MIDWEST RESEARCH INSTITUTE
j Note' This is a reference cited in AP 42, Compilation of Air Pollutant Emission Factors, Volume I Stationary |
I Point and Area Sources. AP42 is located on the EPA web site at www epa gov/ttn/chief/ap42/
?The file name refers to the reference number, the AP42 chapter and section. The file name
i'"ref02_c01s02 pdf" would mean the reference is from AP42 chapter 1 section 2 The reference may be
[from a previous version of the section and no longer cited. The primary source should always be checked
5
UPDATE OF FUGITIVE DUST EMISSION FACTORS IN AP-42 SECTION 11.2
I
FINAL REPORT
EPA Contract No. 68-02-3891
Assignment No. 19
MRI Project No. 8681-L(19)
July 14, 1987
f.
|
§
U
-4
For
a
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Attn: Mr. Frank M. Noonan
' j vc-i t'C-i
4
C'.TY, M':>:
/J
-------
UPDATE OF FUGITIVE DUST EMISSION FACTORS IN AP-42 SECTION 11.2
FINAL REPORT
EPA Contract No. 68-02-3891
Assignment No. 19
MRI Project No. 8681-LU9)
July 14, 1987
For
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Attn: Mr. Frank M. Noonan
MIDWEST RESEARCH INSTITUTE 425 VOLKER BOULEVARD, KANSAS CITY. MISSOURI 64110*816
-------
PREFACE
This final report was prepared for the U.S. Environmental Protection
Agency (EPA), Office of Air Quality Planning and Standards (OAQPS) under
EPA Contract No. 68-02-3981, Assignment No. 19. Mr. Frank M. Noonan, Air
Management Technology Branch, was the requestor of this work. The draft
final report was prepared by Dr. Gregory E. Muleski with assistance from
Dr. Chatten Cowherd and Mr. Phillip Englehart.
Approved for:
MIDWEST RESEARCH INSTITUTE
Chatten Cowherd, Director
Environmental Systems Department
July 14, 1987
i i i
-------
Page
ill
vi
v1
1
7
7
7
13
13
13
14
17
17
22
28
33
41
42
52
66
77
83
83
87
95
99
101
CONTENTS
Introduction
Review of AP-42
2.1 History of AP-42
2.2 Open dust source emission factors in AP-42 . . .
Identification of Candidate Test Reports
3.1 Review of literature
3.2 Screening criteria
3.3 Primary list of test reports
Evaluation Criteria for the Test Reports
4.1 EPA's emission factor quality rating system. . .
4.2 Methods of emission factor determination . . . .
4.3 Emission factor quality rating scheme used in
this study
4.4 Design and rating of control performance
evaluation studies
Candidate Emission Factors and Control Efficiency Data . .
5.1 General testing methodologies
5.2 Unpaved roads (Section 11.2.1)
5.3 Aggregate storage piles (Section 11.2.3) . . . .
5.4 Industrial paved roads (Section 11.2.6)
Discussion and Recommendations
6.1 Unpaved roads (Section 11.2.1)
6.2 Aggregate storage piles (Section 11.2.3) . . . .
6.3 Industrial paved roads (Section 11.2.6)
6.4 Summary
References
v
-------
FIGURES
Number Page
1 Hypothetical decay curves 39
2 Photocopy of Table 5 from Test Report 10a 48
3 Photocopy of Table 8 from Test Report 10a 51
TABLES
Number Page
1 Compilation of AP-42 Open Dust Source Emission Factors . . 8
2 Preliminary List of Test Reports 15
3 Quality Rating Scheme for Single-Valued Emission
Factors 30
4 Quality Rating Scheme for Emission Factor Equations. ... 32
5 List of Test Reports by Pertinent Subsection 41
6 Source Testing Information (Test Reports 10b to lOf) ... 43
7 Range of Conditions and Emission Factors 45
8 Source Testing Information (Test Report 1) 53
9 Range of Conditions and Emission Factors
(Test Report 1) 54
10 Source Testing Information (Test Report 2) 57
11 Range of Conditions and Emission Factors
(Test Report 2) 59
12 Source Testing Information (Test Report 3) 61
13 Range of Conditions and Emission Factors
(Test Report 3) 62
14 Source Testing Information (Test Report 7) 64
15 Range of Conditions, Emission Factors, and Ratings
(Test Report 7) 65
16 Source Testing Information (Test Report 11) 67
17 Range of Conditions, Emission Factors, and Ratings
(Test Report 11) 68
18 Source Testing Information (Test Report 4) 70
19 Range of Conditions, Emission Factors, and Ratings
(Test Report 4) 72
20 Source Testing Information (Test Report 5) 73
21 Range of Conditions, Emission Factors, and Ratings
(Test Report 5) 74
22 Source Testing Information (Test Report 9) 76
23 Range of Conditions, Emission Factors, and Ratings
(Test Report 9) 78
24 Source Testing Information (Test Report 8) 80
25 Range of Conditions and Emission Factors (Test Report 8) . 81
26 Regression Equations Obtained for Materials Handling
Data Sets 90
vi i
-------
SECTION 1
INTRODUCTION
In both the assessment of air quality and the design of control plans
to achieve certain goals in terms of air quality, there is a critical need
for reliable and consistent data on the quantity and physical characteris-
tics of emissions from a variety of sources. The large number of individua
release points and the diversity of source types make field measurements of
emissions at every location impractical. Usually, the only feasible method
of determining pollutant emissions for a given area is to make general
emission estimates typical of each source type.
Calculation of the estimated emission rate for a given source requires
data on source extent, uncontrolled emission factor, and control efficiency
The mathematical expression for this calculation is as follows:
R = Me (1 - c) (1)
where:
R = mass emission rate
M = source extent
e = uncontrolled emission factor, i.e., rate of uncontrolled
emissions per unit of source extent
c = fractional efficiency of control
The emission factor is an estimate of mass of pollutant released to the
atmosphere per unit measure of source activity (e.g., vehicle miles traveled,
tons of material transferred, etc.).
1
-------
The document "Compilation of Air Pollutant Emission Factors" (commonly
known as AP-42) has been published by the U.S. Environmental Protection
Agency (EPA) since February 1972, and represents a compilation of emission
factors for the most significant emission source categories. As more in-
formation about sources and control of emissions has become available,
supplements to AP-42 have been issued for both new emission source cate-
gories and for updating existing emission source categories.
Because the national effort to control industrial sources of pollution
has historically focused on discharge from stacks, ducts or flues, most of
the emission factors reported in AP-42 apply to ducted emission sources.
Over the past 15 years, however, it has become clear that fugitive (non-
ducted) emissions contribute substantially to the Impact of industrial op-
erations and may, in some industries, be greater than the stack emissions.
Industrial sources of fugitive particulate emissions may be divided
into the two classes of process and open dust sources. Process sources are
fully or partially enclosed operations that alter the chemical or physical
properties of a feed material. Examples of process sources are crushers,
sintering machines, and metallurgical furnaces. Open dust sources are
those that entail generation of emissions of solid particles by the forces
of wind and machinery acting on exposed materials. These sources include
open transport, storage and transfer of raw, intermediate, and waste ag-
gregate materials. The remainder of this discussion focuses on the cate-
gory of open dust sources.
Section 11.2 of AP-42 presents open dust emission factors for several
generic source categories. These factors have been used extensively by in-
dustry and regulatory agency personnel during the past decade. Emission
factors are presented for the following open dust sources:
2
-------
Section
Source
11.2.1 Unpaved Roads
11.2.2 Agricultural Tilling
11.2.3 Aggregate Handling and Storage Piles
11.2.4 Heavy Construction Activities
11.2.5 Paved Urban Roads
11.2.6 Industrial Paved Roads
These factors, except that for heavy construction operations, are presented
in the form of predictive equations which relate mass emissions to: (a) mea-
sures of source activity or energy expended; (b) properties of the material
being disturbed; and (c) climatic parameters. As such, these factors becomfc
applicable to a wide range of source conditions, limited only by the extent
of experimental verification.
As part of EPA's anticipated revision of the National Ambient Air
Quality Standards (NAAQS) for particulate matter to address particles less
than or equal to 10 |jm 1n aerodynamic diameter (PM10), the three sections
concerning paved and unpaved roads were revised for inclusion in the Fourth
Edition of AP-42 (September 1985). The three remaining sections were not
updated.
Recent developments support the need for Section 11.2 to be revised an|i
possibly expanded. First, revisions may be warranted simply because new
test data are now available. The data generated in these new studies need
to be reviewed 1n order to determine if revisions of Section 11.2 can be
supported. For example, 1t is likely that recent field tests related to
source categories already addressed in AP-42 may be used to broaden the
applicability of the existing emission factors.
The second development involves increased interest in the control of
fugitive dust emissions. Almost all field studies of road dust emissions
during the 1980s have entailed evaluation of control techniques. As a re-
sult of these studies, considerably more data are now available to estimate
the efficiency of certain control techniques (especially those for unpaved
3
-------
roads). Revision of certain sections in Section 11.2 may be needed in light
of these new results.
The particle size ranges of interest in this report are:
TP - Total airborne particulate matter.
TSP - Total suspended particulate matter, as determined by standard
high volume air sampling.
SP - Particulate matter smaller than 30 in aerodynamic diameter.
This fraction is often used to approximate TSP.
IP - Inhalable particulate matter consisting of particles smaller
than 15 ^m in aerodynamic diameter.
PM10 - Particulate matter consisting of particles smaller than 10 pm
in aerodynamic diameter.
FP - Fine particulate matter consisting of particles smaller than
2.5 |jm in aerodynamic diameter.
Particular attention is devoted to the TSP and SP fractions because of the
current NAAQS and to PMl(> because of the anticipated NAAQS revision pertain-
ing to that fraction.
The purpose of this report 1s to present background information in
support of new and revised AP-42 sections for open dust sources. This re-
port is organized as follows:
Section 2 - Emission Factors Currently Reported in AP-42
Section 3 - Identification of Candidate Test Reports
4
-------
Section 4 - Evaluation Criteria Test Reports
Section 5 - Candidate Emission Factors and Control Efficiency Data
Section 6 - Discussion and Recommendations
Both metric and English units are used in this report. The review of
available test data (Section 5.0) uses the same set of units as does the
test report being evaluated. If both sets are reported, preference is givep
to that set used during the original data reduction (if known).
5
-------
SECTION 2.0
REVIEW OF AP-42
2.1 HISTORY OF AP-42
AP-42 presents data available on pollutant emissions for which adequate
documentation exists to estimate emission factors. The factors given 1n
AP-42 are based on emission data obtained by various methods, of which
source testing, material balance studies, and engineering estimates are the
most common. The primary purpose of the document is for use by individuals
and groups responsible for developing air pollution emission inventories.
AP-42 was first published by the U.S. Public Health Service and, since
1972, by the EPA. The document has been revised on a periodic basis since
that time. Supplements are issued either to revise existing emission fac-
tors or to present information regarding a source not previously included itf
AP-42. The Fourth Edition of AP-42 was issued in September 1985.
2.2 OPEN DUST SOURCE EMISSION FACTORS IN AP-42
In contrast to process sources of fugitive particulate emissions, open
dust sources entail no change of material properties, either physical or
chemical, of a feed material. Examples of open dust sources include mate-
rials transfer and storage piles. Operations which are not open dust
sources include crushing, drying, and screening, all of which deal with a
change in physical properties of a feed material.
Table 1 presents the open dust source emission factors given in
AP-42 (Fourth Edition). Also given is the emission factor rating and the
year that the AP-42 section was introduced or last revised. Ratings are
described in Section 4.0 of this report.
7
-------
TABLE 1. COMPILATION OF AP-42 OPEN DUST SOURCE EMISSION FACTORS
Industrial Fugitive Emission
source category emission source factor (lb/ton)
Adlple add Drying, cooling, and 0.8
production storage
Carbon black Fugitive emissions 0.20
manufacture
Hydrofluoric Spar handling silos 60
acid produc- Spar transfer opera- 6
tion tions
Lead alky! pro- Sludge pits 1.2
ductlon
Grain elevators Receiving 0.6-1
Shipping 0.3-1
Feed mills Receiving 1.3
Shipping 0.5
Handling 3
Wheat milling Receiving 1
Precleaning and 5
handling
Durum milling Receiving 1
Precleaning and S
handling
Rye milling Receiving 1
Precleaning and S
handling
Ory corn milling Receiving 1
Precleaning and S
handling
Rice milling Receiving 0.64
Precleaning and 5
handling
Soybean ml 111ng Receiving 1.6
Handling S
Bulk loading 0.27
wet corn milling Receiving 1
Handling 5
Fermentation Grain handling 3/3
Industry
Ammonium nitrate Bulk loading i 0.02
fertilizer
Phosphate Unloading 0.56
fertilizer
Triple super- Unloading 0 14-0.18
phosphate
fertilizer
Ammonium phos- Product sizing and 0.03
phates material transfer
Emission Year of
factor rating latest revision Comments
B 1977 A process source Included In
factor.
C 1983
D 1980
E 1980
B 1979
B 1977
B 1977
0 1977
0 1977
D 1977
0 1977
0 1977 A process included in factor.
0 1977
D 1977 A process Included in fact|or
0 1977
0 1977 A process Included in factjor
0 1977
0 1977 A process Included in factpr.
0 1977
0 1977 A process Included In factpr.
0 1977
0 1977
0 1977
0 1977
0 1977
E/D 1972/1982 Whiskey/beer making.
B 1984
A 1980
A 1980
A 1980 Process Included; factor
represents one sample.
(Continued)
8
-------
TABLE 1 (Continued)
Industrial
source category
Fugitive
eaissfon source
Emission
factor (lb/ton)
Emission
factor rating
Year of
latest revision
Comments
Urea production
Sagging
0.19
C
1984
Cattle feedlots
Unspecified
27 lb/day/101
head throughput
E
1979
Another factor givfen for lot
capacity.
Cotton harvesting
(picker)
Trailer loading
Field transport
0.4 lb/mile1
2.S lb/mile1
C
C
1979
1979
Cotton harvesting
(stripper)
Trailer loading
Field transport
0.32 lb/mile4
1.6 lb/mile*
C
C
1979
1979
Wheat harvesting
Truck loading
Field transport
0.07 lb/mile2
0.65 lb/mile*
0
0
1980
1980
Sorghum harvesting
Truck loading
Field transport
0.13 lb/mlle2
1.2 lb/mile*
0
0
1980
1980
Primary aluminum
Materials handling
10
A
1973
Metallurgical coke
Charging
Pushing
0.8Sb
0.47
C
A
1982
1982
Iron and steel
varies
See comment
B-E
1983
Table 7.5-1 gives
valued factor for
sources, reader Is
fered to Chapter 1
ilngle-
ieveral
also re-
L.2.
Primary lead
smelters
Sinter transfer to
dump
Sinter product dump
Slag cooling
Materials handling
0.20
0.01
0.47
5.0
E
E
E
B
1980
1980
1980
1981
Gray Iron
foundries
Scrap and charge
handling, heating
Sand handling, prep-
aration, mulling
0.2b
3b
D
0
1981
1981
Process Included ih factor.
Process Included factor.
Asphaltlc concrete
Unloading coarse and
fine aggregate
Aggregate elevator
0.10b
0.20b
E
E
1981
1981
Brick manufactur-
ing
Raw material storage
34
C
1973
Calcium carbide
manufacturing
Circular charging
conveyor
(0.34)
C
1984
Controlled value.
Ceramic clay manu-
facturing
Storage
34
A
1972
Clay/flyash
sintering
Clay/coke crushing,
screening and
storage
Natural clay crush-
ing, screening,
and storage
15
12
c
c
1972
1972
Processes 1ncluded|1n factor.
Processes included]in factor
Glass fiber manu-
facturing
Unloading and convey-
ing
Storage bins
3
0 2
e
B
1985
1985
(Continued)
9
-------
TABLE 1 (Continued)
Industrial
source category
Fugitive
emission source
Emission
factor (lb/ton)
Emission
factor rating
Year of
latest revision
1—
Comments
Phosphate rock
processing
Storage and transfer
Storage piles
2
40
B
B
1980
1980
Sand and gravel
Pile formation by
stack
Batch loading
Storage piles
0.13
0.0E6
3. 5-14.8
E
E
0
1985
1985
1985
Crushed stone
Truck unloading
Truck loading
Conveying
0.0003
0.0003*0.06
0.0034
0
E
E
198S
1985
1983
Taconfte ore
Ore transfer
Bentonlte transfer
Pellet handling
Unpaved roads
0.10
0.04
3.4
9.3-11
0
0
0
C-D
1983
1983
1983
1963
Metallic minerals
Material handling and
transfer
Material handling and
transfer
1.1
0 01-0.12
C
C
1982
1982
Bauxite/alumina (low moisture
ore).
Other materials (low-high
moisture ore).
Western surface
coal mining
Truck loading, bull-
dozing, dragline,
vehicular traffic,
wind erosion
Topsol! removal,
overburden replace-
ment, truck and
train loading, truck
and scraper unload*
1ng, wind erosion
See comment
See concent
A-C
C-E
1983
1983
Predictive equations
Single-valued emission
factors.
Plywood veneer
and layout
Sawdust handling
1.0
E
1980
Woodworking waste
collection
Storage bin vent
Storage bin loadout
1
2
C
C
1979
1979
Unpaved roads
Vehicular traffic
Predictive eqn
A
1985
Agricultural op-
erations
Tilling
Predictive eqn.
A-e
1983
Aggregate storage
piles
Batch and continuous
drop, wind erosion
Predictive eqn.
c
1983
Heavy construc-
tion
Land cleaning, blast-
ing, excavation,
cut and fill. and
construction
1.2 lb/acre/month
1975
Paved urban roads
Vehicular traffic
Predictive eqn.
-
1985
Industrial paved
roads
Vehicular traffic
Predictive eqn
B-0
1985
A rating for separate PM-lOl
equation.
a Values In lb/ton except as noted
6 Revisions made In Supplement A (October 1986).
10
-------
For some open dust source categories described in AP-42, industry-
specific emission factors are given, along with reference to generic emis-
sion factors presented in Section 11.2. For example, in Section 7.5.1 (Iro6
and Steel Production), it is stated that open dust sources contribute to th$
atmospheric particulate burden. It is later mentioned in this section that
empirically derived predictive emission factor equations presented in Sec-
tion 11.2 generally better quantify these sources than do the single-valued
factors given.
11
-------
SECTION 3.0
IDENTIFICATION OF CANDIDATE TEST REPORTS
3.1 REVIEW OF LITERATURE
As the result of a scoping study, Muleski (1986b) discussed recent
fugitive emissions test data as well as recent citations of Section 11.2
that might support revisions and additions to that section of AP-42 leading
to improved open dust source emission and control efficiency estimates.
However, those reports had not undergone review to determine what effect
the new test results might have on Section 11.2. In addition to the test
data discussed in the scoping report, additional data were also identified
by the EPA work assignment manager for consideration in this study.
3.2 SCREENING CRITERIA
In order to reduce the large amount of candidate literature to a final
group of references pertinent to this update, five criteria were used:
1. The information in the reference document must deal with actual
emission factor development and/or control efficiency measurement.)
Many documents discuss emission factors or control efficiencies
but do not derive them.
2. Source testing must be part of the referenced study. Some report^
develop emision factors or control efficiency estimates by apply-
ing assumptions to existing data.
13
-------
3. The referenced study must deal with open dust source emissions of
the types discussed in Chapter 11.2 of AP-42. Process fugitive
emissions such as crushing, screening, and grinding are not per-
tinent to this investigation.
4. The document must constitute the original source of test data.
For example, a convention or symposium paper was not included if
the original study was already contained in a previous document
referenced in the paper.
5. The results of the referenced study must not be presently incor-
porated in AP-42. The purpose of this study is to recommend up-
dating AP-42 with research results not previously contained in
AP-42. If possible, however, new test data are to be combined
with previous data (used to develop the current AP-42 emission
factor) in deriving an updated emission factor.
3.3 PRIMARY LIST OF TEST REPORTS
A set of reference materials, given as Table 2, was gathered using the
criteria outlined above. These documents were then evaluated in terms of
the material presented in the next section.
Note that while Reports 10a through lOf do present actual emission
measurements, these reports were included primarily because they allow inter-
comparison of various source sampling methods that have been used to quan-
tify open dust sources. This is discussed in greater detail in Section 5.0.
14
-------
TABLE 2. PRELIMINARY LIST OF TEST REPORTS
1. T. Cuscino, Jr., et a1., Iron and Steel Plant Open Source Fugitive
Emission Control Evaluation, EPA-600/2-83-11Q, U.S. Environmental Pro-
tection Agency, Research Triangle Park, North Carolina, October 1983.
2. K. D. Rosbury, and R. A. Zlmmer, Cost-Effectiveness of Dust Controls
Used on Unpaved Haul Roads - Volume 1 of 2. Final Report for U.S.
Bureau of Mines, Minneapolis, Minnesota, December 1983.
3. G. E. Muleski et al., Extended Evaluation of Unpaved Road Dust Sup-
pressants in the Iron and Steel Industry, EPA-600/2-84-027, U.S. En-
vironmental Protection Agency, Research Triangle Park, North Carolina,
February 1984.
4. E. T. Brookman et al., Determination of Fugitive Coal Dust Emissions
from Rotary Rail car Dumping, TRC Project No. 1956-L81-00, May 1984.
5. G. E. Muleski, Measurement of Fugitive Dust Emissions from Prilled
Sulfur Handling, Final Report, MRI Project No. 7995-L, Prepared for
Gardinier, Inc., June 1984.
6. PEI Associates, Inc., Handbook - Dust Control at Hazardous Waste
Sites. Draft Final Report prepared for U.S. Environmental Protection
Agency, Cincinnati, Ohio, September 1984.
7. D. Russell and S. C. Caruso, "The Relative Effectiveness of a Dust
Suppressant for Use on Unpaved Roads Within the Iron and Steel Indus-
try," Presented at EPA/AISI Symposium on Iron and Steel Pollution
Abatement, Cleveland, Ohio, October 1984.
8. T. F. Eckle and D. L. Trozzo, "Verification of the Efficiency of a
Road-Dust Emission-Reduction Program by Exposure Profile Measurement,
Presented at EPA/AISI Symposium on Iron and Steel Pollution Abatement
Cleveland, Ohio, October 1984.
9. G. E. Muleski, Fugitive Emission Measurement of Fly Ash Loading at tht
River Rouge Power Plant. Final Report, MRI Project No. 8162-L, Pre-
pared for Detroit Edison, March 1985.
10a. B. E. Pyle and J. D. McCain, Critical Review of Open Source Particu-
late Emission Measurements: Part II - Field Comparison, Final Report
Southern Research Institute, Project No. 5050-4, prepared for the U.S
Environmental Protection Agency, February 1986.
10b. Critical Review of Open Source Particulate Emission Measurements/Phas<
II - Field Tests. Field Data Analysis, and Report, Energy and Environ
mental Management, Inc., prepared for Southern Research Institute,
Birmingham, Alabama, July 1984.
(Continued)
15
-------
TABLE 2 (Continued)
10c. G. E. Muleski, Critical Review of Open Source Particulate Emission Mea-
surements: Part II - Field Comparison. Final Report, MRI Project No.
7993-L, prepared for Southern Research Institute, Birmingham, Alabama,
August 1984.
lOd. K. D. Rosbury and R. A. Zlmmer, Critical Review of Open Source Particu-
late Emission Measurements. Task 2 ~ Field Data Analysis and Report,
PEDCo Environmental, Project No. 4181-67, prepared for Southern Re-
search Institute, Birmingham, Alabama.
lOe. E. T. Brookman, Critical Review of Open Source Particulate Emission
Measurements: Part II - Field Comparison. TRC Environmental Con-
sultants, Project No. 2681-1, prepared for Southern Research Institute,
Birmingham, Alabama, July 1984.
lOf. T. F. Eckle, Critical Review of Open Source Particulate Emission Mea-
surements - Phase II - Field Test, United States Steel Corporation,
September 1984.
11. G. E. Muleski and C. Cowherd, Jr., Evaluation of the Effectiveness of
Chemical Dust Suppressants on Unpaved Roads, Final Report, MRI Project
No. 8127-L, prepared for the U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, November 1986.
16
-------
SECTION 4
EVALUATION CRITERIA FOR THE TEST REPORTS
In selecting candidate open dust source factor and control efficiency
emission data for inclusion in AP-42, primary consideration Is given to the
relative reliability of the new data compared to that currently contained irf
AP-42 for the same source. This section describes: EPA's quality rating
system for AP-42 emission factors; methods for determining open dust source
emission factors; the emission factor rating system used 1n this study; and
the design and rating of sampling strategies used to determine control per-
formance evaluation studies.
4.1 EPA'S EMISSION FACTOR QUALITY RATING SYSTEM
The emission factor rating system developed by the U.S. EPA, Office of
Air Quality Planning and Standards (April 1980) is described in the follow-
ing paragraphs.
Oata obtained from source tests, material balance studies, and engi-
neering estimates are used to calculate the emission factors presented in
AP-42. These data are obtained from a variety of sources, including pub-
lished technical papers and reports, documented emission testing results,
and personal communications. Oata provided by individual sources vary from
single values, to ranges of minimum and maximum values, and finally to
empirical formulas (predictive emission factor equations) which allow for
correction of emission factors to specific source conditions. Some data
sources provide complete details about collection and analysis procedures,
whereas others may provide little information of this type.
17
-------
The rating system for a particular emission factor data set is based
on the following data standards:
A - Tests performed by a sound methodology and reported in enough de-
tail for adequate validation. These tests are not necessarily
EPA reference method tests, although such reference methods are
certainly to be used as a guide.
B - Tests that are performed by a generally sound methodology but
lack enough detail for adequate validation.
C - Tests that are based on an untested or new methodology or that
lack a significant amount of background data.
D - Tests that are based on a generally unacceptable method but may
provide an order-of-magnitude value for the source.
The following criteria are used to evaluate test reports for sound
methodology and adequate detail:
1. Source operation. The manner in which the source was operated is
well documented in the report. The source was operating within
typical parameters during the test.
2. Sampling procedures. If actual procedures deviated from standard
methods, the deviations are well documented. Procedural altera-
tions are often made fn testing an uncommon type of source. When
this occurs, an evaluation is made of how such alternative pro-
cedures could influence test results.
3. Sampling and process data. Many variations can occur without
warning during testing, and sometimes without being noticed.
Such variations can induce wide deviations in sampling results.
If a large spread between test results cannot be explained by
18
-------
Information contained in the test report, the data are suspect
and are given a lower rating.
4. Analysis and calculations. The test reports contain original raw
data sheets. The nomenclature and equations used are compared to
those specified by EPA, to establish equivalency. The depth of
review of the calculations is dictated by the reviewers' confi-
dence in the ability and conscientiousness of the tester, which
in turn is based on factors such as consistency of results and
completeness of other areas of the test report.
An A-rated test may be a source test, a material balance, or some other
methodology, as long as it is generally accepted as a sound method or mea-
suring emissions from that source.
In the ideal situation, a large number of A-rated source test sets
representing a cross section of the industry are reduced to a single value
for each individual source by computing the arithmetic mean of each test
set. The emission factor is then computed by calculating the arithmetic
mean of the individual source values. Alternatively, regression analysis
is used to derive a predictive emission factor equation for the entire
A-rated test set. No B-, C-, or D-rated test sets are used in the calcula-
tion of the emission factor because the number of A-rated tests is suffi-
cient. This ideal method of calculating an emission factor is not always
possible because of lack of A-rated data.
If the number of A-rated tests is so limited that inclusion of B-rated
tests would improve the emission factor, then B-rated test data are include*
in the compilation of the arithmetic mean. No C- or D-rated test data are
averaged with A- or B-rated test data. The rationale for inclusion of any
B-rated test data is documented 1n the background information.
If no A- or B-rated test series are available, the emission factor is
the arithmetic mean of the C- and D-rated test data. The C- and D-rated
19
-------
test data are used only as a last resort, to provide an order-of-magnitude
value.
In AP-42, the reliability of these emission factors is indicated by an
overall Emission Factor Rating ranging from A (excellent) to E (poor).
These ratings take into account the type and amount of data from which the
factors were calculated.
The use of a statistical confidence interval may seem desirable as a
more quantitative measure of the-reliability of an emission factor. Because
of the way an emission factor data base is generated, however, prudent ap-
plication of statistical procedures precludes the use of confidence inter-
vals unless the following conditions are met:
The sample of sources from which the emission factor was deter-
mined is representative of the total population of such sources.
The data collected at an individual source are representative of
that source (i.e., no temporal variability resulting from source
operating conditions could have biased the data).
The method of measurement was properly applied at each source
tested.
Because of the almost Impossible task of assigning a meaningful confidence
limit to the above variables and to other industry-specific variables, the
use of a statistical confidence interval for an emission factor is not
practical.
The following emission factor ratings are applied to the emission fac-
tor table.
A - Excellent. Developed only from A-rated test data taken from many
randomly chosen facilities in the industry population. The source
20
-------
category is specific enough to minimize variability within the source
category population.
B ~ Above average. Developed only from A-rated test data from a rea-
sonable number of facilities. Although no specific bias is evident,
it is not clear if the facilities tested represent a random sample of
the industry. As in the A-rating, the source category is specific
enough to minimize variability within the source category population.
C ~ Average. Developed only from A- and B-rated test data from a
reasonable number of facilities. Although no specific bias is evident
1t is not clear if the facilities tested represent a random sample of
the industry. As in the A rating, the source category Is specific
enough to minimize variability within the source category population.
D - Below average. The emission factor was developed only from A- and
8-rated test data from a small number of facilities, and there may be
reason to suspect that these facilities do not represent a random sam-
ple of the industry. There also may be evidence of variability within
the source category population. Limitations on the use of the emlssior]
factor are footnoted in the emission factor table.
E - Poor. The emission factor was developed from C- and D-rated test
data, and there may be reason to suspect that the facilities tested do
not represent a random sample of the industry. There may be evidence
of variability within the source category population. Limitations on
the use of these factors are always footnoted.
Because the application of these factors is somewhat subjective, the reasons
for each rating are documented in the background information.
21
-------
4.2 METHODS OF EMISSION FACTOR DETERMINATION
Fugitive dust emission rates and particle size distributions are dif-
ficult to quantify because of the diffuse and variable nature of such
sources and the wide range of particle size involved including particles
which deposit immediately adjacent to the source. Standard source testing
methods, which are designed for application to confined flows under steady-
state, forced-flow conditions, are not suitable for measurement of fugitive
emissions unless the plume can be drawn into a forced-flow system.
4.2.1 Mass Emissions Measurement
For field measurement of fugitive mass emissions from sources of in-
terest in this report, basic techniques have been defined:
1. The quasi-stack method involves capturing the entire particulate
emissions stream with enclosures or hoods and applying conven-
tional source testing techniques to the confined flow.
2. The roof monitor method involves measurement of particulate con-
centrations and airflows across well defined building openings
such as roof monitors, ceiling vents, and windows, followed by
calculation of particulate mass flux exiting the building.
3. The upwind-downwind method involves measurement of upwind and
downwind particulate concentrations, utilizing ground based sam-
plers under known meteorological conditions, followed by calcu-
lation of source strength (mass emission rate) with atmospheric
dispersion equations.
4. The exposure profiling method involves simultaneous, multipoint
measurements of particulate concentration and wind speed over the
effective cross-section of the plume, followed by calculation of
net particulate mass flux through integration of the plume pro-
files.
22
-------
Because it 1s usually impractical to enclose open dust sources or to
capture the entire emissions plume, only the upwind-downwind and exposure
profiling methods are suitable for measurement of particulate emissions frorti
most open dust sources. These two methods are discussed separately below.
The basic procedure of the upwind-downwind method involves the measure-!
ment of particulate concentrations both upwind and downwind of the pollutant
source. The number of upwind sampling instruments depends on the degree of
isolation of the source operation of concern (i.e., the absence of Inter-
ference from other sources upwind). Increasing the number of downwind in-
struments improves the reliability in determining the emission rate by pro-
viding better plume definition. In order to reasonably define the plume
emanating from a point source, instruments need to be located at two down-
wind distances and three crosswind distances, at a minimum. The same sam-
pling requirements pertain to line sources except that measurement need not
be made at multiple crosswind distances.
Net downwind (i.e., downwind minus upwind) concentrations are used as
input to dispersion equations (normally of the Gaussian type) to backcalcu-
late the particulate emission rate (i.e., source strength) required to gen-
erate the pollutant concentration measured. Emission factors are obtained
by dividing the calculated emission rate by a source activity rate (e.g.,
number of vehicles, or weight of material transferred per unit time). A
number of meteorological parameters must be concurrently recorded for input
to this dispersion equation. At a minimum the wind direction and speed muslj
be recorded on-site.
While the upwind-downwind method is applicable to virtually all types
of sources, it has significant limitations with regard to development of
source-specific emission factors. The major limitations are as follows:
1. In attempting to quantify a large area source, overlapping of
plumes from upwind (background) sources may preclude the determination of
the specific contribution of the area sourc
23
-------
2. Because of the impracticality of adjusting the locations of the
sampling array for shifts in wind direction during sampling, it cannot be
assumed that plume position is fixed in the application of the dispersion
model.
3. The usual assumption that an area source is uniformly emitting
does not allow for realistic representation of spatial variation in source
activity.
4. The typical use of uncalibrated atmospheric dispersion models in-
troduces the possibility of substantial error (a factor of three according
to Turner, 1970) in the calculated emission rate, even if the stringent
requirement of unobstructed dispersion from a simplified (e.g., constant
emission rate from a single pint) source configuration is met.
The other measurement technique, exposure profiling, offers distinct
advantages for source-specific quantification of fugitive emissions from
open dust sources. The method uses the isokinetic profiling concept that
is the basis for conventional (ducted) source testing. The passage of air-
borne pollutant immediately downwind of the source is measured directly by
means of simultaneous multipoint sampling over the effective cross section
of the fugitive emissions plume. This technique uses a mass-balance cal-
culation scheme similar to EPA Method 5 stack testing rather than requiring
indirect calculation through the application of a generalized atmospheric
dispersion model.
For measurement of nonbuoyant fugitive emissions, profiling sampling
heads are distributed over a vertical network positioned just downwind
(usually about 5'm) from the source. If total particulate emissions are
measured, sampling intakes are pointed into the wind and sampling velocity
is adjusted to match the local mean wind speed, as monitored by anemometers
distributed over height above ground level.
24
-------
The size of the sampling grid needed for exposure profiling of a par-
ticular source may be estimated by observation of the visible size of the
plume or by calculation of plume dispersion. Grid size adjustments may be
required based on the results of preliminary testing. Particulate sampling
heads should be symmetrically distributed over the concentrated portion of
the plume containing about 90% of the total mass flux (exposure). For ex-
ample, assuming that the exposure from a point source is normally distrib-
uted, the exposure values measured by the samplers at the edge of the grid
should be about 25% of the centerline exposure.
To calculate emission rates using the exposure profiling technique,
a conservation of mass approach is used. The passage of airborne particu-
late (i.e., the quantity of emissions per unit of source activity) is ob-
tained by spatial integration of distributed measurements of exposure
(mass/area) over the effective cross section of the plume. The exposure
is the point value of the flux (mass/area-time) of airborne particulate
integrated over the time of measurement. The steps in the calculation
procedure are presented in the paragraphs below.
For directional samplers operated isokinetically, particulate exposures
may be calculated by the following equation:
m - CcQc1
E = 5 = 2.83 x 10 5 -V- (2)
a a
where:
E = particulate exposure, mg/cm2
M = net particulate mass collected by sampler, mg
a = sampler intake area, cm2
Cg = net particulate concentration, pg/m3
Qs = sampler flow rate, CFM
t = duration of sampling, min
25
-------
The coefficient of Equation 2 is a conversion factor. Net mass or concen-
tration refers to that portion which is attributable to the source being
tested after subtraction of the contribution from background.
For nondirectional samplers (with size-specific inlets), exposure must
be calculated by the following equation:
E = 3.05 x 10"8 CfiUst (3)
where the symbols are defined as above and Us is the approaching wind speed
(in fpm). The resulting exposure values represent the specific particle
size range sampled.
The integrated exposure for a given particle size range is found by
numerical integration of the exposure profile over the effective area of
the plume. Mathematically, this is stated as follows:
h L
I - J J 10 E dA = / / 10 E dydz (4)
A 0-L
where:
I = integrated exposure, g
E = particulate exposure, mg/cm2
A = effective area of plume aboveground, m
z = vertical coordinate measured from ground level, m
y = horizontal coordinate measured from center of plume, m
h = effective vertical extent, m
L = one-half of effective horizontal extent, m
Note that, for a line source, exposure is constant with respect to y and
only a single integration over height is required. Physically, I represents
the total passage of airborne particulate matter downwind of the source.
26
-------
4.2.2 Particle Sizing
High-volume cascade impactors with glass fiber impaction substrates,
which are commonly used to measure mass size distribution of atmospheric
particulate, may be adapted for sizing of fugitive particulate emissions.
A cyclone preseparator (or other device) is needed to remove coarse par-
ticles which otherwise would be subject to particle bounce within the 1m-
pactor causing fine particle bias. Once again, the sampling intake should
be pointed into the wind and the sampling velocity adjusted to the mean
local wind speed by fitting the intake with a nozzle of appropriate size.
The size-selective inlet (SSI) for a standard high-volume sampler is
also designed to capture particulate matter smaller than 15 pm in aero-
dynamic diameter. This unit is much less wind sensitive than dichotomous
samplers, but it does not provide a cutpoint at 2.5 pm. However, it can be
adapted for use with a high volume cascade impactor to define a mass size
distribution of smaller than 15 pm in diameter. Recently, size-specific in-^
lets with 10 pm cutpoints have become commercially available in anticipation
of revision of the NAAQS for particulate matter.
Additional methods that have been used to obtain particle size distri-
butions include stacked filtration units (Cahill et al., 1979) and both
optical and electron microscopy. The relative merits of these techniques,
as evaluated by an independent contractor 1n a collaborative field study,
are discussed in Section 5.0.
4.2.3 Emission Factor Derivation
Usually the final emission factor for a given source operation, as pre*
sented in a test report, is derived simply as the arithmetic average of the
individual emission factors calculated from each test of that source. Fre-
quently the range of individual emission factor values is also presented.
27
-------
As an alternative to the presentation of a final emission factor as a
single-valued arithmetic mean, an emission factor may be presented in the
form of a predictive equation derived by regression analysis of test data.
Such an equation mathematically relates emissions to parameters which char-
acterize source conditions. These parameters may be grouped into three
categories:
1. Measures of source activity or energy expended (e.g., the speed
and weight of a vehicle traveling on an unpaved road).
2. Properties of the material being disturbed (e.g., the content of
suspendable fines in the surface material on an unpaved road).
3. Climatic parameters (e.g., number of precipitation-free days per
year on which emissions tend to be at a maximum).
An emission factor equation is useful if it is successful in "explaining"
much of the observed variance in emission factor values on the basis of
corresponding variances in specific source parameters. This enables more
reliable estimates of source emissions on a site-specific basis.
A generic emission factor equation is one that is developed for a
source operation defined on the basis of a single dust generation mechanism
which crosses industry lines. An example would be vehicular traffic on un-
paved roads. To establish its applicability, a generic equation should be
developed from test data obtained in different industries.
4.3 EMISSION FACTOR QUALITY RATING SCHEME USED IN THIS STUDY
The uncontrolled emission factor quality rating scheme used in this
study is identical to that used 1n an earlier update (Cowherd et al., 1983)
and represents a refinement of the rating system developed by EPA for AP-42
emission factors, as described in Section 4.1. The scheme entails the
rating of test data quality followed by the rating of the emission factor(s)
developed from the test data.
28
-------
Test data that were developed from well documented, sound methodologies
were assigned an A rating. Data generated by a methodology that was generally
sound but either did not meet a minimum test system requirements or lacked
enough detail for adequate validation received a B rating.
In evaluating whether an upwind-downwind sampling strategy qualified
as a sound methodology, the following minimum test system requirements were
used. At least five particulate measuring devices must be operated during
a test, with one device located upwind and the others located at two down-
wind and three crosswind distances. The requirements of measurements at
crosswind distances is waived for the case of line sources. Also wind
direction and speed must be concurrently on-site.
The minimum requirements for a sound exposure profiling program were
the following. A vertical line grid of at least three samplers is suffici-
ent for measurement of emissions from line or moving point sources while a
two-dimensional array of at least five samplers is required for quantifi-
cation of fixed virtual point source emissions. At least one upwind sampled
must be operated to measure background concentration, and wind speed must
be measured concurrently on-site.
Neither the upwind-downwind nor the exposure profiling method can be
expected to produce A-rated emissions data when applied to large, poorly
defined area sources, or under very light and variable wind flow conditions,
In these situations, data ratings based on degree of compliance with mini-
mum test system requirements were reduced one letter.
After the test data supporting a particular single-valued emission
factor were evaluated, the criteria presented in Table 3 were used to as-
sign a quality rating to the resulting emission factor. These criteria
were developed to provide objective definition for: (a) industry repre-
sentativeness; and (b) levels of variability within the data set for the
source category. The rating system obviously does not include estimates of
statistical confidence, nor does it reflect the expected accuracy of fugi-
tive dust emission factors relative to conventional stack emission factors.
29
-------
It does, however, serve as useful tool for evaluation of the quality of a
given set of emission factors relative to the entire available fugitive
dust emission factor data base.
TABLE 3. QUALITY RATING SCHEME FOR SINGLE-VALUED EMISSION FACTORS
Code
No. of
test sites
No. of
tests
per site
Total
No. of
tests
Test data
variability
Adjustment
for EFh
rating
1
> 3
> 3
< F2
0
2
> 3
> 3
-
> F2
-1
3
2
> 2
> 5
< F2
-1
4
2
> 2
> 5
> F2
-2
5
-
-
> 3
< F2
-2
6
-
-
> 3
> F2
-3
7
1
2
2
> F2
-3
8
1
2
2
> F2
-4
9
1
1
1
-4
a Data spread in relation to central value. F2 denotes factor of two.
Difference between emission factor rating and test data rating.
Minimum Industry representativeness is defined in terms of number of
test sites and number of tests per site. These criteria were derived from
two principles:
1. Traditionally, three tests of a source represent the minimum re-
quirement for reliable quantification.
2. More than two plant sites are needed to provide minimum industry
representati veness.
30
-------
The level of variability within an emission factor data set was defined
1n terms of the spread of the original emission factor data values about
the mean or median single-valued factor for the source category. The fairly
rigorous criterion that all data points must lie within a factor of two of
the central value was adopted. It is recognized that this criterion is not
insensitive to sample size in that for a sufficiently large test series, at
least one value may be expected to fall outside the factor-of-two limits.
However, this is not considered to be a problem because most of the current
single-valued factors for fugitive dust sources are based on relatively
small sample sizes.
Development of quality ratings for emission factor equations also
required consideration of data representativeness and variability, as in
the case of single-valued emission factors. However, the criteria used
to assign ratings (Table 4) were different, reflecting the more sophisti-
cated model being used to represent the test data. As a general principle,
the quality rating for a given equation should lie between the test data
rating and the rating that would be assigned to a single-valued factor
based on the test data. The following criteria were established for an
emission factor equation to have the same rating as the supporting test
data:
1. At least three test sites and three tests per site, plus an
additional three tests for each independent parameter in the
equation.
2. Quantitative Indication that a significant portion of the emissior]
factor variation 1s attributable to the independent parameter(s)
in the equation.
31
-------
TABLE 4. QUALITY RATING SCHEME FOR EMISSION FACTOR EQUATIONS
No.
of
No. of
Total No.
Adjustment .
Code
test
sites
tests per site
of tests
for EF rating
1
£
3
£ 3
£ (9 + 3P)
0
2
2
£ 3
§ 3P
-1
3
£
1
-
< 3P
-2
a P denotes number of correction parameters in emission factor equa-
tion.
b Difference between emission factor rating and test data rating.
Loss of quality rating in the translation of test data to an emission
factor equation occurs when these criteria are not met. In practice, the
first criterion was far more influential than the second in rating an emis-
sion factor equation, because development of an equation implies that a
substantial portion of the emission factor variation is attributable to the
independent parameter(s). As indicated in Table 4, the rating was reduced
by one level below the test data rating if the number of tests did not meet
the first criterion, but was at least three times greater than the number
of independent parameters in the equation. The rating was reduced two
levels if this supplementary criterion was not met.
The rationale for the supplementary criterion follows from the fact
that the likelihood of including "spurious" relationships between the de-
pendent variable (emissions) and the independent parameters in the equation
increases as the ratio of number of independent parameters to sample size
increases. For example, a four parameter equation based on five tests
would exhibit perfect explanation (R2=1.0) of the emission factor data, but
the relationships contained in such an equation cannot be expected to hold
true in independent applications.
32
-------
4.4 DESIGN AND RATING OF CONTROL PERFORMANCE EVALUATION STUDIES
As noted above, control of open dust sources has recently attracted a
great deal of attention. For instance, almost all field studies of paved
and unpaved road dust emissions since 1980 have entailed the evaluation of
some type of control technique. In general, however, this type of informa-
tion is not included in the current version of AP-42.
Field evaluation of control efficiency requires that the study design
include not only adequate emission measurement techniques (of the type dis-
cussed in Section 4.2) but also a proven "control application plan." In
the past, two major types of plans have been used:
Type 1 - Controlled and uncontrolled emission measurements are ob-
tained simultaneously.
Type 2 - Uncontrolled tests are performed initially, followed by con-
trolled tests.
In order to ensure comparability between the operating characteristics
of the controlled and uncontrolled sources, many evaluations are forced to
employ Type 2 plans. An example would be a wet suppression system used on
a primary crusher. One important exception to this, however, is unpaved
road dust control. In this instance, under a Type 1 plan, testing is con-
ducted on two or more contiguous road segments. One segment is left un-
treated and the others are treated with a separate dust suppressant.
Under a Type 2 plan, uncontrolled testing is initially performed on
one or more road segments, generally under worst-case (dry) conditions.
Each segment is then treated with a different chemical; no segment is left
untreated as a reference. A normalization of emissions may be required to
allow for differences in vehicle characteristics during the uncontrolled
and controlled tests because they do not occur simultaneously. For example
a change in vehicle mix should not be interpreted mistakenly as part of the
efficiency of the control measure being tested.
33
-------
The method used to'normalize emission factors 1s generally based on
the AP-42 predictive emission factor equation for the source under con-
sideration. For example, for unpaved roads, emission factors are scaled
by:
where:
en = normalized value of the emission factor corresponding to
e. = measured emission factor from run 1
Sn = normalizing value for average vehicle speed
S.j = average vehicle speed during run i
Wn = normalizing value for average vehicle weight
Wj = average vehicle weight during run 1
wn =¦ normalizing value for average number of wheels per vehicle
pass
Wj = average number of wheels per vehicle pass during run 1
Note that surface material properties (such as silt content) present in the
AP-42 equations are not considered in the normalization process because the
control measure affects these properties.
Regardless of the control plan selected, it is important that, for the
purpose of estimating annual or seasonal controlled emissions from unpaved
roads, average control efficiency values be based on worst-case (i.e., dry)
uncontrolled emission levels. This is true simply because the AP-42 un-
paved road predictive equation 1s based on source tests conducted under dry
conditions. Extrapolation to annual average emissions estimates is accom-
plished by assuming that emissions are occurring at the estimated rate on
days without measurable precipitation, and conversely are absent on days
(5)
run i
34
-------
with measurable precipitation. This assumption has never been verified in
a rigorous manner; however, MRI's experience with hundreds of field tests
indicate that it is a reasonable assumption if the source operates on a
fairly "continuous" basis.
The uncontrolled emission factor for a specific unpaved road will in-
crease substantially after a precipitation event as the surface dries.
However, in the absence of data sufficient to describe this growth as a
function of traffic parameters, amount of precipitation, time of day,
season, cloud cover, and other variables, uncontrolled emissions are esti-
mated using the simple assumption given above. Thus, in order to defini-
tively estimate emission reductions attributable to a dust suppressant,
control efficiency should be referenced to uncontrolled emissions under dry
conditions. An extended discussion of the interrelationship of control and
natural mitigation is provided elsewhere (Muleski, 1986a).
Finally, it is important that appropriate specification of an efficiency
value depends on the nature of the control. In broad terms, control mea-
sures can be considered as either continuous or periodic, as the following
examples illustrate:
Continuous Controls
Periodic Controls
Wet suppression for
materials handling
Local exhaust hoods
Watering or chemical
treatment of unpaved
roads
Enclosures
Sweeping of paved
travel surfaces
Vegetation of exposed
areas
Chemical stabilization
of exposed areas
35
-------
The major difference between the two types of controls is related to the
time dependency of performance. For continuous controls, efficiency is es-
sentially constant with respect to time. On the other hand, the efficiency
associated with periodic controls tends to decrease (decay) with time. An
example would be chemical treatment of an unpaved road; immediately after
application, the road surface is thoroughly wetted and complete control is
assumed. After curing, a generally high level of control is observed for
approximately 1 week. Thereafter, the efficiency tends to decrease until
the next application (at which time the cycle repeats).
In order to quantify the performance of a specific control, two mea-
sures of control efficiency are required. The first is "instantaneous"
control and is defined by
where:
c(t) - instantaneous control (percent) at t days after applica-
tion
e (t) = emission factor for the controlled source t days after
application
eu = uncontrolled emission factor
For a continuous control technique, the controlled emission factor,
ec(t), is essentially independent of time. Consequently, the control c(t) in
the above expression is also independent of time and is representative of a
continuous technique* On the other hand, for a periodic control, the value
of c(t) in Eq. (5) represents the (instantaneous) level of control over a
specific test period and, hence, at a particular time after application.
(6)
36
-------
The other Important measure of control performance is average effi-
ciency, defined as:
i T
C(T) = ^ J c(t) dt (7)
o
where:
C(T) = average control efficiency during period ending T days
after application (percent)
c(t) = instantaneous control efficiency at t days after applica-
tion (percent)
T = time period over which average control efficiency is de-
sired (days)
Average control efficiency values are needed to estimate emission reduc-
tions due to periodic applications. Note, however, that if c(t) equals a
constant, the two measures are identical.
The rating of reported control efficiency values presents numerous
difficulties. As can be seen from Eq. (5), control efficiency values are
defined as non-linear functions of two emission factors. Both the uncon-
trolled and controlled emission factors may carry their own ratings of A
through E. However, in order to assign a rating to their ratio (I.e., con-
trol efficiency estimate), the interrelationship between the two ratings
must be understood. At the present time, no attempt to investigate this
relationship has been undertaken.
Additional complications arise if the control method being evaluated
is periodic in nature. In this instance, any inherent variability (due to
source conditions, measurement error, etc.) of emission levels about a mean
value at a given point in time is confounded by the fact that the "mean"
controlled emission level varies over time after application. Thus, a
rating applied to a control estimate under these circumstances would in-
volve not only the relative reliability of the ratio of controlled and
uncontrolled emission factors (as discussed above) but also the temporal
variation of the ratio's reliability.
37
-------
As an example, suppose it is known that a control decays linearly from
100% at time zero to 0% at 30 days. Further assume that a sampling method
measures a known unit emission factor as either 0.8 or 1.25 (with equal
probability) and that a single controlled test is performed 15 days after
application. Thus if the "true" uncontrolled emission level was 10 kg/VKT,
then, with equal probability, this emission level would be measured as
8 or 12.5 kg/VKT. At 15 days, the "true" controlled value is 5 kg/VKT,
but would be measured as 4 or 6.25 kg/VKT with equal probability. In this
way, four separate outcomes are possible and each has a probability of
one-fourth:
Measured emission
factor (kg/VKT) Measured control
Outcome Uncontrolled Controlled efficiency (%)
A 8 4 50
8 8 6.25 22
C 12.5 4 68
D 12.5 6.25 50
These four outcomes, as well as the linearly extrapolated lifetime
estimated by each, are shown graphically 1n Figure 1.
Under this very idealized situation, there is a one-fourth probability
that the 30-day lifetime of the control would be estimated as either 19
or 47 days, and only a one-half probability that the lifetime would be
estimated as 30 days. Furthermore, if this experiment were repeated a
large number of times, the mean lifetime would tend towards (1/4 x 19 +
1/4 x 47 + 1/2 x 30) = 31.5 days rather than the known value of 30 days.
38
-------
o
/o
zo
"Pays aftbc Application
Figure 1. Hypothetical decay curves.
-------
The simple example above does not consider replicate tests at a given
time, testing at various times after application, variation in source con-
ditions, or any of the other complicating effects described earlier. How-
ever, the fact that an Idealized situation can lead to such widely different
estimates of control effectiveness does illustrate some of the difficulties
in rating performance data collected in the "real world."
Consequently, no ratings are applied to control performance tests pre-
sented 1n this report. Rather, the recommendations made 1n Section 6.0 are
based on qualitative judgment of the reliability and applicability of
available data in estimating emission reductions with AP-42 methods.
40
-------
SECTION 5
CANDIDATE EMISSION FACTORS AND CONTROL EFFICIENCY DATA
The test reports listed earlier in Table 2 were grouped by relevant
subsections of AP-42 Section 11.2; the resulting list is given in Table 5.
In this section, individual test reports are discussed in terms of:
(a) field sampling methodology (including control efficiency evaluation, If
applicable); (b) number of tests; and (c) location of tests. Quality rat-
ings (based upon the schemes presented in Tables 3 and 4) are assigned to
the emission factor data.
TABLE 5. LIST OF TEST REPORTS BY PERTINENT SUBSECTION
Subsection
Title
Test report
-
General Testing Methods
10a,b,c,d,e,f
11.2.1
Unpaved Roads
1,2,3,7,11
11.2.3
Aggregate Storage Piles
4,5,6,9
11.2.6
Industrial Paved Roads
1,5,8
Note that test reports 10a through lOf have not been assigned to a
specific Section 11.2 subsection. As discussed in Section 3 above, these
collaborative tests were more an examination of testing methodologies than
an evaluation of any type of particulate emission source. Each of the testi
ing organizations participating in this collaborative study also performed
41
-------
field tests presented in Table 2. Because the Implications of this criti-
cal review have direct bearing on the results from the other test reports,
the collaborative studies will be discussed first.
5.1 GENERAL TESTING METHODOLOGIES
Test Report 10a
This report discusses the results of a collaborative field comparison
of exposure profiling techniques. The field study compared the results
obtained by five testing organizations for the same source (a "simulated"
unpaved road formed by artificially loading a paved road at an integrated
iron and steel plant in Indiana). All exposure profiling systems in this
study met the minimum requirements of Section 4.3. Emission factors for
total particulate (TP) as well as four smaller size fractions (less than or
equal to 30, 15, 10, and 2.5 pm in aerodynamic diameter) were reported.
A total of 11 tests were performed during June 1984. Five testing
locations were demarcated for use by different organizations. A different
position was occupied by each testing group on a given day; during the week,
each group conducted at least one test at each position. Note that, at the
start of the week, each organization deployed a standard high-volume (hi-vol)
sampler downwind of the simulated source. These samplers were not moved
during the week of testing and were operated by the organization occupying
the testing location on that day. Filters used in these hi-vols were sup-
plied and analyzed by the independent contractor supervising the collabora-
tive study.
At the start of testing, the road surface loading was estimated to be
approximately 200,000 kg/km (600,000 lb/mile). Vehicle mixes during field
sampling exhibited average weights between 6 and 30 Mg (7 and 33 tons) and
average speeds between 26 and 40 kph (16 and 25 mph). This information is
summarized in Table 6.
42
-------
TABLE 6. SOURCE TESTING INFORMATION (Test Reports 10b to 10f)a
. Test No. of
Operation Equipment Material Location date tests
Vehicle traffic
Vehicle mix
Simulated unpaved road
Indiana
6/84
11
a Because this was a collaborative field test, source parameters were identical in all five test
reports.
^ In order to complete the comparative testing in reasonable time, a paved road was artificially loaded
to increase mass emission levels.
-------
Table 7 lists the average emission factors determined for this source
by the different testing organizations and the range of conditions tested.
Quality ratings were not assigned because a simulated source was evaluated,
and results do not pertain to any actual source category. Additional re-
sults from the collaborative study are presented after individual discus-
sions of the test reports submitted by the five testing organizations.
Test Report 10b
The sampling system for this study was composed of an upwind and down-
wind tower supporting three and five sampling heads, respectively. Each TP
sampling head consisted of a high volume sampler motor, a vertically ori-
ented filter, and a nozzle with a rectangular inlet. Intake flows were
monitored using potentiometers and manometers. (Note that this is essen-
tially the same profiling system used in Test Report 7.)
Additional equipment included: (a) a standard high volume (hi-vol)
sampler (these samples were analyzed by the independent contractor super-
vising the field study); and (b) meteorological instruments for wind speed
and direction. The latter equipment was used to manually set Intake flow
rates based on 15-min averages prior to testing. A uniform velocity dis-
tribution with height was assumed.
Particle sizing employed computer-controlled, scanning electron micros
copy (CCSEM) which was performed for one of the six tests deemed valid. An
additional test was also selected for analysis to estimate size distribu-
tions for tests conducted with wet road surfaces.
Test Report 10c
Major sampling equipment for this study Included a five-head exposure
profiling tower downwind of the source as well as cyclones at two heights
both upwind and downwind of the source. The two downwind cyclones were
fitted with five-stage cascade impactors and'the lower upwind unit included
a three-stage Impactor. The cyclone and impactor units allowed particle
44
-------
TABLE 7. RANGE OF CONDITIONS AND EMISSION FACTORS (Test Reports 10b to lOf)
Range of conditions
Suspended
particulate
PM10
Emission
Test
report
No. of Silt
tests (%)
Moisture
(%>
Wind speed
(mph)
Vehicle speed
(mph)
Vehicle weight
(ton)
emission
factor '
Emission
factor
factor
units
10b
11 6.4-21.0
0.2-1.5
3.6-12.6
20.4-28.0
6.3-32.4
6.56
1.92
kg/VKT
10c
11 6.9-11.1
-
4.8-14.0
16.0-25.0
7.1-33
3.58
1.88
kg/Via
lOd
11 3.9-10.8
2.9-15.2
3.6-11.5
20-30
7.0-34.0
2.37
1.16
kg/VKT
lOe
11 6.39-13.69
-
5.06-11.81
20.19-27.94
7.46-34.98
5.3
1.8
kg/VKT
lOf
11 11.0-16.0
-
6.26-15.26
17.89-24.35
6.5-24.26
7.1
2.6
kg/VKT
- Information not contained in test report.
a Particles < 30 pm aerodynamic diameter.
b Particles < 10 pm aerodynamic diarater.
c Emission factors are arithmetic mean of the 11 test runs. Values are also presented on pages 44 and 62 of Test Report
10a.
-------
size distributions to be determined for Individual tests. (Note that this
1s essentially the same profiling system employed in Test Reports 1, 3, 5,
9, and 11.)
Additional equipment included a standard hi-vol (again analyzed by the
Independent contractor) and wind speed and direction sensors mounted to the
downwind profiling tower. Sampling intake rates were adjusted based on
10-min averages and an assumed logarithmic wind profile. Nozzles were
employed on the cyclone preseparators to adjust Intake velocities while
maintaining a constant volumetric flow.
Test Report lOd
Identical upwind and downwind profiling towers were used in this study.
Each tower supported four sampling heads which employed the stacked filtra-
tion concept (Cahlll et al., 1979). The stacked filtration units (SFUs)
simultaneously provide two separate particle size fractions: 30 and 2.5 umA
cutpoints are assigned to the stainless steel 400 mesh screen and 8.0 |jm
Nuclepore filter, respectively. A 0.8 (jm Nucleopore filter is used as the
backup filter. (Note that this is essentially the same sampling system used
in Test Report 2.)
Additional equipment included a standard hi-vol analyzed by the in-
dependent contractor and a meteorological station to record wind speed and
direction. The latter type of information was used (10-min averages) to
adjust sampling intake rates for isokinesis. It should be noted that,
although Isokinetic sampling allows the determination of TP emission fac-
tors, this type of Information is not typically reported.
Test Report lOe
The downwind profiling tower used in this study supported five sampling
heads. An identical sampling head was deployed upwind of the test road.
The flow rate for each head was servo-controlled, with adjustments made on
the basis of wind speed anemometers located near the head. (Note that this
1s essentially the same sampling system used in Test Report 4.)
46
-------
Additional equipment included standard hi-vols upwind and downwind of
the test road (the downwind unit was analyzed by the independent contractor!)
as well as wind speed/direction systems both upwind and downwind.
Particle size distributions were obtained by CCSEM (using the same
subcontractor as in Test Report 10b). A total of 26 filters were analyzed
and reported.
Test Report lOf
The profiling system used in this study employed essentially the same
type of sampling heads as in Test Report lOe. A total of four heads were
supported by the downwind tower and a single head was deployed upwind.
Additional equipment included a standard hi-vol (analyzed by the independent
contractor) and a wind speed/direction system. Particle size distributions
were obtained by an in-house CCSEM analysis. The development of this pro-
filing is described in Test Report 8.
Results of the Collaborative Study
The five profiling systems evaluated during the collaborative study
were all found to be capable of producing essentially equivalent results in
terms of TP emissions. Note, however, that this statement is based on only
three runs if Test Report lOd is included (as noted above, this organiza-
tion does not usually report total particulate emissions). The other four
organizations, using a variety of techniques (e.g., fixed versus variable
flow rates, manual versus automatic adjustment, different integration
schemes for exposure measurements, etc.), obtained equivalent results over
all 11 tests.
For smaller particle size fractions, however, there was considerably
less agreement among the results obtained by the five organizations. Fig-
ure 2 is a copy of Table 5 from Test Report 10a and presents the correla-
tion matrix between emission factors over the span of the 11 tests.
47
-------
CORRELATION COEFFICIENTS OF TP, TSP, PM-0, AND FP EMISSION FACTORS
REPORTED BY TEST PARTICIPANTS
TP
USS
MRI
EEM
TRC
PEI
USS
-
.770
.802
.787
-
MRI
.770
-
.584
.645
-
EEM
.802
.584
-
.728
-
TRC
.787
.645
.728
-
-
PEI
-
-
-
-
-
HL
0 «10
lim)
USS - .915 .669 .703 .625
MRI .915 - .636 .780 .566
EEM .669 .636 - .718 .564
TRC .703 .780 .718 - .251
PEI .625 .566 .564 .251 -
TSP (<30 urn)
OSS
MRI
EEM
TRC
PEI
-
.904
.772
.795
.504
.904
-
.594
.705
.490
.772
.594
-
.728
.486
.795
.705
.728
-
.286
.504
.490
.464
.286
-
F.
•0
•
A
to
t
.5 pin)
-
.865
.472
.703
.224
.865
-
.612
.764
.434
.472
.612
-
.700
.520
.703
.764
.700
-
.124
.224
.434
.520
.124
-
Figure 2. Photocopy of Table 5 from Test Report 10a. Note that USS,
MRI, EEM, TRC, and PEI correspond to Test Reports lOf,
c, b, e, and d, respectively.
-------
The particle sizing method, stack filtration units (SFUs), employed in|
Test Report lOd purportedly separated collected particulate matter into two
size fractions, with nominal cutpoints of 30 and 2.5 pmA. The Independent
evaluation (Test Report 10a), however, noted that the 30 pmA collection
medium described in Test Report lOd was selected on the basis of pore size
and the collection efficiency was not verified by calibration. In addition!
it was noted that interception and impaction would probably result in a cutf
point substantially smaller than the 30 pmA assigned.
Furthermore, with the increased loading of the collection surface, the
cutpoint would tend to decrease even more from the nominal value assigned.
The latter point was supported by analysis of penetration versus both
(a) h1-vol and (b) profiler catches. In each case, a negative correlation
between penetration and sample catch was found. Thus, with the buildup of
sample mass, the SFUs would tend to become more efficient filtration device?
and, consequently, decreased values for particulate concentrations and emis*
sion factors would be obtained.
Additional substantiation of this behavior is provided in Figure 2.
the two nominal size fractions measured (2.5 and 30 pmA), neither emission
factor presented in Test Report lOd shows a positive correlation with the
other reported results that could be considered significant at the 5% level
Thus, there is no significant tendency for Test Report lOd results to in-
crease when results of the other organizations increase.
It was also noted in the independent evaluation that, because catch
weights were quite close to detection limits in several cases, sampling tim4
could not generally be shortened to avoid variation of cutpoints with load-
ing. Finally, because the sizing system in Test Report lOd did not allow
for a direct measurement of the 10 pmA fraction, two Interpolation schemes
were employed. Note, however, that the use of both schemes on the same data
did not provide independent estimates, as implied in Test Report lOd. Be-
cause of the various problems noted above, the use of stack filtration unit$
(SFUs) could not be recommended by the independent contractor monitoring thd
collaborative study.
49
-------
The CCSEM technique was used by three (Test Reports 10a, lOe, and lOf)
of the five contractors to provide particle size data. Aside from the dif-
ficulties that arise from assigning both volumes and equivalent aerodynamic
diameters to the irregular, inhomogeneous particles encountered, the inde-
pendent contractor also noted fundamental problems with both the upper and
lower ends of the size spectrum. Based on these problems, the Independent
contractor found the CCSEM "methodology, as used by the participants of
this test, unsuitable for applications of this type."
The only sizing technique recommended by the independent contractor
was the inertial separation method by cyclone/lmpactor combinations (Test
Report 10c). Although the independent contractor suggested some variations
to the analysis procedure used 1n Test Report 10c, it was found that the
results from either procedure were generally not significantly different
from one another.
Figure 3 is a reproduction of Table 8 from Test Report 10a and presents
size fractions for 30, 15, 10, and 2.5 pmA. The second entry for each test
is the mass fraction contained in Test Report 10c while the first is the
value recalculated by the independent contractor. A Wilcoxon signed rank
test (McGhee, 1985) indicated that, of all size ranges considered, only the
FP (2.5 pmA) fraction showed differences significant at the 10% level (but
not at the 5% level). Consequently, there 1s no discernible difference be-
tween the two procedures over particle size ranges of current regulatory
interest, although there is reason to suspect that the FP fractions may be
different.
Additional comparisons presented in Test Report 10c indicate that
(a) a 6-m profiling system gives results very comparable to a 7.5 m tower;
(b) a three-stage impactor provides essentially identical 15 and 10 pmA
size fractions as does a five-stage impactor (however, the 2.5 pmA fraction
may be overestimated by a factor of approximately 1.2); and (c) sizing re-
sults obtained by either two- or one-step cyclone washes were essentially
equivalent.
50
-------
1
2
3
4
5
6
7
8
9
1
1
SIZE FRACTIONS OF PARTICULATE AS REPORTED BY MRI AND RECALCULATED Bf
SOqgjgggJjgjSgARCH INSTITUTE |
Size Fraction as Percent Less Than
Organization
TP
(total)
TSP
(<30 ub)
PM1S
(<15 un)
PM10
(<10 urn)
FP
(<2.5 unij
SoRI
MRI
1.00
1.00
0.453
0.395
0.313
0.284
0.240
0.216
0.094
0.066
SoRI
MRI
1.00
1.00
0.412
0.427
0.306
0.263
0.234
0.202
0.098
0.064
SORI
MRI
1.00
1,00
0.487
0.374
0.350
0.243
0.280
0.183
0.113
0.058
SORI
MRI
1.00
1.00
0.527
0.448
0.326
0.303
0.250
0.236
0.107
0.077
SoRI
MRI
1.00
1.00
0.427
0.689
0.274
0.547
0.212
0.449
0.090
0.170
SoRI
MRI
1 .00
1.00
0.507
0.439
0.352
0.294
0.279
0.230
0.120
0.078
SoRI
MRI
1.00
1.00
0.506
0.567
0.357
0.405
0.283
0.320
0.117
0.105
SORI
MRI
1.00
1.00
0.349
0.318
0.233
0.212
0.177
0.164
0.077
0.056
SoRI 1.00 0.304 0.216 0.174 0.070
MRI 1.00 0.373 0.196 0.152 0.054
SoRI 1.00 0.333 0.216 0.166 0.074
MRI 1.00 0.419 0.294 0.230 0.076
SORI 1.00 0.457 0.333 0.240 0.089
MRI 1.00 0.480 0.332 0.253 0.072
Figure 3. Photocopy of Table 8 of Test Report 10a.
51
-------
5.2 UNPAVED ROAQS (SECTION 11.2.1)
Test Report 1
In this study, exposure profiling was used to quantify the performance
of unpaved road dust controls. The control techniques evaluated included
water and a petroleum resin product. Both controls were tasted for heavy-
duty traffic; only the petroleum resin was evaluated for light-duty ve-
hicles. A Type 2 control application plan was employed.
Field tests were conducted at an integrated iron and steel plant in
Ohio. Table 8 shows the number of tests conducted for each source/control
combination.
Total particulate emission measurements were obtained during each test
as part of the exposure profiling technique. Both four and five head pro-
filing systems (4 to 5 m high) were used during this program. Cyclone/
impactor combinations were used for particle size measurements at two down-
wind heights. Additional sampling equipment included: (a) standard h1-vols
both upwind and downwind of the test road; (b) one or two size-selective
inlets (15 m^A cutpoint) upwind of the source; and (c) recording instruments
for wind speed and direction to adjust for isokinetic sampling. This meets
the requirements of a sound, well-documented methodology in Section 4.3,
and emission test data are rated A.
Table 9 presents the average emission factors determined during the
study and the range of test conditions. Note that, because this study was
directed to control performance evaluation rather than emission factor de-
velopment, no quality ratings are assigned.
Chemically controlled unpaved roads were tested during the first
2 days after application; consequently, no long-term average control effi-
ciency values were obtained during this study. In addition, the control
efficiency data obtained for chemically treated heavy-duty unpaved roads
should be considered less reliable because controlled and uncontrolled
52
-------
TABLE 8. SOURCE TESTING INFORMATION (Test Report 1)
Operation
Equipment
Material
Location
Test
dates
No. of
tests
Vehicle
traffic
Heavy-duty vehicles
Unpaved roads un-
controlled
Ohio
11/80
3
Vehicle
traffic
Ligh-duty vehicles
Unpaved roads un-
controlled
Ohio
7/80
4
Vehicle
traffic
Light-duty vehicles
Upaved roads con-
trolled
Ohio
10/80
5
Vehicle
traffic
Heavy-duty vehicles
Unpaved roads con-
controlled
Ohio
11/80
7
Vehicle
traffic
Medium-duty vehicles
Paved roads un-
controlled
Ohio
7/80, 10/80
11/80
7
Vehicle
traffic
Medium-duty vehicles
Paved roads con-
trolled
Ohio
7/80, 10/80
11/80
5
Vehicle
traffic
Medium-duty vehicles
Paved roads un-
controlled
Texas
7/80
4
Vehicle
traffic
Medium-duty vehicles
Paved roads con-
control led
Texas
6/80
7
-------
TABLE 9. RANGE OF CONDITIONS AND EMISSION FACTORS (Test Report 1)
Range of conditions Inhalable
Wind
Vehicle
Vehicle
particulate
PHio
Emission
Material/equipment/
No. of
Silt
Moisture
speed
speed
weight
enisslon
factor
Emission
factor
operation
Location
tests
(X)
(X)
(mph)
(mph)
(ton)
factor
units
Light-duty unpaved h
road/uncontrol1ed
Ohio
4
-
-
1.6-6 2
15
3
3.05
-
lb/VHT
Light-duty unpaved
road/Coherex8
Ohio
5
0 015-1.8
-
4.0-9.1
25
3
0.27
-
lb/VMT
Heavy-duty unpaved .
Ohio
3
14-16
-
7.4-9 5
20
22-53
8.37
-
lb/VHT
road/uncontrolled
Heavy-duty unpaved
Ohio
3
4.5-5.1
-
5.5-6.4
20-25
53-54
2.24
-
lb/VHT
road/water 1 rig
lb/VMT
Heavy-duty unpaved
road/CoherexS
Ohio
4
2.5-5.4
-
5.2-9.3
15-22
19-54
0.53
-
Paved road/uncon-
Ohio
7
10 4-35.7
-
4.0-12
-
14-40
0.68
-
lb/VMT
trolled"
Paved road/vacuum
sweeping
Ohio
4
IB.3-27.9
-
4.5-6.4
-
8.3-18
0.37
-
lb/VMT
Paved road/water
Ohio
1
9.45
-
9.0
-
29
1.32
-
lb/VHT
flushing
Paved road/water
Texas
4
28.2-34.3
-
3 0-5.7
-
9.4-11
0.23
-
lb/VHT
flushing-and broom
sweeping**
Paved road/water
Texas
3
11 2-22.6
-
5 4-8.6
-
9.2-11
0.41
-
lb/VMT
flushing
Paved road/uncon-
trol led
Texas
4
6 45-14.0
-
3.6-6.6
-
11-12
0.95
-
lb/VHT
- = Information not contained in test report
a
Particles
. < 15 pm aerodynamic diameter.
b
Emission
factor
is
the
aritlimetic mean
of
test
runs
F-59, F-60, F-63, and F-64 from page 48, Table 3-11 of test report.
c
Emission
factor
is
the
arithmetic mean
of
test
runs
F-65, F-66, and F-67 from page 48, Table 3-11 of test report.
d
Emission
factor
is
the
arithmetic mean
of
test
runs
F-68, F-69, and F-70 from page 50, Table 3-12 of test report.
e
Emission
factor
is
the
arithmetic mean
of
test
runs
F-65, F-66, and F-67 from page 50, Table 3-12 of test report.
f
Emission
factor
is
the
arithmetic mean
of
test
runs
F-59, F-60, F-63, and F-64 from page 50, Table 3-12 of test report.
g
Emission
factor
is
the
arithmetic mean
of
test
runs
F-34, F-35, F-61, F-62, F-27, F-4S, and F-32 from page 74. Table 3-27 of test report
h
Emission
factor
is
the
arithmetic mean
of
test
runs
F-36, F-37. F-38, and F-39 from page 74, Table 3-27 of test report.
i
Emission
factor
is
the
value obtained 1
from test run
F-74 on page 74, Table 3-27 of test report.
J
Emission
factor
IS
the
arithmetic mean
of
test
runs
B-50, B-51, 8-52, and 8-53 from page 74, Table 3-27 of test report.
k
Emission
factor
is
the
arithmetic mean
of
test
runs
, 6-54, B-55, and B-56 from page 74, Table 3-27 of test report.
1
Emission
factor
is
the
arithmetic mean
Qf
test
runs
B-57, 8-58, 8-59, and 8-60 from page 74, Table 3-27 of test report.
-------
tests were not conducted at the same site. Watering tests of a heavy-duty,
unpaved road, as presented 1n Figure 3-7 of the test report, imply a control
efficiency decay rate of approximately 9%/hr over all size ranges considered
Test Report 2
This study evaluated the performance of several unpaved road dust
controls under heavy-duty traffic at three surface coal mines. Results wer4
obtained using exposure profiling with a Type 1 control application plan.
Tests were conducted at mines in southern Illinois and in southwestern
and northeastern Wyoming. Dust suppressants evaluated included: a salt; a/)
acrylic cement; two emulsified asphalts; an enzyme; a Hgnon sulfonate; and
water. With the exception of the last three, controls were generally
evaluated using both topical and incorporated applications.
Air sampling was primarily accomplished using an exposure profiling
tower which employed SFUs of the type discussed earlier. Consequently, 30
and 2.5 emission factors and control efficiencies are presented in the
report. Additional instruments deployed included: (a) dustfall buckets;
(b) a RAM-1 aerosol monitor; (c) a quartz crystal cascade analyzer (QCCA);
and (d) wind speed/direction recording equipment. The QCCA was not found tq
be suitable for the field tests; as a result, 15 pmA SSIs were substituted
at the second and third mines. The test report states that IP control ef-
ficiencies were based on measured concentrations (rather than mass emission
rates used for TSP and FP); however, no summary information about IP con-
trol 1s provided in the report.
The profiling system employed meets the minimum requirements of Sec-
tion 4.3. General documentation is adequate. However, 1n light of the
findings by the Independent contractor in Test Report 10a, the SFUs employe*)
in this study cannot be expected to yield reliable measurements for emis-
sions in the two nominal particle size fractions. Consequently, the emis-
sions data obtained in this study should be downgraded from a rating of A
to C. Although the emission measurements cannot be considered reliable, th$
55
-------
ratio of controlled and uncontrolled emission levels might be expected to
provide control efficiency values of greater reliability.
Table 10 identifies the number of tests and source/control combina-
tions evaluated at each mine. Average emission factors and range of source
conditions are presented in Table 11.
Control efficiency values over time were reported for each mine/control
combination. However, many of these combinations exhibit apparent Increases
in efficiency over time. This anomalous behavior is possibly due to the
fact that efficiency values were not referenced to dry, uncontrolled emis-
sions (cf. Section 4.4).
Test Report 3
This study represents a continuation of the control performance evalu-
ations begun in Test Report 1. Three unpaved road dust controls were evalu-
ated: an emulsified asphalt; a petroleum resin; and water. The evaluations
were conducted under medium to heavy duty traffic at two steel plants in
Indiana and Missouri. Unlike Test Report 1, the primary focus in this study
was the long-term decay of chemical dust suppressants applied to unpaved
roads. Table 12 summarizes the source testing information presented in the
report. A Type 2 control plan was employed.
The primary sampling equipment for this study Included a four-head
(6 m tall) exposure profiling tower with cyclone/impactor combinations at
two downwind heights. Total particulate concentrations upwind of the source
were measured using a cyclone/impactor combination. An additional cyclone
was deployed upwind for controlled tests. Wind speed and direction were
continuously monitored, and 10 min averages were used to maintain isokinetic
sampling. As 1n Test Report 1, the test data are rated A. Table 13 pre-
sents average emission factors and ranges of source conditions.
56
-------
est
6
12
2
8
12
20
18
16
16
20
12
41
TABLE 10. SOURCE TESTING INFORMATION (Test Report 2)
Equipment
Material'
Location
Heavy-duty vehicles
Heavy-duty vehicles
Heavy-duty vehicles
Heavy-duty vehicles
Heavy-duty vehicles
Heavy-duty vehicles
Heavy-duty vehicles
Heavy-duty vehicles
Heavy-duty vehicles
Heavy-duty vehicles
Heavy-duty vehicles
Heavy-duty vehicles
Unpaved road with CaCl2
Unpaved road with
acrylic
Unpaved road witji emul-
sified asphalt
Unpaved road with
lignon
Unpaved road with water
Unpaved road uncon-
trolled
Unpaved road with CaCl2d
Unpaved road with emul-
sified asphalt
Unpaved raad with
acrylic
Unpaved goad with
lignon
Unpaved road with waterc
Unpaved road uncon-
trolled
Southern Illinois
Southern Illinois
Southern Illinois
Southern Illinois
Southern Illinois
Southern Illinois
Southwest Wyoming
Southwest Wyoming
Southwest Wyoming
Southwest Wyoming
Southwest Wyoming
Southwest Wyoming
(Continued)
-------
TABLE 10. (Continued)
Test No. of
Operation Equipment Material Location dates tests
Vehicle
traffic
Heavy-duty vehicles
Unpaved road with CaCl2^
Northeast Wyoming
10/83-11/83
8
Vehicle
traffic
Heavy-duty vehicles
Unpaved road with biocat9
Northeast Wyoming
10/83
5
Vehicle
traffic
Heavy-duty vehicles
Unpaved road wit^ emul-
sified asphalt
Northeast Wyoming
10/83
4
Vehicle
traffic
Heavy-duty vehicles
Unpaved road with
lignon sulfonate
Northeast tyyoming
10/83
8
Vehicle
traffic
Heavy-duty vehicles
Unpaved roadhuncon-
controlled
Northeast Wyoming
10/83
19
a Oust suppressants were applied to unpaved roads to determine the control effectiveness.
The dust suppressant was applied in two sections of the unpaved road using different application
techniques. See Table 6.2, page 6-5 of test report for dust suppressent application.
c After each dust suppressant was applied to the unpaved road, a section was left for no control.
^ The dust suppressant was applied in two sections of the unpaved road using different application
techniques. See Table 6.7, page 6-13 of test report for dust suppressant application.
0
After each dust suppressant was applied to the unpaved road, a section was left for no control.
^ The dust suppressant was applied in two sections of the unpaved road using different application
techniques. See Table 6.12, page 6-21 of test report for dust suppressant application.
9 Oust suppressant was applied in one section of the unpaved road using one application technique.
See Table 6.12, page 6-21 of test report for dust suppressant application.
^ After each dust suppressant was applied to the unpaved road, a section was left for no control.
-------
TABLE 11. RANGE OF CONDITIONS AND EMISSION FACTORS (Test Report 2)
Materia)/equipment/
operation
Location
No. of
tests
Silt
(*>
Range of conditions
Wind Vehicle
Moisture
(%)
speed
(ntph)
speed
(mph)
Vehicle
weight
(ton)
Suspended
particulate
emission,
factor *
Fine
particulate
Emission
emission factor
factor 1 units
Unpaved road/ Southern 6
with CaCL2 II1inois
Unpaged road/with Southern 12
acrylic Illinois
Unpaved road/with Southern 2
enulsified asphalt Illinois
Unpaved road/with Southern 8
1ignon Illinois
Unpaved road/water Southern 12
water Illinois
Unpaved road/uncon- Southern 20
trolled Illinois
Unpaved road/wlth Southwest 18
CaCl2 Wyootng
Unpaved road/with Southwest 16
emulsified asphalt Wyoming
Unpaved road/with Southwest 16
acrylic Wyoming
Unpaved road/with Southwest 20
1 ignon Wyoming
Unpaved road/with Southwest 12
water Wyctning
Unpaved road/uncon- Southwest 41
trolled Wyoming
2 20-5.32 0 16-3 07
1.64-3.59 0.06-0.55
3.0-5.15 0.31-1.23
2.05-3.69 0.08-0.17
1.29-8.21 0.17-0.79
1 82-8.03 0 2-2.2
3.37-6.0 0.4-4.80
2.06-4.28 1 0-4.6
2 16-4 92 0 B-3.4
33 9-40.3 28.2-65.9
2.01
5.69
1.0
20 8-41.1 22.2-88.7 3.42
33 6
60.8
8.6S
2.33-13.6 0.4-5.4
32 4-43.2 16 0-65 2 6 08
37.1-49.3 38.6-61.6 2.77
20.8-49.3 16.0-8B.7 4.47
27.0-38.3 43.6-83.3 7.71
25.0-36.9 18.1-67.4 13.84
34.9-45.1 37.5-87.4 7.28
37.8-48.9 51.3-69.2 7.14
29 2-40.0 13.5-89.9 6.22
25.0-48.9 13.5-82.4 14.42
0 12
0 68
0 62
0.48
0.64
0.80
0.66
1.32
0.89
0.82
0.74
1.27
lb/VHT
lb/VHT
lb/VHT
lh/VMT
lb/VMT
lb/VHT
lb/VHT
lb/VWT
Ib/VMT
lb/VHT
lb/VHT
lb/VMT
(Continued)
-------
TABLE 11. (Continued)
Material/equipment/
operation
Location
Mo. of
tests
Silt
(X)
Range of conditions
Wind Vehicle
Moisture
(*)
speed
(mph)
speed
(mph)
Vehicle
weight
(ton)
Suspended
particulate
emission
factor '
Fine
particulate
emission
factor *c
Emission
factor
units
Unpaved road/with
CaCl2
Unpaved road with/
biocat
Unpaved road/with
emulsified
asphalt
Unpaved road/with
1 igon
Unpaved roadiun-
controlled
Northeast
Wyoming
Northeast
Wyoming
Northeast
Wyoming
Northeast
Wyoming
Northeast
Wyoming
8
5
4
8
19
2.0-5.3 1.0-2 80
29.4-35.6 47.2-155.0 3.03
6.7-8.3
4.5-5.7
0.4-0.8
2.6-4.8 0.8-1.6
0.4-1.4
23.1
20.2
13.4-169.9 3.58
76.4-132.9 1.79
23.0-36.6 17.0-157.0 1.84
20.0-48.9 13.4-169.9 3.36
0.53
0.29
0.16
0.23
0.45
Ib/VHT
lb/VMT
lb/VMT
lb/VHT
lb/VMT
- = Information not contained in test report.
a Particles < 30 pm aerodynamic diameter See discussion in Section 5 1.
b Particles < 2.5 pa aerodynamic diameter. See discussion in Section 5.1
C Emission factors are arithmetic mean of the test runs from Tables B-l, B-2, and B-3 of test report.
^ Emission factors are the arithmetic mean from 12 of the 16 test runs. Data were missing for the four test runs as found in Table B-2 of test report.
Emission factors are the arithmetic mean from three of the five test runs. Data were missing for the two test runs as found in Table 8-3 of test
report.
Emission factors are the arithmetic mean from 17 of the 19 test runs. Data were missing for the two test runs as found in Table B-3 of test report.
-------
TABLE 12. SOURCE TESTING INFORMATION (Test Report 3)
Operation
Equipment
Material
Location
Test
dates
No. of
tests
Vehicle traffic
Medium-duty vehicles
Uncontrolled unpaved
road
Indiana
6/82
3
Vehicle traffic
Medium-duty vehicles
Unpaved road with
asphalt emulsion
Indiana
6/82-10/82
8
Vehicle traffic
Medium-duty vehicles
Uncontrolled unpaved
road
Missouri
9/82
3
Vehicle traffic
Medium-duty vehicles
Unpaved road with
water
Missouri
9/82
3
Vehicle traffic
Medium-duty vehicles
Unpaved road with
petroleum resin
Missouri
9/82-11/82
8
Vehicle traffic
Medium to heavy-duty
vehicles
Unpaved road with
petroleum resin
reappli ed
Missouri
11/82-12/82
4
-------
TABLE 13. RANGE OF CONDITIONS AUG EMISSION FACTORS (Test Report 3)
Total
Materlal/equipment/
operation
Location
No of
tests
Silt
(X)
Moisture
(X)
Wind
speed
(mph)
Vehicle
speed
(mph)
Vehicle
weight
(ton)
particulate
emission
factor
PM.o
Emission
factor
Emission
factor
units
Uncontrolled unpaved
road
Indiana
3
5.8-7.5
-
4.2-5.8
15-17
25-2B
18.9
3.5
lb/VHT
Unpaved road with f
asphalt emulsion
Indiana
8
0.28-13
-
2.2-6.6
13-15
23-34
4.13
0 50
lb/VHT
Uncontrolled unpaved
road9
Missouri
3
6.3-7.7
-
2.0-4.2
15
50-54
16.7
2.98
lb/VHT
Unpaved^road with
water
Missouri
3
4.9-5.3a
-
4.4-6.1
15
48-50
3.27
0.37
Ib/VMT
Unpaved road with,
petroleum resin
Missouri
6
1.5-7.1
-
2.8-12
15-22
27-56
8.97
1.03
lb/VHT
Unpaved road with
petroleum .resin
reapplied-1
Missouri
4
0.034-1.7b
4.9-8.8
28-49
31-54
1.53
0.22
lb/VHT
: Information not contained in test report.
One sample missing.
The oass of one sample was so small as to be undetectable.
Airborne particles regardless of size.
Particles < 10 pm aerodynamic diameter.
Emission factors are arlthnetic mean of test runs AG-1, AG-2, and AG-3 from page 53, Table 3-6 of test report.
Emission factors are arithmetic mean of test runs AG-4 through AG-II from page 53, Table 3-6 of test report.
® Emission factors are arithmetic mean of test runs AJ-l, AJ-2, and AJ-3 from page 53, Table 3-6 of test report.
^ Emission factors are arithmetic mean of test runs AJ-4,'AJ-5, and AJ-6 from page S3, Table 3-6 of test report.
Emission factors are arithmetic mean of test runs AJ-7 through AJ-12 from page 53, Table 3-6 of test report. Tests AJ-16, -17 excluded because of high
moisture contents.
Emission factors are arithmetic mean of test runs AJ-13, -14, -15, and -18 from page 53, Table 3-6 of test report.
-------
Efficiency decay rates were obtained for each control application. Th$
decay rate for watering was found to be comparable to that in Test Report 1
despite differences 1n ambient temperature and application intensity.
Test Report 7
This paper was primarily directed to a discussion of a generic unpavedl
road dust suppressant that could be produced on-site at iron and steel
plants. Exposure profiling of both controlled and uncontrolled emissions
from unpaved roads is also described; however, little information about
the field tests is presented. Profiling was conducted under subcontract
by the same testing organization 1n Test Report 10b. Because no reference
to an earlier test report is given in the paper, it is assumed that this
paper represents the original source of the test data. Test data are
rated B because of the lack of adequate documentation and because individual
test results are not presented.
Both commercially available and generic petroleum resin products were
evaluated in this study. Testing took place at a coke and iron facility ini
Pennsylvania during the fall of 1983. Available testing information and
test results are presented in Tables 14 and 15, respectively. For both sup4
pressants, the test report presented control efficiency values for three
times after an Initial application and once after a repeat application.
Test Report 11
Largely representing a continuation of the control evaluations con-
ducted in Test Reports 1 and 3, this study discussed the long-term perfor-
mance of five unpaved road dust suppressants used in the iron and steel
industry. Field tests were conducted at the same test sites as in Test
Report 3.
The suppressants evaluated were; an emulsified asphalt; an acrylic
cement; both commercial and generic petroleum resin products; and a salt.
63
-------
TABLE 14. SOURCE TESTING INFORMATION (Test Report 7)
Operation
Equipment
Material
Location
Test
dates
No. of
tests
Vehicle traffic
Heavy-haul trucks
Unpaved road with oil
resin/water emulsion
Pennsylvania
10/83-11/83
4C
Vehicle traffic
Heavy-haul trucks
Unpaved road with, com-
mercial product
Pennsylvania
10/83-11/83
4C
Vehicle traffic
Heavy-haul trucks
Uncontrolled unpaved
road
Pennsylvania
10/83
9
Dust suppressant is the generic formula developed for the study.
Commercial petroleum resin dust suppressant.
Four controlled emission factors were reported. However, report states that each is based on two tests
where possible.
-------
TABLE 15. RANGE OF CONDITIONS, EMISSION FACTORS, AND RATINGS
(Test Report 7)
Range of conditions
Wind Vehicle Vehicle Emission
Material/equipment/ No. of Silt Moisture speed speed weight Emission factor
operation tests (%) (%) (niph) (mph) (ton) factor units
Unpaved road with oil 4 - - - - - 0.69
resin/water emul-
sion
Unpaved road with 4 - - - - - 0.37a
commercial
product
Uncontrolled unpaved 9 ~ 5.99^
road
Information not contained in report.
Emission factor is arithmetic mean of four test runs from Table 5. See footnote c in Table 14. Units
not given but believed to be Ib/VMT.
Uncontrolled emission factor based on nine test runs. Units not stated but believed to be lb/VMT.
-------
The report used code letters for each suppressant to discourage selective
citation of test results.
The basic study design incorporated exposure profiling with a Type 1
control application plan. Because of difficulties encountered at the
Indiana site, however, a Type 2 plan was later adopted. Testing informa-
tion 1s presented in Table 16.
The primary sampling equipment included four-head profiling towers
(6 m height), a downwind eye1 one/1mpactor combination, and an upwind stan-
dard hi-vol fitted with a cascade Impactor. The height at which downwind
particle size measurements were obtained was selected on the basis of prior
testing and approximated the point 1n the dust plume at which half the mass
emissions pass above and half below. Additional instrumentation included
recording wind speed and direction sensors used to maintain isokinetic
sampling.
The profiling tests used a sound methodology and were well documented.
As noted above, results from the Indiana site were potentially influenced by
upwind sources and nearby structures. For this reason, exposure profiling
test data from plant AP are rated B while data from plant AQ are rated A.
Test results and ranges of source conditions are presented in Table 17.
Additional studies were conducted prior to field testing in order to
characterize both unpaved road traffic and dust suppressant usage in the
iron and steel industry. The results of these surveys were used to evaluate
the suppressants under service conditions representative of unpaved roads
in the iron and steel industry. In addition, the test report combines con-
trol efficiency decay rates obtained during the field study with earlier
results to develop a control performance model for petroleum resin products.
5.3 AGGREGATE STORAGE PILES (SECTION 11.2.3)
Three test reports were Identified that pertain to the source activi-
ties currently discussed in Section 11.2.3 of AP-42. In contrast to the
66
-------
es
6
6
3
5
9
11
11
5
6
2
TABLE 16. SOURCE TESTING INFORMATION (Test Report 11)
Equipment
Material
Location Test dates
Medium to heavy-duty
vehicles
Unpaved road with Coherex® Indiana 5/85-8/85
Medium to heavy-duty
vehicles
Unpaved road with Petro Tac® Indiana 5/85-8/85
Medium to heavy-duty
vehicles
Unpaved road with calcium Indiana 5/85-8/85
chloride
Medium to heavy-duty
vehicles
Uncontrolled unpaved road Indiana 5/85-8/85
Medium to heavy-duty
vehicles
Unpaved road with Coherex® Missouri 9/85-11/85
Medium to heavy-duty
vehicles
Unpaved road with Soil
Seraent®
Missouri 9/85-11/85
Medium to heavy-duty
vehicles
Unpaved road with Generic 2 Missouri 9/85-11/85
Medium to heavy-duty
vehicles
Unpaved road with Petro Tac® Missouri 9/85-11/85
Medium to heavy-duty
vehicles
Unpaved road with calcium Missouri 9/85-11/85
chloride
Medium to heavy-duty
vehicles
Uncontrolled unpaved road Missouri 9/85-11/85
-------
TABLE 17. RANGE OF CONDITIONS, EMISSION FACTORS. AND RATINGS
(Test Report 11)
Hange of conditions
Total
Wind
Vehicle
Vehicle
particulate
PHio
Emission
Mat er 1 a 1 / equ I potent/
No. of
Silt
Moisture
speed
speed
weight
emission
Emission
factor
operation
Location
tests
(X)
C%)
(mph)
{mph)
(ton)
factor
factor
units
Unpaved road/ with
Indiana
-------
studies discussed above, these reports largely deal with emission factor
development. These reports are discussed below.
Test Report 4
This report presents the results of field tests conducted on dust
emissions from rotary coal car dumping at a power plant in Maryland. The
car dumper is enclosed in an east-west shed and thus should be considered
at least partially controlled. Source testing information is provided in
Table 18.
Mass flux measurements were taken at both the entrance and exit doors
to characterize mass emissions leaving the shed. However, only one doorway
was sampled during the first phase of the program. Additional upwind/
downwind and dustfall measurements were taken within the shed. Sampling
equipment/operation met the requirements of Section 4.3.
The mass flux samplers were the same isokinetic units described in
Test Report lOe. Particle size measurements were obtained for most tests
which exhibited westerly winds; measurements were made using five-stage
cascade impactors at three of the sampling locations. These units were
also fitted with preseparating settling chambers (which are referred to as
horizontal elutriators in Brookman, 1983). The preseparator cutpoints
are given as 30 and 42 pm for 20 and 40 cfm flow rates, respectively.
During Phase I, the impactors units were operated at 20 cfm. However,
because difficulties were encountered in maintaining that flow rate, Phase tl
employed a 40 cfm flow rate. Additional problems were reported involving
particle bounce through the impactor. It should be noted that the choice
of preseparator in this study actually compounded particle bounce problems
because it allowed larger particles to enter (and potentially bounce throug(i)
the impactor at higher flow rates. For this reason, particulate emissions
data for the smaller size ranges should be downgraded to a B rating.
Final size distributions reported were obtained by multiplying the
average impactor data obtained during Phase II by the ratio of optical
69
-------
TABLE 18. SOURCE TESTING INFORMATION (Test Report 4)
Operation
Equipment
Material
Location
Test
date
No. of
tests
Batch-drop
Rotary railcar
dumper
Coal
Maryland
7/83-11/83
62
-------
microscopy to impaction results for one test. Results of testing are pre-
sented in Table 19.
Additional regression analysis was performed on the test results, with
separate predictive models, based on moisture content, developed for washed
and unwashed coal. However, the relationships were considered weak and un-
suitable for prediction purposes. Average suspended particulate emission
factors for washed and unwashed coal were 0.0006 and 0.0016 lb/ton, respec-
tively.
Test Report 5
This report describes the results of field tests conducted at a prillejl
sulfur facility in California. Particulate emissions from the batch drop of
prill into a partial enclosure were quantified. The enclosure was con-
structed to simulate a melter feed hopper. This simulation was necessary
because the client did not use prill at its facility; consequently, test-
ing was performed at a prill production plant.
Freshly produced wet prill was allowed to dry prior to testing. The
same batch of sulfur was dropped repeatedly to simulate various handling
operations and to determine any increase in emissions with increased fines
content. Over six tests, the material was transferred a total of 18 times.
Additional source testing information is presented in Table 20.
Total particulate mass measurements were obtained using profiling
heads of the type discussed in Test Report 10c. A total of six units were
arranged in a two-dimensional array 2 m downwind of the hopper. Cyclone/
impactor combinations were used both upwind and downwind of the source.
Additional equipment included wind speed/direction Instruments used to
maintain isokinetic sampling. The test data are rated A.
Test results, range of source conditions, and quality ratings assigned
are given in Table 21. Note that the tests were undertaken to extend the
applicability of current AP-42 estimation methods and not to develop s1ngle-|
valued emission factors for the source.
71
-------
TABLE 19. RANGE OF CONDITIONS, EMISSION FACTORS, AND RATINGS
(Test Report 4)
Material/equipment/ No. of
operation tests
Range of conditions
Silt
(%)
Moisture
(%)
Wind
speed
(mph)
Suspended
particulate
emission
factor
PM10 Emission
Emission factor
factor units
Coal batch-drop
railcar dumper
62 0.7-4.8° 2.7-7.4 0.5-2.9*
1.0-3.4
0.0011e 0.00024 lb/ton
NA = Not applicable.
- = Information not found in test report.
a Wind speed at exit doorway of dumper shed from page 47-49, Table 5-3 of test report.
k Wind speed at entrance doorway of dumper shed from page 47-49, Table 6-3 of test report.
C Average moisture content from page 55-56, Table 5-6 of test report.
d Average silt content from page 55-56, Table 5-6 of test report.
e Emission factor is arithmetic mean for 62 test runs from Table 5-8, page 59-60 of test re-
port. Value is also given on page 72 of test report. Emission factor rated D, based on
Table 3. See discussion in text for additional emission factors for washed and unwashed
coal.
^ Individual factors not presented; mean PM10 fraction reported as 0.22 on page 72. Emission
factor rated E, based on Table 3.
-------
TABLE 20. SOURCE TESTING INFORMATION (Test Report 5)
Operation
Equipment
Material
Location
Test
dates
No. of
tests
Batch drop3
Front-end loader'5
Wet formed prilled
sulfur
California
4/84
6
Loader travel
Ford CL65 front-
end loader
Paved surface within
plant
California
4/84
3
Prilled sulfur was dumped at a constant height of 5 ft.
One cubic yard front-end loader.
-------
TABLE 21. RANGE OF CONDITIONS, EMISSION FACTORS, AND RATINGS
(Test Report 5)
Material/equipment/ No. of
operation tests
Silt
(%)
Range of conditions
Moisture
(%)
Wind
speed
(mph)
Vehicle
speed
(mph)
Vehicle
weight
(ton)
Suspended
particulate
emission
factor
PM1C
Emission
factor
Emission
factor
units
Prilled surface/
batch-drop
Paved surface/
loader travel
0.44-2.5a 1.1-2.73 5.1-17.5 NA
5.2-48
NA
9.8-13.4
NA
3.2
0.040
0.77
lb/ton
lb/VMT
NA = Not applicable.
- = Information not contained in test report.
a
b
c
d
Average of two values.
All three tests had the same value.
Particles < 30 pm aerodynamic diameter.
Emission factor is the arithmetic mean of test runs AK1 to AK6 from page 26, Table 3-5 of test report. Single-valued
factor rated D on basis of Table 3. See discussion in text.
Emission factor is the arithmetic mean of test runs AK7, AK8, and AK9 from page 26, Table 3-5 of test report.
Single-valued factor rated D on basis of Table 3. See also discussion in text.
-------
Test Report 6
This report presents dust control efficiency values for materials
handling operations involving a front endloader and dump trucks. However,
only control efficiency values are presented; no mass emission rates or
factors are given. A subsequent telephone conversation with the testing
organization indicated that the efficiency values were based on relative
measures of emissions. Because the study design did not concern absolute
measurement of emissions, such as relating total suspended emissions to a
unit activity level, this report was deleted from consideration.
Test Report 9
This study presented size-specific particulate emission factors for
the loading of fly ash into open trucks. Tests were conducted at a coal-
fired power plant 1n Michigan. Load-out was from an enclosed loading bay
below a silo used to store ash collected by an ESP. Because the truck/
trailer combinations used to haul the fly ash were slightly longer than the
loading bay, the bay doorways were alternately open and blocked by the ve-
hicle during loading operations.
To facilitate testing, a slight modification to the operation was made
to ensure that the downwind doorway was continuously blocked during the
loading process. The only additional modification Involved keeping the
bay's overhead door at a constant 12 ft height. This change reduced the
necessary height of the downwind sampling array and avoided any difficulty
In isokinetically sampling particulate mass through a flexible plastic cur-
tain at the bay doors. Because none of these changes altered the physical
operation of the load-out process, mass emissions were said to be unaffected
and merely rerouted through the open areas of the doorway. A summary of thd
testing information is provided in Table 22.
Downwind mass flux measurements were obtained using TP samplers of the
type used in Test Report XOc. A total of six downwind samplers of this typd
were used to quantify mass flux values. Particle size measurements were
75
-------
TABLE 22. SOURCE TESTING INFORMATION (Test Report 9)
Operation
Equipment
Material
Location
Test
date
No. of
tests
Batch loading
Truck/trailer combination
Fly ash
Michigan
9/84
4
-------
made using cyclone/lmpactor combinations at two locations (which were co-
located with TP samplers). Upwind concentration and size distribution mea-
surements were obtained with another cyclone/impactor combination.
Additional instrumentation included anemometers at five of the six
downwind sampling locations as well as a wind vane at the sixth point. Oner
to f1ve-m1nute averages from these instruments were used to maintain iso-
kinetic sampling. All downwind air sampling and ancillary equipment were
mounted in the plane of the doorway and were rotated into fixed locations
after the truck/trailer entered the bay. A two-parameter wind station was
deployed outside of the loading bay to record ambient wind speed and
direction.
The sampling methodology was sound and wel1-documented; because the
source modifications were not considered to affect mass emissions, the test
data are rated A. Test results, ranges of conditions, and quality ratings
assigned to the data are presented in Table 23.
5.4 INDUSTRIAL PAVED ROADS (SECTION 11.2.6)
Three test reports dealing with Section 11.2.6 of AP-42, Industrial
Paved Roads, were identified. Those reports are discussed below.
Test Report 1
This study evaluated paved road control techniques at two integrated
iron and steel plants in Ohio and Texas. Sampling methodologies used were
identical to those discussed earlier in the section on unpaved roads.
Source testing information and test results were given previously as
Tables 8 and 9, respectively. Control efficiencies are presented for
vacuum sweeping, water flushing, and flushing with broom sweeping. Re-
ported efficiencies for the latter two should be considered less reliable
because controlled and uncontrolled tests were not always performed on the
same road because of meteorological and logistical constraints.
77
-------
TABLE 23. RANGE OF CONDITIONS, EMISSION FACTORS, AND RATINGS
(Test Report 9)
Ranae of conditions
Suspended
Material/equipment/ No. of
operation tests
Silt
(%)
Moisture
(%)
Wind
speed
(raph)
particulate
emissiog
factor
PM10
Emissign
factor
Emission
factor
units
Fly a|h batch load- 4
ing
25-40d
26-30
4-6
0.0044
0.0017
1b/ton
a Particles < 30 pm aerodynamic diameter.
Particles < 10 pm aerodynamic diameter.
Emission factors are arithraetric mean of test runs AN-1, AN-2, AN-3, and AN-4 from
page 22, Table 3-4 of test report. Factors rated D on basis of Table 3.
Report states on page 18 that silt contents are believed to be lower bounds on material
finer than 200 mesh.
-------
Test Report 5
In addition to the materials handling tests described earlier, this
study also quantified particulate emissions from vehicle travel on paved
surfaces. A total of three tests were performed. The first two were con-
ducted on the rather heavily loaded surface resulting from the repeated
transfer of prill during the handling tests. The final test was conducted
after the surface had been manually cleaned by flushing and sweeping.
Source testing information was given previously in Table 20.
A five-head profiling tower (5 m tall) was used to sample total mass
flux. Additional concentration and particle size measurements were ob-
tained from cyclone/impactor combinations at two downwind and one upwind
heights. As before, test data are rated A. Testing results were given
earlier in Table 21.
Test Report 8
This paper discusses the development of an exposure profiling system
as well as an evaluation of the effectiveness of paved road vacuum sweeping,
Note that, because no reference is given to an earlier test report, this
paper is considered to be the original source of the test data.
The exposure profiling and particle sizing systems used in the study
were essentially identical to those discussed in Test Report lOf. Test dati
are B-rated because of inadequate detail in the report. Source testing in-
formation and results of the tests are presented in Tables 24 and 25, re-
spectively.
79
-------
TABLE 24. SOURCE TESTING INFORMATION (Test Report 8)
Test No. of
Operation Equipment Material Location dates tests
Vehicle traffic Average vehicle mix Paved road uncontrolled Pennsylvania - 10
Vehicle traffic Average vehicle mix Paved road controlled3 Pennsylvania - 5
- = Information not contained in Test Report.
a Control of the paved road was a twice per week vacuum sweeping program.
-------
TABLE 25. RANGE OF CONDITIONS AND EMISSION FACTORS (Test Report 8)
Range of conditions
Wind Vehicle Vehicle SP PNl0 Emission
Material/equipment/ No. of Si 1tc Moisture speed speed weight Emission Emission factor
operation tests (%) (%) (mph) (mph) (ton) factor factor units
Uncontrolled paved 10 8.95 - - - 0.5-35 0.18 - lb/VMT
road/vehicle mixa
Controlled paved b 5 16.2 - 0.5-35 0.096 - lb/VMT
road/vehicle mix
NA = Not applicable
- = Information not contained in test report.
a Emission rate is the area under the exposure profile curve from Figure 2, page 5 of test report
divided by the average number of vehicles for six of the 10 test runs.
k Emission rate is the area under the exposure profile curve from Figure 3, page 8 of test report
divided by the average number of vehicles for four of the five test runs.
Q
Values are the average silt content of the test runs.
-------
SECTION 6
DISCUSSION ANO RECOMMENDATIONS
The preceding section discussed available test reports. The test re-
port results are used in this section to assess AP-42 Section 11.2 This
assessment can be divided into three main categories: (a) use of the test
results as independent data in order to measure the accuracy of the current
AP-42 predictive emission factor equations; (b) use of the test data to ex-
pand the range of source conditions underlying the current AP-42 equations
and possibly recommend revised equations; and (c) use of the test data to
expand the current AP-42 discussions of control methods.
6.1 UNPAVED ROADS (SECTION 11.2.1)
The particle size data in Test Report 1 was employed in preparing Sec-
tion 11.2.1 for inclusion in the Fourth Edition of AP-42. Consequently, the
uncontrolled emission factors given in that report cannot be used to assess
the performance of the current AP-42 unpaved road equations when applied to
independent data sets.
The uncontrolled emission factors in Test Reports 2 and 7 were not con*
sidered for further evaluation. As discussed in Section 5.1, an independent
evaluation found that the SFUs used in Test Report 2 are incapable of pro-
viding reliable emission factors for the size ranges assigned to the device
Furthermore, Test Report 7 provides only TP emission factors; because this
size range 1s not included in Section 11.2.1, no comparisons are possible.
General agreement between predicted (using the current AP-42 equation)
and observed emissions was found to be good for the 10 tests of uncontrolled
83
-------
emissions in Test Reports 3 and 11. Summary statistics for the ratio of
predicted to observed emissions are presented belowr
< 30 umSa < 15 pmA < 10 umA < 2.5 urnA
Geometric mean 1.01 1.65 1.55 2.51
Standard geometric 1.65 1.78 1.73 1.90
deviation
Stokes diameter based on a particle density of 2.5 g/cm3. Results for
the six tests in Test Report 3 only; Test Report 11 did not present
emission factors in this size range.
In terms of particle size ranges of current regulatory interest, it would
appear that independent applications of the current AP-42 unpaved road equa-
tion yield estimates within acceptable limits.
While the travel surfaces described in Test Reports 10a through lOf
and 5 were paved, surface loadings were far above (from 3 to 80 times greater)
those underlying the current AP-42 paved road equations. As reported in
Test Reports 5 and 10c, in those instances where loose material essentially
covers a paved surface, the travel surface is perhaps better characterized
as unpaved than paved in terms of particulate emissions. Although it is un-
likely that this type of source is of common importance, slight revisions
to AP-42 Sections 11.2.1 and 11.2.6 to incorporate this finding may be war-
ranted. Section 6.3 of this report provides an additional discussion of
emissions from heavily loaded paved surfaces.
As noted throughout this report, the control of unpaved road dust
emissions has attracted considerable attention during the 1980s. Many in-
dustries have implemented plant-wide dust control programs for unpaved
roads. Independent applications of the unpaved road equation have indicated
that estimates of the emission factor (e in Eq. (1)) are generally quite
good; however, the lack of guidance on estimating the effective control term
in that equation has hindered reliable unpaved road emission estimates for
84
-------
many Industries. Note that, because most control measures applied to un-
paved roads are periodic in nature (as discussed in Section 4.3), use of thfe
time-averaged value of efficiency given in Eq. (6) is appropriate.
Owing to the wide range of: (a) available dust suppressants; (b) ap-
plication parameters (such as the intensity, dilution, and frequency of
chemical application) used for these suppressants; and (c) service environ-
ments including traffic parameters (such as average vehicle weight or daily
number of passes) and meteorological parameters (such as rainfall), it is
often very difficult to transfer prior field test results obtained under onfe
set of conditions to estimate average control efficiency values for another
set of conditions. To date, only one attempt has been made to develop a
model of average control performance. This model was developed in Test
Report 11 and is presented below.
It is important to realize that any given test series can provide only
one estimate of average control efficiency. This is true simply because the
various values of instantaneous control efficiency obtained at different
times after application must be combined to obtain an effective decay rate.
For example, although there may be a data base of 100 controlled tests, thife
data base may provide only 10 average control values.
As discussed in Test Report 11, only petroleum resin products have beefi
evaluated in enough detail to warrant an attempt at a control performance
model. Many suppressants (such as asphalt emulsions and acrylic cements)
have been evaluated under only two or three very different sets of condi-
tions.
Test Report 11 combined the results of seven long-term field evalua-
tions of petroleum resins to obtain the following models of average control
efficiency:
85
-------
Size
fraction
Nominal
averaging
period
Sample
size
Estimated average
efficiency (%)
Correlation
coefficient
TP
14 day
7
37 + 44 g
0.948
30 day
5
28 + 52 g
0.939
PM10
14 day
6
64 + 23 g
0.755
30 day
4
50 + 36 g
0.915
The variable "g" represents ground inventory (L/m2). See text for a
discussion of g. The slopes and intercepts for each of the models
were obtained from a least-squares, linear regression analysis of
seven long-term field studies.
The TP models all show correlations significant at the 2% level, while for
PM10, the corresponding level is only 10%.
The factor, g, is termed the (cumulative) ground inventory and is found
by adding together the total volume (per unit area) of chemical concentrate
(not solution) applied since the start of the dust control season. For ex-
ample, if a plant originally applied 2 L/m2 of a 20% solution on April 1,
and followed with 1.5 L/m2 of a 16% solution on the first of each following
month, then after the June 1 application, g = 0.88 L/m2. In this example,
because applications occur once a month, the nominal averaging period 1s
30 days. Thus, between June 1 and July 1, an average TP control of 28 + 52
(0.88) = 74% and PM10 control of 50 + 36 (0.88) = 82% are estimated by the
models given above.
The average TP and PM10 control performance models for petroleum resins
presented above were designed to meet typical needs in the iron and steel
industry. Application (intensity and frequency of treatment) and traffic
parameters inherent in the model reasonably span common conditions in that
industry. How well the model performs under different service conditions
(e.g., rural roads) cannot be assessed at this time. However, because
roughly two-thirds of the field tests supporting the current unpaved road
equation were conducted in the iron and steel industry-, it is reasonable to
include this model in Section 11.2.1.
86
-------
It Is recommended that AP-42 Section 11.2.1.3 (Control Methods) be re-
vised in order to include this average control performance model. Because
so many industrial unpaved roads are currently controlled, it is important
that users of AP-42 be able to estimate control efficiency values. In addi-
tion, it is recommended that any revision also include numerous references
to test reports and manuals that contain data on other control techniques
(e.g., Cowherd and Klnsey, 1986).
6.2 AGGREGATE STORAGE PILES (SECTION 11.2.3)
Unlike the case of unpaved roads, the recent test data that pertain to
this section of AP-42 Section 11.2 potentially represent a substantial in-
crease to the data base underlying current AP-42 predictive equations. Howi
ever, only the results of Test Report 5 could be used in a reexamination of
the predictive equations for materials handling. Note that run AK-5 was ex-|
eluded because the test report stated that there was strong evidence that
test was biased by wind erosion.
The results from Test Reports 4 and 9 were not included for several
reasons. The material of interest in Test Report 9—fly ash—is not
generally considered an aggregate material because it consists of fine
particles of relatively uniform size. The test report discussed problems
encountered 1n the size classification of this material by dry sieving. Re-
ported silt contents were said to represent lower bounds on the fraction of
material that would pass a 200 mesh screen during the sieving procedure dis-
cussed in Section 11.2.3. In addition, the source operation was enclosed
and partially controlled by water. Because of these differences in both
material and source operation parameters, the results from the test report
were not included.
Results from Test Report 4 were excluded for a variety of reasons. As
in Test Report 9, the source was enclosed. An additional consideration is
related to the problems encountered with particle size measurements. As
discussed 1n Section 5.0, particle bounce problems were compounded by the
choice of preseparation device, and results from impaction were essentially
87
-------
changed to results from optical microscopy. Because the independent con-
tractor in Test Report 10a could recommend only impaction for exposure pro-
filing particle size measurements and because of the obvious need for reli-
able particle size information in light of the anticipated NAAQS revision,
it was believed that problems encountered with sizing, 1n addition to the
difference in source operation, made this data set unsuitable for inclusion
(Note, however, that these test results are employed later as independent
data to assess the predictive accuracy of a revised emission factor equa-
tion. )
A preliminary step in incorporating results from Test Report 5 into
the materials handling data base involved correction to identical size
fractions. This step was necessary because the current AP-42 equations
are based on particles £ 30 jjijiS (Stokes diameter based on a density of
2.5 g/cm3), and results in Test Report 5 are for 30 pmA. Correction of
results from Test Report 5 to 30 ymS (or 50 pmA) was accomplished using a
log-normal size distribution and the data contained in the test report.
The results are given below:
Ratio for stated size3
Data base
30 |jmA
15 umA
10 pmA
5 miiiA
2.5 umA
Current batch drop'1
0.73
0.48
0.36
0.23
0.13
Current continuous dropb
0.77
0.49
0.37
0.21
0.11
Test Report 5, Runs AK-1
and -2C
0.73
0.49
0.36
0.22
0.11
Test Report 5, Runs AK-3
through-6
0.70
0.41
0.23
0.11
0.06
a Ratio of stated size fraction to fraction * 50 pmA.
b Values taken from current AP-42 Section 11.2.3.
c Values obtained from test report size data and assumed log-normal
distribution.
Note that the size distributions for the data sets are very comparable.
86
-------
Three predictive equations (for emissions £ 50 (jmA) were obtained by
stepwise linear regression (Nie et al., 1975), corresponding to the three
data subsets:
A. Current batch drop data base supplemented with Test Report 5 re-
sults (13 data points).
B. Current continuous drop data base (9 data points).
C. A and B combined (22 data points).
Potential correction parameters included:
Silt content, s (%)
Moisture content, M (9>)
Wind speed, U (mph)
Drop height, h (ft)
Dumping device capacity, Y (yd3)
Note that the last parameter pertains only to batch drop tests. The depen-
dent variable was the $ 50 pmA emission factor in pounds of emissions per
thousand tons of material transferred. All variables were log transformed
in order to obtain a multiplicative model.
Resulting equations are presented in Table 26. Note that moisture
content appears in each model; in fact, moisture was the first variable
to enter in each of the three stepwise linear regression analyses.
An analysis of variance indicated that the regression results for data
sets A and C are both significant at the 0.01% level; the corresponding
level for set B is 2.3%. Each equation obtained was further examined using
a standard cross-validation (CV) technique (Mosteller and Tukey, 1977).
Using this technique, each point in the underlying data base is excluded
one at a time, and the equation generated from the reduced data base is used
to estimate the missing value.
89
-------
TABLE 26. REGRESSION EQUATIONS OBTAINED FOR MATERIALS
HANDLING DATA SETS3
Data set
Sample
size
Predictive equation
Multiple R2
13
9
22
0.74(s)0,64(U)1,4/(M)1,0
0.14(h)1,2/(M)1,6
1.0(U)1-3/(M)1'4
0.896
0.716
0.858
See text for data set and variable identification.
In this way, n quasi-Independent estimates are obtained from a data base of
n tests, and the validity of using stepwise regression to obtain a model is
evaluated. Summary information on this process is provided below:
Data set/equation
A (13 tests)
8 (9 tests)
C (22 tests)
Ratio of quasi-independent
Estimate to observed emission factor
Range
0.35-2.54
0.12-12.6
0.15-4.38
Geometric
mean
0.96
0.91
1.01
Geometric
std. dev.
1.79
4.55
2.48
The most important results from the cross-validation analysis pertain
to the model for continuous drop operations (Data Set B). In addition to
its poor performance 1n generating quasi-independent estimates for the
missing data points, the equations obtained from the reduced data base were
largely "unstable." From the nine reduced data sets, equations involving
six different sets of correction parameters were obtained. For example,
90
-------
silt content entered six of the nine equations with exponents ranging from
-0.953 to 1.07; height, on the other hand, entered only three equations.
Because the elimination of only one data point results in such drastic
changes in the predictive model, it can be concluded that the equation for
Data Set 8 in Table 26 is of little merit.
The cross-validation results for Data Sets A and C are much more favorf
able. For Set A, all three variables 1n Table 26 (s, U, and M) entered
into each equation obtained from the 13 reduced data bases; in -addition,
all exponents obtained showed relatively little variation. For Set C,
the two variables in Table 26 (II and M) entered each equation generated
from the reduced data sets; additionally, silt entered once and drop height
seven times during the 22 trials.
The predictive equations for Data Sets A and C may be said to be fairl^
stable and, as shown earlier, are of relatively high accuracy in providing
quasi-independent estimates of missing data. Although the equation for
Set A exhibits less variation in terms of predictive accuracy, the equation
for C encompasses far greater source and material property variation (e.g.,
both batch and continuous drop operations, wider range of moisture contents
etc.). Furthermore, a Kruskal-Wallis analysis of variance (McGhee, 1985)
indicated no significant difference in residuals as a function of the three
data subsets (i.e., current AP-42 batch and continuous drop bases and Test
Report 5).
Consequently, it is recommended that the equation developed for Data
Set C replace the two materials handling equations currently contained in
Section 11.2.3. Thus, a single equation is recommended in the form:
91
-------
(a)1'3
E = k(0.0016) ' ^ 4 (kg/Mg)
U>1.3
(8)
E = k(0.0032) (lb/ton)
2
where:
E = emission factor
k = particle size multiplier (dimensionless)
U = mean wind speed, m/s (mph)
M = material moisture content (%)
Note that this equation is identical to that given earlier in Table 26.
The particle size multiplier k varies with aerodynamic particle diameter
as shown below:
Aerodynamic particle size multiplier (k)
< 30 ufii <15 urn < 1U urn < S urn < Z.S pm
0.74 0.48 0.35 0.20 0.11
These size fractions represent averages of the data given earlier weighted
by the number of tests in Data Set C. Based on the criteria presented in
Table 4, the above equation would be rated A.
Note that Eq. (7), unlike the current expressions in Section 11.2.3,
does not include silt as a correction parameter. There are a number of
reasons why this parameter is excluded. In Data Set C, there is a negative
(but insignificant) correlation between silt and emission factor. Although
it is reasonable to expect an increase in emissions as silt increases, this
92
-------
was not found 1n the data set. It is presumed that this can be attributed
to high correlation between silt and wind speed, and between wind speed and|
emission factor (with both correlations significant at the 5% level). The
second correlation is expected; the first Intercorrelation, however, is not|
supported by any physical reason, but rather is largely due to the fact that
most tests with high silt contents in the data set were conducted under
lower wind speeds. This confounding of test conditions is the cause of thei
probably spurious relationship between silt and emissions.
In addition, it is possible that there Is an important relationship bef
tween the silt and moisture contents for aggregate materials. The tests
supporting the current AP-42 batch drop equation yield a negative correla-
tion between these two parameters which 1s significant at the 10£ level.
However, as other tests (with different materials under consideration) are
included, silt and moisture exhibit insignificant, positive intercorrela-
tions. At present, it cannot be said with any degree of certainty whether
silt and moisture, for a given material or aggregates in general, are inter*
related, and additional experimentation is needed. With the data currently
available, however, the high degree of intercorrelation between silt and
wind speed precludes silt content as a correction parameter.
As a final remark in this regard, it should be noted that the relation*
ships expressed in Eq. (7) are generally comparable to those in the current
AP-42 batch and continuous drop equations. The recommended equation is
based on a reexamination of the relationship between the emission factor
and Independent source parameters using stepwise linear regression. As
noted in Cowherd et al. (1983), the current AP-42 predictive equations for
batch and continuous drop operations were developed by fitting available
test data to a functional relationship. Because only relatively few data
were available, relationships for these particular equations were not de-
veloped by regression analysis (as were the equations for paved and unpaved
roads). The form of the relationships underlying the current equations was
based on analogy with those for other fugitive dust sources. In addition,
the batch drop equation was further modified in Cowherd et al. (1983) prior
to inclusion in AP-42.
93
-------
An additional examination of the recommended equation employed the
data contained in Test Report 4. Of the 62 emissions tests conducted,
45 were associated with moisture and wind speed data. The equation recom-
mended above was used to estimate the reported emission factor, and results
are presented below.
Ratio of predicted to
reported emission factors
Geometric
Size range9 Geometric mean standard deviation
Suspended 1.51 2.53
particulate
S 30 pmA 1.45
£ 10 ymA 2.40
Particle size data shown on page 72 of Test Report 4. Because
only suspended values are reported for each test, only the
mean ratio can be compared for smaller size ranges. Estimate
for suspended particulate assumes a size multiplier k equal
to 1.
Several items about this comparison should be noted. First, only one
significant figure was used to report emission factors in about 70% of the
45 tests. If the additional digits were truncated rather than rounded
(which may be more likely if the data were generated by a computer), the
ratio of predicted to reported emissions would be systematically biased
toward higher values. In addition, the source evaluated in Test Report 4
was enclosed while Eq. (7) is based only on open tests, and the dumping
capacity of the rail car is roughly one order of magnitude greater than the
largest value for the batch drop tests supporting Eq. (7). Finally, as
noted earlier, that test report discussed several problems encountered
with particle sizing, and test data for smaller size ranges are B rated.
Despite these problems, the above comparison indicates that the estimated
emission factors agree fairly well with the reported values. This is espe-
cially true for the suspended particulate and 30 |jmA size fractions. The
94
-------
agreement for the PM10 is not as good, although it is unknown what effect
the problems encountered with particle sizing in the test report have on
the ratio.
6.3 INDUSTRIAL PAVEO ROADS (SECTION 11.2.6)
The uncontrolled tests of particulate emissions from paved roads pre-
sented in Test Report 1 were included In the data base used in developing
the current AP-42 emission factor equations for IP, PM10, and FP. Conse-
quently, results from that test report cannot provide information on the
relative accuracy in independent applications of the current predictive
equations.
Test Report 5, however, does present independent applications of the
TSP predictive equation. Three tests of emissions from traffic on paved
surfaces were conducted. The first two tests were performed with heavily
loaded surfaces and the third was conducted after the surface had been
cleaned. None of the loading values fall into the range of (149 to
7,100 lb/mile) of the tests supporting the current TSP equation. Compari-
sons of predicted to observed emissions (taken from page 31 of the report)
are summarized below:
Loading Ratio of predicted
Run (1b/mi1e) to actual emission factors
AK-7 48,000 8.70
AK-8 23,000 1.44
AK-9 90 1.23
The test report noted that the agreement between predicted and observed
emissions was generally good and became better as source conditions ap-
proached the conditions of the tests used to develop the predictive equa-
tion. The report also suggested that heavily loaded paved surfaces may be
better considered as unpaved in terms of emission estimates. Note that, hi
95
-------
the unpaved road equation been used to estimate runs AK-7 and AK-8, measured
emission levels would be underestimated by a mean factor of 2.2, while the
paved road equation overpredlcts by a mean factor of 3.5. The difficulty
of estimating emissions from heavily loaded paved surfaces was also
addressed in Test Report 10c.
In order to assess the current single-valued PM10 emission factor for
light duty traffic on heavily loaded roads, emission factors for this size
range were generated using the data and calculation scheme presented
in Test Report 5. The results of the comparison are presented below:
PMki emission factor (lb/VMT)
Run Predicted Actual8
AK-7 0.33 0.17
AK-8 0.33 0.67
a
Values determined using size data presented in Test Report 5
with calculation scheme described.
Although only two runs are available for comparison, agreement is generally
acceptable. Finally, note that Run AK-9 could not be used to assess the
current PM10 equation in AP-42 Section 11.2.6 because: (a) the vehicle
weight in this test was far below the range used in developing the equation;
and (b) adjustment of the loading value 90 lb/mile to a mass per unit area
would require information not contained in the test report. Recall, how-
ever, that the earlier comparison between predicted and observed SP emis-
sions showed good agreement for this test.
In general, although there are few independent data available to assess
the accuracy of the current AP-42 paved road estimates, the comparison pre-
sented in this section would indicate good agreement in cases where source
conditions are comparable to those in the underlying data bases. Additional
Predicted
-r actual
1.94
0.49
96
-------
data (Test Report 8) also showed good agreement between predicted and ob-
served emissions. This report is discussed later in this section in con-
nection with control methods used for Industrial paved roads.
As is the case for unpaved roads, many industries have recently insti-
tuted plant-wide control programs for paved roads. However, various paved
road control techniques have not been studied in as much detail as those foJ
unpaved roads. This limitation is not as restrictive as it may first appeay
because available paved road control measures reduce (silt) loading on the
travel surface. Thus, controlled emission factors can be estimated by sub-
stituting these reduced loading values in the current AP-42 predictive equa-{
tions.
The results of controlled paved road tests in Test Report 1 were esti-
mated using the current paved road equations; the results of this comparisoij
are summarized below:
Ratio of predicted to
<. observed emission factors
Sample r
Control method size PM1rt FP
Vacuum sweeping0 4 1.31/1.92 1.02/1.97 1.24/2.07
Water flushing 4 1.67/1.84 1.38/2.00 2.05/3.22
Flushing and broom 3 5.61/1.95 4.34/1.83 9.54/4.25
sweeping
a
b
c
First entry is geometric mean ratio, second is standard geometric
deviation.
Observed PM10 obtained using log-normal interpolation between re-
ported IP and FP emission factors.
All silt loading values for vacuum sweeping are well below values in
data base supporting AP-42 equation.
Agreement between estimated and observed emissions is quite good for both
vacuum sweeping and water flushing. For these controls, the limited data
97
-------
available would Indicate that controlled emission estimates can be obtained
using the current AP-42 (uncontrolled) predictive equations. For flushing
and broom sweeping, however, the industrial paved road equation tends to
substantially overpredict observed emission levels.
A similar comparison for suspended particulate emissions from paved
roads was presented 1n Test Report 8. Measured emissions from a paved road
both before and after vacuum sweeping were compared to estimates obtained
from an earlier version of the current AP-42 equation for TSP. This
version, originally presented in Supplement 14 (May 1983), differs from the
current version only 1n that the constant term was changed from 0.090 to
0.077 lb/VMT. Emission rates presented in Test Report 8 are summarized
below:
Emission rate (lb/mile/hr)
Estimated
Measured3 Previous equation Current equation*3
Before vacuum sweeping 8.43 8.46 7.24
After vacuum sweeping 4.54 4.44 3.80
a Measured emissions reflect particles smaller than 30 pm as determined
by CCSEM. Before and after measurements based on average from 6 and
4 tests, respectively. Test Report 8 does not present results for
individual tests.
b Obtained by multiplying results in Test Report 8 by (0.077/0.090). See
discussion 1n text.
Although only average emission levels can be compared, the limited data
available support the contention that controlled TSP emissions may be esti-
mated within acceptable limits by the current AP-42 uncontrolled emission
factor equation.
It is recommended that Section 11.2.6.4 be revised to reflect the find-
ings that, based on the available data, emissions from certain controlled
98
-------
industrial paved roads can be estimated using the equations currently pre-
sented in AP-42. Although the underlying data base is limited, adequate
estimates were obtained for vacuum sweeping, over all particle size ranges
of interest in Section 11.2.6, and for water flushing, over the IP, PMl0,
and FP size fractions. The results for vacuum sweeping are particularly
noteworthy because all controlled (silt) loading values used in the com-
parisons presented earlier were considerably lower than the values support-
ing the predictive equations for uncontrolled emissions.
6.4 SUMMARY
An important shift in emphasis has been noted in fugitive dust studies
over the past 5 or 6 years compared to those performed in the 1970s. The
previous decade witnessed numerous field studies yielding results that
formed the basis of most open dust emission factors presented in AP-42.
When these emission factors were used to inventory particulate emissions in
various industries, it became apparent that open dust sources in general
(and traffic on paved and unpaved roads in particular) often account for a
considerable portion of the total release of particulate emissions at a
facility. Recognition of the importance of major open dust sources in turn
led to interest in the control of these sources, and field studies were
undertaken to quantify the effectiveness of various control techniques.
The revisions to AP-42 Section 11.2 that are recommended in this report
mirror the developments discussed above. Although there were relatively few
new, independent data for uncontrolled paved and unpaved road emissions,
the data available indicated that the current AP-42 methods yield estimates
with acceptable accuracy. The majority of new data for these types of
sources pertain to control performance evaluation. Consequently, the revi-
sions recommended for AP-42 Sections 11.2.1 and 11.2.6 address the new area
of interest.
The only exception is the recommended revision to AP-42 Section 11.2.3
for aggregate storage piles. Here, new data were available that extended
the range of silt and moisture contents and source operation parameters.
99
-------
These additional data allowed a reexamination of the current AP-42 estima-
tion procedures, and the revised equation presented in Section 6.2 is the
result of this reexamination.
100
-------
7.0 REFERENCES
Brookman, E. T. 1983. Critical Review of Open Source Particulate Emission
Measurements - Current Procedures. TRC Environmental Consultants.
Cahlll, T. A., et al. 1979. Ambient Aerosol Sampling with Stacked Filter
Units. FHWA-RD-78-178, Federal Highway Administration, Washington,
D.C.
Cowherd, C., et al. 1983. Fugitive Dust Emission Factor Update for AP-42.
EPA Contract No. 68-02-3177, Assignment 25, U.S. Environmental Protec-
tion Agency, Research Triangle Park, North Carolina.
Cowherd, C., and J. S. Kinsey. 1986. Identification, Assessment, and Con-
trol of Fugitive Particulate Emissions. EPA-600/8-86-023, U.S. Envi-
ronmental Protection Agency, Research Triangle Park, North Carolina.
McGhee, J. W. 1985. Introductory Statistics. West Publishing, St. Paul,
Minnesota.
Mostfeller, F. , and J. W. Tukey. 1977. Data Analysis and Regression.
Addison-Wesley, Reading, Massachusetts.
Muleskl, G. E. 1986a. Estimation of Unpaved Road Emission Reductions at
Iron and Steel Plants. Draft Supplemental Report. Midwest Research
Institute.
101
-------
Muleski, G. E. 1986b. Update of Fugitive Emission Factors in AP-42. EPA
Contract No. 68-02-3891, Assignment 11, Final Report prepared for
the U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina.
Nie, N. H., et al. 1975. Statistical Package for the Social Sciences,
Second Edition. McGraw-Hill, New York, New York.
Turner, D. B. 1970. Workbook of Atmospheric Dispersion Estimates.
AP-26, U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina.
102
------- |