactions are acid, iron and
sulfate. The acid further reacts with the coal by dissolving trace metals and
elements. During wet weather, these materials are washed off the surface and
moved through the interior of the coal pile. Runoff during storm events and
seepage from the pile base during wet and dry weather carry these products of the
pyrite oxidation to produce a Ughly acidic wastewater. The qualitative component
of the model simulates these reactions; using the water balance produced in the
hydraulic model, it will predict the loading of acid, sulfates and trace elements
in the runoff.
The model looks at both dry weather reactions and the transport of the
products of oxidation during wet weather. For dry weather, the model simulates
the total pollutants in the coal and those in a dissolved state available for
transport. For wet weather periods, the model simulates the distribution of a
component of the dissolved metals, sulfates and acid from the surface of the
pile to the interior and directly to surface runoff. In addition, the removal
21-17
-------
of materials from the interior of the pile by dry weather seepage is simulated.
The results of the qualitative model will be a calculation of daily, monthly and
totals of acid, metals and sulfates.
HIGHLIGHTS OF A WORKING QUALITATIVE MODEL
Once the qualitative component of the model has been calibrated and veri-
fied, it will be available for use. Input to this model are from tvo sources:
1. A magnetic tape created by the hydraulic model containing data on
rainfall, transport routes, flow rates from seepage and runoff and
average daily air temperature.
2. Card input describing factors of pyrite oxidation including: percent
of pyrite in the coal, density of the coal pile, amount of trace metals
in the coal and the solubility of sulfate and iron compounds.
As data from the ongoing field program are evaluated, the oxidation reaction
rates will be adjusted. Output of the working qualitative model will be:
1. A calendar for any selected year showing dry versus wet days,
2. A calendar listing daily runoff flow volumes,
3. Daily total loadings of acid, sulfate, iron and trace metals,
4. Optional plots on a detailed storm basis showing the relationships
of precipitation, flows and pollutant loadings.
CALIBRATION EFFORT TO DATE
The calibration effort being completed has found that the model can easily
simulate the water balance of large storm but usually overpredicts the runoff
volume from short intense storms. In addition, gully erosion, which is very
sensitive to the intensity of precipitation, is difficult to simulate using
precipitation amounts averaged over an hourly period.
21-18
-------
SUMMARY
The ultimate purpose of simulating coal pile drainage is for use in the
design of systems to treat acid drainage before discharge. Industries storing
coal may base treatment system design on short-term monitoring programs; however,
that data may not have any relationship to average or worst-case runoff condi-
tions. The runoff model will allow an industry to simulate several runoff
conditions and thereby have a better basis for design.
The current coal pile runoff program in New York State has two objects:
1. To provide detailed information on meteorological, coal storage, pile
moisture, runoff quality and hydrologic conditions to the developing
model so that key coefficients may be adjusted;
2. To provide approximately 10 weeks of continuous data from each of two
coal pile.monitoring sites as a data bank useful for comparative
studies and treatment design.
One important issue to be resolved during this program is whether a runoff
condition occurs producing a discharge "clean" enough to bypass treatment.
Runoff data collected at two utility sites in Pennsylvania in 1976 indicates
that a first flush did indeed occur. However, initial data collected to date
showed that runoff acidity and sulfate were very high during all phases of the
storm and during dry weather seepage. The analysis of storms of greater volume
may help resolve this first flush issue.
The field data collection programs are to be completed in May of 1982 with
the runoff model completed in September of 1982.
21-19
-------
*« U'5' Environmental
Agency and n0 oici i-et "" "™ oi
-------
INTRODUCTION
Various approaches for characterizing coal pile leachate and runoff have
been used by other researchers; these approaches have included sprinkling
pans of coal with a rainfall simulator (7), recirculating distilled water
over an inclined box packed with sample (4), and collecting actual
leachate and runoff on site (2). As the methodologies and coals used in
different tests varied, so did the results.
One general conclusion of a study (5) concerning the treatability of coal
pile leachate and runoff was that heavy metals and other trace elements in
leachate were direct functions of acidity. As acidity increased, so did
the concentration of heavy metals and trace elements in coal pile leachate
and runoff. Verification of this conclusion was an important objective of
the study reported herein.
Generally, it would be preferable to collect samples from actual
facilities in order to assess the characteristics of runoff and leachate.
However, because the coal pile under consideration did not yet exist, it
was simulated by the test methodology described in this paper. These
efforts were designed to represent actual conditions as closely as
possible, including actual dimensions and design density of the full-scale
coal pile.
This study was conducted for a facility converting to coal as a primary
fuel source. An on-site coal storage pile was therefore required. A coal
pile simulation test was conducted to determine the quality of leachate
and runoff resultant from rainfall contacting the facility's coal pile.
Testing was conducted on four types of coal being considered as fuel
sources: Coals 1, 2, 3, and 4. The simulated leachate and runoff were
compared to applicable effluent criteria.
TEST METHODOLOGY
A coal pile simulation apparatus was built to the same scale as the
proposed coal pile: 10 ft (3 m) in height, 14-ft (4-m) sides sloped
45 degrees to the horizontal. The test apparatus, shown in Fig. 1,
consisted of vertical columns and open-top, sloped troughs. Only
polyvinyl chloride (PVC) pipe, rubber hoses, fiberglass netting, plastic
duct tape, nylon cord, and wood were used to prevent metal contamination,
since heavy metals often are the primary concern in coal pile runoff and
leachate. Ten-ft (3-ra) high vertical columns, constructed of
4-in. (102-mm) diameter PVC pipe and packed with coal at the design
density, were intended to simulate rainfall passing through the entire
10-ft (3-m) depth of the coal pile. Fourteen-ft (4-m) long troughs,
constructed of 10-in. (254-mm) diameter PVC pipe cut in half and packed
with coal at the design density of 55 pounds per cubic foot (Ib/cu ft)
(881 kg/m-3), were designed to simulate the side slopes of the coal
pile. Thus, the troughs simulated runoff, the columns leachate.
22-2
-------
SOURCE: ENVIRONMENTAL SCIENCE AND ENGINEERING, INC., 1980
Rgure 1
COAL PILE SIMULATION APPARATUS
Coal Pile Simulation Study
22-3
-------
Simulated runoff was drained by two hoses at the end of the trough. The
troughs were under-drained at a coal depth of 3 to 3.5 in. (76 to 89 nun)
to remove the leachate, as depicted in Fig. 2. If the troughs had not
been under-drained, water which infiltrated into the coal pile would have
continued contacting the coal as it drained along the bottom of the trough
and contributed to the volume of runoff. The under-drains also allowed
separation of the volume of water which would infiltrate into a coal pile
from the volume of water defined as runoff. These trough leachate samples
were collected and analyzed to provide additional data describing runoff
characteristics.
Prior to testing, the coal samples were passed through a 2.25-in. (57-mm)
sieve to ensure that the tested coal met the size specifications of the
actual coal to be stored in the pile. The coal was then packed into both
the PVC columns and the PVC troughs at the full-scale design density and
secured with fiberglass netting, plastic duct tape, and nylon cord.
Small amounts of certain inorganic salts were added to raw ground water to
approximate the characteristics of site-specific rainfall. An average
site-specific rainfall event of 0.66 in. (17 mm) was selected as the
rainfall volume to be applied for the simulation. The synthetic rainfall
was distributed over the PVC troughs via a sprinkling system consisting of
plastic lawn sprinkler hoses. The majority of the fiberglass netting
securing the coal was removed after initial wetting had occurred and the
coal had settled into the trough. Simulated runoff and trough leachate
were collected in separate glass containers, as shown in Fig. 3. Several
rainfall gauges placed under the apparatus were used to measure the amount
of applied synthetic rain water falling on the coal pile apparatus.
Sprinkling was continued until the average reading of these gauges was
0.66 in. (17 mm)-
To generate simulated leachate from the vertical PVC columns, a measured
volume of synthetic rain water was poured on the coal at the top of the
columns until leachate appeared at the bottom. All resulting simulated
leachate was collected in a glass container. Details of leachate
generation are shown in Fig. 4.
Each coal type produced three streams: runoff, trough leachate, and
column leachate. Each sample was analyzed for various parameters
addressed in applicable effluent criteria. All analyses were performed
according to United States Environmental Protection Agency (EPA)
Method 60014-79-020 and Standard Methods for Analysis of Water and
Wastewater (1).
RESULTS
Table 1 presents the analytical results of the coal pile simulation. The
applicable effluent criteria are also presented for comparison.
No consistent difference in concentration was observed between simulated
runoff and simulated leachate from the four coal types in the test
apparatus. The rainfall event applied to the troughs and the amount of
synthetic rain water applied to the columns to obtain leachate resulted in
22-4
-------
FRONT VIEW
RUBBER
DRAIN HOSES
COAL
PVC PIPE
PERFORATED
PVC DRAIN
SIDE VIEW
LEACHATE
COLLECTION'
RUNOFF
COLLECTION
SOURCE: ENVIRONMENTAL SCIENCE AND ENGINEERING, INC., 1980.
Figure 2
DETAILS OF TROUGH
Coal Pile Simulation Study
22-5
-------
SOURCE: ENVIRONMENTAL SCIENCE AND ENGINEERING, INC., 1980.
Figure 3
TROUGH UNDERDRAINS AND
RUNOFF COLLECTION
Coal Pile Simulation Study
22-6
-------
SOURCE: ENVIRONMENTAL SCIENCE AND ENGINEERING, INC., 1980.
Figure 4
COLUMN LEACHATE GENERATION
AND COLLECTION
Coal Pile Simulation Study
22-7
-------
Table 1. Results of Goal Pile Simitation Test and Applicable Effluent
Standards (all values ng/L except pH)
Parameter
(1)
PH
fj Arsenic
to
' Total
CO
Chromium
lead
Mercury
Chlorides
Copper
Iron
Zinc
Effluent
Standard
(2)
6.0-8.5
0.05
1.0
0.05
H/D
500.0
0.5
0.3
1.0
Synthetic
Rain Hater
(3)
6.6
H/D
H/D
0.002
H/D
81.0
H/D
H/D
0.188
Coal Type 1
TO
(4)
6.3
H/D
0.003
H/D
H/D
116.0
H/D
H/D
0.016
CL
(5)
6.4
0.013
0.004
0.002
H/D
99.0
H/D
0.204
0.016
TL
(6)
6.5
0.01
0.004
0.002
H/D
124.0
H/D
0.204
0.024
SAMPLE nENTIFlGATICN
Coal Type 2 Coal Type 3
TO
(7)
6.1
H/D
0.004
0.012
H/D
109.0
H/D
H/D
0.013
a
(8)
6.2
H/D
0.002
H/D
H/D
92.0
H/D
H/D
0.01
TL
(9)
6.5
H/D
0.002
0.002
H/D
106.0
H/D
H/D
0.016
TO
(10)
6.1
H/D
0.002
H/D
H/D
103.0
H/D
H/D
0.013
a
(11)
6.2
H/D
H/D
H/D
H/D
88.0
H/D
H/D
0.013
TL
(12)
6.5
H/D
0.001
0.002
H/D
114.0
H/D
H/D
0.013
Coal Type 4
TO
(13)
2.7
0.221
0.068
0.132
H/D
109.0
0.687
361.0
4.52
CL
(14)
2.8
0.114
0.048
0.018
H/D
90.0
0.395
190.0
6.04
TL
(15)
2.8
0.131
0.040
0.406
H/D
96.0
0.469
214.0
4.93
TO • Trough simulated runoff
CL * Colum simulated leachate
TL " Trough simjlated leachate
N/D - Hone detectable
NOTE: 1 mg/L - 1 ppra.
Source: ESE. 1980.
-------
approximately the same solid-to-liquid ratio. -This ratio was 5.6:1 for
the troughs and 4.5:1 to 5.8:1 for the columns.
Samples of simulated leachate and runoff from Coal 4 in the test apparatus
exceeded the criteria for several parameters. Moreover, the simulated
leachate and runoff resulting from this coal were highly acidic, with pH
values between 2.7 and 2.8.
SUMMARY AND CONCLUSIONS
Coal pile simulation tests were performed on four coal types in order to
predict leachate and runoff quality. Samples generated from the testing
were compared to applicable effluent criteria. One of the four coal types
(Coal Type 4) tested resulted in simulated runoff and leachate which
consistently violated several criteria. The other three coal leachates
met all effluent criteria.
The testing agreed with previous research, showing that runoff and
leachate with low pH values could be expected to have high concentrations
of heavy metals. The testing also demonstrated a procedure for use in
evaluating alternative coal supplies by assessing their pollution
potential.
22-9
-------
CONTROL OF ACID PROBLEMS IN DRAINAGE
FROM COAL STORAGE PILES
by
Harvey Olem, Tracey L. Bell, and Jeffrey J. Longaker
Office of Natural Resources
Tennessee Valley Authority
Chattanooga, TN 37401
The work described in this paper was not funded by the U.S. Environmental
Protection Agency. The contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
23-1
-------
Abstract
A method has been identified for controlling acid production and sub-
sequent dissolution of toxic pollutants in drainage from coal storage
piles. Results of laboratory and field experiments indicate that it may
be possible to prevent, rather than treat, acid drainage by applying an
environmentally safe detergent formulation periodically to the coal. A
mild solution of sodium lauryl sulfate (SLS) was found to effectively
block the activity of the bacteria that promote acid formation and chemi-
cal leaching. Drainage from coal treated once with 50 mg/L of SLS remained
neutral for 60 days, about three times longer than the untreated control
sample. Extrapolating results to an industrial-scale application revealed
that the cost of the SLS needed for a single application might be as high
as $500 per hectare of coal storage area ($200 per acre).
Introduction
As more and more utilities and industries throughout the United
States shift from the consumption of oil and natural gas to the more
abundant supplies of coal, there should be an increase in onsite stor-
age of the coal. It has been projected that by 1985 the amount of coal
in storage in the United States will increase from its current level of
128 million metric tons (1978 estimate) to 227 million metric tons. By
the year 2000, the total coal stockpiled could reach 680 million metric
tons (1). This projected increase in coal storage has generated renewed
interest in the environmental effects of contaminated drainage from coal
storage facilities.
23-2
-------
The most efficient method currently used for storing large quanti-
ties of coal is placement on the ground. Because the coal is exposed to
the elements, rainwater falling on the pile can become contaminated by
the action of chemolithotrophic bacteria on pyritic materials (usually
iron disulfides). This occurs by the same series of reactions that are
known to produce acid drainage from coal mines. Consequently, pollution
of both surface and ground water is usually the major problem associated
with storage of this coal.
Drainage from coal storage areas generally contains high concentra-
tions of metallic ions (aluminum, iron, manganese, mercury, nickel, zinc,
etc.), sulfates, and suspended solids. It also usually has an extremely
low pH (pH 2 to 3). Currently this drainage is typically collected in
a central pond and treated, usually by addition of hydrated lime.
This paper presents results of laboratory experiments which indicate
that it may be possible to prevent, rather than treat, acid drainage by
applying an environmentally safe detergent formulation periodically to
the coal. Laboratory experiments consisted of applying two anionic deter-
gents to columns filled with coal which had been rinsed with water until
the pH of drainage was neutral. This was done to simulate freshly mined
coal or more nearly, coal that had been processed by coal washing. A
third column was filled with coal and did not receive any detergents as
a control. Due to the limited number of tests performed to date, the
results are considered preliminary.
Background
One of the most common and most troublesome impurities in coal is a
metallic sulfide, iron pyrite. When mining exposes pyrite (FeS2) to water
and oxygen, a chain of chemical reactions begins.
23-3
-------
2FeS2 + 702 + 2H20 = 2Fe2+ + 4S042" + 4H+ (D
4Fe2+ + 02 + 4H+ = 4Fe3+ + 2H20 (2)
Fe3+ + 3H20 = Fe(OH)3 + 3H+ (3)
3+ 2+ 2" +
FeS2 + l4Fe+ + 8H20 = 15Fe + 2S04" + 16H (4)
2+
In reaction 1 pyrite is oxidized to produce ferrous iron (Fe ) , sul-
fate (SO,2"), and some acid (H+) . The ferrous iron can then be oxidized
3+
with this acid and some oxygen (reaction 2) to form ferric ions (Fe ).
Under certain conditions, the ferric irons will hydrolyze (reaction 3)
to form the yellow-orange ferric hydroxide precipitate (Fe(OH),j.) that
typically appears in acid mine drainage streams. Ferric iron can also
react with more pyrite (reaction 4) to form more ferrous iron and acidity.
These reactions produce acid at a constant rate.
Acid drainage would develop very slowly if this strictly chemical
chain of events were not greatly accelerated by the biochemical action
of certain bacteria. First isolated in 1947, Thiobacillus ferrooxidans
were then believed to play some role in acid mine drainage formation, but
their exact role was not well understood (1). More recently, Singer and
Stumm (2) showed that direct oxidation of pyrite by oxygen (reaction 1)
was too slow to generate the amount of acidity observed in nature and con-
cluded that the oxidation of ferrous iron to ferric (reaction 2) created
a cycle whereby the ferric ions produced in this reaction directly attack
pyrite to form considerable amounts of acid (reaction 4). Reaction 2,
however, is extremely slow chemically under acid conditions. The half
time for spontaneous ferrous iron oxidation in a sterile solution of
pH 3.5 has been estimated to be 2,000 days (3). In the presence of
T. ferrooxidans, however, this reaction was shown to accelerate by a
23-4
-------
factor of 10 , thus producing significant quantities of ferric iron.
The subsequent oxidation of pyrite by these ferric ions produces more
ferrous ions to allow the cycle to continue, producing more and more
acidity until all the reactive pyrite is leached.
Bacterial Inhibition
It is generally considered preferable to prevent the formation of
pollutants through at-source control rather than provide continuous treat-
ment of the resulting contaminated water. It is with this objective in
mind that bacterial inhibition of acid formation was investigated to pre-
vent the contamination of rainfall-runoff draining through coal piles.
Anionic detergents, while normally considered as cleansers rather than
bactericides, do possess bactericidal properties. Dugan (4) found that
anionic detergents can effectively stop iron oxidation by T. ferrooxidans
at concentrations as low as 2 mg/L. Kleinmann (5,6) demonstrated the
•*
effectiveness of one one type of detergent, sodium lauryl sulfate, in
reducing acid formation in coal refuse piles and surface coal mines at
somewhat higher detergent concentrations. The reason for this inhibition
is not yet clear, but evidence indicates that the semipermeable properties
of the cytoplasmic membrane is altered so that H is allowed to seep into
the .normally neutral interior of the cell (7). Thus, the bacteria are
appropriately attacked by the acid they helped produce.
In this study, two types of anionic detergents were evaluated, sodium
lauryl sulfate (SLS) and neutralized benzene sulfonate (NBS). The common
use for each detergent is shown in Table 1.
Materials and Methods
Three acrylic plastic columns were set up in the laboratory to simu-
late a coal pile (Figure 1). The columns were approximately 2 meters high
23-5
-------
Table 1. Detergents Tested to Prevent
Coal Pile Drainage Contamination
Detergent
Common Use
Percent Active Ingredient
in Product Tested
Sodium Lauryl
Sulfate
Neutralized
Benzene
Sulfonate
Active ingredient in
many hair shampoos
Active ingredient in
many laundry detergents
30
15
PLASTIC
CAPS-
s COLLECT!-.
BEAKERS
* COLUMN C
CUT AT
MIOSECTION
Figure 1. Laboratory columns used to simulate
coal pile drainage.
23-6
-------
and 30 cm in diameter. A 1.5-cm diameter drain valve was located in the
bottom of each column to collect leachate. The underdrain system in the
columns consisted of a 20-cm depth of thoroughly washed pea gravel under
25 cm of washed sand. To prevent the gravel from clogging the drain, a
capped 10-cm diameter, 13-cm long perforated plastic column was glued
over each drain hold. Compressed air was passed through.water and fed
through an "L" shaped glass tube to simulate atmospheric conditions within
a coal pile. The tube was located at mid-height in the sand layer. The
sand also acted as a diffuser to distribute the air across the coal.
Coal was obtained from the Tennessee Valley Authority (TVA) Cumberland
Power Plant located near Nashville, Tennessee. Small samples of the coal
were collected from different locations and depths around the coal pile.
The samples were then composited. The composite was rinsed with water
until the pH of drainage was neutral in order to simulate freshly mined
coal or coal that had been processed by coal washing. A portion of the
washed coal was analyzed for total sulfur, moisture, volatile matter, ash,
fixed carbon, and Btu content.
Approximately 0.057 cubic meters of the coal was divided equally
between the three columns. The procedure of the three coal samples was
identical except for the different solutions applied.
Each coal sample was initially placed in a 5-gallon bucket and satu-
rated with the necessary solution for one hour. The excess water was
poured off and the resultant slurry poured into the appropriate column.
In the first column, coal was treated with deionized water only. Coal in
another column was treated with a 50 rag/L solution of SLS and deionized
water. A 50 mg/L solution of NBS and deionized water was added to coal
in the remaining column.
23-7
-------
The columns were dosed twice a week with 1,500 ml of deionized water
to simulate a 2.5-cm (1-inch) rainfall contacting a coal pile. Leachate
was then collected from the drain valves at the bottom of the columns for
analysis.
Acidity, pH, conductivity, iron, manganese, and the most probable
number (MPN) of iron-oxidizing bacteria were routinely analyzed on the
collected samples. Selected trace metals and detergent concentrations
were also analyzed periodically during the experiment. The deionized
water used in the experiment was routinely analyzed for pH. All chemical
analyses were performed according to procedures outlined in "Standard
Methods" (8). Enumeration of iron-oxidizing bacteria was estimated by
employing the inorganic salts culture medium and the combination multiple
tube and microtitre plate procedure described by Olem and Unz (9).
Results and Discussion
Coal Characteristics—Characteristics of the coal used in these
experiments are shown in Table 2. The total sulfur content was typical
for coal burned at most TVA power plants. Coal with these characteristics
has been found to produce strongly acidic drainage exposed to the elements.
Simulated Rainwater—The deionized water used to simulate rainfall
had a pH of approximately 5.5. The acidity present in the deionized water
was due to contact with atmospheric carbon dioxide.
Leachate Characteristics—The quality of leachate from each coal col-
umn three days after initial application of detergents is presented in
Table 3. As expected, the drainage was initially pH 7.2 for all three
columns. Most other water chemistry characteristics were similar for
the drainage from all three columns, although certain parameters varied
widely. After remaining nearly neutral for about 20 days, the leachate
23-8
-------
Table 2. Analysis of Coal Used in Laboratory
Simulation of Coal Pile Drainage
Parameter Units Value
Total Moisture
Volatile Matter
Ash
Fixed Carbon
Total Sulfur
Energy Content
%
% Dry Basis
% Dry Basis
% Dry Basis
% Dry Basis
Btu/lb Dry Basis
3.8
39.8
9 f\
.8
r* f\ 1
50.4
3.1
13,162
23-9
-------
Table 3. Quality of Leachate from Laboratory Coal Columns
Three Days After Initial Application of Detergents
Parameter
Concentration In Leachate (MgA unless indicated)
Control SLS Treated NBS Treated
pH (S.U.)
Acidity (mg/1 CaCO )
Conductivity (jJmhos/cm)
Fe
Mn
Pb
Hg
As
Cd
Cu
Se
Cr
Ni
Zn
7.2
0.6
480
1,020
96
4
0.7
1
0.2
<1
1
2
<50
<5
7.2
0.3
210
299
22
<1
0.3
<1
<0.1
<1
<1
320
<50
<5
7.2
0.3
600
1,680
100
11
<0.2
1
<0. 1
4
2
1
<50
<5
-------
suddenly became ten-thousandfold more acid in the control column, as pH
decreased from 7 to about 3 (Figure!;. The pH continued to decline for
another 20 days until it leveled off at about pH 2.
Drainage from the test column treated with NBS became slightly more
alkaline at first but remained approximately neutral only about five days
longer than the control. Then it paralleled the same pattern of rapidly
increasing acidity, reaching pH 2 after about 50 days.
Drainage from the coal treated with SLS became even slightly more
alkaline for the first 20 days. After a slight decrease in pH, it
remained neutral for two and one-half times longer than the control.
Even when the pH finally dropped after about 60 days, it hovered for
another 25 days between pH 4 and 5. The acidity at this pH was approxi-
mately 100 mg/L, compared to 5,000 mg/L for the control column. This
difference could be extremely important in terms of the treatment- that
would be needed to neutralize it.
The coal treated with SLS eventually produced drainage with the same
low pH of about 2, but only after 140 days, three and one-half times longer
than the 40 days for the untreated control.
Other constituents generally followed the same pattern as the changes
in pH. The quality of leachate from each column 56 days after initial
detergent" application is presented in Table 4. At this point in the
experiment, only drainage from the SLS-treated coal was neutral. Acidity
in drainage was over 100 times higher in the control and NBS-treated coal.
Significantly lower concentrations were also observed in drainage from
the SLS-treated coal for conductivity, Fe, Mn, As, Cd, Cu, Cr, Ni, and
Zn. Only Pb and Se were higher in drainage from SLS-treated coal compared
to the control. However, these concentrations were almost at the detec-
tion limits for these constituents.
23-11
-------
8
o CONTROL
•SLS TREATED
A NBS TREATED
PH
to
CO
>—'
N5
20
40 60 80 100 120
DAYS AFTER TREATMENT
140
Figure 2. Changes in pH in laboratory-generated
coal pile drainage following application
of 50 mg/L of sodium lauryl sulfate (SLS)
and neutralized benzene sulfonate (NBS).
-------
Table 4. Quality of Leachate from Laboratory Coal Columns
Fifty Six Days After Initial Application of Detergents
to
CO
Parameter
pH (S.U.)
Acidity (mg/1 CaCO )
Conductivity (jjmhos/cm)
Fe
Mn
Pb
Hg
As
Cd
Cu
Se
Cr
Ni
Zn
Concentration
Control
2.0
5,000
6,000
1,560,000
48,900
<1
0.3
200
310
1,050
<1
295
1,500
13,700
In Leachate (pg/1
SLS Treated
6.9
10
1,950
365
1,780
2
<0.2
1
2.3
7
4
<1
<50
<5
unless indicated)
NBS Treated
2.1
5,500
5,500
1,004,000
74,500
7
<0.2
220
340
1,160
1
2
3,300
16,700
-------
Drainage from both treated columns had detergent concentrations
below detection limits (<0.1 mg/L measured as methylene blue active sub-
stance) . This is important because these detergents are toxic to aquatic
life. For example, Dailela et al. (10) found the LC5Q value for a cer-
tain species of fish exposed to SLS to be 11.2 rag/L. Apparently, SLS was
not washed much out of the coal. Most of it was probably adsorbed onto
the coal particles and remained there until its chemical structure was
broken down into simple degradation products such as carbon dioxide, water
and sodium and sulfate ions.
Kleinmann and Erickson (6) performed a detergent adsorption capacity
test on refuse from a coal-cleaning plant and determined that one gram
of refuse adsorbed between 50 and 68 |jg of SLS. These values were an
order of magnitude higher than results for similar tests performed on
surface-mined overburden material. The same test was performed on the
coal used in the laboratory column experiments. Adsorption capacity was
found to be 60 jjg SLS/g coal. Because of the lower adsorption capacity
of overburden materials, Kleinmann (6) incorporated SLS into rubber
pellets which gradually release the detergent into infiltrating rainwater.
This method may not be needed for coal storage and coal refuse disposal
areas.
The "NBS and SLS detergent applications apparently did not kill all
the iron-oxidizing bacteria, because small numbers of live, but appar-
ently inactive, bacteria were present even in samples collected after' the
coal was treated. For example, bacterial numbers in the drainage 28 days
after treatment were 160,000 per 100 mL for the control column and 7,900
per 100 mL for the SLS-treated column. Toward the end of the experiment,
when drainage from the treated and control columns were acid, bacterial
23-14
-------
numbers were more similar. One hundred nineteen days after treatment,
drainage from the control and SLS-treated columns contained 1,100,000
and 920,000 iron-oxidizing bacteria per 100 ml.
The reasons for the relative effectiveness of SLS treatment compared
to NBS are not known. If detergents alter the semipermeable properties
of the cytoplasmic membrane and allow H+ to seep into the interior of
the cell, as has been reported, then it is possible that a detergent
that better attacks protein would be more effective. SLS, being the
active ingredient in many hair shampoos, is designed to attack protein-
type buildup on hair. On the other hand, NBS is the active ingredient in
many laundry detergents and is designed to remove soil particles, not pro-
teinaccous matter. This may be one. reason for the relative ineffectiveness
of NBS.
Projected Costs—Roughly extrapolating results to an industrial-scale
application, the cost of SLS needed for a single application was estimated
at about $500 per hectare of coal storage area ($200 per acre). This
figure assumes a delivered cost of $1.65 per Kg ($0.75 per Ib) for the
SLS used in the experiments and allows for a 50-percent decrease in the
efficiency of detergent application in the field. Because of the
limited number of experiments performed to date, these costs must be
considered preliminary.
Methods of Application—Results of laboratory experiments showed
that SLS may be effective in preventing problems of acidity and dissolved
metals in coal pile drainage for about two months. If an operating
facility stored less than a 2-month supply of coal and always burned the
oldest coal first, a single application to each new coal delivery might
prevent the development of acid drainage. However, there are probably
few plants where these conditions would be met at all times. In the
23-15
-------
tests performed thus far, it was assumed that it would be necessary to
treat all the coal from the top to the bottom of an entire pile each time
it is required. There are some indications, however, that bacteria are
active mainly in the top layer of coal, where the temperature and avail-
ability of air best suit their needs. If that is the case, it might be
necessary to treat only the surface of the pile, or it might be possible
to extend significantly the time between full-pile treatments by means
of supplementary surface layer treatment.
It might even be economical to treat each new coal delivery before
it is added to the pile, either as a supplementary treatment or as the
only treatment required. At many plants, the surface of the coal pile
is routinely sprinkled with water to control dust, and a small amount
of SLS might easily be added to the tanks of the sprinkler trucks. A
spray system might be installed to treat every truckload or trainload
of coal with a small amount of SLS, either as it is dumped or as it moves
along conveyors to the pile so that the surface layer of new coal would
continually be protected against acid formation. If significant time
is expected to elapse between mining and delivery, the coal might even
be protected with SLS as soon as it is mined and then treated again when
it is delivered.
Additional Research Needs—Since drainage from existing coal piles is
often already acid, there is a question as to whether it would be possible
to correct an existing problem by application of detergents. A modifica-
tion of the methods described in these studies would likely be necessary
because SLS breaks down rapidly in strong acid solutions. It may be possi-
ble to correct an existing problem by treating the coal with SLS in combi-
nation with some buffering agent. Further experiments using several buffers
23-16
-------
such as sodium bicarbinate and phosphate are currently in progress to
address this concern.
It is not known whether SLS affects the burning characteristics of
coal. This must be determined because laboratory results indicated that
SLS was readily adsorbed onto the coal particles. Another potential con-
cern, the added sulfur due to the SLS, should not present a. problem. The
sulfur content theoretically added to the coal due to the sulfate ions in
SLS is insignificant relative to the high-sulfur content of coal.
Conclusions
1. Controlled application of anionic detergents may be useful.in pre-
venting acid drainage from coal storage piles.
2. Sodium lauryl sulfate was more effective than NBS in preventing acid
production in laboratory-simulated coal piles.
3. Based on laboratory studies, it appears that a single application
may last up to 60 days.
4. Extrapolating laboratory results to an industrial-scale application,
the cost of SLS needed for a single treatment was calculated to be
$500 per hectare of coal storage area ($200 per acre).
Acknowledgements
We thank Robert L. P. Kleinmann and Patricia M. Erickson for per-
forming adsorption tests and Donald J. Rucker for critical review of the
manuscript.
Literature Cited
(1) Colmer, A. R.; Hinkle, M. E., Science, 1947, 106. 253.
(2) Singer, P. C.; Stumm, W., Science. 1970, 167, 1121.
23-17
-------
(3) Singer, P. C.; Stumm, W., "Oxygenation of Ferrous Iron: The
Rate-determining Step in the Formation of Acid Mine Drainage."
Federal Water Pollution Control Administration, Report 14010,
1969.
(4) Dugan, R. P., Ohio J. Sci., 1975, 75_, 266.
(5) Kleinmann, R.L.P., In "Proceedings, Symposium on Surface Mining
Hydrology, Sedimentology, and Reclamation1!; Graves, D. H., Ed.;
Report No. UKY BU123, University of Kentucky, Lexington, 1980;
pp 333-337.
(6) Kleinmann, R.L.P., Erickson, P. M., In "Proceedings, Symposium
on Surface Mining Hydrology, Sedimentology, and Reclamation";
Graves, D. H., Ed.; Report No. UKY BU126, University of Kentucky,
Lexington, 1981; pp 325-330.
(7) Hugo, W. B., J. Bacteriol., 1967, 30, 17.
(8) "Standard Methods for the Examination of Water and Wastewater."
15th Ed.; American Public Health Association, Washington, B.C.,
1980.
(9) Olem, H.; Unz, R. F., Biotechnol. Bioeng., 1977, l£, 1475.
(10) Dailela, R. C., et 4!., Water, Air and Soil Poll., 1981, 15_ 1.
63588/6650-B/6-82/214
23-18
-------
ATTENDEES
Antisavage, Joseph P.
Betz Laboratories, Inc.
4636 Somerton Rd
Trevose, PA 19047
215/355-3300 x377
Armstrong, James A.
Denver Research Institute
University of Denver
Denver, CO 80208
303/753-2892
Bohn, Russel
Environmental Services
and Technology
11 E. 69th Ter
Kansas City, MO 64113
Boudreau, Richard M.
Central Illinois Light Co.
300 Liberty St
Peoria, IL 61602
309/672-5471
Bowman, W. Alan
Applied Meteorology, Inc.
Suite 326
9000 Southwest Frwy
Houston, TX 77074
713/777-0106
Bramson, Mark
Syntech Products Corp.
520 E. Woodruff
Toledo, OH 43624
419/241-1215
Brass, Dwight
George A. Rolfes Co.
Box 458
Boone, IA 50036
515/432-3300
Brookman, Edward T.
TRC Environmental
Consultants, Inc.
800 Connecticut Blvd
E. Hartford, CT 06108
203/289-8631
Campbell, Ivor E.
Smelter Control
Research Associates
150 E. Broad St
Columbus, OH 43215
Chalmers, Jerry
S. C. Dept of Health and
Environmental Control
2600 Bull St
Columbia, SC 29201
803/758-5406
Cole, Clifford F.
TRC Environmental
Consultants, Inc.
8775 E. Orchard Rd,
Suite 816
Englewood, CO 80231
303/779-4940
Connolly, Michael
Minnesota Pollution
Control Agency
1935 W. County Rd B-Z
Roseville, MN 55118
Courtney, Peter
Law Engineering
2749 Delk Rd
Marietta, GA 30067
Cowherd, Chatten
Midwest Research Institute
425 Volker Blvd
Kansas City, MO 64110
816/753-7600 x531
Cox, H. B.
U.S. Steel Research
(U. S. Mining, Inc.)
B St
Pittsburgh, PA 15235
412/372-1212 x4291
Craig, Alfred B. (MD-54)
U.S. EPA/AEERL (lERL)
Research Triangle Park, NC 27711
919/541-2824
A-l
-------
Attendees (cont.)
Crockett, E. P.
American Petroleum Institute
2101 L St, N.W.
Washington, DC 20036
202/457-7084
Dorsey, James A. (MD-62B)
U.S. EPA/AEERL (IERL)
Research Triangle Park, NC 27711
919/541-2509
Dubrowski, John J.
Exxon Research and Engineering
Florham Park, NJ 07932
201/765-2101
Fletcher, James K.
Kennecott Minerals Co.
PO Box 11248
Salt Lake City, UT 84147
801/322-8262
Gatz, Donald F.
Illinois State Water Survey
PO Box 5050, Station A
Champaign, IL 61820
217/333-2512
Goebel, Gerald R.
DNR and EP/Div of Air
Pollution Control
Ft. Boone Plaza,
18 Reilly Rd
Frankfort, KY 40601
502/564-3382
Hague, William
Julius Koch USA, Inc.
PO Box A-995
New Bedford, MA 02741
617/995-9565
Harris, D. Bruce (MD-54)
U.S. EPA/AEERL (IERL)
Research Triangle Park, NC 27711
919/541-7807
Harrison, Paul R.
Envirosol
1700 N. Fiske
Pasadena, CA 91104
818/797-9581
Hartshorn, Wayne
Sonic Development Corp.
305 Island Rd
Mahwah, NJ 07430
201/825-3030
Holton, Greg (ORNL)
First Environment
314 W. Broadway
Lenoir City, TN 37771
Hyatt, James
Tennessee Valley Authority
940 CSTZ-C
Chattanooga, TN 37401
Hyatt, John G.
Exxon Co., USA
Colony Shale Oil Project
PO Box 440342
Aurora, CO 80044
303/695-2213
Hyde, Raymond G.
Atlantic Research Corp.
5390 Cherokee Ave
Alexandria. VA 22314
"703/642-4193
Ives, Jim
Anaconda Minerals Co.
PO Box 5300
Denver, CO 80217
303/575-7504
Kaiser, Robert A.
Ohio Edison Co.
76 S. Main St
Akron, OH 44308
216/384-5770
Kawaters, Woody
TRC Environmental
Consultants, Inc.
800 Connecticut Blvd
E. Hartford, CT 06108
Kretch, Frank
Sonic Development Corp.
305 Island Rd
Mahwah, NJ 07430
201/885-3030
A-2
-------
Attendees (cont.)
Kuykendal, William B. (MD-15)
U.S. EPA/OAQPS/AQMD (IERL)
Research Triangle Park, NC 27711
919/541-5372
Larson, Alan G. (TRC)
The Kent-wood Moore Co.
5690 Denver Tech Ctr Blvd
Englewood, CO 80111
303/773-3399
Lawrence, Robert
KPN International, Inc.
19 Pebble Rd
Newtown, CT 06470
Letizia, Anthony P.
Envirosphere Co.
2 World Trade Center
New York, NY 10048
212/839-1088
Lewis, Ted P.
Iowa Electric Light and
Power Co.
PO Box 351
Cedar Rapids, LA 52406
319/398-4601
Lubas, Thomas
Port Authority of NY and NJ
1 World Trade Center,
64 Floor E.
New York, NY 10048
Manning, James
US EPA Region 4
345 Courtland St, N. E.
Atlanta, GA 30365
404/347-3286
Martin, Dennis
TRC Environmental
Consultants, Inc.
800 Connecticut Blvd
E. Hartford, CT 06108
Mathai, C. V.
Arizona Public Service Co.
(AeroVironment, Inc.)
PO Box 53999 (Sta. 5680)
Phoenix, AZ 85072-399S
602/371-6467
Milhous, Madison N.
Long Island Lighting Co.
1775 E. Old Country Rd
Hicksville, NY 11801
516/733-4385
Mohr, Ray
Colorado Health Dept
1101 Bellaire St
Denver, CO 80220
303/320-4180
Moldovan, John
Anaconda Minerals Co.
5555 17th St
Denver, CO 80433
303/575-4312
Mulloy, Frederick W.
Phillips Petroleum Co.
7 D 3 Phillips Bldg
Bartlesville, OK 74004
Nicholson, Brock M. (MD-15)
U.S. EPA/OAQPS/AQMD (CPDD)
Research Triangle Park, NC 27711
919/541-5517
Nilsen, Glennyce C.
United Illuminating Co.
80 Temple St
New Haven, CT 06506
Olem, Harvey
Tennessee Valley Authority
245 401 Bldg
Chattanooga, TN 37401
615/751-7338
Pace, Thompson G. (MD-15)
U.S. EPA/OAQPS/AQMD
Research Triangle Park, NC 27711
919/541-5634
Plaks, Norman (MD-61)
U.S. EPA/AEERL (IERL)
Research Triangle Park, NC 27711
919/541-3084
Porter, Richard A.
NAI
1710 Firman Dr
Richardson, TX 75081
214/644-1616
A-3
-------
Attendees (cont.)
Rakes, Samuel L. (MD-62)
U.S. EPA/AEERL (IERL)
Research Triangle Park, NC
919/541-2828
27711
Ray, B. Michael
Northrop Environmental Training
PO Box 12313
Research Triangle Park, NC 27709
919/549-0652
Reddy, Tupili S.
Tennessee Air Pollution Control Div
150 - 9th Ave, N.
Nashville, TN 37203
615/741-3651
Ripp, John A.
TRC Environmental
Consultants, Inc.
800 Connecticut Blvd
E. Hartford, CT 06108
203/289-8631
Romaine, Demarest (Dave)
Consolidated Edison Co. of
New York, Inc.
4 Irving PI
New York, NY 10003
212/460-4600
Rosbury, Keith D.
PEDCo Environmental,
2420 Pershing Rd
Kansas City, MO 64108
Inc.
Rovell-Rixx, David C.
National Council for Air and
Stream Improvement. Inc.
PO Box 11483
Gainesville, FL 36204
Sando, Dick
Sonic Development Corp.
305 Island Rd
Mahwah, NJ 07430
201/825-3030
Schumacher, Aileen
Environmental Science and
Engineering
PO Box ESE
Gainesville, FL 32602
904/372-3318
Sehmel, George A.
Pacific Northwest Laboratory
24000 Stevens 1100 Area
Richland, WA 99352
509/375-6161
Shrock, John
Illinois Environmental
Protection Agency
2200 Churchill Rd
Springfield, IL 62706
217/782-1830
Simpson, Charles
SynTech Products Corp.
520 E. Woodruff
Toledo, OH 43624
419/241-1215
Skoney, Davi d
Erie Co. Dept of Environment
and Planning
95 Franklin St
Buffalo, NY 14225
716/846-8556
Smith, M. L.
Andersen Samplers, Inc.
4215 Wendell Dr
Atlanta, GA 30315
404/691-1910
Staley, Laurel J.
U.S. EPA/RREL (IERL)
Cincinnati, OH 45268
513/569-7863
Stark, Phillip E.
Shell Oil Co. - Mining
PO Box 2906
Houston, TX 77001
713/870-4263
A-4
-------
Attendees (cont.)
Stone, F. Kenneth
S. C. Dept of Health and
Environmental Control
2600 Bull St
Columbia, SC 29201
803/758-5406
Tucker, A. L.
Jones and Laughlin Steel Co.
3001 Dickey Rd
E. Chicago, IN 46312
219/391-2571
Turetsky, William S.
Allied Chemical Co.
PO Box 1139R
Morristown, NJ 07960
201/455-2690
Wells, Robert C.
Enviroplan, Inc.
59 Main St
West Orange, NJ 07052
201/325-1544
Williams, Colin J.
Rowan, Williams, Davies,
and Irwin, Inc.
(MHTR Ltd)
650 Woodlawn Rd, West
Guelph, Ontario N1K1B8
519/823-1311
Wootten, John M.
Peabody Coal Co.
PO Box 14495
St. Louis, MO 63178
Yocom, John E.
TRC Environmental
Consultants, Inc.
800 Connecticut Blvd
E. Hartford, CT 06108
203/289-8631
A-5
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/9-89-085
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Fifth Symposium on Fugitive Emissions:
Measurement and Control (May 3-5, 1982,
Charleston, South Carolina)
5. REPORT DATE
September 1989
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
D.Bruce Harris and William B. Kuykendal,
General Chairmen
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
See Block 12
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
NA (Inhouse)
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings; May 1982
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES ^EERL project officer was D. Bruce Harris, Mail Drop 54, 919/
541-7807. Symposium was held May 3-5, 1982, in Charleston, SC.
is. ABSTRACT
proceedj.ngs document presentations at the Fifth Symposium on Fugitive
Emissions: Measurement and Control, May 3- 5, 1982, in Charleston, SC. The Sym-
posium was sponsored by the U. S. EPA's Air and Energy Engineering Research Lab-
oratory (known then as the Industrial Environmental Research Laboratory) in Re-
search Triangle Park, NC, as part of the Agency's continuing effort to develop me-
thods for measuring and controlling airborne and waterborne fugitive emissions
from energy and industrial processes. The objective of the symposium was to
bring together people from industrial, academic, research, and government orga-
nizations with experience or interest in fugitive emissions problems to exchange
information of mutual potential benefit. The program included presentations by
individuals from a variety of organizations describing their experience and view-
points regarding the impact, measurement, and control of fugitive emissions. An
international flavor was provided by presentations by authors from Belgium, Canada,
and Sweden.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Pollution
Measurement
Emission
Processing
Leakage
Hydrocarbons
Coal Storage
Coal Dust
Roads
Dust Control
Aerosols
Particles
Pollution Control
Stationary Sources
Fugitive Emissions
Particulate
Dust Suppressants
Street Sweepers
13 B
14G
13H
07C
081
21D
05E
07D
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
20. SECURITY CLASS (This page)
Unclassified
21. NO. OF PAGES
306
22. PRICE
EPA Form 2220-1 (9-73)
A-6
------- |