El P A U.S. Environmental Protection Agency Industrial Environmental Research
"• •» Off ice of Research and Development Laboratory
Research Triangle Park, North Carolina 27711
EPA-600/7-77-148
December 1977
SECOND SYMPOSIUM
ON FUGITIVE EMISSIONS:
MEASUREMENT AND CONTROL
(May 1977, Houston, Texas)
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into seven series. These seven broad categories
were established to facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously planned to foster
technology transfer and a maximum interface in related fields. The seven series
are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the effort
funded under the 17-agency Federal Energy/Environment Research and Development
Program. These studies relate to EPA's mission to protect the public health and welfare
from adverse effects of pollutants associated with energy systems. The goal of the
Program is to assure the rapid development of domestic energy supplies in an environ-
mentally-compatible manner by providing the necessary environmental data and
control technology. Investigations include analyses of the transport of energy-related
pollutants and their health and ecological effects; assessments of, and development
of, control technologies for energy systems; and integrated assessments of a wide
range of energy-related environmental issues.
REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect the
views and policies of the Government, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/7-77-148
December 1977
SECOND SYMPOSIUM
ON FUGITIVE EMISSIONS:
MEASUREMENT AND CONTROL
(May 1977, Houston, Texas)
J. King, Compiler
TRC—The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
Contract No. 68-02-2133
Program Element No. INE624
EPA Project Officer: Robert M. Statnick
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
FOREWORD
The technical papers included in this volume were prepared for presentation
at the Second Symposium on "Fugitive Emissions: Measurement and Control"
on May 23-25, 1977 at Houston/Texas.
The Symposium was sponsored jointly by the Environmental Protection Agency's
Industrial Environmental Research Laboratory at Research Triangle Park,
North Carolina and the Air Pollution Control Association's TS-5.2 Committee
on Fugitive Emissions. The objectives of the Symposium were to disseminate
information relative to the impact of fugitive emissions from industrial
processes on our atmosphere and ground waters, and to bring together con-
cerned representatives of industrial, research-oriented and governmental
organizations to consider strategies for fugitive emission control.
Dr. Robert M. Statnick of the Industrial Environmental Research Laboratory,
Environmental Protection Agency, Research Triangle Park, North Carolina,
was the Project Officer and General Chairman of the Symposium.
Joanne King of TRC - The Research Corporation of New England, Wethersfield,
Connecticut, was the Symposium Coordinator and Compiler of the Proceedings.
-------
TABLE OF CONTENTS
(* indicates speaker)
May 23. 1977 Paqe
SESSION I: INTRODUCTION *_
OVERVIEW OF FUGITIVE EMISSIONS 3
Gary McCutchen, EPA
THE ROLE OF FUGITIVE EMISSION FACTORS IN MEETING NATIONAL
AMBIENT AIR QUALITY STANDARDS 25
James H. Southerland* and Charles C. Masser, EPA
SESSION II: MEASUREMENT
MEASUREMENT OF FUGITIVE PARTICULATE 47
Chatten Cowherd, Jr., Ph.D., Midwest Research Institute
FUGITIVE MONITORING AT A COAL CLEANING PLANT SITE 63
D. Ambrose, D. Brown* and R. Clark,
Battelle Columbus Laboratories
PARTICULATE SAMPLING TECHNIQUES FOR A COKE QUENCH TOWER 115
A. H. Laube, John Jeffery*, York Research Corporation,
and Carl Edlund, EPA
MONITORING INDUSTRIAL FUGITIVE EMISSIONS - AN OCCUPATIONAL
HEALTH PERSPECTIVE 141
James L. Oser*, Ronald J. Young, John M. Dement, and
Howard R. Ludwig, NIOSH
May 24, 1977
SESSION III: CONTROL
MODELING FUGITIVE PARTICULATES FOR CONTROL STRATEGY DEVELOPMENT -
A CASE HISTORY 167
George A. Jutze and John M. Zoller*, PEDCo
Environmentalt Inc.
ASBESTOS WASTE EMISSION CONTROL 187
Mary K. Stinson*, EPA, Colin F. Harwood, Pall Corporation,
and Paul Ase, IIT Research Institute
-------
TABLE OF CONTENTS (continued)
Page
FUGITIVE EMISSIONS FROM CHEMICAL FERTILIZER MINING 205
J. C. Ochsner and T. R. Blackwood*, Monsanto Research
Corporation
FUGITIVE EMISSION CONTROL IN THE IRON & STEEL INDUSTRY 225
Arthur G. Nicola, Pennsylvania Engineering Corporation
SESSION IV: WATER
STORMWATER: ONE STEEL MILL'S EVALUATION 255
John W. Luton* and William B. Chadick, Armoo Steel
Corporation
CHARACTERIZATION OF COAL PILE DRAINAGE 269
Doye B. Cox and R. J. Ruane*, Tennessee Valley Authority
MEASUREMENT OF NON-POINT SOURCES FROM A COAL-FIRED UTILITY
AND THE IMPACT ON RECEIVING WATERS 299
Gordon T. Brookman* and Willard A. Wade III, TRC - The
Research Corporation of New England
THE DEVELOPMENT OF A MATHEMATICAL MODEL TO SIMULATE INDUSTRIAL
NON-POINT SOURCE POLLUTION 337
James J. Binder* and Gordon T. Brookman, TRC - The
Research Corporation of New England
May 25. 1977
SESSION V: INDUSTRY
CURRENT STATUS OF PROCESS FUGITIVE PARTICULATE EMISSION ESTIMATING
TECHNIQUES 381
John M. Zoller, Thomas A. Janszen, PEDCo Environmental..
Inc., and Gilbert H. Wood, EPA
AIR EMISSIONS IN IRON ORE MINING AND ENRICHMENT 457
G. V. Jorgenson*, J. P. Pilney, and E. E. Erickson,
Midwest Research Institute
DESIGN OF A STUDY TO MEASURE FUGITIVE EMISSIONS FROM PETROLEUM
REFINING 479
Donald D. Rosebrook, Radian Corporation
SOME AIR QUALITY AND ENERGY CONSERVATION CONSIDERATIONS FOR THE USE
OF EMULSIONS TO REPLACE ASPHALT CUTBACKS IN CERTAIN PAVING
OPERATIONS 509
Francis M. Kirwan*, EPA, and Clarence Maday, Consultant
iv
-------
Monday Morn-ing - May 233 1977
SESSION I: INTRODUCTION
Chairman: Robert M. Statnick, Ph.D., EPA
Page
OVERVIEW OF FUGITIVE EMISSIONS 3
Gary McCutchen, EPA
THE ROLE OF FUGITIVE EMISSION FACTORS IN MEETING NATIONAL AMBIENT 25
AIR QUALITY STANDARDS
James H. Southerland* and Charles C. Masser, EPA
* indicates speaker
-1-
-------
OVERVIEW OF FUGITIVE EMISSIONS
By
Gary McCutchen
Emission Standards and Engineering Division
Environmental Protection Agency
Research Triangle Park, North Carolina
-3-
-------
ABSTRACT
OVERVIEW OF FUGITIVE EMISSIONS
New information on particulate and hydrocarbon fugitive emission
quantities has confirmed their significance. Near sources, fugitive
emission impact on air quality predominates, especially if stack
emissions are reasonably well-controlled. The effect of including
fugitive emissions in SIP revisions and New Source Review and emission
offset determinations is expected to be great, but the implications
of this policy have not been fully realized. Opacity measurements have
the potential to substitute for difficult mass emission tests.
-4-
-------
OVERVIEW OF FUGITIVE EMISSIONS
Gary McCutchen
At last year's first fugitive emissions symposium, several speakers
indicated that the quantification and regulation of these emissions would
likely require a major control agency effort during the next five years.
Looking back over developments in the past year, I think you will agree
that these predictions, if anything, understated the attention fugitive
sources are receiving.
Two developments have spurred control agency efforts: (a) non-
attainment of National Ambient Air Quality Standards (NAAQS) in many
Air Quality Control Regions (AQCR) and (b) an accumulation of additional
information indicating that fugitive emissions and dust are significant
in terms of both mass and ambient air impact. These and related develop-
ments are the topic of this paper.
Definitions^1*
The term fugitive emissions as used in this paper includes either
gaseous or particulate emissions from industry-related operations that
escape to the atmosphere without passing through a primary exhaust system.
This includes emissions from loading, unloading, and transport of materials;
storage piles; and roads and open areas within the plant boundaries.
As distinguished from fugitive emissions, fugitive dust includes natural
dusts (e.g., dust storms), and agricultural and other non-industry activi-
ties (e.g., unpaved roads, urban street dust, commercial construction sites),
Fugitive sources refers to both fugitive dust and fugitive emission sources.
-5-
-------
Non-Attainment
Of the 313 AQCRs and State portions of interstate AQCRs, 245 have not
attained NAAQS for one or more criteria pollutants.^2' This has resulted
in 246 pollutant-specific calls for attainment/maintenance revisions to
SIPs and 25 pollutant-specific calls for maintenance revisions only (Table
1). The majority of the plan revisions are for oxidants and total suspended
particulate (TSP). The large number of oxidant revisions is partly a result
of calls for statewide revisions in Regions II, III, and V. Transport
problems necessitate hydrocarbon controls over wide geographic areas to
attain the oxidant standard.
The second largest number of SIP calls for revision were made for TSP.
Additional point source control and fugitive dust control were the measures
most frequently recommended. Fugitive dust control is a new program for
most control agencies, including EPA, but a comprehensive EPA policy paper
(31
is now undergoing final internal review/ ' Fugitive emissions, especially
particulate matter from point sources, are being closely studied in areas
where more stringent control measures are required. Nearly 70 percent of
all Regional Office contract assistance requests concern particulate
sources/ ' and these studies concentrate on fugitive sources. As an
example, GCA inspected a ferrochrome plant in Charleston, South Carolina,
/5\
for Region IV. ; Their draft final report includes estimates that particu-
late emissions total 92 pounds per hour, of which only eight pounds per
hour is emitted from control device stacks. The remaining 84 pounds per
hour are fugitive emissions.
Fugitive Emission Measurement
Selected fugitive emission estimates are summarized in Table 2. In
many cases they constitute a large percentage of total emissions from a
-6-
-------
well-controlled plant. Fugitive emissions represent 75 percent of total
vinyl chloride emissions from a polyvinyl chloride (PVC) plant meeting
NESHAPS (which require an estimated 90 percent control of fugitive and 99
percent control of stack emissions).^ The surprising thing is that fugi-
tive emissions can also be significant at uncontrolled plants, e.g., 40
percent of total vinyl chloride emissions from uncontrolled PVC manufacture.
Emissions from some fugitive sources are astonishingly large. Fugitive
hydrocarbon emissions from cutback asphalts are equivalent to 2.3 percent
of total national hydrocarbon emissions: 742,000 tons in 1975.^''
Air Quality Impact
Previous air quality studies have indicated that fugitive emission
ambient air concentrations around a plant are greatest at the fenceline
and decrease rapidly with distance from the source. High fenceline concen-
trations are expected because fugitive emissions usually occur (a) near the
ground with (b) little or no thermal lift. They thus reach ground level
with much less plume dilution than do stack emissions. Since fugitive
emissions don't go through a control device, faster-settling coarse particles
are expected to comprise a larger portion of the fugitive emission mass.
The result would be a rapid decrease in fugitive emission ambient air con-
centrations downwind of the source as the large particles settle out. Now,
as a result of efforts to model ambient concentrations near stationary
sources of lead, there are strong indications that fugitive emission impacts
can extent several kilometers beyond plant fencelines. Even more unexpected
is the possibility that fugitive emissions from tall structures may not
reach ground level until well beyond the fenceline. Ambient air monitors
located at property lines wouldn't measure the impact such emissions would
have further downwind.
-7-
-------
The original NAAQS strategies established in the early 1970's were
concerned almost exclusively with stack emissions. There wasn't even
enough fugitive emission information available to assess the importance
of such sources, although Federal guidelines (40 CFR 51, Appendix B) did
recommend a "reasonable precautions" approach, including general techniques
for minimizing both fugitive emissions and fugitive dust. The lead NAAQS
strategies being developed, in contrast, reflect an increased awareness of
fugitive emissions as an ambient air problem.
As a part of the lead NAAQS development, EPA is attempting to deter-
mine lead ambient air concentrations in the vicinity of lead sources using
diffusion modeling. The model used to estimate air quality around lead
sources is the only wel1-documented one able to accommodate deposition
phenomena. This is the Atmospheric Transport and Diffusion Model (ATM)
developed at Oak Ridge National Laboratory. It is designed to consider both
deposition and gravitational settling, so requires particle size data in
addition to the usual modeling parameters. For the first time, fugitive
as well as stack emissions from point sources can be modeled.
Using ATM, EPA identified three stationary sources of lead around
which 90-day average ambient air lead concentrations could reach in excess
of five micrograms per cubic meter (ug/m3) despite good stack air pollution
control. These sources are (a) primary lead smelters, (b) primary copper
smelters, and (c) tetraethyl lead (TEL) plants/ ' For the smelters,
preliminary results show that fugitive emissions contribute more to ambient
air concentrations than do stack emissions, as Figure 1 shows. The solid
lines in Figure 1 represent total emissions impact in terms of "worst case"
90-day average ambient air concentrations. Stack emissions impact is shown
by a dashed line for the lead smelter; for the copper smelter, the 90-day
-8-
-------
average stack contribution is essentially zero. Shaded areas represent
the fugitive emissions impact.
The preliminary parameters used to model the lead and copper smelter
ambient air impacts are summarized in Tables 3 and 4, respectively. In
terms of size, the copper smelter produces 365 tons of copper per day, while
the lead smelter produces 185 tons of lead per day. Lead emissions from the
copper smelter total 3.6 grams per second, half of it fugitive. The lead
smelter emits 2.73 grams lead per second, only 15 percent of it fugitive.
Stack emissions for both smelters represent good control efficiency, while
fugitive emissions are assumed to be uncontrolled.
The striking differences between the total concentration curves in
Figure 1 are a result of emission rates and assumed release height. At the
lead smelter, fugitive emissions are released near the ground. Ambient air
Impact near the plant is large, but decreases rapidly with distance.
Emissions from the reverberatory furnace building keep concentrations fairly
high from about one kilometer (km) to nearly 3 km. This exponential decay
curve is typical of fugitive emissions released near ground level.
Less typical is the concentration curve for the copper smelter.
Fugitive emissions are, in this case, released 30 meters above ground; their
full impact occurs nearly 2 km downwind of the plant. The effect is charac-
teristic of a short stack. Note that the lead ambient concentration is
above 15 ug/m3 even 5 km from the plant, a result of fugitive lead emission
rates four times that of the lead smelter.
These results are, of course, preliminary. Not all fugitive emission
sources at these plants are included and some of the modeling parameters
may change, but it is clean that EPA needs to take fugitive sources into
account.
-9-
-------
A less complex program for modeling short-term (24-hour average)
ambient concentrations resulting from stack and fugitive particulate
emissions will soon be available from EPA. It is a revision of an existing
point source model, and is being prepared by Phil Youngblood, MDAD, EPA,
in conjunction with a document on industrial process fugitive emissions
being prepared by PEDCo Environmental for Gil Wood, ESED, EPA.
The effect of fugitive emissions from numerous sources on urban air
quality is even more difficult to determine that the effect of a single
source. In one analysis, the long-term contribution of fugitive dust and
emissions to hi-vol sampler readings was determined using data from seven
U.S. urban areas. ' Near industries, fugitive sources added 15 micrograms
per cubic meter (ug/m3) to the central city total suspected particulate
(TSP) concentrations and 23 yg/m3 in suburban areas. This is probably a
combination of industry-related fugitive emissions, reentrained street dust,
and motor vehicle emissions. In commercial areas, fugitive sources added
12 and 20 yg/m3 at urban and suburban sites, respectively. These values
are only slightly lower than those for industrial areas, but the difference
could represent the direct impact of industrial fugitive emissions.
NSPS and NESHAPS
Realization of the difficulties in attainment of NAAQS caused by
fugitive emissions has also strengthened concern that new facilities could
add to the problem. Some new source performance standards (NSPS) require
fugitive emission control (Table 5), but there is now a much greater
emphasis on the problem. This shift in program emphasis is most apparent
in EPA's oxidant-related hydrocarbon studies. Within the past year,
Emission Standards and Engineering Division (ESED) program resources shifted
from nearly no hydrocarbon work to a commitment of over 50 percent of ESED
-10-
-------
resources to this task. Many sources, from oil field through refining and
organic chemical production to end use, are under study. Present plans are
to first issue guideline documents for States to use in revising SIP's,
then develop NSPS. Some of these studies are already providing data.
Emissions from miscellaneous refinery sources (valves, pumps, etc.) make
up significant fractions of total refinery hydrocarbon emissions. The
first five guidance documents, on different coating operations (auto assem-
bly and coil, can, fabric, and paper coating), are being issued for use by
States with oxidant problems.
These studies also illustrate the difficult task emission testing often
poses. Efforts to quantify emissions from floating roof storage vessels,
in four separate studies, have cost $700,000;' ' another $500,000 to
$750,000 commitment is expected over the next two years. Nearly all of this
is industry, rather than EPA, funded. The cost of testing is one reason
why so little is known about fugitive emission quantities, and why control
agencies (and, usually, industry) prefer specifying equipment and/or
operating procedures to field testing.
Controlling especially dangerous pollutants, national emission stan-
dards for hazardous air pollutants (NESHAPS) have historically reflected
greater concern for fugitive emissions than have NSPS. Asbestos regulations,
which already require control measures for storage piles, are being expanded
or studied for decorative sprays, quarries, and paving roads with asbestos-
containing gravel (asbestos tailings for roadway use are already prohibited).
The vinyl chloride NESHAPS includes requirements for 90 percent control of
nine different fugitive emission sources at manufacturing plants.' '
Finally, the need for and feasibility of a NESHAPS for benzene emissions
from coke ovens and other sources is being studied.
-11-
-------
New Source Review
Many stationary source New Source Review (NSR) policies, including
prevention of significant deterioration and emission offsets, are so new
that their full implications have not yet been realized, but every inter-
(3)
pretation to date considers fugitive emissions a part of total emissions.
The issues resulting from this decision are far too complex to be discussed
in this overview, so only two brief examples will be given:
Major sources - only "major" new sources, in terms of tons per year
of allowable emissions, are affected by most NSR procedures. Owners and
operators should be aware that this includes allowable fugitive, as well
as stack, emissions.
Fugitive source offsets - The quantity of new emissions that must be
offset in non-attainment areas includes fugitive emissions; credit, of
course, is also given for reducing fugitive sources.
Fugitive Emission Studies
Probably the most studied of any source category, iron and steel
mills contain numerous fugitive pollution sources. There are over 20 EPA-
funded studies of the iron and steel industry underway, and over half
concern fugitive air pollution source evaluation. Attention has been
focused on these sources because 80 percent of iron and steel mill air
pollution problems are fugitive emission related. ' Coke oven coal and
coke handling, door and topside leaks, and pushing and quenching emissions
can all be classified as fugitive, as can storage pile, roof monitor, and
transportation-related emissions. The studies have resulted in improved
emission factors for many difficult-to-quantify steel mill sources. Quench
towers and coke oven emissions have been source sampled. Factors for paved
-12-
-------
and unpaved roads, storage piles, and material handling have been developed
for heavy-duty vehicles and specific materials. Some of these estimates
are given in Table 2; the others appear in a special fugitive process
emissions study.
The primary purpose of the special study is to produce a guideline
document for the States to use in estimating fugitive emission rates and
ambient air impacts for SIP revisions, NSR review, and other air management
programs.' ' The document, which is based on a review of the literature,
comprises approximately 400 pages of information on fugitive particulate
sources and should be distributed this month. Factors from the document
are already being used to evaluate fugitive source emission rates and
emission reduction capabilities.
Most of the document consists of information on various source cate-
gories, including (a) a description of each process, (b) fugitive emission
rates and characteristics, (c) example plant emission inventories, and
(d) control techniques. Coverage of iron and steel, lime, primary lead,
and materials extraction (hard and soft rock and earthy materials) source
categories is fairly good. Existing information on other sources is much
less satisfactory.
In addition to emission rates, the document discusses capture and
control options and costs. Impact on air quality is examined, and the
24-hour model described earlier will be tailored for use with the emission
information provided. Finally, strategies for integrating this information
into the SIPs is outlined.
Regulatory Aspects
At the last count, 17 States lacked regulations pertaining to fugitive
emissions.^ ' The other States have one or more of three general types of
-13-
-------
regulations: (a) non-specific nuisance, (b) quantitative property line,
and (c) specific control measures. ' The most common is the nuisance, or
"reasonable precautions" type. It is usually more effective in dealing
with diverse sources of fugitive dust, but some States, like Vermont and
West Virginia, have modified such regulations to effectively apply to
control of fugitive emissions.
Property line regulations take many forms. Texas, for example, has
successfully applied an upwind-downwind method, limiting the difference
in measured ambient concentrations. Illinois and Missouri, in addition,
consider particulate size.
Colorado is considered the best example of a State which requires
implementation of specific control techniques for specific fugitive dust
and emission sources, such as unpaved roads, explosives detonation, open
mining, and unenclosed operations. Many other States specify control
techniques for specific processes, such as Alabama's coke oven regulation:
There are as many variations as there are States, but the primary
regulatory and enforcement tool continues to be opacity. The development
of detailed coke oven opacity reading procedures over the past year has
helped establish a firm methodology for non-stack opacity observations. '
Similar techniques may prove applicable for other difficult-to-read fugitive
sources.
Another development is the slow but steady correlation of optical
density with emission mass for specific sources. Sources for which an
opacity/mass relationship exists include asphalt concrete plants, sludge
incineration, catalytic cracking unit regenerators (petroleum refineries),
secondary brass and bronze and lead smelters, and coal-fired power plants.
-14-
-------
One unsatisfactory attempt has been made to obtain a similar correlation
for oil-fired power plant emissions.
These studies are only for one stack at one plant over a few days'
operation. A much broader data base is needed before general correlations
for each source category can be attempted, but this is a start. If success-
ful , such efforts may eventually lead to optical measurement of fugitive
plumes, thereby avoiding difficult and expensive mass measurement techniques.
Summary
New information on particulate and hydrocarbon fugitive emission
quantities has confirmed their significance. Near sources, fugitive
emission impact on air quality predominates, especially if stack emissions
are reasonably well-controlled. The effect of including fugitive emissions
in SIP revisions and New Source Review and emission offset determinations
is expected to be great, but the implications of this policy have not been
fully realized. Opacity measurements have the potential to substitute for
difficult mass emission tests.
-15-
-------
References
1. Adapted from: Lillis, E.G., and D. Young, "EPA Looks at 'Fugitive
Emissions,1" JAPCA, 25:1015 (1975).
2. EPA-450/2-76-026, State Air Pollution Implementation Plan Progress
Report, January 1 to June 30, 1976, EPA, Research Triangle Park, NC 27711,
pp. b-9 (Oct. 1976).
3. Conversations with Dave Dunbar and Mike Trutna, CPDD, OAQPS, EPA,
March and April 1977.
4. Conversation with Tom Williams, CPDD, OAQPS, EPA, April 8, 1977.
5. Source Inspections in Selected Region IV Nonattainment Areas to Deter-
mine Capabilities of Reducing TSP Emissions. Volume I. (draft) GCA Corpora-
tion, submitted March 18, 1977.
6. Conversation with Susan Wyatt, ESED, OAQPS, EPA, April 11, 1977.
7. Conversation with Frank Kirwan, SASD, OAQPS, EPA, March 23, 1977. For
additional details, refer to the Kirwan and Maday paper presented at this
symposium.
8. Information obtained from Mark Scruggs, MDAD, OAQPS, EPA. Emission and
modeling factors compiled by Mark Scruggs from several sources.
9. EPA 450/3-76-024, National Assessment of the Urban Particulate Problem,
Volume 1, National Assessment, p. 59, July 1976.
10. Conversation with Richard Burr, ESED, OAQPS, EPA, April 14, 1977.
11. Conversation with Carl Edlund, DSSE, OGC, EPA, April 21, 1977.
12. Conversation with Gilbert Wood, ESED, OAQPS, EPA, April 14, 1977.
The document is: Control Program Guideline for Industrial Process Fugitive
Particulate Emissions.
13. Conversation with Lance Granger, Midwest Research Institute, 125 Volker
Avenue, Kansas City, MO, April 1, 1977.
14. Conversation with Gary Young, NEIC, EPA, and John Hepola, Region III,
EPA, February 24, 1977.
15. Conversation with William D. Connor, IERL, EPA, April 27, 1977 See
also: EPA-650/2-74-120, In-Stack Transmissometer Measurement nf P^tjr.^t0
Opacity and Mass Concentration (Nov. 1974) and EPA-650/2-/4-128, Measurement
of the Opacity and Mass Concentration of Particulate Emissions hy^
Transmissometer.
-16-
-------
50
40
OJ
01
to
s-
T3
O
co
^
Oi
1C
O)
u
c
O
O
-o
1C
(U
c
0)
30
20
10
FIGURE 1. Ambient Air Lead Concentrations
Near Pb and Cu Smelters
Using ATM8
Lead Smelter
• Total
O Stack
0 Fugitive
Copper Smelter
• Total
n Stack
II Fugitive
12345
Distance From Center of Plant, km
-17-
-------
TABLE 1
NATIONAL ATTAINMENT/NON-ATTAINMENT STATUS OF AQCRs AND
STATE PORTIONS OF INTERSTATE AQCRs
oo
I
SIP Status
Attainment
Revisions for Attainment/
Maintenance
Revisions for Maintenance
Non-Attainment -
no revisions
TSP
124
78
11
100
313
S02
256
12
11
34
313
CO
233
41
2
37
313
Ox
143
114
0
56
313
NOX
308
1
1
3
313
Total s
1064
246
25
230
-------
VO
I
TABLE 2
FUGITIVE/STACK EMISSION ESTIMATES
Sources
Primary Lead Smelter3
Reverberatory furnace
Blast furnace
Sintering
Fugitive Emissions
0.35 gm Pb/sec
0.02
0.03
Well -Control led
Stack Emissions
0.27 gm Pb/sec
0.38
1.65
Ratio:
Fug. /Stack
1.30
0.05
0.02
Phosphate Fertilizer Plant
WPPA, SPA, DAP, & TSP
Pond and TSP storage
Ferroalloy Production
1975 actual emissions
1975 emissions if SIPs met
Cement Plant0
Iron and Steel Millsf
BOF - open hood
- closed hood
EAF - carbon steel
- alloy steel
Sintering
Open hearth
Scarfing
Windblown9
N/A
0.49 Ib F/ton ?2®5 equiv.
15,400 tons PT/yr
15,400 tons PT/yr
14.64 g/sec
0.5 Ib/ton product
0,5
3
1.5
0.7
0.29
0.011
5 Ib/ton steel produced
0.29 Ib F/ton P2^ equiv. 1 6g
N/A
113,846 tons PT/yr
4,650 tons PT/yr
49.26 g/sec
0.29 Ib/ton product
0.02
1.35
,675
2.6
0.35
0.015
N/A
0.13
3.31
0.30
1.72
25
2.22
2.22
0.27
0.83
0.73
N/A
Page 1 of 2
-------
TABLE 2 (Cont.)
FUGITIVE/STACK EMISSION ESTIMATES
Footnotes:
aData compiled by Mark Scruggs, MDAD, OAQPS, EPA, from various sources.
Total NSPS emission limitation for the sources listed.
CTSP storage (.05 Ib/ton) is NSPS limit for 1500 ton/4 day storage; pond emissions derived from
data presented by W.R. King at last year's symposium. Assumed a 125 acre pond corresponding
to 500 TPD production, using square root of pond area ratio to scale up from 100 acre pond
emissions of 0.39 Ib/ton P205: y^JS/ToTx .39 = .44 Ib/ton P205.
Conversation with Lance Granger, MRI, April 1, 1977.
^ eExample from draft Control Program Guideline for Industrial Process Fugitive Particulate
<7> Emissions, Gilbert Wood, OAQPS, EPA Project Officer.
Preliminary emission factors from various sources, compiled by Reid Iversen, ESED, OAQPS, EPA.
gPreliminar.y average emissions from roads, storage piles, and open areas. Based on a study
conducted by MRI for EPA, Robert Hendricks, ORD,EPA, Project Officer.
Page 2 of 2
-------
TABLE 3
CHARACTERISTICS FOR A HYPOTHETICAL 185 TON/DAY PRIMARY LEAD SMELTER
(8)
Source
Sinter Machine
Stack
Blast Furnace
Stack
Reverberatory
( Furnace Stack
to
7" Fugitive Emissions
From Sinter Bldg.
Lead
Emission
Rate
qm/s
1.65
0.38
0.27
0.03
Control
Efficiency
99
99
94
0
Stack
Height**
m
128
38
38
12
Stack
Diameter
m
2.9
4.1
4.1
NA
Stack
Gas
Velocity
m/s
14.6
2.4
2.4
NA
Stack
Gas
Tempera-
ture °K
389
329
329
290
Area of
Building
m2
NA
NA
NA
1290
Fugitive Emissions
From Blast Fur-
nace Building 0.02
Fugitive Emissions
From Reverbera-
tory Building 0.35
Fugitive Emissions
From Zinc Fuming
Area 0.01
Fugitive Emissions
From Zinc Fur-
nace Building 0.02
0
0
14
17
18
15
NA
NA
NA
NA
NA
NA
NA
NA
290
290
290
290
**The stack height for fugitive emissions represents the average height of release
NA - Not Applicable
1390
770
430
310
-------
TABLE 4
CHARACTERISTICS FOR A HYPOTHETICAL 365 TON/DAY COPPER SMELTING PLANT
(8)
NJ
I
Roaster Stack
Smelter Stack
Converter Stack
Fugitive Emissions
From Roasting
Fugitive Emissions
From Smelting
Fugitive Emissions
From Converters
Lead
Emission
Rate
gm/s
0.62
0.71
0.44
0.13
0.50
1.20
Control
Efficiency
%
97
94
98
0
0
0
Stack
Height**
m
166
166
172
30
30
30
Stack
Diameter
m
6.7
6.7
5.5
NA
NA
NA
Stack
Gas
Velocity
m/s
10.4
10.4
7.6
NA
NA
NA
Stack
Gas
Tempera-
ture °K
480
480
464
300
300
300
Area of
Building
m2
NA
NA
NA
1800
5100
2100
**The stack height for fugitive emissions represents the average height of release.
NA = Not Applicable.
-------
TABLE 5
NSPS and NESHAPS FUGITIVE EMISSION REGULATIONS
Source Category
Type of Standard
Emissions Limited To
Pollutant
i
N>
U>
Portland Cement
Asphalt Concrete Plants
Storage Vessels for Petroleum
Liquids (>40,000 gallons)
Secondary Lead Smelters
(pot furnaces)
Secondary Brass & Bronze
(blast & electric furnaces)
Iron and Steel Plants
Electric Arc Furnace Shop
Primary Aluminum Reduction
Phosphate Fertilizer Plants
(GTSP storage)
Coal Preparation Plants
Ferroalloy Production
(dust-handling system)
Asbestos Sources
Visible emissions
Visible emissions
Equipment
Visible emissions
Visible emissions
Visible emissions
Mass and visible
emissions
Mass emissions
Visible emissions
Visible emissions
Equi pment/operati ng
standards
10% opacity
20% opacity
N/A
10% opacity
10% opacity
Furnace:
10% opacity except:
20% during charging
40% during tapping
Dust handling system:
10% opacity
Mass: 2 Ib/ton product
Opacity: 10% from potroom
20% from bake plant
0.25 g F/hr/metric ton
equiv. P2®5 stored
20% opacity
10% opacity
N/A
Page 1 of 2
Particulate
Particul ate
HC
Particul ate
Particulate
Particulate
Fluorides
Particulate
Particul ate
Fluorides
Particulate
Particulate
Asbestos
-------
TABLE 5 (Cont.)
NSPS and NESHAPS FUGITIVE EMISSION REGULATIONS
Source Category
Type of Standard
Emissions Limited To
Pollutant
Beryllium Processors/Users
Rocket Motor Firing
Other Sources
Chior-Alkali Plants
Ethylene Dichloride, Vinyl
Chloride, and Polyvinyl
Chloride Plants
Ambient concentration
Ambient concentration
Mass emissions
Mass and equipment/
operating standards
75 ug-min/m3 over
specified time period
0.1 yg/m3, 30-day avg.
1300 g Hg/day from
cell room
Generally, 10 ppm or
equivalent
Beryllium
Beryllium
Mercury
Vinyl Chloride
Page 2 of 2
-------
THE ROLE OF FUGITIVE EMISSION FACTORS IN MEETING NATIONAL
AMBIENT AIR QUALITY STANDARDS
By
James H. Souther!and and Charles C. Masser
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Monitoring and Data Analysis Division
Air Management Technology Branch
-25-
-------
Abstract
Studies have shown the local and regional impact of fugitive dust and
emissions on air quality. These studies also explain the basis for conclu-
sions that such sources must be considered in developing action strategies
to meet the National Ambient Air Quality Standards. A perspective illustrat-
ing the impact of fugitive sources of particulate and organics being omitted
in base year State Implementation Plan (SIP) inventories is given. The
status and role of emission factors useful in preparation of revised SIP's
is presented with a brief summary of the major data needs particularly
in the context of availability and applicability for generalized factors
to be utilized in AP-42, Compilation of Air Pollutant Emission Factors.
Plans to fill identified data gaps are presented with a call for assistance
to the technical community in meeting the objective of making state of
the art emission factors available to the State and local agencies who
rely upon them as an integral part of their air resource management
programs.
-26-
-------
THE ROLE OF FUGITIVE EMISSION FACTORS IN MEETING NATIONAL AMBIENT
AIR QUALITY STANDARDS
By
James H. Souther!and and Charles C. Masser
Background
The Clean Air Act , as amended in 1970, provides that EPA develop air
quality criteria, adopt National Ambient Air Quality Standards (NAAQS) for
major air pollutants, define Air Quality Control Regions (AQCR's) (i.e.,
areas of similar air pollution characteristics), and provides for a joint
Federal-State control program to ensure that the air quality in the various
AQCR's throughout the country attains and maintains the NAAQS. The Act
also provides for State Implementation Plans (SIP's) to be prepared by
the States and approved by EPA. These plans present each State's intended
course of action to implement the national standards within the periods
specified in the Act - 3 years after approval for attaining the primary
or health-related standards and a "reasonable time" for attaining the more
stringent (welfare-related) secondary standards.
In April 1971, EPA promulgated NAAQS for five major pollutants for
which air quality criteria has been developed. These five pollutants were
total suspended particulates (TSP), sulfur oxides, nitrogen dioxide, carbon
monoxide, and photochemical oxidants. A non-methane hydrocarbon standard
was promulgated as a guide only for States to use in developing plans to
meet the oxidant standard. Subsequently in early 1972, the States prepared
-27-
-------
and submitted implementation plans that in most cases were designed to
attain and maintain standards by mid-1975.
Progress and Problems
The bulk of the provisions of the SIP's relating to TSP control (which
is of particular interest to us at this point) were approved by EPA on
May 31, 1972. Thus, under the timetable provided by the Act, the date of
May 31, 19752 became the scheduled date for most AQCR's to attain the
primary standard for TSP. A general improvement in ambient air quality
with respect to total suspended particulate is continuing at the rate of
3 to 4 percent pe-" year based upon data from nearly 1800 sites.^ This
improvement is shown graphically in Figure 1. The present rate of progress,
however, may not be sustained since fewer traditional sources remain to
be controlled and since fugitive dust, fugitive process emissions, and
re-entrained urban particulates are more difficult to control. Approximately
30 percent of the nation's population is still living in areas above the
long-term primary annual particulate standard.
On July 1, 1976, calls for State Implementation Plan (SIP) revisions
were made that require States to adopt new regulations for such areas where
problems exist. Failure to attain the national ambient air quality standards
for particulate matter in numerous urban areas as reflected by this call for
SIP revisions has resulted in a re-examination of the urban particulate
problem. Basically, the particulate control strategy developed as part
of the original SIP's included an analysis of the contribution of conven-
tional point and area sources without much consideration of other less
conventional sources of particulate. That these sources were not considered
-28-
-------
I
S3
no
100 -
90-
80-
z
S 70-
£ a.
u 50-
2 .
S qo-
30-
20-
10-
IV
A v
,
r~i
ex. ^^
T V V
t 1
*~ — ~o
i 1
1971 1972 1973 1974 1975
YEAR
I— 90TH PERCENTILE
-75THPERCENTILE
COMPOSITE AVERAGE
mean)
T
-25TH PERCENTILE
-10TH PERCENTILE
Legend
Figure 1. Trends of annual mean total suspended
particulate concentrations from 1971
to 1975 at 1792 sampling sites.3
-------
was due in part to the fact that no emission factors were available. It
now appears that previously "neglected" sources of fugitive dust and emis-
sions are of major importance and must receive additional attention.
Volatile organics ("hydrocarbons") and oxides of nitrogen are
precursors of oxidant. A preliminary analysis of short-term trends
(1973-1975)3 suggests a decline in summer oxidant/ozone violations in the
eastern part of the United States (eight sites decreasing and three sites
increasing) and a general increase in metropolitan Denver. There are too
few sites with sufficient historical data to characterize trends adequately,
but control of fugitive emissions of volatile organics is believed to be a
necessary component of control strategies to reduce oxidant precursor
levels to the extent necessary to meet the national oxidant standard.
Generally fugitive emissions of criteria pollutants other than particulate
and hydrocarbons are not currently believed to be significant, but there
are exceptions such as S02 from smelters. With the promulgation of a lead
standard later this year, fugitive lead emissions no doubt will also receive
additional attention and the demand for emission factors will increase.
Need for Control of Fugitive Emissions
With increased implementation of conventional stack controls on
stationary sources, traditionally estimated particulate emissions have
decreased substantially as seen in Table I3, which for the most part does
not include fugitives. Failure of conventional control efforts to provide
the decreases in ambient levels that were anticipated have lead to a
re-examination of reasons for nonattainment. In so doing, various agencies
have noted the apparent significant air quality impact of fugitive emission
-30-
-------
Table 1. Summary of National Emissions Estimates, 1970-1975*
(106 tons/yr)
Year
i
UJ
Participates
SOX
NOX
HC
CO
1970
1971
1972
1973
1974
1975
26.8
24.9
23.4
21.9
20.3
18.0
34.2
32.3
26.7
35.6
34.1
32.9
22.7
23.4
24.6
25.7
25.0
24.4
33.9
33.3
34.1
34.0
32.9
30.9
113.7
113.7
115.8
111.5
103.3
96.2
*These figures generally do not include fugitive process
emissions or fugitive dust
-------
sources. Some initial estimates have indicated that they often appear to
have a greater effect on air quality in the immediate vicinity of a source
than do controlled (and sometimes uncontrolled) stack emissions. One recent
study of 14 cities^ found that these unconventional sources contribute
25 to 35 micrograms per cubic meter to the citywide TSP levels, thereby
preventing the attainment of standards.
At this time, it may be necessary to define the differences between
and discuss pertinent facts concerning two separate but, in some cases,
strikingly similar air pollution problems (i.e., fugitive process emissions
and fugitive dust emissions). When using the term fugitive (process)
emissions, we are referring to both gaseous and particulate emissions that
result from industrial-related operations and which escape to the atmosphere
through windows, doors, vents, etc. but not through a primary exhaust
system such as a stack, flue, or control system. Fugitive process emissions
may result from metallurgical furnace operations, materials handling,
transfer and storage operations where emissions escape to the atmosphere.
Fugitive emissions, by their very nature, generally occur at or near
ground level and tend to remain at low levels where the localized impact
on air quality in the area is greatest.
Fugitive dust emissions, on the other hand, are generally related
to natural or man-associated dusts (particulate only) that become airborne
due to the forces of wind, man's activity, or both. Fugitive dust
emissions may include windblown and traffic-generated particulate matter
from unpaved dirt roads, tilled farm lands, exposed surface areas at
construction sites, etc. Naturally "erodible" dusts that become airborne
are also generally included in the fugitive dust definition.
-32-
-------
While measurements of fugitive dust and fugitive emissions (both
particulate and gaseous) have proven difficult, approximations have been
made which indicate that such sources may comprise a significant portion
of nationwide emissions and localized sources impacting upon local air
quality are sufficiently widespread to be of national significance. Thus,
if the NAAQS (for TSP and oxidant in particular) are to be attained
nationwide, serious attention must be paid to the role which fugitive
dust and fugitive process emissions play.
The Role and Uses of AP-42 Emission Factors
In the assessment of community air pollution, there is a critical
need for accurate data on the quantity and characteristics of emissions
from the numerous sources that contribute to the problem. The large
number of individual sources and the diversity of source types generally
make conducting field measurements of emissions on a source-by-source
basis at each point of release impractical. The most feasible method of
determining pollutant emissions for a given community is to make generalized
estimates of typical emissions from each of the source types. Emissions
from all sources, as totaled and tabulated in an emission inventory,
provide a valuable technique for assessing the character of the local
problems and for development of workable, reasonable, and cost-effective
control strategies. They, along with ambient data, constitute the basic
"corner stones" of a viable air pollution control and air resource
management program.
One of the most useful (and critical) tools for estimating typical
emissions for an emission inventory is the emission factor. This is an
-33-
-------
estimate of the rate at which a pollutant is released to the atmosphere
from a specific source category as a result of some activity such as
combustion or industrial production divided by the level of the activity.
In other words, the emission factor relates the quantity of pollutants
emitted to some measurable surrogate indicator (activity level) such as
production capacity, quantity of fuel burned, or vehicle miles traveled.
EPA has collected and has a continuing program underway to develop
and update such emission factors in published form; "Compilation of Air
Pollutant Emission Factors,"5 commonly referred to as AP-42. This document
has become an internationally accepted primary reference that is used by
State and local agency officials and others in their emission inventory
efforts and related activities which require estimation of emissions (such
as preparation of environmental impact statements).
The emission factors presented in AP-42 have been developed
using a wide spectrum of techniques. A program to periodically update
and revise this document has been in effect and is continuing. This
effort involves the search for and procurement of all the available and
applicable written information on a source category. After these data
and literature are reviewed, organized, and analyzed, the process descrip-
tions, process flowsheets, and other background portions of the section
are prepared. Using information compiled, representative emission factors
are developed for each pollutant emitted at each emission point of the
particular process. These factors are usually obtained by application of
a combination of engineering judgment and statistical manipulation in an
attempt to describe the representative or typical plant. When feasible,
the ranges in the factors are presented for further clarity. The statistical
-34-
-------
manipulations involved may range in complexity from simple averaging of
limited simple measurements to a rather vigorous regression and correlation
analysis with various process variables that may exist. Occasionally,
enough data exist to permit the development of either empirical or theoreti-
cal formulas (or graphs) relating emission factors to various process
parameters such as stream temperature, sulfur content, or catalyst deple-
tion. In these cases, representative values of these process parameters
are selected and substituted into the formulas or graphs that, in turn,
yield representative emission factors which are then tabulated. The
pertinent formulas and graphical data are also normally included to provide
the flexibility needed for estimation of "custom" emission factors when
the process conditions differ from those present in AP-42.
The reader must be cautioned not to use emission factors indiscrimi-
nately. That is, the factors generally will not permit the calculation
of accurate emissions measurements from an individual installation.
Only an on-site source test can normally provide data sifficiently accurate
and precise to use in such undertakings as the design and purchase of
control equipment or the initiation of a legal action. Factors are more
valid when applied to a large number of processes as, for example, when
emission inventories are conducted as part of community or nationwide air
pollution studies. Emission inventories and emission factors are thus
probably best described as technical tools for use in management analysis
activities such as control strategy development, program assessment, or
other facets of air resource management.
Emission factors are normally the most sensitive components of emission
inventories in terms of inventory error. Thus, it is important that
-35-
-------
published factors in AP-42 be comprehensive and up-to-date with the
"state of knowledge" regarding various sources. Consequently, AP-42 is
updated with periodical supplements which add, or otherwise revise factors
to reflect recent information on newly identified and measured sources;
to reflect subtle or gradual changes in process design and operation;
to reflect a new review of accepted factors (generally for major sources)
which have been in use for many years; and occasionally, to correct errors
or unclear passages that may have been discovered.
Continuing efforts at EPA to collect data for emission factor enhance-
ment are both active and passive. New emission factors may be acquired
as a direct result of source testing funded and initiated solely or
primarily for the purpose of development of emission factors (active), or
as a "by-product" of data collection (testing) for purposes such as enforce-
ment, research, etc. (passive). Due to the extremely high costs for source
testing and the many testing efforts that are carried out independent of
emission factor development, most data collection tends to be under the
passive mode. Future attempts to obtain results from the many testing pro-
grams being carried out by EPA, State and local agencies, other government
agencies, and private industry are also undertaken. Maintaining cognizance
of these activities in itself is often a somewhat monumental task.
To assist EPA in collecting and keeping track of source test data
that may be available and to maximize the utilization capabilities if such
data, the SOurce lest DATa (SOTDAT) system has been developed. This
system keyed to the National Emission Data System (NEDS - a nationwide
emission inventory system, is maintained by data supplied by the States to
EPA through the EPA Regional Offices) allows for storage and retrieval of
-36-
-------
various air pollution source tests which are conducted by a variety of
sources. Data stored in this system provides a supplementary source of
raw test data for use with literature and other data available in develop-
ment of emission factors. With this background in mind, let me discuss
how this applies to fugitive emissions and fugitive dust sources.
Status of Fugitive Process Emission Factors
Reflecting the concern of EPA with regard to "fugitives" and their
impact upon air quality, the U. S. Environmental Protection Agency has
initiated development of various measurement techniques and programs to pro-
vide quantitative assessment of various sources. As evidenced by this
meeting, the Agency is continuing its effort towards improving the measure-
ment techniques and quality and comprehensiveness of the data base. Many
groups within EPA have undertaken various projects to attempt to quantify
or develop methods to quantify fugitive dust and fugitive process emission
sources even beyond what one might surmise from examining the agenda
for this meeting. A complete list would be too long for inclusion in
this paper and, in fact, it would be very difficult to assemble because of
the number and diversity of such efforts.
As an indication of the status of our assessment efforts, let us
examine as an example an effort which EPA is now completing (under con-
tract) entitled "Control Program Guideline for Industrial Process Fugitive
Particulate Emissions" (discussed in more detail on the program for this
meeting). This essentially is a distillation and summary of the "state-
of-the-art" for estimating fugitive particulate emissions for a wide range
of sources. With the upcoming revision to State Implementation Plans (SIPS's)
-37-
-------
which have been requested by EPA in many areas not meeting the NAAQS for
TSP, air pollution control agencies at all levels have expressed concern
relative to the proper fashion in which approach, evaluate, and handle
fugitive particulate emissions from industrial operations. Thus, this
project was developed to pull together all available information on the
subject.
The guideline addresses:
. Industrial process fugitive particulate emission sources
. Control technology for these defined sources
. Conceptual procedures to use to estimate their impact on
ambient air quality
. Guidance to the States for integration of source impact for
such sources into the SIP process
Although the results of this project are very timely and useful, one of
the major findings of the study and review was that there is a general
lack of adequate test data for development of comprehensive and defensible
emission factors for fugitive process emission sources. Although the
estimates and emission factors derived through this effort are scheduled
to be included in a supplement to AP-42, many of these factors are
recognized as in need of additional investigation. Weak factors and gap
areas have been defined which need to be systematically approached,
examined in more detail, and specific sampling accomplished to fill in
gaps and strengthen areas of weakness in data.
Fugitive emissions, of course, are not limited to TSP though this
pollutant has received major emphasis in this discussion. Other pollutants
may surface as areas of concern as our knowledge broadens. The Air
-38-
-------
Management Technology Branch (in cooperation with the National Air Data
Branch which currently has publication responsibilities), Monitoring and
Data Analysis Division, Office of Air Quality Planning and Standards, is
now involved in incorporating available fugitive emission factors into
AP-42 as they become available. Unfortunately, the mechanism for accom-
plishing this objective in each case is usually unique. In some instances,
it may be possible to incorporate development of an AP-42 section with an
accompanying background support document into a larger assessment contract
effort. In other cases, it may be necessary to issue an independent
contract to develop AP-42 sections from existing data or from source
test data collected from specific field work. In other instances, we
may develop a test procedure, gather data, analyze it, and write the
appropriate AP-42 section in-house. In-house resources available for
this activity, however, are limited and must be utilized very efficiently.
Table 2 contains a list of selected sources of fugitive emission
and fugitive dust for which we currently need additional data for
improvement. The list is by no means meant to be all inclusive or
complete or to say that existing factors in use are not useful, but is
shown to give you an idea of the direction in which we are headed and to
give some indication of areas for which we expect to be seeking data and/or
publishing revised factors in the next several months. Other related
fugitive sources will be characterized concurrently with those shown if
data become available.
At the present time, there are many efforts underway to characterize
fugitive emissions which hopefully will be amenable to incorporate into
AP-42 and fill some of the identified gaps. Both government and private
-39-
-------
Table 2
Selected List of Fugitive Emission and Fugitive Dust Categories
Most Critically in Need of Emission Factor Improvement
o
I
Fugitive Process
. Iron and Steel
. Primary Non-Ferrous
. Secondary Non-Ferrous
. Foundries
. Portland Cement
. Lime Manufacturing
. Petroleum Industry
- Extraction
- Field Processing
- Drains
- Cooling Towers
- Valves and Seals
Material Mining, Handling, and Storage
. Coal
. Sand and Gravel
. Stone
. Grains
. Ferti1i zers
. Pesticides
. Metallic Ores
. Fly Ash
. Wood Chips
Fugitive Dust
. Open Land
. Roads
. Parking Lots
(NOTE: The need for particle size data is becoming more critical
for accurately depicting deposition in diffusion models which are
frequently used to determine impact of fugitive emissions.)
-------
industry are involved in these efforts. Table 3 is a selected list of
several efforts, results of which are expected to provide us with input
to our goal of making the latest information available to various control
officials who rely upon AP-42 as a primary data source. Input from other
such studies and organizations in establishing data needs and filling
these gaps is solicited.
As these data become available and are incorporated into AP-42, and
guideline documents containing additional background information are
prepared, the State and local air pollution agencies will be better
informed and able to cope with their local situations.
People in government and industry recognize the need for more accurate,
up-to-date information in the area of emission characterization. There
will be a continuing need for cooperation on the parts of government and
private industry to see that data representing realistic emission character-
ization are available for use by agencies as required to develop and carry
out air pollution implementation plans. This cooperative spirit is
evidenced by this meeting and forum it provides. The number of such
meetings seems to be increasing. As a matter of interest, a workshop
specifically on emission inventories and emission factors is to be
presented at Research Triangle Park, N.C., September 13-15. The focus
of this meeting will be on organic emission factors and will address
i
fugitive emissions as well as "stack" emissions.
Summary
Although the trend in air quality levels is generally downward,
studies have shown that the impact of fugitive dust and emissions on
-41-
-------
Table 3
List of Selected Projects Expected to Provide Input for
AP-42 Emission Factor Development
Subject of Project
Iron, Steel and Gray Iron
Control Technology Assessment
of Fugitive Emissions
Sources Assessment of Fifty
Industries for Environmental
Risk
Emission Factor Analysis of
the Iron and Steel Industry
Fugitive Hydrocarbon Emissions
from Vegetation
Re-entrainment of Dust from
Paved Roads
Fugitive Dust Emissions from
Coal Strip Mining
Fugitive Emissions from the
Rock Crushing Industry
Hydrocarbon Source Testing in
the Los Angeles Air Basis
Hydrocarbon Emissions from
Storage Tanks
Hydrocarbon Emissions from
Loading Operations
Hydrocarbon Emissions from
Storage Tanks
Fugitive Emissions from
On- and Off-Shore Production
Facilities
Responsible Organization
EPA/Industrial Environmental
Research Laboratory, Research
Triangle Park, N.C.
EPA/Industrial Environmental
Research Laboratory, Research
Triangle Park, N.C.
American Iron and Steel Institute
EPA/Office of Air Quality Planning
and Standards, Research Triangle
Park, N.C.
EPA/Office of Air Quality Planning
and Standards, Research Triangle
Park, N.C.
EPA/Office of Air Quality Planning
and Standards, Research Triangle
Park, N.C.
EPA/Region VIII, Denver, Colorado
EPA/Division of Stationary Source
Enforcement, Washington, D.C.
California Air Resources Board,
Sacramento, California
Western Oil and Gas Association,
Los Angeles, California
Western Oil and Gas Association,
Los Angeles, California
American Petroleum Institute,
Washington, D.C.
American Petroleum Institute
Washington, D.C.
-42-
-------
local air quality may be significant. Analysis of results of such studies
lead to conclusions that fugitive sources must be considered in developing
action strategies to meet the National Ambient Air Quality Standards. The
impact of such sources has not been adequately addressed in most areas
prior to this time, and were generally omitted in base year State Implemen-
tation Plan (SIP) inventories in part due to lack of existing emission
factors. An examination of the status and role of emission factors needed
for preparation of revised SIP's indicates that there are many areas with
major data needs particularly in the context of availability and
applicability of generalized factors to be utilized in AP-42, Compilation
of Air Pollutant Emission Factors. A program to gather and incorporate
data from various agencies and other organizations is underway and will
hopefully provide State agencies and others with the data base needed
for program development. This can only be accomplished by the continued
cooperative interface between various government and private organizations.
-43-
-------
References
1. The Clean Air Act (42 U.S.C. 1857 et seq.), including the "Clean
Air Amendments of 1970", P.L. 91-604, December 31, 1970.
2. State Air Pollution Implementation Plan Progress Report. June 1
to June 30. 1976. EPA-450/2-76-026. October 1976, U.S. Environ-
mental Protection Agency, Research Triangle Park, N.C. 27711.
3. National Air Quality and Emission Trends Report. 1975, EPA-450/1-76-002,
November 1976, U.S. Environmental Protection Agency, Research
Triangle Park, N.C. 27711.
4. National Assessment of the Urban Particulate Problem, Volume I -
Summary of National Assessment, EPA-450/3-76-Q24, July 1976,
U.S. Environmental Protection Agency, Research Triangle Park,
N.C. 27711.
5. Compilation of Air Pollutant Emission Factors, AP-42, April 1973,
(Second Edition with Supplements 1-6, April 1976), U.S. Environ-
mental Protection Agency, Research Triangle Park, N.C. 27711.
-44-
-------
Monday Afternoon - May 23, 1977
SESSION II: MEASUREMENT
Chairman: James A. Dorsey, EPA
Page
MEASUREMENT OF FUGITIVE PARTICULATE 47
Chatten Cowherd, Jr., Ph.D., Midwest Research. Institute
FUGITIVE MONITORING AT A COAL CLEANING PLANT SITE 63
D. Ambrose, D. Brown* and R. Clark,
Battelle Columbus Laboratories
PARTICULATE SAMPLING TECHNIQUES FOR A COKE QUENCH TOWER 115
A. H. Laube, John Jeffery*, York Research Corporation,
and Carl Edlund, EPA
MONITORING INDUSTRIAL FUGITIVE EMISSIONS - AN OCCUPATIONAL
HEALTH PERSPECTIVE 141
James L. Oser*, Ronald J. Young, John M. Dement,
Howard R. Ludwig, NTOffff
* indicates speaker
-45-
-------
MEASUREMENT OF FUGITIVE PARTICULATE
By
Chatten Cowherd, Jr., Ph.D.
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
-47-
-------
MEASUREMENT OF FUGITIVE PARTICIPATE
by
Chatten Cowherd, Jr.
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
ABSTRACT
This paper describes an alternative to the upwind/downwind sampling
strategy for use in quantifying particulate emissions from open (unconfinable)
sources. Exposure profiling consists of the direct measurement of the total
passage of airborne particulate downwind of the source by means of simulta-
neous multipoint sampling over the effective cross-section of the fugitive
emissions plume. Samples are collected isokinetically by directing sampling
intakes into the wind and by matching intake velocity to the local mean wind
speed.
Exposure profiling offers distinct advantages over the upwind/
downwind method in quantifying fugitive particulate emissions from open
sources as required for control strategy development. The method yields
source-specific emissions data needed to evaluate the effectiveness of control
measures. Moreover, based on field tests of several types of open dust
sources, the accuracy of measurements obtained by exposure profiling is bet-
ter than that achievable by the upwind/downwind method, even with site-specific
calibration of the dispersion model used in the latter method.
-48-
-------
INTRODUCTION
Failure to attain national ambient air quality standards for total
suspended particulates (TSP) in both urban and rural areas has spurred a
detailed reexamination of the national TSP problem. As a result of this
assessment, it has become evident that fugitive emissions sources contribute
substantially to TSP levels in many of the nonattainment areas. The regional
impact of fugitive sources of particulate matter is the result of emissions
of fine particles which have the potential for atmospheric transport over
substantial distances from the source.
Analysis of the impact of fugitive emissions of particulate matter
must take into account the ranges of particle size associated with the ad-
verse effects of particulate air pollution. Adverse health effects are
generally attributable to particles smaller than about 5 micrometers (urn)
in aerodynamic (equivalent sphere) diameter. Coarse particles tend to settle
out over short drift distances, creating localized nuisance problems. How-
ever, dustfall on roadways or other unprotected areas is a source of fine
particle emissions through the action of traffic and wind erosion.
The quantification of particulate emissions from fugitive sources
is necessary to the rational assessment of the environmental impact created
by these sources and to the development of effective control technology. In
addition to mass rates of emission, information is needed on particle size
distribution and the presence of toxic constituents within specific size
ranges.
-49-
-------
In large part, proven methods for quantifying fugitive emissions
have not been fully developed. Atypical quantification problems are pre-
sented by the diffuse and variable nature of fugitive sources. Standard
source testing methods, as written, strictly apply only to well defined, con-
strained flow fields with velocities above about 2 m/sec. Such methods are
applicable to fugitive emissions only if it is possible to capture the entire
plume by means of an enclosure or hooding device.
The Research Corporation of New England (TRC) ' has defined the
following three basic sampling strategies for measurement of fugitive emis-
sions from most potentially significant sources:
1. The quasi-stack method involves capturing the entire emissions
stream with enclosures or hoods and applying conventional source testing
techniques to the confined flow.
2. The roof-monitor method involves measurement of concentrations
and air flows across well-defined building openings such as roof monitors,
ceiling vents and windows.
3. The upwind-downwind method involves measurement of upwind and
downwind air quality, utilizing ground-based samplers under known meteoro-
logical conditions, and calculation of source strength with atmospheric dis-
persion equations.
This paper focuses on an alternative to the upwind/downwind sampl-
ing strategy for use in quantification of particulate emissions from open
sources such as traffic on unpaved roads and aggregate materials handling
operations. Exposure profiling consists of the direct measurement of the
-50-
-------
total passage of airborne particulate immediately downwind of the source by
means of simultaneous multipoint sampling over the effective cross-section
of the fugitive emissions plume. This strategy utilizes a calculation scheme
similar to the quasi-stack or roof-monitor method rather than requiring the
application of a dispersion model.
OPEN SOURCE QUANTIFICATION BY UPWIND/DOWNWIND METHOD
The upwind/downwind method has frequently been used to measure
fugitive particulate emissions from open (unconfinable) sources. Typically,
particulate concentration samplers (most often high-volume filtration sam-
plers) are positioned at a considerable distance from the source (for example,
at the property line around an industrial operation) in order to measure the
highest particulate levels to which the public might be exposed. The calcu-
lation of the emission rate by dispersion modeling is often treated as having
secondary importance, especially because of the difficult problem of identify-
ing the contributions of elements of the mix of open (and possibly confinable)
sources.
While the above strategy is useful in characterizing the air
quality impact of an open source mix, it has significant limitations with
regard to control strategy development. The major limitations are as follows:
1. Overlapping of source plumes precludes the determination of
source specific contributions on the basis of particulate concentration
alone.
-51-
-------
2. Air samplers with poorly defined intake flow structure (in-
cluding the conventional high-volume sampler) exhibit diffuse cutoff size
characteristics for particle capture, which tend to be affected by wind
conditions/2^
3. Uncalibrated atmospheric dispersion models introduce the possi-
bility of substantial error (a factor of three^ ') in the calculated emission
rate, even if the stringent requirement of unobstructed dispersion from a
simplified source configuration is met.
The first two limitations are not a direct consequence of the
upwind/downwind method but of the way it is used. These limitations could
be removed by using samplers designed to capture all or a known size fraction
of the atmospheric particulate, and by designing sampler placement to isolate
the air quality impact of a well-defined source operation.
However, there would remain the need to improve method accuracy by
calibration of the dispersion model for the specific conditions of wind,
surface roughness, and so on, which influence the near-surface dispersion
process. This need is evident from the significant size of the variation
in model-calculated emission rates for aggregate process operations, based
on data from individual samplers operated simultaneously at different down-
(4)
wind locations. The suggested use of tracers for this purpose is com-
plicated by the characteristically diffuse and variable nature of an open
dust source and the need for a polydisperse tracer test approximating the
particle size distribution of the source emissions.
-52-
-------
EXPOSURE PROFILING METHOD
The exposure (total mass flux) profiling method was developed by
Midwest Research Institute ' to measure particulate emissions from spe-
cific open sources, utilizing the isokinetic profiling concept which is the
basis for conventional source testing. For measurement on nonbuoyant fugi-
tive emissions, sampling heads are distributed uver a vertical network posi-
tioned just downwind (usually about 5 meters) from the source. Sampling
intakes are pointed into the wind and sampling velocity is adjusted to match
the local mean wind speed, as monitored by distributed anemometers. A
vertical line grid of samplers is sufficient for measurement of emissions
from line or moving point sources while a two-dimensional array of samplers
is required for quantification of area source or point source emissions.
GRID SIZE AND SAMPLING DURATION
Sampling heads are distributed over a sufficiently large portion
of the plume so that vertical and lateral plume boundaries may be located
by spatial extrapolation of exposure measurements. The size limit of area
sources for which exposure profiling is practical is determined by the
feasibility of erecting sampling towers of sufficient height and number to
characterize the plume. This problem is minimized by sampling when the wind
direction is parallel to the direction of the minimum dimension of the area
source.
The size of the sampling grid needed for exposure profiling of a
particular source may be estimated by observation of the visible size of the
-53-
-------
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 example, if the exposure from a point source is normally
distributed, as shown in Figure 1, the exposure values measured by the sam-
plers at the edge of the grid should be about 25% of the centerline exposure.
Sampling time should be long enough to provide sufficient partic-
ulate mass and to average over several units of cyclic fluctuation in the
emission rate (for example, vehicle passes on an unpaved road). The first
condition is easily met because of the proximity of the sampling grid to the
source.
Assuming that sample collection media do not overload, the upper
limit on sampling time is dictated by the need to sample under conditions
of relatively constant wind direction and speed. In the absence of passage
of weather fronts through the area, acceptable wind conditions might be
anticipated to persist for a period of 1 to 6 hr.
CALCULATION PROCEDURE
The passage of airborne particulate, i.e., the quantity of emissions
per unit of source activity, can be obtained by spatial integration (over the
effective cross-section of the plume) of distributed measurements of ex-
posure (mass/area). The exposure is the point value of the flux (mass/area-
time) of airborne particulate integrated over the time of measurement.
-54-
-------
Virtual Point Source
I
Ul
*• X
Exposure
Profiles
Wind Direction
Figure 1. Example Exposure Profiling Arrangement
-------
Mathematically stated, the total mass emission rate (R) is given by:
R = I f f m(h,w) dhdw
t Jh
where m = dust catch by exposure sampler after subtraction of background
a = intake area of sampler
t = sampling time
h = vertical distance coordinate
w = lateral distance coordinate
A = effective cross-sectional area of plume
In the case of a line source with an emission height near ground
level, the mass emission rate per source length unit being sampled is given
by:
R = w I minj. dh
m(h)
a
where W = width of the sampling intake
H = effective extent of the plume above ground
In order to obtain an accurate measurement of airborne particulate
exposure, sampling must be conducted isokinetically, i.e., flow streamlines
enter the sampler rectilinearly. This means that the sampling intake must
be aimed directly into the wind and, to the extent possible, the sampling
velocity must equal the local wind speed. The first condition is by far the
more critical.
If it is necessary to sample at a nonisokinetic flow rate (for
example, to obtain sufficient sample under light wind conditions), multi-
plicative factors may be used to correct measured exposures to corresponding
-56-
-------
isokinetic values.^ »°) These corrections require information on the particle
size distribution of the emissions.
High-volume cascade impactors with glass fiber impaction substrates,
which are commonly used to measure particle size distribution of atmospheric
particulate, may be adapted for sizing of fugitive particulate. A cyclone
preseparator (or other device) is needed to remove coarse particles which
otherwise would be subject to particle bounce within the impactor causing
fine particle bias. Once again, the sampling intake should be pointed
into the wind and the sampling velocity matched to the mean local wind speed.
Midwest Research Institute has used the exposure profiling method
(s f\\
to develop emission factors for unpaved (dirt and gravel) roads, * ' agri-
cultural tilling, aggregate storage piles (crushed stone, iron ore
and slag ), and dust entrainment from paved roadways. Emission factors
have been derived for particles smaller than 30 Jim in diameter (the approxi-
mate aerodynamic cutoff diameter for capture of fugitive dust by a standard
high-volume filtration sampler, (based on a particle density of 2.0 - 2.5
g/cm ) and for particles smaller than 5 um.
Figure 2 shows the intake design used in profiling atmospheric dust
entrained by traffic on paved urban roadways. Intake velocity is adjustable
up to 15 mph. The sample stream passes through a settling chamber and then
upward through a standard 8 in. by 10 in. glass fiber filter positioned
horizontally. Figure 3 gives particulate concentration profiles measured
5 meters downwind of a 4-lane paved roadway with a traffic density of about
(6)
500 vehicles/hour.
-57-
-------
ConcenfraMon
Mode
To Vacuum
Source and
Controls
Figure 2. MRI Exposure Profiler
-58-
-------
vo
I
£ 3
0)
^ 2
IU
X
0
Site: 37th Street
Kansas City, Mo.
O Run 3
O Run 5
V Run 6
Exposure Profiler positioned
5m downwind from curbing
50 100 150 200
ISOKINETIC PARTICULATE CONCENTRATION (yug/m3) Above Background
250
Figure 3. Vertical Profiles of Participate Concentration (Street Dust Entrainment)
-------
Based on replicate exposure profiling of open dust sources under
varying conditions of source activity and properties of the emitting surface,
emission factor formulae have been derived that successfully predict test
results with a maximum error of 20%.^ ' These formulae account for the frac-
tion of suspendable fines (silt) in the emitting surface, the surface moisture
content, and the rate of mechanical energy expended in the process which
generates the emissions. Based on the above results, the accuracy of expo-
sure profiling is considerably better than the + 50% range given for the
upwind/downwind method with site-specific dispersion model calibration.
CONCLUSIONS
Exposure profiling offers distinct advantages over the standard upwind/
downwind method for quantification of particulate emissions from specific
open sources for purposes of control strategy development. The major advan-
tages are as follows:
1. Placement of isokinetic samplers near the source yields defin-
itive information on total particulate emissions from a specific operation,
which can be used to test the effectiveness of specific control measures.
2. Utilization of total plume profiling eliminates the need for
reliance on a conventional dispersion model in the calculation of emission
rate.
3. Exposure profiling yields results which are more accurate than
those achievable by the upwind/downwind method even if the difficult require-
ment of site-specific dispersion model calibration is met.
-60-
-------
These conclusions establish the value of exposure profiling in
quantifying particulate emissions from open sources as required for effec-
tive source-specific control strategy development.
-61-
-------
REFERENCES
1. Kalika, P. W., R. E. Kenson, and P. T. Bartlett, Development of Proce-
dures for the Measurement of Fugitive Emissions. Final Report, The
Research Corporation of New England for U.S. Environmental Protection
Agency Publication No. EPA-600/2-76-284, December 1976.
2. Turner, D. B., Workbook of Atmospheric Dispersion Estimates. U.S.
Environmental Protection Agency, Publication No. AP-26, 1970.
3. Lundgren, D. A. and H. J. Paulus, "The Mass Distribution of Large
Atmospheric Particles," Paper No. 73-163, presented at the 66th Annual
Meeting of the Air Pollution Control Association, Chicago, Illinois,
June 24-28, 1973.
4. Blackwood, T. R., T. F. Boyle, T. L. Peltier, E. C. Eimutis, and D. L.
Zanders, Fugitive Dust from Mining Operations. Final Report, Monsanto
Research Company for the U.S. Environmental Protection Agency,
Contract No. 68-02-1320, Task 6, May 1975.
5. Cowherd, C. Jr., K. Axetell, Jr., C. M. Guenther, and G. Jutze, Develop-
ment of Emission Factors for Fugitive Dust Sources. Final Report,
Midwest Research Institute for U.S. Environmental Protection Agency,
Publication No. EPA-450/3-74-037, June 1974.
6. Midwest Research Institute, Quantification of Dust Entrainment from Paved
Roadways. Final Report, U.S. Environmental Protection Agency, Contract
No. 68-02-1403, Task 25, (in preparation).
7. Midwest Research Institute, A Study of Fugitive Emissions from Metal-
lurgical Processes. Final Report, U.S. Environmental Protection Agency,
Contract No. 68-02-2120, (in preparation).
-62-
-------
FUGITIVE MONITORING AT A
COAL CLEANING PLANT SITE
By
D. Ambrose, D. Brown, and R. Clark
BATTELLE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
-63-
-------
ABSTRACT
This paper presents preliminary procedures and findings of a fugi-
tive dust monitoring effort about a large coal cleaning plant. Hi-volume
samplers were the primary instrument of measurement; limited data concerning
particle sizing is presented. Extensive meteorological data is offered.
Microscopic and chemical analysis was performed. Data suggesting
geographical migration of select metals are displayed.
Preliminary data is presented which compares chemical concentra-
tions with mass weights obtained under a specific meteorological setting.
The coal cleaning plant was not in operation during the monitoring
efforts. The main thrust was to define pre-operational loadings (fugitive)
so as to permit a comparison with the near future loadings when this large
coal cleaning plant becomes operational.
-64-
-------
INTRODUCTION
This grab sampling effort was conducted as part of a program for
environmental assessment of coal cleaning processes under a current contract
from the Fuel Process Branch, Energy Assessment and Control Division,
Industrial Environmental Research Laboratory, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina.
This paper presents an outline of particulate measurement activities
and the requisite computer modeling/simulation. Analytical results of the
chemical nature of the captured particulates are given. The micro-meteorology
that occurred during the field effort is provided and interpreted. The
results of a microscopic examination of the particulates are included in this
paper.
A detailed discussion of the theory and mechanics of wide area
source modeling procedure is outlined herein.
The last segment of this paper concludes the efforts of this grab
sampling experience. This summary provides information about the apparent
fall-out rates as compared with geographical distances downwind from target
sources. Trends in the relationships of chemical concentration inherent
in particulates from coal source areas are given. Loading values of weekend
periods vs work week periods are provided. Suggestions as to the diurnal
loading variations are included.
-65-
-------
BACKGROUND
Study Objectives
An intense environmental assessment of a coal cleaning facility
has been undertaken by Battelle's Columbus Laboratories at the site of an,
Appalachian generating station.
One portion of the environmental assessment was devoted to eval-
uating air quality impacts deriving from sources of fugitive dust located
in the area of plant operations prior to the start up of the coal cleaning
facility at the Appalachian power station. Battelle's objectives in this
portion of the assessment were twofold: one, to identify the types of
pollutants emitted from the fugitive dust sources, and two, to determine
through site monitoring the amount of material that is being deposited on
downwind sites.
Objectives of Air Monitoring
Three air monitoring campaigns, each extending for 48 hours,
commencing in December, 1976, and ending in April, 1977, were conducted by
Battelle. At the outset of the study the purpose for on-site air monitoring
was to determine ambient or background levels of total suspended coal and
ash particulates before the start-up of the coal cleaning plant. However,
due to time constraints imposed and to the limitations of data analysis
imposed by non-continuous on-site monitoring, the objectives were later
modified to collecting grab samples from which a more comprehensive
sampling program could be developed for future studies during the coal
cleaning plant's interim and final plant configuration.
-66-
-------
All three air monitoring campaigns were conducted prior to the
operation of the coal cleaning facility. The first monitoring period took
place from 2000 December 17, 1976, to 2000 December 19, 1976. The second
and third air monitoring campaigns were also of 48-hour duration and occurred
from 2000 January 5 to 2000 January 7, 1977, and from 2000 April 5 to 2000
April 7, 1977, respectively.
Objectives of Diffusion Modeling
In studying potential areas for locating air monitors it was
decided that a diffusion analysis be undertaken to indicate a range of dis-
tances where the maximum concentrations may be found from each of the
fugitive dust sources. An area source mathematical diffusion model was
developed not only to validate the final determination of the areas selected
for monitoring but also to predict with reasonable accuracy concentrations
of fugitive dust at any desired receptor location.
General Description of Site
An intense environmental assessment of a coal cleaning facility
was undertaken by Battelle's Columbus Laboratories at the site of an
Appalachian generating station. The Appalachian power station is intended
to supply electrical energy to thousands of consumers throughout the
region.
The power station consists of two generator units that utilize
superheated steam from two boilers, each 40' x 66' x 151', to transform
mechanical energy to electrical energy. The boilers are fired by tons of
pulverized coal. Air pollutants are discharged through two tall stacks.
The stacks were constructed to these heights to help assure an effective
-67-
-------
level of dispersion.
MEASUREMENTS
Criteria in Selecting Monitoring Sites
In designing an air monitoring system, the first consideration is
generally given to the intent for monitoring. Are maximum concentrations of
primary interest or is a spatial representation of average concentrations of
special importance? Battelle, in monitoring fugitive dust about this coal
cleaning facility, wished to obtain both a spatial distribution of the dust
deposition and the magnitude of the maximum concentrations. The second
important consideration in designing a monitoring system is the nature of
emissions. Since the sources in the study area are fugitive in nature, it
is reasonable to assume that the greatest impacts will occur within short
distances downwind and therefore monitoring efforts should be concentrated
in close proximity of the sources.
Ten monitoring sites were carefully selected in accordance with
the criteria established above. Depicted in Figure 1 are the location of
the monitoring sites. Five sites (Nos, 1, 2, 3, 4, and 7) are located in
the immediate vicinity of the power plant operations. The remaining sites
(Nos. 5, 6, 8, 9, and 10) are well removed from the center of plant oper-
ations. Ideally, a monitoring system is designed to produce a set of
samples that represent a spatial distribution of pollutant concentration.
The monitoring system established at the Appalachian power plant may be
somewhat less than an ideal system in that power availability and climato-
logical features of the area must be realistically factored into such field
studies. For future studies the monitoring system may include additional
-68-
-------
t
x Hi volume sampler
• Wind speed and direction
recorder
Existing ash
disposal area
Coal cleaning
refuse disposal
U*— Coal cleaning plant
i Substation
i J
Coal storage pile
VxPower plant
O Cooling towers
Unpaved parking lot
Ta I!est Stack
FIGURE 1. LOCATION OF SOURCES AND MONITORING SITES AT COAL
CLEANING AND POWER PLANT (Not Drawn to Scale)
-69-
-------
monitoring sites and additional particle sizing monitors.
Shown below are the results of Battelle's Fugitive Dust Emission
Model which was utilized in studying potential sites for monitoring. The
distances summarized in the table below agree favorably with the locations
of the monitoring sites situated around plant operations. The term "major
axis" is used in reference to the configuration of the source.
Distance of Maximum
Fallout Concentrations
Source Wind Direction (meters)
Coal Pile Plant Perpendicular to Major Axis 250-400
Parallel to Major Axis 400-500
Boney Pile Perpendicular to Major Axis 150-300
Parallel to Major Axis 200-400
Ash Disposal Area Perpendicular to Major Axis 0-100
Parallel to Major Axis 0-100
Description of Sampling Site
Listed below is a brief description of the ten sampling sites:
Site #1 - Located about 175 meters northeast of coal storage pile and 300
meters northeast of the coal cleaning plant.
Site #2 - Located about 200 meters southeast of the coal storage pile.
Site #3 - Located about 150 meters northwest of the coal storage pile
and 50 meters northeast of the coal cleaning plant.
Site #4 - Located about 400 meters north of the coal storage pile and
about 300 meters east of the boney pile.
Site #5 - Located about 200 meters east of the ash disposal area.
Site #6 - Located on a farm about 1600 meters southwest of the coal
cleaning plant.
Site #7 - Located about 500 meters southwest of coal storage pile and
adjacent to an employee parking lot at the company's property line.
-70-
-------
Site #8 - Located about 1200 meters north of coal storage pile in an area
that is being readied to handle the refuse from the coal cleaning
plant.
Site #9 - Located about 1800 meters north of the coal storage pile near the
company property line.
Site #10 - Located on a farm about 2200 meters northeast of the coal storage
pile.
Figure 1 is a drawing of the plant site and location of the hi-
volume samplers.
Description of the Sources
All of the fugitive dust sources, as inferred by the term fugitive,
are partially or fully exposed to dispersion by the wind. Those sources on
the power plant property include an ash disposal area, where flyash and
bottom ash are routinely deposited by trucks; a boney pile, which serves as
a refuse area for undesired coal from the mining operations; and a large coal
pile from which coal is fed into the coal-fired boilers. Fugitive dust moni-
toring also occurred at a proposed refuse area for the coal cleaning plant.
In addition, a modern coal cleaning facility adjacent to the power plant was
also considered as a source of fugitive dust. These sources are depicted in
-71-
-------
Figure 1. Notice the amount of land area some of these sources occupy.
The ash disposal area occupies an area of about 17.7 acres and is
of irregular configuration. Here bottom and flyash are spread uniformly
over the area; the height of the ash disposal area is only slightly above
ground level.
The boney pile is also of irregular shape and covers an area of
about 11.1 acres. It, however, extends well above the ground with an average
height of 60 feet.
The largest source in both land area and expected airborne coal
particles is the coal storage pile. It occupies an area of approximately
20 acres and is of rectangular configuration. The coal pile towers well
above the ground averaging about 100 feet throughout its length.
The coal cleaning plant is a ten-story structure which will soon
become completely enclosed. In the future, after the structure has been
enclosed with sheet metal, it is expected that the rate of emissions that
are common to older coal cleaning plants will be reduced.
Other fugitive sources near the coal cleaning plant operations
include the dust generated by both vehicular traffic and construction activi-
ties. Construction has been ongoing for a number of years and is projected
to remain so for at least two more years. There are also two coal storage
silos which deposit coal carried by conveyor belts onto the coal pile. A
significant amount of coal particles are released during this operation.
The conveyor belts were not identified as a principal source of fugitive dust
since they are almost fully enclosed, However, coal dust does emit from the
bottoms of these conveyor belts.
-72-
-------
Description of Fugitive Dust Sampling Campaigns
The fugitive dust sampling study consisted of three campaigns;
namely Campaign I, December 17 to 19, 1976j Campaign II, January 5 to 7,
1977; and Campaign III, April 5 to 7, 1977. All three campaigns had the
same sampling schedule which consisted of 6 samplers that were operated for
a twelve-hour period beginning at 2000 hours and ending at 0800 hours, and
4 samplers that were operated for a 24-hour period beginning at 2000 hours
and ending at 2000 hours. In addition to the 10 hi-volume samplers, a
weather station was operated on site which recorded wind speed and wind
direction during each campaign, and an Andersen sampler was operated during
Campaign III. Three sites were supplied with electrical power by portable
gasoline-fueled generators.
Sites were maintained by Battelle-Columbus and personnel provided
by the power company. Operators maintained records of equipment malfunctions,
such as generator failure, power failure, and charts failing to ink properly.
All samples were returned to Battelle-Columbus for analysis. Lab
work included the weighing of hi-volume filter and microscopic analysis of
the filter pads, and reduction of the meteorological data. After weighing
and microscopic examination of each filter, selected hi-vol filters were
forwarded to an independent lab for chemical analysis. Standard analytical
procedures were used for all analysis.
Sampler Results
Presented below is a discussion and tabulation of analytical
loading values and/or physical and chemical properties of the captured
particulates.
-73-
-------
Mass Concentrations
The general objective of this effort is to obtain quantitative and
qualitative data on particulate matter generated by various fugitive dust
sources.
Another objective was to determine if the fugitive dust was leaving
the property of the coal cleaning plant (under construction) and impacting
on the surrounding environment. Preliminary data indicate that fugitive
dust has little impact on the surrounding environment.
All of the samples were collected on 8" x 10" fiberglass, hi volume
filters, for either a 12-hour or 24-hour-period, Filters with minimum
amounts of trace metals were used in Campaign III, The filters were weighed
before and after each sampling to determine the mass of the particles
collected. The weighing of samples was performed by Battelle-Columbus
personnel.
Generally, the maximum concentrations were observed at Sites 1, 2,
and 3. These sites are located adjacent to the coal storage pile; also the
maximum concentrations occurred during the 0800-2000 hours sampling periods.
During the first campaign (Friday, December 17 - Sunday, December 19, 1976)
sampling was conducted on the weekend to reduce the impact of fugitive
emissions from construction activities. Table 1 presents the results of mass
concentrations found at each sampling site for each time frame. The average
o
(arithmetic) concentrations for the 12-hour samples was 114,0 yg/m , while
2
the average for 24-hour samples was 52 yg/m . The maximum and minimum con-
o
centrations for the 12-hour samples were 564.0 and 30,0 yg/m , respectively,
while the maximum and minimum concentrations for the 24-hour samples were
3 3
11.0 yg/m and 28.0 yg/m , respectively.
-74-
-------
TABLE 1. EXPERIMENTAL PARTICULATE CONCENTRATIONS
NEAR COAL CLEANING PLANT OBTAINED BY HI
VOLUME SAMPLER
Sampler
Location
Date
Time
Suspended
Particulate Concentration
12-Hour Samples
17-18 Dec 76
18 Dec 76
18-19 Dec 76
19 Dec 76
5-6 Jan 77
6 Jan 77
6-7 Jan 77
7 Jan 77
5-6 Apr 77
6 Apr 77
6-7 Apr 77
7 Apr 77
17-18 Dec 76
18 Dec 76
18-19 Dec 76
19 Dec 76
5-6 Jan 77
6 Jan 77
6-7 Jan 77
7 Jan 77
5-6 Apr 77
6 Apr 77
6-7 Apr 77
7 Apr 77
17-18 Dec 76
18 Dec 76
18-19 Dec 76
19 Dec 76
5-6 Jan 77
6 Jan 77
6-7 Jan 77
7 Jan 77
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
78.0
Invalid data
Invalid data
Invalid data
54.0
136.0
60.0
56.0
149.0
124.0
252.0
398.0
107.0
69.0
60.0
306.0
59.0
230.0
63.0
294.0
79.0
168.0
209.0
291.0
75.0
87.0
72.0
70.0
82.0
189.0
58.0
92.0
-75-
-------
TABLE 1 . (Continued)
Sampler
Location Date
5-6 Apr 77
6 Apr 77
6-7 Apr 77
7 Apr 77
4 17-18 Dec 76
18 Dec 76
18-19 Dec 76
19 Dec 76
5-6 Jan 77
6 Jan 77
6-7 Jan 77
7 Jan 77
5-6 Apr 77
6 Apr 77
6-7 Apr 77
7 Apr 77
5 17-18 Dec 76
18 Dec 76
18-19 Dec 76
19 Dec 76
5-6 Jan 77
6 Jan 77
6-7 Jan 77
7 Jan 77
5-6 Apr 77
6 Apr 77
6-7 Apr 77
7 Apr 77
8 17-18 Dec 76
18 Dec 76
18-19 Dec 76
19 Dec 76
5-6 Jan 77
6 Jan 77
6-7 Jan 77
7 Jan 77
Time
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
2000-0800
0800-2000
Suspended 3
Particulate Concentration (ug/n» )
164,0
149,0
220,0
564.0
30,0
48.0
75.0
100.0
75.0
59.0
52,0
50,0
106.0
46.0
143.0
145.0
31.0
47,0
71,0
45.0
106.0
86.0
43.0
46.0
46.0
49.0
87.0
69.0
39,0
58,0
63,0
59.0
76.0
115.0
62.0
45,0
-76-
-------
TABLE 1. (Continued)
Sampler
Location
Date
Time
Suspended ,
Particulate Concentration (ug/m )
5-6 Apr 77
6 Apr 77
6-7 Apr 77
7 Apr 77
2000-0800
0800-2000
2000-0800
0800-2000
85.0
105.0
98.0
191.0
24-Hour Samples
10
17-18 Dec 76
18-19 Dec 76
5-6 Jan 77
6-7 Jan 77
5-6 Apr 77
6-7 Apr 77
17-18 Dec 76
18-19 Dec 76
5-6 Jan 77
6-7 Jan 77
5-6 Apr 77
6-7 Apr 77
17-18 Dec 77
18-19 Dec 77
5-6 Jan 77
6-7 Jan 77
5-6 Apr 77
6-7 Apr 77
17-18 Dec 76
18-19 Dec 76
5-6 Jan 77
6-7 Jan 77
. 5-6 Apr 77
6-7 Apr 77
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
2000-2000
33.0
47.0
76,0
39.0
28.0
38.0
64.0
88.0
110.0
48,0
54,0
67.0
33.0
Invalid data
45,0
41,0
39.0
42,0
33,0
57,0
82,0
41,0
37,0
57.0
-77-
-------
There was one valid measurement conducted with the particle sizing
sampler. These data are presented in Table 2.
Microscopic Analysis
The objective of the microscopic analysis was to provide a distinc-
tion between the various fugitive dusts such as pollen, construction dust,
coal dust, and coal ash dust.
Each pad was examined under a low-power stereo microscope to see
if the particulate matter was evenly distributed and if the pads were smeared
during handling. Once these conditions were ascertained, a small piece,
approximately 1/2 inch square, was cut from the pad and further analyses were
conducted on these small pieces. The small pieces of filter pad were placed
on a glass slide and a drop of lens immersion oil (index refraction = 1,5150)
and a cover slip placed over the filter. The immersion oil matted the glass
fibers in the pad and made them transparent.
Optical examinations were made using two instruments: (1) a Leitz
microscope using transmitted light, and (2) a Zeiss metallograph using
reflected light either with dark field or polarized light illumination. All
optical examinations were conducted at magnifications of approximately 500
diameters.
The estimates of coal and ash were in 10% intervals, i.e., 0-10,
10-20, 20-30, etc. No attempt was made to estimate the percentage of the
tertiary category, although morphological and dark field/polarized light
observations were made»
-78-
-------
TABLE 2.
Stage
1
2
3
4
Backup Filter
DATA PRESENTATION OF
SIZING SAMPLER^
Percent Loading
In Size Range
33.3
13.2
5.9
3.4
44.2
PARTICLE
Size Range
(microns)
7.0 and above
3.3 - 7.0
2.0 - 3.3
1.1 - 2.0
0 - 1.1
(a) Sampler was located at Site #3.
-79-
-------
Chemical Analysis
The chemical analysis of hi-volume filters was designed for coal
or coal ash which we assumed to be a significant portion of the particulate,
particularly the 12-hour samples. Data from the microscopic analysis were
used to determine which filters would be analyzed for trace elements. Gener-
ally, the filter that exhibited the highest percentage of coal or ash from
each site was selected for chemical analysis. The following parameters were
analyzed on each candidate hi volume filter.
Antimony Mercury
Arsenic Nickel
Beryllium Titanium
Cadmium Vanadium
Chromium Zinc
Copper Total Organic Carbon
Iron Fluoride
Lead Chloride
Manganese
Beryllium, vanadium, and antimony were not observed at any of the
sites and titanium was observed at two locations (Sites 1 and 2).
In order to provide a general indication of the trace metal levels
encountered, brief summaries have been prepared. The reader is advised
that these presentations are insufficient to establish a sufficient data
base. Table 3 lists the average values of 12-hour samples and 24-hour
samples as well as minimum and maximum levels for each time frame. Table 4
lists the average values of Campaign I and II for selected sites. Results
of the analysis of Campaign III is not available at this time. These sites
-80-
-------
TABLE 3. COMPARISON OF TRACE ELEMENT CONCENTRATION BETWEEN
12-HOUR SAMPLES AND 24-HOUR SAMPLES^
i
oo
3
Concentrations, yg/m
Average
Arithmetic
Maximum
Minimum
Average
Arithmetic
Maximum
Minimum
As Cd Cr Cu Fe
12-Hour
.015 .015 .031 .194 2.73
.053 .039 .064 .484 8.09
.005 .001 .016 .011 1.04
24-Hour
.013 .009 .024 .119 1.84
.021 .020 .022 .230 2.34
.007 .003 .013 .112 1.40
Pb Mn
Samples
.550 .036
1.50 .093
.230 .031
Samples
.377 .029
.488 .063
.268 .026
Hg Ni Ti Zn
.00017 .015 .86a .25
.00041 .050 1.18 .43
.00003 .006 .58 .08
.00005 .010 - .13
.00008 .018 - .17
.00002 .001 - .06
Cl F
1.59 2.71
3.40 7.48
.82 .39
.70 1.22
1.30 2.40
.33 .06
(a) The average of two observed values,
-------
TABLE A. COMPARISON OF TRACE ELEMENT CONCENTRATION OF
SELECTED SITES(a)
I
00
i-o
3
Concentration. ue/n
Sampler Location
l
3
4
8
10
6
As
.021
.016
.008
.014
.017
.013
Cd
.006
.014
.008
.023
.012
.007
Cr
.064
.032
.016
.018
.017
.015
Cu
.243
.402
.048
.045
.200
.179
Fe
2.56
2.81
1.47
1.62
1.97
1.66
Pb
.598
.893
.283
.405
.437
.310
Mn
.071
.078
.029
.033
.034
.042
Hg
.00041
.00019
.00009
.00013
.00007
.00004
Ni Ti V
.007 1.18 -
.019
.018
.004
.006
.010
Zn
.41
.37
.14
.21
.15
.12
Cl
.95
2.31
.98
2.35
.29
.26
F
7.48
3.47
1.39
2.35
.67
.88
(a) These values listed are the averages of Campaigns I and II.
(b) These values are for Campaign II only.
-------
are generally in a line from the coal pile at distances from 175 meters to
approximately 2200 meters. Site No. 6 is off property and generally not
affected by the emissions from the coal pile or power plant stacks and
therefore serves as a background or a control site.
Trace element values vary widely among mass concentration ranges;
f\
i.e., a filter that weighed between 251 and 300 yg/m had a reported value
3 3
of .007 yg/m of As and .001 yg/m of Cd, whereas a filter that weighed
3 3 ^
between 0 and 50 yg/m had reported values of .007 yg/m of As and .030 yg/m
of Cd. These data indicate that filters exhibiting the highest mass concen-
trations may not exhibit the highest trace element concentrations, Table 5
illustrates how reported values for trace elements vary with mass concen-
trations and sampling time (12-hour/24-hour).
Weather Conditions Observed
During Each Campaign
What follows is a chronological review of the weather conditions
associated with each of the three experimental air monitoring campaigns.
In addition, hourly averages of wind directions and wind speeds are tabu-
lated in Tables 6-8 for each campaign. The wind speeds and wind direction
shown in Tables 6-8 were determined by averaging data on weather strip charts
taken near the site of the large coal storage pile. It became increasingly
difficult to average a wind direction which fluctuated between all points of
the compass. This type of fluctuation occurred primarily with calm wind
conditions during two sampling nights, one in each of the first two monitoring
campaigns. Fortunately, the fugitive dust loading was minimal during this
environmental set.
-83-
-------
TABLE 5. TRACE ELEMENT CONCENTRATIONS AS THEY RELATE TO MASS CONCENTRATION RANGES
12-Hour Samples
Ug/m 3 AS
0-50 .007
.005
51-100 .017
.015
.012
.015
.011
.008
101-150 .006
1 .021
oo
-c-
' 151-200 .024
201-250 .053
251-300 .007
24-Hour Samples
0-50 .011
.006
.008
51-100 . 015
.007
.013
.022
Cd
.030
.003
.002
.039
.007
.021
.012
.023
_
.006
.005
.005
.001
.010
.013
-
.003
.012
.020
.003
Cr
.022
.019
.023
.020
.016
.021
.013
.020
.060
.064
.043
.045
.041
.014
.022
.016
.016
.020
.013
.020
Concentrations , ug/tn
Cu Fe Pb Mn
.089
.011
.163
.050
.040
.230
.084
.320
.130
.243
.484
.437
.246
.180
.200
.112
.178
.230
.230
.163
2.20
1.24
2.79
2.20
1.04
2.40
1.70
2.50
1.80
2.56
3.13
8.09
3.88
1.40
1.90
1.41
1.92
1.80
1.90
2.04
.320
.285
.500
.810
.351
.730
.280
1.50
.r~o
.598
.585
.614
.369
.320
..420
.268
.310
.340
.500
.373
.046
.046
.052
.034
.031
.046
.029
.055
.060
.071
.101
.093
.058
.028
.060
.026
.056
.055
.026
.041
Hg
.00014
.0003
.00015
.00012
.00013
.00014
.00017
.00025
.00016
.00041
.00013
.00023
.00012
.00002
.00007
.00003
.00005
.00002
.00008
.00006
Ni Ti
.050
.006
-
.007
-
.027
.028
.020
.007 1.18
.014
.021 .58
.010
.013
.018
.003
.007
.001
.010
.002
Zn
.29
.26
.10
.20
.22
.29
.20
.31
.25
.41
.43
.41
.11
.09
.17
.06
.09
.14
.17
.13
Cl
2.0
2.80
3.27
1.10
.36
.20
.82
3.40
1.50
.95
1.21
.93
1.23
1.92
.90
.58
1.92
1.30
.58
-
F
2.0
2.50
.39
2.20
2.49
.20
1.50
3.30
2.5
7.48
3.63
1.58
3.41
1.69
1.4
1.75
1.69
2.4
.94
.40
101-150
.021
.003
.022
.137 2.34
.488 .063 .00006 .011
.15
.42 1.12
-------
TABLE 6. WIND DIRECTION-WIND SPEED DATA, CAMPAIGN I
(2000 December 17 - 2000 December 19, 1976)
Time
*2000-2100Dec.
2100-2200
2200-2300
2300-0000
0000-0100 Dec.
0100-0200
0200-0300
0300-0400
0400-0500
1 0500-0600
£S 0600-0700
0700-0800
*0800-0900Dec.
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
1500-1600
1600-1700
1700-1800
1800-1900
1900-2000
Wind Direction,
degrees
17 285
290
285
290
18 290
285
285
290
315
290 (270-330)
260 (210-300)
295 (270-360)
18 295 (270-330)
295 (270-330)
305 (270-360)
295 (250-315)
295 (250-350)
290 (250-360)
280
285
290
330
80
60
Wind Speed.
ra.p.h.
10
9
8
8.5
8
5
4
4.5
5.5
3.5
3
3
4
7
7
7.5
7
7
8
6,5
3
2
2.5
2.0
Time
*2000-2100Dec. 18
2100-2200
2200-2300
2300-0000
0000-0100 Dec. 19
0100-0200
0200-0300
030C-0400
0400-0500
0500-0600
0600-0700
0700-0800
*OSOO-0900Dec. 19
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
1500-1600
1600-1700
1700-1800
1800-1900
1900-2000
Wind Direction,
degrees
10 (315-90) Variable
310 (225-350) "
70 (0-135)
60 (10-135)
45 "
70 (45-110) "
70
200 "
330 "
250
250 "
250 "
220
225
210
245
240
255
250
235
235
235
240
235
Wind Speed.
m.p.h.
1
1.5
1.5
1
1
1
1.5
1
1.5
1.5
1.5
2
3.5
4
5
10
10
13
10.5
12
9.5
7.5
8
3
Indicates initiation of Hi-Vol sampling period.
-------
TABLE 7. WIND DIRECTION-WIND SPEED DATA, CAMPAIGN II
(2000 January 5 - 2000 January 7, 1977)
i
oo
Time
*2000-2100 Jan. 5
2100-2200
2200-2300
2300-2400
2400-0100
0100-0200
0200-0300
0300-0400
0400-0500
0500-0600
0600-0700
0700-0800
*0800-0900 June 6
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500 '
1500-1600
1600-1700
1700-1800
1800-1900
1900-2000
Wind Direction,
degrees
50
315
165
150
80
135
140
85
60
50
35
225
230
242
235
230
240
230
237
225
225
225
225
220
Wind Speed,
m.p.h.
2.0
3.0
4.0
2.0
1.5
1.0
3.0
2.0
2.5
3.5
2.0
2.5
2.0
2.0
2.5
4.0
4.0
5.0
7.0
10.0
8.5
7.5
8.5
7.5
Time
*2000-2100 Jan. 6
2100-2200
2200-2300
2300-2400
2400-0100
0100-0200
0200-0300
0300-0400
0400-0500
0500-0600
0600-0700
0700-0800
0800-0900 June 7
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
1500-1600
1600-1700
1700-1800
1800-1900
1900-2000
Wind Direction,
degrees
215
215
230
225
235
235
255
275
290
270
295
325
305
310
290
295
315
310
315
290
290
280
265
260
Wind Speed,
m.p.h.
6.5
4.5
3.0
3.5
2.0
2.0
2.5
3.5
5.5
7.0
5.5
8.0
8.5
11.5
12.5
13.0
14.0
14.5
14.5
14.0
12.0
9.0
7.5
5.0
* Indicates initiation of Hi-Vol sampling period.
-------
TABLE 8. WIND DIRECTION-WIND SPEED DATA, CAMPAIGN III
(2000 April 5 - 2000 April 7, 1977)
CO
•-J
I
Time
2000-2100 April 5
2100-2200
2200-2300
2300-0000
0000-0100 April 6
0100-0200
0200-0300
0300-0400
0400-0500
0500-0600
0600-0700
0700-0800
*0800-0900 April 6
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
1500-1600
1600-1700
1700-1800
1SOO-1900
1900-2000
Wind Direction,
degrees
260
265
270
285
275
265
265
270
265
270
270
270
265
275
275
295
270
275
288
298
330
310
305
310
Wind Speed,
m.p.h.
16.0
17.0
18.0
9.0
11.5
9.5
18.5
15.0
14.0
18.5
19.0
17.5
22.5
17.0
16,0
13.5
19.0
14.0
17.0
12.5
15.0
15.0
12.0
11.5
Time
2000-2100 April 6
2100-2200
2200-2300
2300-0000
0000-0100 April 7
0100-0200
0200-0300
0300-0400
0400-0500
0500-0600
0600-0700
0700-0800
*0800-0900 April 7
0900-1000
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
1500-1600
1600-1700
1700-1800
1800-1900
1900-2000
Wind Direction,
degrees
300
280
275
270
260
260
260
260
225
210
225
225
230
235
220
240
250
230
230
235
230
225
215
140
Wind Speed,
m.p.h.
6.5
8.0
4.5
8.0
6.5
6.0
7.0
6.0
3.5
4.0
3.5
8.0
12.0
14.5
16.5
15.5
13.5
14.0
16.0
16.0
9.0
9.5
4.0
3.5
* Indicates initiation of Hi-Vol sampling period.
-------
Campaign I
2000 December 17. 1976 -
2000 December 19. 1976
December 17, 1976 - Friday. Light snow fell throughout the day
accumulating between 1-2 inches by nightfall. Cloudy skies prevailed
throughout most of the night; temperatures were well below freezing
throughout the day and night.
December 18, 1976 - Saturday. By dawn, skies had cleared and some
warming occurred throughout the day. The snow that fell the previous day
melted by noon and by mid-afternoon much of the surface moisture had evapor-
ated. Skies remained clear throughout the day and temperatures rose to about
32°F. Temperatures fell to the mid twenties during the night. Wind speeds
were especially low due to the influences of a high pressure area situated
100 miles south of the power plant and also to a nighttime temperature
inversion which developed from radiational cooling.
December 19, 1976 - Sunday. Warm air advection began before
sunrise and continued throughout the day causing an increase in wind speed
and temperature. Skies were mostly sunny. Temperatures rose into the 50's
by late afternoon. A good deal of soil drying occurred during the past two
days. The first campaign ended at 8:00 PM Sunday evening.
Campaign II
2000 January 5 -
2000 January 7, 1977
January 5, 1977 - Wednesday. Variable cloudiness and cold temper-
atures were the agenda for the day. Snow depth of about six inches covered
most of the study area. However there had been some earlier drifting of
snow thus leaving some areas exposed, especially within the grounds of the
-88-
-------
power plant. Temperatures rose to a tropical 15°. Skies were partially
overcast during the night.
January 6, 1977 - Thursday. It continued very cold today and
skies became increasingly overcast with a threat of precipitation appearing '
in the west horizon by darkness. Temperatures, however, rose to about freezing.
Snow began at 7:00 PM and persisted through the night.
January 7, 1977 - Friday. Snow continued to fall steadily to about
9:00 AM. About five inches had fallen during the night, A period of snow
squalls occurred during the late morning and early afternoon hours. Winds
were very strong, gusting to about 40 mph throughout the daylight hours and
causing blowing ?,nd drafting of snow. Temperatures dropped to 15° by evening.
Campaign II came to an end at 8:00 PM.
Campaign III
2000 April 5 -
2000 April 7. 1977
April 5, 1977 - Tuesday. Rain which fell during the previous two
days created saturated soil conditions. Soil evaporation occurred this day
associated with partly cloudy skies through late afternoon. Temperatures
were unseasonably cold reaching to only 44°F. Showers changing to snow
flurries began in the late afternoon. Snow flurries and snow squalls
continued through the night.
April 6, 1977 - Wednesday. A snow blanket of about one inch
covered the ground during the morning hours. Heavy snow squalls lasting
5-30 minutes accompanied by strong gusty winds persisted throughout the day.
Intervals of sunshine melted any accumulation of snow resulting from the
snow squall. Flurries continued into the night.
-89-
-------
April 7, 1977 - Thursday. Clear skies gave way to cloudy
conditions by late morning. Temperatures were slightly above freezing.
Showers began around noon and persisted to the early evening hours. Winds
continued to be gusty as it had been throughout the monitoring period.
Campaign III ended at 8:00 PM this evening.
A Comparison of Meteorological
Conditions and Hi-Vol Readings
Shown in Table 4 are the mass weight concentrations of each hi-vol
site for each monitoring campaign. Notice that some of the sites (Nos. 1,
2, 3, 4, 5, and 8) collected 12-hour samples while the other sites (Nos. 6,
7, 9, and 10) collected 24-hour samples. The 12-hour samples were intended
to show diurnal variations in particle loading. Displayed in Tables 1-3
are the hourly wind directions and wind speed averages for the corresponding
time periods. Using the observed fugitive dust concentrations, the hourly
wind direction and speed averages and the general weather and soil conditions,
a qualitative comparison was made to identify the relationships between
existing weather conditions and mass concentration.
Campaign I - 2000 December 17 -
2000 December 19, 1976
2000 December 17 - 0800 December 18, 1976. The first campaign
was conducted over a weekend during off-peak hours when both loading and
construction activities were low. The direction of the wind during the first
12-hour period fluctuated between 260° (WSW) and 330° (NW). The dominant
wind direction (290° - WNW) and an average wind speed of 6.0 mph were
inducive in causing the greatest loading at Site No. 2 (substation). The
substation was directly in line for the largest amount of deposition from
-90-
-------
the coal pile. The remaining sites were either further downwind from the
sources or were significantly displaced from a direct impact of fallout.
0800 December 18 - 2000 December 18. 1976 (24-Hour Stations). Since
the winds were primarily from the W and NW during the first 24 hours of this
campaign, only the 24-hour sampling station, No. 7, southeast of the power
plant showed significant concentration (64 ug/m^) of fugitive dust (Nos. 6
and 9).
0800 December 18 - 2000 December 18, 1976 (12-Hour Stations). Drier
soil conditions lead to higher 12-hour readings during the 0800-2000 hour
o
period. Site No. 3 collected the largest amount, 87 yg/m . Winds fluctuated
with a greater frequency than the previous 12 hours blowing more from the NW.
Thus, except for Site No. 3, all of the hi-vols. received a minimal impact of
fugitive dust.
2000 December 18 - 0800 December 19, 1976. Weather conditions were
extremely calm. The wind was light and variable during the night, thus
resulting in similar mass readings at all the 12-hour hi-vol. stations. The
3
range of fugitive dust concentration was from 60 yg/m at Site No. 2 near
2
the coal pile to 75 yg/m at Site No. 4.
0800 December 19 - 2000 December 19, 1976 (24-Hour Stations). In
the absence of any precipitation and with an increase in wind speeds to an
average of 12 mph during the second 12 hours, surface conditions continued
to dry. This lead to higher readings at all 24-hour stations (Nos. 6, 7,
and 10).
0800 December 19 - 2000 December 19, 1976 (12-Hour Stations). The
soil continued to dry aided by a noticeable increase in wind speeds blowing
-91-
-------
from the southwest (235°). As in the first 12 hours of this campaign, Site
o
No. 2 collected the greatest amount (306 yg/m ). The wind direction for both
the first and this last 12 hours of Campaign I was from the SW. Wind speeds
were greater this 12 hours, therefore causing a larger concentration. The
coal pile a few hundred feet to the southwest of Site No. 2 is the primary
source of coal dust at Site No. 2. Considering its proximity to the ash
disposal area, a rather small amount (45 yg/m3) was collected at Site No. 5.
This may simply be due to the fact that the ash's content is higher in mois-
ture and therefore less disturbed by the action of the wind. Site No. 4
received a mass concentration of 100 yg/m3.
Campaign II - 2000 January 5 -
2000 January 7, 1977
Stable atmospheric conditions resulted in light winds and variable
wind directions. Under these conditions it becomes difficult to assess the
source or sources responsible for contributing dust to the hi-vol. samplers.
o
The highest reading (106 yg/m ) was recorded at Site No. 5 located northwest
of the power plant.
0800 January 6 - 2000 January 6. 1977 (24-Hour Stations). The
same weather conditions that occurred in Campaign I (calm night followed by
increase in wind speeds from the SW the following day) resulted in a rather
high particulate concentration at Site No. 7 (110 yg/m3). The principal
source of fugitive dust may be due to general construction activities in the
area and to a parking lot located SW of Site No. 7. Since weather conditions
were quite calm during the first 12 hours, Sites Nos. 6 and 10 may have
collected more than a background reading.
-92-
-------
0800 January 6 - 2000 January 6, 1977 (12-Hour Stations). As in
the previous campaign, a southwest wind again led to high readings at Sites
No. 1 (136 yg/m3), No. 2 (230 yg/m3), No. 3 (189 yg/m3), and No. 8 (115 yg/m3).
The coal pile has a direct effect on Site No. 2 when the wind is from the WSW.
General dusty conditions around Site No. 2 could also be adding a rather
large amount to Site No. 2. Site No. 1 is directly downwind of the coal pile
under a SW wind and therefore is the likely source of coal dust. The construc-
tion activity at the coal cleaning facility and the dusty surface conditions
in that area appear to be the major sources of fugitive dust to Site No. 3.
The relatively high reading at Site No. 8, which is well removed from the
center of the power plant operation to the SE, would be indicative of a
transport of fugitive dust from the power plant area.
2000 January 6 - 0800 January 7, 1977. Snow fell throughout the
entire 12-hour period. Thus all the readings were as expected, low ranging
•3
from 43 to 63 yg/m .
0800 January 7 - 2000 January 7, 1977 (24-Hour Stations). Readings
at all of these sites were low as expected due to the accumulation of four
inches of snow during the previous night.
0800 January 7 - 2000 January 7, 1977 (12-Hour Stations). Although
the winds were frequently gusty from the NW, the readings were low at Sites
Nos. 1, 4, 5, and 8. These sites were not downwind of any major source.
o
However, hi-volume sampler No. 2 recorded a mass concentration of 294 yg/m .
The coal pile is the obvious source of fugitive dust. The gusty winds had
exposed large portions of coal that had snow accumulated on top of it the
previous night.
-93-
-------
Campaign III - 2000 April 5
2000 April 7, 1977
Fugitive dust concentrations throughout the entirety of this cam-
paign were high at hi-vol. Nos. 1, 2, 3, 4, and 8. Winds from the SW, W,
and WNW averaged 12-20 mph with gusts up to 50 mph. The prevailing wind
direction was at 270°, West. The mass concentrations result again illustrates
that a southwest-west wind produces high concentrations at all the monitoring
sites (Sites Nos. 1, 2, 3, 4, 5, and 8) located in and around the plant
operations. The data collected at Site No. 5 again indicated that the ash
disposal area is not a major source of fugitive dust during winter and spring
weather conditions. It may become one where an extensive period of dryness
occurs. All of the 24-hour samples including No. 7 collected small amounts
of fugitive dust. Site No. 6 received small amounts of fugitive dust through-
out this campaign and acts as a background sampler when the wind is from the
NW, W and SW. Data (concentrations low) collected at Sites Nos. 9 and 10
showed that the dust (coal, ash, etc.) airborne from the plant operations
impacts before reaching these sites.
-94-
-------
DIFFUSION MODELING
Description of Model and Modeling Approach
Battelle's Fugitive Dust Emission Model is based on U.S. EPA's
Multiple Point Source Model (PTMTP).2 It predicts hourly mass concentrations
of fugitive dust at any number of downwind distances. Several major alter-
ations were made to EPA's PTMTP.
First, sources of fugitive dust were treated as area sources due
to the dimensions of the area occupied by these types of sources. For
example, the dimensions of the coal pile are 1,36 km x .6 km or 21.5 acres.
To consider each source as an area source, the concept of a virtual point
3
source was programmed in the model. This assumes that the area sources are
squared and required the user to input the length of one side. In the study
at the power plant, each fugitive dust source was subdivided into smaller
area sources approximately square in configuration.
To mathematically permit the deposition of fugitive dust particles,
a simple particle deposition function was incorporated in the model. The
amount of material deposited on a receptor location is a function of wind
speed, emission rate, particle size, and distance. As wind speed increases
more dust becomes entrained in the surface circulation (i.e., emission rate
increases) thus creating a greater likelihood for a large amount of dust
deposition. The smaller the size of the particles released into the wind,
/
the further downwind they will be transported before settling to the ground.
The range of distances where the maximum amount of fallout will occur is
obviously dependent on the wind speed and particle size.
The particle deposition function (Vx/u) was programmed in the
-95-
-------
model as a distance. V, the settling velocity, was assigned a velocity
which reflected a range of particle sizes. In the modeling work at the
power plant, it was assumed that most of the particles released from fugitive
sources were greater than 30 microns in diameter. The deposition function
was applied to all receptor distances and the resulting value was assigned
to receptor height. This approach effectively increased receptor heights
at all downwind distances for a flat terrain. By defining a deposition
function in this manner, one does not have to manipulate plume centerline
calculations, which coincidentally is normally not relevant where dealing
with coal storage and coal refuse sources. The height of the local terrain
is then added to the height determined by the deposition function.
A third modification of EPA's PTMTP involved the formulation of an
emission rate rather than a direct input of emissions for each fugitive dust
source. First approximations of emission rates for each of the sources in
the study area were based on an article written by S. L. Vekris, M,Sc., and
4
published in the Ontario Hydro Research quarterly. The study on dispersion
of coal particles from a storage pile was conducted at the Lakeview Gener-
ating Station on the north shore of Lake Ontario. Mr. Vekris found that for
a coal pile covering 40 acres and extending from 0 to 30 meters above the
ground, an emission rate of 50 grams per second was reasonable to assume in
modeling work for particles greater than 30 microns in size.
Subsequently, a constant was calculated which expressed the ratio
of the acreage occupied by each source of fugitive dust at the Appalachian
power station to the acreage of the coal pile at ,the Lakeview Generating
i
Station. The value of this ratio is then multiplied ±rt the emission rate
equation by the square of the wind speed. The amount of dust released
whether it originated from a coal, ash, or terrestrial source, has been
-96-
-------
programmed to be directly proportional to the amount of energy imparted by
the wind or the wind speed squared.
Finally, a constant, K, was formulated in the emission rate
equation which is used after the first computer run in estimating the
emission rate for all of the sources. The constant represents the value of
the slope of the regression line. The general form of the emission rate
equation is
n v 2 Area (acres)
Q = K x u x —77*— 4-
x 40 acres
Additional modifications to EPA's PTMTP included the elimination
of the plume rise routine. It was assumed that none of the sources of
fugitive dust considered in this study imparted any thermal buoyancy to the
airborne particles.
Methodology of Determining Coefficient of Emission
To determine the value of the coefficient of emission, formulated
in the emission model and discussed in a previous section, a linear regres-
sion analysis was performed using the observed data and the model's first
approximations. A linear regression analysis is a statistical method which
correlates one variable to another. It involves the theory of the minimi-
zation of the sum of least squares.
If it is assumed that a straight line approximately describes the
observed data to the models predictions the observed concentration, then
this relationship takes the following form and is commonly referred to as the
line of regression;
Y = a + bx
-97-
-------
where a and b represent the intercept and slope, respectively. In deter-
mining the values of a and b, it becomes necessary to decide which line best
describes the observed data and model first approximations. The principle
of least squares aids in calculating the values of the intercept, a, and the
slope, b. It states that the line of best fit to a series of values is that
line about which the sum of the squares of the deviations (the difference
between the line and the actual values) is a minimum.
The results of the sum of least squares is given in the following
form of the line of regression:
P = 59.5 + 1.47 x
The value of the slope of the line, 1,47, now becomes the value of
the coefficient of emission, K. This value was substituted into the emission
rate equation and the results of Battelle's Fugitive Dust Emissions Model
are shown in Figure 2. A line that best fits the predicted vs observed data
has been drawn.
/ Possible Causes of Scattering About Regression Line
The graph displayed in Figure 2 shows a significant amount of
scatter about the line of regression. Several reasons could be attributed
to this fact.
First, a number of sources were not included in the modeling work
due to the nature of the sources. These include dust generated by construc-
tion activities, dust generated by vehicular traffic and parking lots, and
several storage silos which dump granulated coal onto the coal storage pile.
After careful examination of the observed data and the model's predictions,
it is believed that all of these sources plus the general dusty surfaces of
the plant are contributing in some instances significant amounts of dust.
-98-
-------
175
150
125
100
c
8
I 75
•o
o>
h.
o>
to
-Q
O
50
25
25
50 75 100 125
Predicted Concentration (/j.g/m3)
150
175
FIGURE 2. RESULTS OF LINEAR REGRESSION ANALYSIS
(PREDICTED VS OBSERVED)
-99-
-------
Second, in addition to these sources, stack emissions are also
believed to have made a substantial contribution to several of the monitoring
sites. Tables 9 and 10 show the results of an optical review of the sampling
filters for Campaigns I and II. On numerous occasions there were at least
some deposits of soot on the sampling filters. The soot either was released
from the stacks, was emitted as a product of combustion associated with the
diesel-fired equipment, such as coal trucks, or was re-entrained off the
grounds of the plant and onto the filters. For example, during Campaign II
at Sites Nos. 5, 6, 9, and 10 almost half of the total mass concentration
constituted ash. It is highly unlikely to expect the ash disposal area to
be the principal source of the ash since all of the filters at Sites Nos. 5,
6, 9, and 10 had a moderate to large amount of soot deposits. In particular,
one filter (Site No. 10 - Wed. 2000-2000, 1-5-77) was reported to be sooty.
Thus it is believed that stack emissions, under certain meteorological condi-
tions, contributed a significant amount to the total mass particulate concen-
trations collected at these sites.
Soot deposits were also reported at the sites (Nos. 1, 2, 3, 4, 7,
and 8) located on plant property. Here soot may be originating off the
plant grounds and/or originating from vehicular traffic. Notice too that on
none of the filters was there a significant amount of tertiary deposits,
thus indicating that construction dust, in itself, it not a major source of
fugitive dust at the power plant and that most of the dust on the plant's
grounds originated from coal or ash sources.
In conclusion, there appears to be some impaction of stack emissions
onto several of the monitoring sites, particularly on Sites Nos. 9 and 10,
Data for the stack parameters and stack emissions were not available during
the study period and therefore analysis of stack emissions was not included
-100-
-------
TABLE 9. RESULTS OF OPTICAL EXAMINATION OF FILTER PADS
CAMPAIGN I
Location
Time/Date
Coal
%
Coal Particle Size
(microns)
1, Frl-2000
12-17-76
1, Sat-0800
12-18-76
1, Sat-2000
12-18-76
1, Sun-0800
12-19-76
2, Frl-2000
12-17-76
2, Sat-0800
12-18-76
2, Sat-2000
12-18-76
2, Sun-0800
12-19-76
3, Fri-2000
12-17-76
3, Sat-0800
12-18-76
0-20
0-20
0-20
0-20
0-20
0-20
(low side)
0-20
0-20
0-20
trace
> 1
10-40
>trace < 5-30
>trace < 5-20
>trace < 5-30/40
0-20 >trace 5-30
(high side)
>trace 5-40
< 5-20
< 5-30
trace < 5-60/70
0.5-1
5-40
3, Sat-2000
12-18-76
0-20
5-30
3, Sun-0800
12-19-76
0-20
trace < 5-30
-101-
-------
TABLE 9- (Continued)
Location
Time/Date
4, Fri-2000
12-17-76
4, Sat-0800
12-18-76
4, Sat-2000
12-18-76
4, Sun-0800
12-19-76
Coal
%
0-20
(low side)
0-20
(low side)
0-20
(low side)
0-20
Ash Coal Particle Size
% (microns)
1-2
2-3
1-2
< 1
-..,.. ,
5-20
5-20
5-20
5-40
(high side)
5, Fri-2000
12-17-76
5, Sat-0800
12-18-76
5, Sat-2000
12-18-76
5, Sun-0800
12-19-76
6, Fri-2000
12-17-76
6, Sat-2000
12-18-76
7, Fri-2000
12-17-76
7, Sat-2000
12-18-76
8, Fri-2000
12-17-76
8, Sat-0800
12-18-76
< 1-2
< 1-2
2-3
1-2
1-2
1-2
1-3
0-20
L-M
1
trace
0-20
(low to med)
0-20
(low to med)
0-20
(low to med)
0-20
(low to med)
0-20
L to M
0-20
(med to high)
0-20
M to H
0-20
L-M
0-20
L-M
0-20
L-M
5-10
5-20
5-15
5-15
5-15
5-15
5-20
5-40
5-15
5-15
-102-
-------
TABLE 9. (Continued)
Location
Time/Date
Coal
%
%
Coal Particle Size
(microns)
8, Sat-2000
12-18-76
8, Sun-0800
12-19-76
1-2
< 1
20-40 5-15
(low to med)
0-20 5-15
(high side)
9, Fri-2000
12-17-76 < 1
10, Fri-2000
12-17-76 < 1
10, Sat-2000
12-18-76 > 1
0-20
(high side)
0-20
(high side)
20-40
(low side)
5-15/20
5-20
5-15/20
-103-
-------
TABLE 10. RESULTS OF OPTICAL EXAMINATION OF FILTER PADS
CAMPAIGN II
Location
Time/Date
Coal Ai'li Ti-rttnry Avt:rn|fi? Cniil Size
%* 7. t (Microns) R«n(;e
II-Pump House
Wed. 2000-0800
1-5-77
0-10
Remarks
(Under Polarized Llf.lit/Dnrk Field)
Trnrc <5,20-30
II
Thurs. 0800-2000
1-6-77
11
Thurs. 2000-0800
1-6-77
II
Fri. 0800-2000
1-7-77
12-Suhstation
Ved. 2000-0800
1-5-88
12
Thurs. 0800-2000
1-6-77
*2
Thurs. 200-0800
1-6-77
12
Fri. 0800-2000
1-7-77
03 - On-site
Trailer
Wed. 2000-0800
1-5-77
13
Thurs. 0800-2000
1-6-77
13
Thurs. 2000-0800
1-6-77
13
Fri. 0900-2000
1/7/77
ft, Mobile Van
Wed. 2000-0800
1-5-77
14
Thurs. 0800-2000
1-6-77
14
Thurfi. 2000-0800
1-6-77
14
Fri. 0800-2000
1-7-77
0-10 <1 Trace <5,30-40 (some 100)
0-10 <1 Trace <5.10-20 (some 40450)
0-10 <1 Trace 5-10-30 (some 40&50)
10-20 1-2
Low Side
0-10
Hi Side
>7
0-10
Trace <5,10-30
0-10 <1 Trace
5,20-30 (some larger)
(ones, —100)
0-10 <1 Trace <5-70
0-10 »1 Trace <5-30
0-10 <1 Trace <5-30 (some larger)
(ones, ~60)
0-10 <1 Trace <5-30 (some larger)
Trace <5-30-40
Soot.
0-10 <1 Trace <5,10-20 (few larf-er) Soot.
Trace <5,10-20 (some larger) Soot.
0-10 2-3 Trace <5,10-30 (few larger) Soot.
10-20 <1 Trace <5,10-30 (some larger)
0-10 <1 Trace <5,10-30 (some larger) Soot.
Soot.
Soot.
Soot.
Soot.
Some soot.
Some soot.
Very little soot.
-104-
-------
TABLE 10. (Continued)
Location
T 1 mo /Da I c
15 - Ash
Disposal
Wed. 2000-0800
1-5-77
15
Thurs. 0800-2000
1-6-77
15
Thurn. 2000-0800
1-6-77
*5
Frl. 0800-2000
1-7-77
16 Schtrf Farm **
Wed. 2000-2000
1-5-77
16
Thurs. 2000-2000
1-6-77
17 Unit S3 **
Wed. 2000-2000
1-5-77
17
Thurs. 2000-2000
1-6-77
18 Refuse Coal
Area
Wed. 2000-0800
1-5-77
»8 **
Thurs. 0800-2000
1-6-77
*8
Thurs. 2000-0800
1-6-77
IB
Frl. 0800-2000
1-7-77
09 Recreation,
Area **
Sat. 2000-0800
12-18-76
*9 **
Wed. 2000-2000
1-5-77
*9
Thurs. 2000-2000
1-6-77
»10 Stiles Farm**
Wed. 2000-2000
1-5-77
110
Thurs. 2000
1-5-77
Cofll Anil Tcrt lury
I* % X
0-10 1-2 ' Trace
0-10 <1 Trace
0-10 <1 Trace
0-10 <5 Trace
Low Side
<5
0-10 <1 Trace
Low
<5
0-10 <1 Trace
Low
<5
0-10 1-2 Trace
0-10 «•! Trace
0-10 <1 Trace
Low
<5
0-10 1-2 Trace
0-10 <1 Trace
Low
<5
0-10 1 Trace
0-10 3-4 Trace
Low
<5
0-10 1-2 Trace
Low
<5
0-10 2-3 Trace
Low
<5
0-10 5-10 Trace
Low
<5
0-10 5-10 Trace
Low
<5
AvrrnRf CoaJ Sln«
(Mlrroiis) Kunp.e
<5, 30 (some larger)
<5-30-40
<5-20
<5-20 (come larger)
<5-40 (some larp.er)
<5-20 (few larger)
<5-30
<5-30
<5-10-20
<5-20
<5-30 (few larger)
(ones 70)
<5-30 (few larger)
<5-10-15
<5-20
<5-10
<5-10-15
<5-20
Remarks
(Under Polarized Uc.nt/Dark Fit- Id)
Soot.
Some soot.
Some soot.
Some soot.
Soot.
Soot.
Very sooty.
Soot.
Soot.
Soot,
Soot.
Some soot.
Soot.
Soot.
•
Soot.
Very sooty.
Soot.
* Z Reported in based upon an area/F.I id impact
** Filter forwarded for chemical analysis.
or 2 area of available filter pad.
-105-
-------
in the diffusion modeling. Therefore, a small percentage of the scattering
about the line of regression can be attributed to stack emissions.
A third reason for the scatter about the line of regression can
be attributed to the method used in averaging hourly wind direction and wind
speed data. During a given hour the direction of the wind may and did
fluctuate quite severely. It is difficult to include these fluctuations
when representing the wind directions for a given hour with a single value.
To remedy this, wind direction data will be averaged over fifteen-minute
intervals in future studies.
The model, as expected, is quite sensitive to wind speed. At wind
speeds less than approximately 3 mph the model predicts a very small
amount of dust deposition. However, except for residual dust left in the air
from operations made during windy conditions, one would expect that
contributions of dust during calm wind conditions from fugitive dust sources
would be minimal.
The model was quite effective in predicting concentrations during
windy conditions. As the wind speed increases the model becomes more
successful. A listing of the model's strengths and weaknesses is summarized
below. Some of the model's weaknesses are inherited from the base model
(EPA-PTMTP) and are common to nearly all Gaussian-type diffusion models.
1. The model's capability in predicting dust concen-
trations is sensitive to wind direction. Wind
direction sector averaging is not included in the
model's make-up and therefore it becomes extremely
important to give more consideration to wind
directions that occur with higher wind speeds than
those with lower wind speeds.
-106-
-------
2. The model is very sensitive to wind speed. For low
wind speeds, it predicts extremely low concentrations.
However, when dealing with fugitive dust sources,
one expects the amount of airborne particles to
be insignificant during calm wind conditions.
3. The spread of the plume of dust vertically and
horizontally from a storage pile may not in fact
be reasonably described by the Gaussian distribution.
4. The model is quite reliable in predicting concen-
trations associated with windy surface conditions.
It's reliability increases with increasing wind speed.
5. The model is better able to preduct long-term averages
(i.e., 24 hours and greater) than short-term averages
(12-hour averages and less).
-107-
-------
CONCLUSIONS
The overall objective of this study was to identify fugitive dust
sources and quantify the particulate emissions from these sources.
The maximum mass concentrations were observed at Sites 1, 2, and 3.
These sites are located at distances from 150 to 200 meters from the coal
pile. The extent of a significant fugitive dust impast was measured as far
downwind as Site No. 8, which was approximately 1200 meters from the coal
pile. Generally, heaviest concentrations occurred during 0800-2000 hour
sampling period. During much of the sampling period in Campaign II the
3
ground was covered with snow, and measurements were recorded over 100 yg/m
and coal particles were observed on the snow by the investigators of this
study. Of the 24-hour sampling stations, a maximum concentration of 110
yg/m was measured at Site No. 7. This site was located approximately 500
meters south upwind from the coal pile and adjacent to an unpaved employee
parking lot, which probably contributed to the mass loading of this site.
There was one valid measurement conducted using the particle sizing sampler.
Approximately 44 percent of the particles were in the 0-1.1 micron range.
The microscropic analysis of the filter pads concluded that the
majority of coal particles were observed in 5 - 70 micron range while the
majority of the ash particles were observed in the 5-20 micron range.
Also, the analysis indicated the coal was the dominant particle on the
majority of the pads.
The chemical analysis indicates that beryllium, vanadium, and
antimony were not observed at any of the monitoring sites and titanium
was observed at two sites (1 and 2).
This study indicates that trace element concentrations are not
-108-
-------
always directly proportional to the mass loading of the filters, i.e., the
maximum observed concentration for nickel was reported on a filter that
•3
weighed between 0-50 yg/m and the maximum concentrations for cadmium and
3
lead were reported on filters that weighed between 51 - 100 yg/m . The
minimum reported value for cadmium was observed on a filter that weighed
o
between 251 - 300 yg/m . Lead appears to have high levels of ambient concen-
trations. This may be attributed to the large amount of vehicular traffic
in the area. The average trace element concentrations were higher at 12-hour
sampling sites with the exception of arsenic, where the average concentration
3
of the 12-hour site was -015 yg/m , while the average concentration of the
24-hour site was .013 yg/m .
To conclude some of the more important observations of weather
conditions and its relationship with observed sampling data, a listing is
given below:
(1) Precipitation in the form of rain or snow occurred
during 4 of the 6 sampling days.
(2) Southwest winds, the prevailing wind direction in this
Appalachian area, occurred with the highest frequency
(0.7 or 70%) during the three monitoring campaigns.
Fugitive dust concentrations at most of the sampling
sites observed with southwest winds were higher
than those associated with any of the other observed
wind conditions.
(3) On two of the six sampling nights, weather conditions
were extremely stable characterized by low wind speeds
(<2 mph) and variable wind direction. Under these
conditions, stack emissions are believed to have
-109-
-------
contributed to total mass concentrations collected
at Sites Nos. 5, 8, 9, and 10. Total mass concen-
trations were similar at the sites during calm
weather conditions.
(4) Low fugitive dust readings were observed "ith northwest
winds at all sampling sites except at Site No. 2.
The coal pile is a major source at Site No. 2 when
the wind is from the NW.
(5) The ash disposal area was not a principal source of
fugitive dust. This is probably due to the moisture
content of the ash in situ.
(6) The coal pile was the primary source of coal dust
collected at sampling sites No. 1 and 2.
(7) Where the x^ind blew from the NW, W, and SW, which
occurred 80% of the time, sampling data collected at
Site No. 6 are of background nature. Except for one
24-hour period, total mass concentrations observed
3
at Site No. 6 ranged from 33-47 yg/m .
(8) The general dusty surfaces of the plant grounds are
a significant source of coal and ash dust collected
at all sites located around plant operations and
especially at Site No. 3. The dust lying on plant
ground is believed to have derived from the coal pile,
mining operations and possibly the stacks.
(9) Vehicular traffic around the plant also appears to
have made a significant contribution to some of the
-110-
-------
sampling sites.
In addition to the above conclusions, below is a summary listing of
a number of reasons for the scattering in the diffusion model's predictions.
(1) A number of sources were not incorporated into the
diffusion modeling due to the nature of the sources.
These included dust generated by vehicular traffic,
parking lots, construction activities, several storage
silos, and especially the general dusty surface of
the plant grounds.
(2) Data for power plant stack emissions were not available
during the study period. Thus, an analysis of stack
emissions was not incorporated in the diffusion modeling.
(3) Averaging wind direction data over a given hour will
cause inconsistencies in the model's prediction during
variable wind conditions. For future studies, wind
direction data will be averaged over fifteen-minute
intervals or perhaps a smaller time interval.
Finally, a list is given below which summarizes the strengths and
weaknesses of Battelle's Fugitive Dust Emission Model;
(1) The model is sensitive to wind speeds. It predicts
essentially zero concentrations from fugitive dust
sources at wind speeds less than 2 mph. As mentioned
earlier, one expects, however, to find an insignificant
amount of coal dust on hi-volume filters during calm
wind conditions.
(2) The model is quite reliable in predicting concentrations
of fugitive dust associated with wind surface conditions.
-Ill-
-------
(3) The model is better able to predict long-term averages
(i.e., 24 hours and greater) and less valuable for
short-term averages (i.e., 12 hours and less).
(4) The applications of the model are best suited for
use in computing fugitive dust concentrations
originating from exposed storage areas (coal piles,
boney piles, debris piles, etc.) associated with
mining and power plant operations.
-112-
-------
BIBLIOGRAPHY
1. Dr. Theodore J. Thomas, David P. Ambrose, Battelle's Fugitive Dust Model,
Battelle's Columbus Laboratories (Internal Files), 1976.
2. Multiple Point Source Model (PTMTP), DBT 51, U.S. Environmental Protection
Agency, TJNAMAP Tape (PB 229-771), Research Triangle Park, N.C., November
29, 1974.
3. Turner, D. Bruce, Workbook of Atmospheric Dispersion Estimates, PHS Publ.
No. 999-AD-26, 1967, 84 pp.
4. S. L. Vekris, M.Sc., "Dispersion of Coal Particles from Storage Piles",
Hydro Research Quarterly, Second Quarter, Vol. 23, No. 2, 1971.
-113-
-------
PARTICULATE SAMPLING TECHNIQUES
FOR A COKE QUENCH TOWER
By
A. H. Laube, John Jeffery
York Research Corporation
One Research Drive
Stamford, Connecticut 06906
and
Carl Edlund
U. S. Environmental Protection Agency
Office of Enforcement
Washington, D.C.
-115-
-------
ABSTRACT:
Although pollutants from quenching coke are emitted from
towers these emissions have frequently been termed fugi-
tive due to their intermittent nature, high turbulence
and moisture content makes conventional sampling
approaches extremely difficult. This presentation
summarizes the results of a study conducted by the
Division of Stationary Source Enforcement and its con-
tractor, York Research Corporation to determine the size
of quench tower particulate emissions and methods used to
sample such emissions.
-116-
-------
In the Spring of 1976 York Research Corporation was award-
ed a Task Order by the Environmental Protection Agency
to determine the magnitude and character of quench tower
emissions with respect to quench water quality.
A common practice in this country's steel industry is the
use of contaminated water to quench incandescent coke.
Typically, 10 to 20 ton loads of hot (2000 + °F) coke are
quenched by 6,000 to 12,000 gallons of water. Each quench
takes about 2 to 3 minutes and produces huge billowing
clouds of steam, water droplets, and air contaminants.
In order to draft these emissions out of the work area,
quenching takes place under towers which are open at the
bottom to admit the coke car. Baffles are often fitted
above the quench car to reduce the amount of large diameter
particulates emitted and amount of make-up water needed,
by returning water not expelled from the tower back to
the sump.
Previous test methodologies to determine the magnitude and
nature of air pollutants from quench towers were often
flawed by sampling difficulties such as the exclusion of
certain size particles or limited in the measurement of
necessary process parameters. Early sampling attempts using
greased plates and petri dishes measured only large
-117-
-------
diameter, heavy aerosols (droplets, solid, or both). Many
pollution control engineers have theorized that significant
amounts of coke plant water contaminants directed to quench
tower sumps for disposal are transformed by the quenching
process into air pollutants which escape the baffles.
Further attempts by state and local agencies to measure
particulate emissions utilizing standard EPA sampling equip-
ment were also limited in their scope and encountered a
number of sampling difficulties. Major problems
occurred because the short, violent, rush of steam did not
allow accurate reading and adjustment of the sampling equip-
ment to maintain isokinetic conditions (the EPA method con-
templates measuring velocities and making adjustments every 3
to 5 minutes—more than the total time of the quench). Also,
droplets in the exhaust stream plugged filters, and made
determinations of the molecular weight of the gases nearly
impossible. In addition, the short duration of the quenches
made capturing the prescribed volume of gases extremely
difficult. The square shape of most quench towers, and the
use of internal partitions probably compounded these problems
by causing uneven flows across the tower cross-section.
A main objective of this study was, therefore, to design and
implement a sampling program which accurately measured parti-
culate emissions. Table 1 lists the parameters, method of
-118-
-------
measurement and locations employed in this test program. This
paper concentrates on the particulate sampling portion of
the test program. In order to measure quench tower emissions
in a manner that incorporated the techniques of 40 CFR, Part
60 (EPA Method 5) and that also gave reproducible results,
major problems in (A) site selection, (B) sampling train
configuration, and (C) sampling techniques had to be overcome.
Table 2 lists the major problems and their possible effect
on the sampling results and our solutions.
A. SITE SELECTION
In order to most accurately determine the impact of
water quality on quench emissions, criteria for
selecting a test site included: tower configuration
which would produce uniform cross-sectional flows and
the ability to quench with a wide range of Total
Dissolved Solids concentrations in the water. After
a general survey of quench towers, a tower at U.S.
Steel's Lorain Works was selected for testing. This
tower was a circular brick chimney which, after the
baffles, did not have internal buttressing. In addi-
tion, the facility was constructed with the ability
to re-route contaminated water to other quench towers
in the plant and use service water for quenching at
-119-
-------
the selected towers.
A schematic of the tower is shown in Figure 1. Tests
were conducted approximately 100 feet above ground
and 20 feet from the top of the tower.
B. SAMPLE TRAIN CONFIGURATION
Emission testing was conducted with a Hi-volume
Aerotherm sample train to minimize the amount of
time required to collect sufficient volume of gas
and weight of particulate matter. (Figure 2)
In order to allow a more accurate determination of
moisture in the stack gas, prevent filter plugging,
and provide a rough indication of aerosol sizing,
a cyclone was fitted on the front of an eleven
foot probe. The cyclone was designed to have a 50%
cut size of about 10 microns at anticipated flow
rates. Particulate and gas were drawn through the
unheated cyclone where large diameter particles and
water droplets were collected. After passing through
the heated (250°F) probe at about 2 cubic feet per
minute and a heated (250°F) Spectrograde type AE
glass fiber filter where particles larger than 0.3
microns were collected, the gas sample passed through
-120-
-------
an impinger train-ice bath assembly where additional
particulate matter and some condensibles were removed.
The gas was then further dried prior to entering the
flow control module by means of an impinger filled
with 300 grams of silica gel.
A continuous strip chart recorder was connected to
a Hastings meter (see Figure 2) to record the velo-
city head pressure (A p) and to aid in verifying any
uneven flow patterns. The unit operated with air
continually purging the pitot lines thus preventing
stack gas from entering the meter. Differential
pressure across the pitot tube was measured as a
change in the pneumatic bridge, formed by the velo-
city transducer, manifold, and pitot tube. As flow
across the tip occurred a differential pressure devel-
oped, imbalanced the bridge and caused a small amount
of purge gas to flow through the transducer. The
transducer measured this flow which was related to
the main gas flow at the tip of the pitot tube.
Purge gas still exhausted through both openings,
but at a slightly unequal rate. The instrument
allowed accurate measurement of flow in the steam
plume even though large diameter solids (coke breeze)
and water droplets were present.
-121-
-------
Preliminary Field Tests
In August 1976, a pretest survey was conducted to determine:
1. If the circular tower selected produced uniform veloci-
ties across the sampling points required by EPA Method 1.
2. If the planned sampling train modification would enable
isokinetic sampling.
3. If emission results would be reproducible.
4. Procedures for the official tests.
Velocity profiles were obtained for at least two quenches at
each of the required 36 points. These preliminary profiles
were analyzed for each one second time interval for magni-
tude of A p. The Ap's of each interval were then added
together and averaged by dividing the time of each quench
into the total. Visual comparison of overlaid profiles
showed that roughly the same pattern, time, and velocity
heads were found for each quench, suggesting that velocity
heads for a particular point in the stack could be pre-
dicted.
Slopes (caused by the heat rise generated by the incandescent
-122-
-------
coke; the rising steam plume and the declining steam plume)
were also analyzed to aid in obtaining sampling rate adjust-
ments. Slopes of the initial increase in the velocity head
(ramp-up) were found to approximate a 45° angle. The declining
slope of the profile never fell to a zero velocity pressure
as the steam plume continued after the coke car exited the
tower.
The three sections of the profile (Figure 3) were defined
as ramp-up - from the time the car entered the tower until
the time water hit the incandescent coke. (Average time 30
seconds). Plateau or middle - from the time the peak velo-
city was reached to the time the water spray ended (average
time 60 seconds). And ramp down - from the time the water
spray ended until the coke car left the tower (average time
45 seconds). From these Ap's, velocity (feet per minute)
calculations were prepared utilizing moisture data obtained
from the previous tests, results from a limited
number of wet bulb-dry bulb runs and, Orsat samples obtained
in accordance with EPA Method 3 with the use of two condenser
units to trap entrained water droplets. In addition to the
velocity traverses, initial measurements were performed
with a Fecheimer probe to determine the maximum angle of
flow in the tower. Data from these findings were limited
but indicated that future traverses and sampling should be
conducted near 110 degrees measured from the horizontal.
-123-
-------
These were later confirmed by further testing conducted with
a vane type device.
Based on this preliminary test information, it was decided to
perform three tests of four quenches each at one point in the
stack. In order to reflect the total quench emissions,
sampling began when the quench car entered the tower and
ended when it left. The purpose of these tests was to deter-
mine how many quenches would be sampled per run in the
official tests and provide some indication of reproducibility
and precision. Process parameters such as water quality,
coke greenness, etc. were not recorded. Laboratory results
from these three tests showed sufficient weights of parti-
culate (over 100 mg)and volumes of dry stack gas (over 15
cubic feet) could be obtained from four quenches instead
of 'the twenty or more quenches which might have been required
by EPA Method 5. Variability of these tests ranged from
approximately 0.7 gr/SCFD to 1.2 gr/SCFD (which was accept-
ably close) and'isokinetic rates were within 10% of 100%.
Since the number of quenches could be reduced from twenty or
more to four and flows did not vary much from point to point,
the original number of tests (three to be performed at 36
points in the tower) was increased to twelve contaminated and
twelve clean. This allowed statistical verification of the
results and the ability to assess the impact of other vari-
-124-
-------
ables (e.g. fugitive emissions, greenness of coke, water
quality changes, wind direction...) which might be hidden by
sampling a large number of quenches or by shifting from
traverse point to traverse point. Based on the preliminary
velocity profiles and the preliminary particulate tests,
points on each axis of three rings, twelve points total, were
selected as being most likely to give mass emission rates
representative of the tower as a whole. See Figure 2.
(C) SAMPLING TECHNIQUE
Data from the preliminary field tests indicated that velocity
head ( A p) was predictable and values for sample rates (AH)
ramp-up, etc. could be determined based on a preliminary
velocity profile taken before each test run. Data from the
velocity profile was then utilized to determine:
1. Absence of erratic flow pattern(s). (If present, the
test was delayed).
2. Velocity pressure (AP)
3. Velocity (feet per minute)
4. Required nozzle size
-125-
-------
5. Predicted sampling rates
Test Sequence
The initial traverse was used to establish the rate (length
of time) of rise of the velocity head ( A p), the maximum
A p, and the rate of decline of A p to an end point greater
than zero. For each quench the sampling rate (AH) was gradu-
ally increased from the signal ear-in to approximately 0.50
inches of water over a thirty (30) second period. This maximum
sampling rate was maintained until the signal water-off
was given and a ramp down procedure was initiated to a pre-
determined rate (approximately 0.26 inches of water)
established from the averaging of the ramp down portion of
the velocity profile. Sampling was terminated when the
car was completely out of the tower.
RESULTS
A sampling program has been designed to avoid or overcome
the hostile environment of a coke quench tower. Although
a full evaluation of the thousands of data elements
collected in the study is not yet complete enough analysis
has been performed to conclude that:
The particulate emission tests were as
-126-
-------
precise as, and correlatable to
Standard EPA measurement techniques.
ACKNOWLEDGEMENT
The authors acknowledge the advice of R. V. Hendricks and
R. M. Statnick of the Environmental Protection Agency. Work
supported by U. S. EPA Office of Enforcement under Contracts
68-01-3161 and 68-01-4138 and Industrial Environmental
Research Laboratory (IERL) under Contract 68-02-1401.
The efforts of York Research's stack testing crew in over-
coming the hostile environment of a quench tower at night
are appreciated.
-127-
-------
REFERENCES
1. Fullerton, R.W., "Impingement Baffles to reduce Emissions
from Coke Quenching", J. Air Pol. Control Assoc.,
r?:807-809, December, 1967.
2. "Air Pollution Control by Coking Plants", United Nations
Report: Economic Commission for Europe, ST/ECE/Coal/
26.1968, pp 3-27.
3. Sidlow, A. F., "Source Test Report, Kaiser Steel Plant,
Fontana, California", San Bernardino County Air
Pollution Control District, February 29, 1972.
4. Memorandum from Robert A. Armbrust, Region IX, New York
State Department of Environmental Conservation, to
Bernard Bloom, Division of Stationary Source Enforcement,
Office of Enforcement, U.S.E.P.A., November 6, 1975.
-128-
-------
PARAMETERS AND MEASUREMENT METHODS
TABLE 1
I
M
ho
I
Parameter
A. Particulate Emission
1. Mass and concentra-
tions (cyclone-sus-
pended and dissolved
solids, water and
acetone wash of
nozzle, cyclone and
probe, filter, and
impinyers).
2. Size distribution of
participates and water
droplets.
3. Fugitive emissions
4. Elemental Analysis
5. Mutagenicity
Number of Tests
and
Number of Quenches
44 contaminated
96-10 u
2
2
96
6. High Molecular weight
Hydrocarbons
Contaminated Clean
1 2
3
Water spray on coke.
Water supply, return 1
and sump.
Field
Measurement Method
EPA Method 5
modified
Cyclone 10 u and
Method 5 train
Sensitized paper
tests
KLD probe
Visible
observation
c
Method 5 train "
Method 5 train
with Battelle
absorber
Method 5 train
Batte]le Absorber
Water Sampled
Water Sampled
Laboratory
Analys is
Method
Method 5
Coulter
Counter and
Chemical
Composi tion
Spark Source
Spectromftry
Level I
Analysis
Levol I
Analysis
Level I
Analysis
Levol T
Analysis
Snmpli ng
Location
(Figure I)
(10)
(10)
(10)
(10)
(10)
(4 -8 f, 12)
(5-8-12 fi
11 (conibi ned)
-------
TABLE 1 (Continued)
U>
o
I
Parameter
Number of Tests
and
Number of Quenches
n.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Gaseous Emission
Contaminated
27
36
Cyanide
Ammonia
Sulfide
Phenol 27
Sulfur Dioxide; 36
Oxyqen (Ornat 40) (GFM-2)
Carbon Dioxide(Orsat
40)
Hydrocarbon GFM-2
Carbon Monoxide "
Total Volatilized
Hydrocarbons (70OF,
29.92 inches of mercury)
C. Stack Flow Information
1. Velocity Head
2. Flow direction
3. Gas temperature
D. Water Volumetric flow and
mass rates
1. Stack - large droplets
vapor and small droplets
2. Spray - on to coke
3. Return - to r.ump
4. Sump - make-up
Clean
36
36
36
36
36
Orsat-10
96
15
40
96
96
96
90
32
Field
Measurement Method
Laboratory
Analysis
Method
EPA Method 6 train Method 6
modified with mois-
ture traps
Orsat fi Grab Flask Mpthod G.C.
tl II It rt 11 11
Grab Flask Method "
Hastings- Rayd.ist moter
Vane-Fecheimer
Six thermocouples positioned
approx. 15 feet apart up the
tower.
Method 5 train
Cyclone
Impinger assembly
Sonic flow meter
Level of sump measurements
Level of sump measurement.!?
Sampling
Location
(10)
(10)
(10)
(8)
(11)
(11)
-------
TABLE 1 (Continued)
Paramoher
Number of Tests
and
Number of Quenches
Field
Measurement Method
Laboratory
Analysis
Me thod
E. Flattery and Tower Processes Contaminated Clean
1. Battery, oven coking time 96
2. Greenness ratings 96
3. Oven heater temperature 96
4. Temperature of Jncandescent 38
quenched-cofce 12
5. Duration and sequence of 96
quench operation
6- Coke Composition
Contaminated
Quenched Incan-
descent
Information from
plant data
Visible observation
Plant data
Infrared thermometer
Thermocouple
Noted on velocity profili
Sampling
Location
(Figure 1)
From pumphoune roof(R)
From
Wharf area
roof(fl)
Clean
Quenched Incan-
descent
fc Nitrogen
% Carbon
% Hydrogen
% Moisture*
% Volatile*
% Sulfur*
Btu/lb.
3
3
3
1
1
3
3
3
3
3
2
2
3
3
3
3
3
-
-
4
4
4
4
4
2
2
4
4
Proximnte-
ASTM 1)3172-73
Samples taken from Ultimate-ASTM oven doors
oven doors(incand.) D3176-74 wharf area
and wharf area
(quenched). Coke
was placed in
metal containers,
sealed and shipped
to Stamford, CT for
analysis.
* Analytical results from plant data are available for 16 coke samples during cloan water tests.
7 . Contaminant
Deposition
on coke
Cyanide
TOC
Dissolved solids
COD
Phenol
Chloride
Calcium
Sulfato
Sulfide
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4 Samples taken from Samples washed Wharf area
4 wharf area. Coke and solutions
4 was placed in metal analyzed by EPA
4 containers.sealed Standard water
4 and shipped to and waste methods
4 Stamford,CT for
4 analysis
-------
TABLE 1 (Continued)
to
NS
Parameter
8. Coal
Volatile
Ash
Free Moisture
Sulfur
Number of Tests
and
Number of Quenches
18
18
18
18
F. Ambient Conditions
1. Wind direction, speed, dewpoint, Each hour
relative humidity, temperature,
and barometric pressure.
G. Water Quality
1. Phenol : 80 contaminated
2. Cyanide 104 - clean
3. Sulfate
4. Sulfide
5. Ammonia
G. Calcium
7. Oils
8. Chloride
9. COD
10. TOC
11. Electrical Conductivity
12. Total Solids
13. Dissolved solids
14. Suspended solids
15. Volatile solids
Field
Meosurement_ Method
Plant data
Plant data
Plant data
Plant data
Plant data
Laboratory
Analysis
Method
Composite- samples
taken from lines
tapped into inlet
header line and
return trough.
Make-up samples
obtained by:(con-
taminated) bucket in
open stream,(clean)
service water in
wharf area.
Two samples taken
each test day.
One sample for 8
quenches.
F.ri\
"Methods for
Chemi c a 1
Analysis of
Water and
Wastes.
Sampling
Location
(Figure 1)
—.•£
Airport l
1.4 miles
from plant
pipe (Fl) '
Return-trouqh (12)
Mako-np-pLpr;
(4) to sump—
cont.imj nated.
(5) service water
in wharf area-clean.
-------
TABLE 1 (Continued)
Parameter^
II. Sump Sediment
1. Conductivity
2. TOC
3. Phenol
4. Total Solids
5. Volatile Solids
6. Sulfate
7. Cyanide
8. Ammonia
9. Calcium
10. COD
11. Oil
12. Chloride
Number of Tests
and
Number of Quenches^
3 samples
Field
Measurement Method
Weighted bucket
dragged along
sump bottom
Laboratory
Analysis
Me thod
Sampling
Location
(Figure 1)
EPA "Methods Sump (13)
for Chemical
Analysis of
Water and Wastes"
OJ
-------
Potential Problems
Small range of TDS
Concentrations
Tower configuration
Short duration of quench result-
ed in velocity profiles changing
faster than could be measured
or recorded by usual methods.
Droplets in stack sample
Droplets and grit in stack
gases
Fugitive emissions
Sampling locations
Number of quenches
POTENTIAL PROBLEMS AND SOLUTIONS
IN TESTING QUENCH TOWER EMISSIONS
TABLE 2
Effect on Accuracy
Desired a wide range (at least
3-5 multiple of concentration
to show effect)
Square, squat, sectionalized
towers are likely to produce
non-uniform flow
Velocities move up and down so
rapidly as to make accuracy
difficult to achieve.
Inability to measure actual
moisture content, molecular
weight of stack gas, filter
plugging.
Plugging of pitot tubes
Emissions exited from entrances
of tower causing non-representa-
tive sampling.
Possible uneven velocity patterns
throughout stack.
Method 5 would require samplinq
20 quenches to get a one hour
test, however, this would have
resulted in plugging and other
errors.
Solution
Ratio of TDS concontrntion
at Lorn in was 10:1 in I.he
same quench tower.
Tall circular tow«?r with no
internal buttressinq was
se]octed.
Use oE Hastingn-Uayelisf.
velocity moter with con-
tinuous recordinq.
Cyclone (50% cut size about
7-13 microns) was fitted
on the front of the probe.
Used Hastings Rayilist velo-
city meter with continuous
purge.
Kntrance emissions monitored
and testing halted whon ex-
cossive (lonqer than one
minute per quench or at
Project Director's discretion).
Performed traverse (if 3f>
points-velocity was not
affocted by position.
Pre-survey tests showed thnlr
four (4) quenches would pro-
vide sufficient weight of
particulate (over 100 mg) and
volumes of dry stack gas (over
I1* cubic ft) using a high
volume sampler.
-------
TAHLE 2 (Continued)
Potential Problems Effect on jVccuracy Solution
Skewed Flow The vertical component of Vane measurements wore made
velocity is different from bo determine probe position-
the velocity of a skewed flow. ing.
Some water flow measurements Mass balance of water flows in Sonic measurements device
were difficult to obtain the quench processes were not installed on header pipe,
because of the physical related directly to each stack sump level measurements
layout of piping and return measurement. performed and results for a
flow ditches. series of quenches were
averaged.
-------
FIGURE 1
QUENCH TOWER
Ox
I
1 INCANDESCENT COKE
2 QUENCHED COKE
3 EXHAUST GASES
4 CONTAMINATED WATER
5 SERVICE WATER
6 HEAD TANK - OVERFLOW
7 HEAD TANK - STAND PIPE
8 NOZZLE HEADER
9 NOZZLE HEADER DRAIN
10 EMISSIONS TESTING STATION
11 SUMP
12 RETURN DRAIN DITCH
13 BAFFLES
-------
OVEN
FILTER
u>
-J
I
CYCLONES FOR PARTICULATE
AND WATER DROPLET SIZING
IMPINGER TRAIN
FINE ADJ. VALVE
MASS TESTS
PERFORMED AT POINTS
2,5, & 8 ON Oo
EACH AXIS 0°
o o
o o
ORIFICE P
MAGNEHELIC
DRY GAS METER
^VACUUM GAUGE
\ COARSE
ADJ.VALVE
VACUUM PUMP
FIGURE 2
HIGH VOLUME (AEROTHERM) PARTICULATE
SAMPLING TRAIN
AND
VELOCITY TRAVERSE AND MASS
SAMPLING LOCATIONS
-------
00
LU
FIGURE 3
VELOCITY PROFILE
10 SECONDS/CM
-------
I
M
OJ
I
FIGURE 4
TOTAL QUENCH TOWER PARTICULATE EMISSIONS
LESS ACETONE WASH AND FILTERABLE CYCLONE SOLIDS
VERSUS
DISSOLVED SOLIDS IN QUENCH WATER
(95% CONFIDENCE LIMITS SHOWN)
LEAST SQUARE EQUATION
Y=0.175x+0.56
QUENCHING WITH "CLEAN" MAKE-UP WATER
QUENCHING WITH CONTAMINATED MAKE-UP WATER
18
20
QUENCH TDS FACTOR *(POUNDS OF TDS PER TON COAL)
*PRODUCT OF TDS CONCENTRATION AND AMOUNT OF WATER EJECTED FROM THE TOWER
-------
o
6.0
•a:
o
5.0
~ 4.0
o;
§ 3.0
Q.
2.0
t/o
oo
1.0
FIGURE 5
TOTAL QUENCH TOWER PARTICULATE EMISSIONS
VERSUS
DISSOLVED SOLIDS IN QUENCH WATER
(95% CONFIDENCE LIMITS SHOWN)
V=0.18x + 1.4
QUENCHING WITH CONTAMINATED MAKE-UP WATER
I
QUENCHING WITH "CLEAN" MAKE-UP WATER
1
0 1.0 2.0
18.0
4.0 6.0 8.0 10.0 12.0 14.0 16.0
QUENCH TDS FACTOR *( POUNDS OF TDS PER TON COAL)
*PRODUCT OF TDS CONCENTRATION AND AMOUNT OF WATER EJECTED FROM THE TOWER
20.0
-------
MONITORING INDUSTRIAL FUGITIVE EMISSIONS
—AN OCCUPATIONAL HEALTH PERSPECTIVE
By
James L. Oser
Ronald J. Young
John M. Dement
Howard R. Ludwig
Industrial Hygiene Section
Industry-Wide Studies Branch
Division of Surveillance, Hazard Evaluations and Field Studies
National Institute for Occupational Safety and Health
-141-
-------
ABSTRACT
Fugitive emissions from industrial processes are a primary monitoring
concern of NIOSH field research investigators. The NIOSH-OSHA
Standards Completion Program (SCP) has tested about 400 and validated
some 200 sampling and analytical methods. In-plant worker exposures
are routinely monitored and evaluated based on personal sampling
(breathing zone samples) which provide 8-hour time weighted average
exposure concentrations. Low volume sampling (0.010 to 2.5 Jlpm)
utilizing filters, and liquid and solid sorbents are used for a variety
of industrial contaminants. Many contaminants are adequately evaluated
3
in the ppb and yg/m range using existing analytical techniques em-
ploying these miniaturized sampling methods. Occupational exposure
data for benzene, PCBs, asbestos, epichlorohydrin, butadiene, and
styrene are presented.
-142-
-------
INTRODUCTION
Industrial fugitive emissions constitute a continuing problem not only
to environmentalists but also to the industrial hygienists as they
constitute exposures to workers within the industrial plant. The use
of the term "fugitive emissions" to convey the concept of pollutant
discharges that enter man's environment from other than well-defined
sources, such as stacks, ducts and exhausts for air pollutants or
pipes, channels, and culverts for water pollutants, can also be
applied to general air contaminants of concern within industrial
plants. This places emphasis on the similarity of concern for these
culprit pollutants resulting from industrial related operations.
Despite mutual concern with the same industrial contaminants, the
overall monitoring objectives of the Environmental Protection Agency
(EPA) and the National Institute for Occupational Safety and Health
(NIOSH) differ. Generally, EPA has as its purpose monitoring emission
rates from industrial sources and assessing the impact on the environ-
ment including man, animal, and plant life; NIOSH has as its purpose
monitoring worker exposures and assessing the impact on his health,
safety, and personal well-being. Over the years the emphasis in
monitoring worker exposures has been placed on fugitive emissions which
are inhaled by workers in the normal performance of duties. The
methods of controlling these exposures include 1) engineering controls -
enclosures, equipment design, ventilation systems, process changes;
2) administrative controls - limiting exposure times, rotation of
employees, work practices, and medical and biological evaluations;
-143-
-------
and 3) personal protective equipment - respirators, protective clothing,
eye, ear and face protectors. The major problems encountered are with
multiple exposures and stresses, synergistic effects, pre-existing
conditions, individual susceptibilities, and off-duty activities.
OCCUPATIONAL HEALTH STANDARDS
The concepts of a safe concentration for exposure to toxic substances
originated in the late 1930"s and were topics of discussion in 1938
at the first annual meeting of the American Conference of Governmental
Industrial Hygienist (ACGIH). The state of Massachusetts published a
list of suggested maximum concentrations for 41 substances in 1940.
Similar lists were set up by a number of states and values were
recommended by the U.S. Public Health Service and the American National
Standards Institute (ANSI). The first formal list of Threshold Limit
Values (TLVs, then called maximum allowable concentrations) was
published in mimeographed form in 1946 at the 8th annual meeting of
ACGIH. Subsequent suggested values and lists were published in the
American Industrial Hygiene Journal commencing in 1950. At present,
TLVs for chemical substances are reviewed annually and published in
booklet form available from ACGIH.
The TLVs established by the ACGIH for 1968 were promulgated as
official Federal standards in 1969 under the Walsh-Healy Public
(2)
Contracts Act. These 400 TLVs constitute the mainstay of health
standards applicable under the Occupational Safety and Health (OSH)
Act of 1970 with the exception of twenty-two health consensus standards
-144-
-------
established by ANSI/ ' Standards for asbestos/4^ vinyl chloride/5^
coke oven emissions/ 14 chemical carcinogens/7^ and benzene^
have subsequently been promulgated by the Occupational Safety and
Health Administration (OSHA).
Since the OSH Act, NIOSH has developed criteria for recommended
standards for over 50 chemical substances including recommended
monitoring and analytical procedures. Under a joint NIOSH and OSHA
effort referred to as the Standards Completion Program (SCP), the
development of accurate methods for compliance monitoring and analysis
(g\
for the OSHA health standards was undertaken in early 1974. To
date, over 200 methods have been tested and validated. These methods
are presently available from the National Technical Information
Service (NTIS) in sets of about 18 substances per set, costing
approximately $6.00 in hardcopy. At final program completion, NIOSH
plans include publication of these procedures.
MONITORING AND ANALYTICAL TECHNIQUES
Over the years, industrial hygienists have adopted personal sampling
methods as the preferred approach for evaluating worker exposure to
airborne contaminants. The use of miniaturized battery operated
pumps which can easily be strapped to the belt of workers and sample
collectors that can be clipped to the shirt lapels has become accepted
practice. Generally, these pumps are operated at about 1.0 to 2.0 £pm
for aerosols in the form of dusts, fumes and mists and 10 to 1000 m£
per min. for gases and vapors. Flow rate through a 37 mm membrane
-145-
-------
filter used to collect particulate samples should probably not operate
at less than 1.5 Jlpm so as not to sacrifice collection efficiency.
However, reliable collection of gases can be effectively accomplished
with solid sorbents and flow rates as low as 10 m£ per minute.
Since the existing Federal standards are based on 8-10 hour working
days, every attempt is made to collect representative samples for a
working shift, usually up to 6 hour samples. Some standards also
have ceiling limits which would permit evaluation based on 15 minute
periods for peak exposures to contaminants. Depending on the con-
taminant of concern, sampling methods have been developed using midget
impingers or bubblers, filter media, solid sorbent materials and in
some instances sampling trains consisting of combinations of the
above.
Midget impingers or bubblers are used as liquid sorbent collectors
and recommended for collection of about 7 percent of the substances
tested under the SCP. Impingers or bubblers using 10 to 15 m£ of
absorbing solution of liquid sorbent are attached to the lapel in
an upright position in a special holder for sampling purposes. Although
they are more difficult for sampling because of restricted wearer
movement, possible spillage, evaporation, and breakge, they are used
for sampling where other methods fail or have not been developed,-such
as for hydrogen sulfide, ozone, hydrogen fluoride and many organo-
phosphate pesticides. For example, ozone is collected by aspirating
a measured volume of air (45 liters recommended) through 10 mSL of
alkaline potassium iodide solution. Hydrogen sulfide is collected
-146-
-------
by drawing a measured volume of air through 10 m£ of an alkaline
suspension of cadmium hydroxide absorbing solution. In this case
the impinger is wrapped in aluminum foil to prevent photodecomposition
of the cadmium sulfide formed during sampling.
Filter media collectors are recommended for 10 percent of the sub-
stances tested under SCP. Several types of filter media are commonly
used for collection of aerosols such as metal dust and fume, fibers and
mists. The membrane filters and glass fiber filters are most common.
They are routinely used for silica, lead, cadmium, beryllium, asbestos,
cotton dust, sulfuric acid, arsenic and the terphenyls with success.
Respirable sampling for dusts is accomplished by using a cyclone
presampler to remove non-respirable dust particles. Gravimetric or
chemical analysis can be performed from filter media samples. For
example, sulfuric acid mist is sampled with a 37 mm, 0.8 micrometer
mixed cellulose ester membrane filter supported by a cellulose backup
pad in a three-piece polystyrene cassette filter holder held together
with a shrinkable band or tape.
Solid sorbent media used in sampling tubes are the most popular new
method in use and under continued development for a variety of sub-
stances encountered in industrial situations. Over 60 percent of the
substances validated under SCP utilize solid sorbent collection media.
Of these, 50 percent employ charcoal sampling tubes. Charcoal has
gained universal acceptance because of its effective sorbent charac-
teristic - essentially 100 percent for most organic solvent vapors
and many gaseous substances. It has also become the most economical
-147-
-------
because of wide commercial applications. It cannot be universally
utilized, however, because of the limitations of:
• adsorption with more polar compounds
• desorption, particularly with higher molecular weight
compounds
• migration, particularly with some gaseous substances.
The standard (150 mg) charcoal tube is 7 cm long with an inside
diameter of 4 mm and contains two sections of 20/40 mesh activated
coconut charcoal. The front section contains 100 mg of charcoal while
the backup section contains 50 mg of charcoal separated with a 3 mm
piece of urethane foam (see Figure I). The purpose of the backup
section is to determine the possibility of sample breakthrough and
consequent loss. Two tubes in series are sometimes necessary to
evaluate possible sample migration, e.g. butadiene. More recently,
larger tubes, a 600 mg charcoal (400 mg front with 200 mg back section)
and a jumbo 1000 mg (800 mg front and 200 mg back) have been developed
to accommodate more volatile substances such as trifluoromonobromo-
methane. A wide variety of chemical types including the aliphatic
and aromatic hydrocarbons, halogenated hydrocarbons, ketones, esters,
and ethers can be evaluated with charcoal tube sampling techniques.
Silica gel solid sorbent collectors are recommended for use in about
7 percent of the chemical substances validated under the SCP. Silica
gel is inexpensive and is an acceptable sorbent for the more polar
chemical substances. The standard silica gel tube (225 mg) is 7 cm
-148-
-------
long with 4 mm inside diameter and contains two sections of 20/40 mesh
silica gel. The front section contains 150 mg of silica gel while
the backup section contains 75 mg of silica gel separated by a 2 mm
portion of urethane foam. Silica gel sampling methods have been
developed for acetylene tetrabromide, nitrobenzene, diethylamine,
ethylamine, aniline, ethylmorpholine, xylidene, and dimethylamine.
Larger tubes are also available and are recommended for sampling methyl
alcohol. Recognized limitations with silica gel sampling tubes in-
clude a high affinity for water so that vapors may not be adequately
trapped in the presence of high relative humidity.
Other types of solid sorbent collection media have been validated for
only 2 percent of the substances included under the SCP. For example,
the XAD-2 resin is used for ethyl silicate and nicotine sampling.
Other types of solid sorbents in experimental stages are fluorosil
for polychlorinated biphenyls, Porapak Q for furfural alcohol,
Chromosorb 104 for mercaptans, and Tenax for the polycyclic aromatic
hydrocarbons. Several different types of solid sorbent materials
are being tested under laboratory conditions for reliability and
applicability.(10)
Combining sampling types is sometimes required to assure adequate
contaminant collection. For example, the method for sampling penta-
chloronaphthalene includes a glass fiber filter connected in series
with a midget bubbler containing 15 m£ of iso-octane. Nitric oxide
and nitrogen dioxide may be collected on different sections of a
solid sorbent tube using a triethanolamine (TEA) - impregnated
-149-
-------
molecular sieve collection media. The nitrogen dioxide is absorbed
on the first section of the tube. The nitric oxide is converted to
nitrogen dioxide by an oxidizer and subsequently adsorbed on the
second section of the tube.
The primary analytical tool is the gas chromatograph equipped with
flame ionization, electron capture, flame photometry, alkali flame
ionization or electrolytic conductivity detectors. About 70 percent
of the analytical methods validated under the SCP employ gas
chromatographic techniques. The gas chromatograph is especially
useful in testing for a wide variety of gases, vapors, and mists
encountered in industrial situations. Atomic absorption used for
identification and evaluation of metal dust and fume accounts for
about 4 percent of the methods validated. Other analytical techniques
include potentiametric, colorimetric, titrimetric, and gravimetric
methods.
EXPOSURE DATA
NIOSH is responsible for monitoring and evaluating industrial
exposures primarily related to mortality and morbidity studies.
Present studies entail monitoring of over 40 different agents.
Some current studies of interest include documentation of occupational
exposure to benzene, PCBs, asbestos, epichlorohydrin, styrene and
butadiene. Exposure data are reported as 8-hr, time weighted
average exposures or ceiling concentrations (15 minutes) depending
on the exposure situation and the appropriate standard.
-150-
-------
A recent NIOSH mortality study of workers has shown that benzene is
leukemogenic. Based on this study and other available information,
OSHA has promulgated an emergency temporary standard for benzene with
a permissible exposure level of one part per million as an 8 hr.
( 8)
TWA. Benzene is present in gasoline at about 2 percent by volume.
Exposure levels to retail automobile service station attendants will
vary depending on the liquid volume percent, air temperature, amount
of gas pumped and wind direction. Ongoing NIOSH benzene monitoring
studies indicate exposures to service station attendants as high as
0.8 ppm (782 ppb). Benzene exposure data are shown in Table I.
Polychlorinated biphenyls have been of concern to both EPA and NIOSH
because of the potential for carcinogenic activity and the' build-up
(12)
of concentrations in the environment. The existing Federal
standard is 500 yg/m for compounds containing 54% chlorine. NIOSH
has monitored worker exposures to PCBs in manufacturing processes and
3
have found levels to be about 50 iig/m . PCBs exposure data are shown
in Table II.
The health hazards associated with exposure to asbestos fibers are
well documented in the literature. The existing Federal standard is
2 fibers per cc greater than 5 microns in length as determined by
(13)
optical microscopy as of July 1976. NIOSH industrial hygiene
studies have demonstrated the feasibility of compliance with the
(14)
current standard. Asbestos fiber exposures in operations
where asbestos is present as contaminants of other products such
as talcs are of major concern. Elevated fiber exposures have been
-151-
-------
demonstrated in these operations. A summary of asbestos exposure
data obtained by optical microscopy is presented in Table III.
Certain problems are evident in the current occupational asbestos
exposure monitoring process, not the least of which is the fact
that only fibers larger than 5 ym in length are considered. Studies
of airborne dusts in industrial asbestos exposure operations have
shown that a large majority of airborne fibers (1-51%) are shorter
than 5 pm with median fiber lengths ranging from <1 to 5 ym (see
Table IV). A major effort is underway by NIOSH to develop and/or
evaluate alternative techniques such as electron microscopy for
routine monitoring.
Epichlorohydrin is well recognized to be a highly toxic chemical,
a primary skin irritant and sensitizer, an upper respiratory
irritant and has demonstrated both acute and chronic toxicity by
every route of entry. Animal testing provides incriminating
evidence as to possible mutagenic, teratogenic and carcinogenic
effects. The existing Federal standard is 5 ppm with a NIOSH
proposal to reduce the level to 0.5 ppm. Monitoring levels in
manufacturing facilities indicate the levels generally to be
within the existing and proposed limits. Table V shows a summary
of exposure data for several epichlorohydrin manufacturing operations.
NIOSH is currently conducting an environmental-epidemiological study
of the styrene-butadiene rubber idnsutry. This study was prompted
by industry reporting 8 cases of leukemia at their Port Neches,
-152-
-------
Texas facilities in March, 1976. McMichael et al have reported
that workers employed in a facility where synthetic rubber was
manufactured had a six-fold relative risk of dying from lymphatic
and hemopoietic malignancies as compared to other workers in the
same tire manufacturing plant who had not worked in the synthetic
rubber production. The current Federal standard for styrene
is 100 ppm and for 1,3 butadiene is 1000 ppm. Monitoring revealed
very low exposure levels, well within the standards and generally
less than 5 ppm. A summary of exposure data for styrene and
butadiene is shown in Table VI.
SUMMARY AND CONCLUSIONS
Fugitive emissions from industrial sources are a mutual concern to
NIOSH in evaluating worker exposure and to EPA in evaluating the
environmental effects. However, the approach to evaluation is often
different because the purposes for monitoring and monitoring locations
are usually different. Personal sampling does not usually reflect
short-term exposure to high concentrations, but rather averages of the
exposure over a full shift. Workers who must perform a job in areas
of possible high exposure concentrations such as maintenance workers
under extreme exposure conditions, must be provided with control
devices or must limit exposure times to meet existing standards.
Environmental sampling has traditionally been based on high volume
sampling techniques. With the development of more sophisticated
analytical technologies, in some instances, environmental monitoring
-153-
-------
to detect fugitive emissions may be accomplished by employing some
of the more recently developed miniaturized sampling techniques.
-154-
-------
REFERENCES
1. Bowditch, M., Drinker, C.K., Drinker, P., Haggard, H.W., and
Hamilton, A., "Code for Safe Concentrations of Certain Common
Toxic Substances Used in Industry", J. Ind. Hvg. Toxicol. 22:251,
1940.
2. U.S. Dept. of Labor, Safety and Health Standards, Federal
Register. Part II. Vol. 34:96, May 20, 1969.
3. U.S. Dept of Labor, Part 1910 - Occupational Safety & Health
Standards, Emergency Standard for Exposure Asbestos Dust,
Fed. Reg., Vol. 36, May 29, 1971.
4. U.S. Dept. of Labor, Part 1910 - Occupational Safety and Health
Standards, Emergency Standard for Exposure Asbestos Dust,
Fed. Reg., Vol. 36:235, Dec. 7, 1971.
5. U.S. Dept. of Labor, Part 1910 - Occupational Safety and Health
Standards, Exposure to Vinyl Chloride, Fed. Reg., Vol. 39:194,
Oct. 4, 1974.
6. U.S. Dept. of Labor, Part 1910 - Occupational Safety and Health
Standard, Exposure to Coke Oven Emissions, Fed. Reg., Part III,
Vol. 41:206, Oct. 29, 1976.
7. U.S. Dept. of Labor, Part 1910 - Occupational Safety and Health
Standards, Carcinogens, Fed. Reg., Part III, Vol. 39:20, Jan. 29,
1974.
8. U.S. Dept. of Labor, Part 1910 - Occupational Safety and Health
Standards, Occupational Exposure to Benzene, Fed. Reg., Part IV,
Vol. 42:85, May 3, 1977.
9. Taylor, D.G., "NIOSH Sampling and Analytical Methods Validation",
Presented at the ISA meeting, October 1976, Houston.
10. "Second NIOSH Solid Sorbents Roundtable" held on December 5-6,
1973, NIOSH publication No. 76-193, July 1976.
11. Infante, P.F., Rinsky, R.A., Wagoner, J.K., and Young, R.J.,
"Leukemia Among Workers Exposed to Benzene", The Lancet (in press)
12. Ayer, F.A., compiler, "National Conference on Polychlorinated
Biphenyls" (Nov. 19-21, 1975, Chicago, 111.), Environmental
Protection Agency, EPA-560/6-75-004, March 1976.
13. U.S. Dept. of Labor, Part 1910 - Occupational Safety and Health
Standards, Fed. Reg.. Part II. Vol. 39:125, June 27, 1974,
pg. 23543.
-155-
-------
REFERENCES (continued)
14. Curtis, R.A., Bierbaum, P.J., "Technological Feasibility of the
2 Fiber/cc Asbestos Standard in Asbestos Textile Facilities",
AIHA Jr., Feb. 1975.
15. Criteria for a Recommended Standard - Occ. Exposure to Epichloro-
hydrin, National Institute for Safety and Health, Sept., 1976,
NIOSH Pub. No. 76-206.
16. McMichael, A.J. et al, "Mortality Among Rubber Workers: Rela-
tionship to Specific Jobs", JOM. Vol. 18:3, p. 178-185, March
1976.
-156-
-------
7cm
Glass Tube
2
Ul
-J
I
Fiberglass
Urethane Foam
Primary
Section
Backup
Section
20-40 Mesh Activated
Coconut Charcoal
4mm
O.d.) _
I
vs
6mm
FIGURE 1, CHARCOAL SAMPLING TUBE
-------
TABLE I
SERVICE STATION ATTENDANTS' EXPOSURE TO BENZENE
TEMP,. °F
-24
59
49
42
71
niiiu v^rctJW
MPH
—
2,1
3,6
1,2
3,7
NOZZLE TIME, SEC.
1221
206
902
1016
52
FUEL PUMPED, GAL.
173,7
144,6
158,6
225,9
58,3
2 HR. EXPOSURE,
14
158
782
205
35
PPB
SOURCE: HARTLE, R,, UNPUBLISHED DATA
-------
TABLE II
RGB'S - SUWARY PLANT DATA
EMPLOYEE'S 8 HOUR TV/A EXPOSURE TO PCB AT MANUFACTURING
FACILITIES, JULY, 1976
OCCUPATIONAL TITLE 8 HR. TW\
CHLORINATOR OPERATOR 33
TANK FARM AND SHIPMENT W
DISTILLATION COLUMN OPER, 57
STILL OPERATOR 63
INCINERATOR OPERATOR 45
SOURCE: JONES, M,, UNPUBLISHED DATA,
-159-
-------
TABLE III
SUMMARY OF OVER 10,000 SAMPLES ANALYZED
FOR ASBESTOS FIBER LEVELS BY INDUSTRIAL CATEGORY,
CONCENTRATIONS DETERMINED BY OPTICAL MICROSCOPY
INDUSTRY
TEXTILE
INSULATION
PAPER PACKING AND ASPHALT
PRODUCTION
CEMENT SHINGLES, MILLBOARD
AND GASKET
FRICTION
CEMENT PIPE
TALC PRODUCTION
GARAGES
RANGE OF MEANS
CFiBRES > 5 )JM/ML)
0,1 - 29,9
0,1 - 74,4
0,2 - 13,6
0,1 - 4,4
0,1 - 14.4
0,2 - 6,3
2,1 - 7,0
0,1 - 0,2
RANGE OF
INDIVIDUAL SAMPLES
(FIBRES > 5 JJM/ML)
0,0 - 143,9
0,0 - 208,4
0,0 - 18,9
0,0 - 16,6
0,1 - 32,4
0,0 - 13,4
0,3 - 29,1*
0,0 - 15,3*
SOURCE: DEMENT, J, ET AL, AN, N,Y, Ac, OF Sc,, VOL, 271, p, 345-352,
1976
* DEMENT, J,, (UNPUBLISHED DATA)
-160-
-------
TABLE IV
SUMMARY OF TYPICAL AIRBORNE FIBER LENGTHS FOR
TYPICAL INDUSTRIAL OPERATIONS AS DETERMINED BY ELECTRON MICROSCOPY
OPERATION
> 5)JM
TEXTILE
FIBRE PREPARATION & CARDING
SPINNING, TWISTING, WEAVING
FRICTION
ASBESTOS-CEMENT PIPE
MIXING
FINISHING
PIPE INSULATION
HARDROCK MINING
TALC PRODUCTION
2
2
2
1
51
7*
3*
SOURCE: DEMENT, J, ET AL, AN, N,Y, Ac, OF Sc,, VoL/271
p, 345-352, 1976
*DEMENT, J, (UNPUBLISHED DATA)
-161-
-------
TABLE V
EPICHLJOROHYDRIN-SUflARY PLANT DATA
EMPLOYEE'S 8 HOUR TWA EXPOSURE TO EPICHLORO-
HYDRIN AT MANUFACTURING FACILITIES, 1976
OCCUPATIONAL TITLE 8 HR TVW (PPM)
GLYCERIN CONTROL OPER, 0,4
GLYCERIN SHIFT FORMAN 0,2
TANK CAR LOADING 0,3
EPOXY RESIN OPERATOR 0,1
SOURCE: OSER, J,, UNPUBLISHED DATA
-162-
-------
TABLE VI
SBR - SUMMARY PLANT DATA
EMPLOYEE'S 8 HOUR TWA EXPOSURE TO
STYRENE AND BUTADIENE AT A SBR FACILITY, JULY 1976
OCCUPATIONAL TITLE 8 HR. TWA (PPM)
STYRENE BUTADIENE
REACTOR OPERATOR 0,3 1,8
TANK FARM OPERATOR 0,7 1,9
REACTOR OPERATOR 1,3 0.8
COAGULATION OPERATOR 1,0 0,1
BLEND OPERATOR 1,1 0,03
SOURCE: YOUNG, R, ET AL, UNPUBLISHED DATA
-163-
-------
Tuesday Morning - May 24* 1977
SESSION III: CONTROL
Chairman: Henry J. Kolnsberg, TRC
Page
MODELING FUGITIVE PARTICULATES FOR CONTROL STRATEGY DEVELOPMENT - 167
A CASE HISTORY
George A. Jutze and John M. Zoller*, PEDCo
Environmentalj Ina.
ASBESTOS WASTE EMISSION CONTROL 187
Mary K. Stinson*, £P4, Colin F. Harwood, Pall Corporation,
and Paul Ase, IIT Research Institute
FUGITIVE EMISSIONS FROM CHEMICAL FERTILIZER MINING 205
J. C. Ochsner and T. R. Blackwood*, Monsanto Research
Corporation
FUGITIVE EMISSION CONTROL IN THE IRON & STEEL INDUSTRY 225
Arthur G. Nicola, Pennsylvania Engineering Corporation
* indicates speaker
-165-
-------
MODELING FUGITIVE PARTICULATES FOR
CONTROL STRATEGY DEVELOPMENT - A CASE HISTORY
By
George A. Jutze
and
John M. Zoller
PEDCo Environmental, Inc.
-167-
-------
MODELING FUGITIVE PARTICULATES FOR CONTROL
STRATEGY DEVELOPMENT - A CASE HISTORY
George A. Jutze
PEDCo Environmental, Inc,
John M. Zoller
PEDCo Environmental, Inc.
ABSTRACT
This presentation discusses the various data gathering
efforts and modeling requirements which were utilized to
evaluate a series of particulate control strategies appli-
cable to the Clark-Mohave-Yuma (Las Vegas) Air Quality
Control Region. Since relatively few uncontrolled point
sources and conventional area sources of total suspended
particulates (TSP) exist in this region, fugitive particu-
lates/dust sources were identified as producing an impact on
air quality which affects non-attainment of the NAAQS for
TSP. The results of control strategy testing utilizing the
Hanna-Gifford simulation are presented.
-168-
-------
BACKGROUND
At the present time the National Ambient Air Quality
Standard (NAAQS) for Total Suspended Particulate (TSP) is
being exceeded in the Clark-Mohave-Yuma (Las Vegas) and
Northwest Nevada Intrastate CReno) Air Quality Control
Regions (AQCR). Since relatively few uncontrolled point
sources and conventional area sources of TSP have been
identified in these regions, the probability existed that
the impact on TSP air quality from fugitive dust sources is
causing non-attainment. PEDCo Environmental was contracted
by the U.S. Environmental Protection Agency, Region IX, to
assist the state and local agencies in evaluating this
matter.
PURPOSE AND SCOPE
The purpose of this project was to:
(1) identify.the sources impacting or projected to impact
in the future on sampling sites and other receptor locations
with high particulate concentrations;
(2) determine the relative contribution of each of the
identified sources and source categories;
(3) recommend control strategies; and
i
(4) estimate the expected effectiveness of the recommended
control strategies on emission reductions and on ambient air
quality.
To accomplish these goals, six specific tasks were
performed as described below:
-169-
-------
A. Air Quality Monitoring Data
All necessary air quality and meteorological data in
the study areas for the base year 1975 were assembled and
analyzed to better identify the nature of the particulate
problem. "Micro-inventories" of the areas surrounding each
hi-vol location were also conducted to determine specific
sources and source categories influencing the measurements
at the sites.
B. Factors Influencing Emissions from Fugitive Sources
All fugitive dust and naturally occurring dust source
categories were identified and available emission factors
and data relative to variations in and particle size distri-
bution of the emission rates were assembled. From these
data, estimates of emission rates by particle size range for
each fugitive dust source category were derived. Functions
were developed to describe variations in these expected
emission rates.
C. Emission Inventory and Emission Projections
An emission inventory of all point and area sources of
particulate, including fugitive dust and background sources,
was prepared for each study area. Emissions were accurately
allocated into a grid system. The emission estimate sub-
totals in each category were provided by season and particle
size range. Projected emission inventories for 1980 and
1985 were prepared in the same format as the base-year
inventory.
D. Emission Modeling
A relationship between emissions and ambient air
quality was derived for the study areas. The model was then
used to calculate concentrations at each of the specified
receptor sites for the base year and 1985.
E. Control Strategy Development
Potential control measures for significant source
categories were reviewed. As a result, five control stra-
tegies were independently evaluated in the Las Vegas area
and three strategies in the Reno area. The recommended
strategies were tested by input of the estimated reduced
emissions to the source-receptor model developed in Task D.
-170-
-------
P. Particulate Characterization
An analysis of hi-vol glass fiber, hi-vol cellulose
acetate, and Andersen glass fiber were compared with inven-
tory results for each sampling site for determination of
source contribution. The filters analyzed were both his-
torical (hi-vol glass fiber) and the results of a four-week
special sampling program (hi-vol glass fiber, cellulose
acetate, and Andersen glass fiber). In addition, approxi-
mately fifteen particulate grab samples of soil, deposited
dust, and the like were analyzed from each of the project
study areas. Analysis included microscopic, silica frac-
tion, combustible fraction, and two suspect metals (calcium
and lead).
A project report was prepared for each task and are
entitled as follows:
0 Task A Report - Air Quality Monitoring Data
0 Task B Report - Factors Influencing Emissions from
Fugitive Sources
0 Task C Report - Emission Inventory and Emission
Projections
0 Task D/E Report - Emission Modeling and Control
Strategy Development (Las Vegas)
0 Task D/E Report - Emission Modeling and Control
Strategy Development (Reno)
0 Task F Report - Particulate Characterization Study
At certain points in this presentation, the various
"Task Reports" will be cited. However, the material dis-
cussed herein basically deals with the Task D/E Reports for
Las Vegas and Reno.
-171-
-------
MODEL DESCRIPTION
To evaluate the impact of control strategies, it is
necessary first to develop a relationship between emissions
and air quality. Several diffusion modeling techniques have
been suggested by the U.S. Environmental Protection Agency
for this purpose, including: Miller-Holzworth, AQDM, and
Hanna-Gifford. The Hanna-Gifford approach was evaluated for
applicability based on several reasons. The emission density
vs. air quality relationship developed in the Task A Report
showed a reasonable correlation, but not high enough to be
used to estimate air quality. The Hanna-Gifford simulation
is a more sophisticated approach to emission density and air
quality relationships and is expected to give a better
correlation. Few major (tall stack) point sources are
present in the study area. These sources would be evaluated
poorly by Hanna-Gifford. This model best accomodates sources
with low-level emission heights. This is typical of the
source categories in the study area. Fugitive dust sources
are generally ground level.
The basic model is an expression developed by S.R.
Hanna and F.A. Gifford to calculate the impact of various
sources on a central receptor over a long period of time.
The expression, the result of integrating the receptor-
oriented Gaussian plume equation over upwind area sources
(grids), is:
_/2 1 (Ax/2)1"b /d + n Q. [(2i+l)1"b-(2i-l)1"b][
x ~ITT u a(l-b) ^ ° i=l 1 )
-172-
-------
where
X - concentration in a grid, yg/m
u = wind speed, m/sec
Ax = grid width, m
a,b = stability coefficients (a = 0.15, b = 0.75 for
urban areas)
2
Q = source strength of receptor grid, yg/m -sec
•Hi
Q. = source strength of i grid away from the receptor
2
grid, ug/m -sec
n = number of grids in any one direction which will
have a significant impact on the receptor grid, for
this study assume n = 2.
For an area, marked off in grids of a given size, the quantity
Jl
For a 2 km grid, Ax = 2000 m and C = 119.6.
For a 5 km grid, Ax = 5000 m and C = 150.5.
For a 6 km grid, Ax = 6000 m and C = 157.5.
It was assumed that the air quality of each receptor
grid is affected not only by the emissions of that grid but
also by the emissions from two rings of surrounding source
grids. The quantity within the square brackets [ ] in the
first equation expresses the relative contribution of these
emissions based on distance in terms of number of grids (i)
away from the receptor. This scheme has been modified for
this study by assuming that the impact of the source grid
emissions on a receptor grid is proportional to the distance
between the centers of these grids in relative grid widths.
In the original model, for each grid in the ring of
grids nearest the receptor the relative impact of the emis-
sions is the same, since for each of these grids i = 1. In
the modified version i = 1.00 for grids which have a side in
common with the receptor grid. However, for grids which lie
-173-
-------
on the diagonal of the central grid, i = 1.41. Consequently,
emissions from these grids have a lesser impact than those
from grids having a common side with the receptor.
Note, the grid sizes cannot be mixed. Impact on a
receptor must be calculated from source grids of an equiva-
lent size. Therefore, in order to estimate air quality in
those 2 km grids which are nearest the 5 km grids (in Las
Vegas) or 6 km grids (in Reno), it was necessary to create
two rows of artificial 2 km grids whose emissions were
fractional parts of the emissions of the corresponding
larger km grids. Two rows of artificial 5 or 6 km grids
must, likewise, be created around the entire grid network.
Since an emission inventory for these areas does not exist,
the emissions of the perimeter grids were repeated. This
appears to be a valid assumption in light of the fact that
the two areas, gridded and non-gridded, have a similar make-
up, and that if these areas were assumed to have zero
emissions the air quality impact might be severly under-
estimated. Also, within the central part of the grid
system, artificial 5 or 6 km grids were composed of com-
binations of 2 km grids.
Meteorological conditions will also have an effect on
air quality. The impact of the emissions from each source
grid was weighted based on the frequency of occurrence of
the wind out of the direction that grid lies from the recep-
tor. Wind speed was also varied with wind direction.
Historical data and a sixteen-point compass were used.
The emission inventory data were used as input to the
model. Measured annual geometric mean concentrations were
correlated against the modeled air quality at each of the
ten sampling stations in Las Vegas and nine sampling sta-
-174-
-------
tions in Reno. Very few sampling stations were located at
the center of a grid. Therefore, the modeled air quality
was obtained by weighting the modeled impact from each grid
which at least partially occupies an imaginary grid of equal
size having a sampling site at its center. The grids at
each sampling site, their allocation factors, and the re-
sults of this correlation procedure are shown in Tables 1
and 2. Correlation results (0.69 for Las Vegas and 0.79
for Reno) show that the Hanna-Gifford approach is an ac-
ceptable diffusion model to use in this study. Statistical
analyses support the assumption that this correlation does
indeed show that there is a significant relationship between
measured and modeled air quality using the Hanna-Gifford
model. The slope and intercept are the parameters required
to calibrate the model for comparison to actual data.
The intercept of the correlation is normally assumed to
be an indication of background concentration which is
unaccounted for by any known source of emissions within a
given area. Although the intercept of the correlation is
provided by the best fit of a line through measurei versus
modeled data, the value 51.3 shown in Table 1 is not repre-
sentative of background levels in Las Vegas. The Clark
County District Health Department suggested using a back-
ground value of 35 yg/m based on three year running av-
erages from their background station. Due to the large
intercept obtained from the correlation, several other
methods were investigated in an attempt to develop an
improved relationship between measured and modeled data.
Separate correlations for high and low concentrations, and
for urban and rural concentrations, however, provided no
better relationship. The average difference between mea-
sured and modeled data, which may also be used as an indi-
-175-
-------
Tab He 1. CORRELATION OF LAS VEGAS ANNUAL GEOMETRIC MEANS
WITH MODELED AIR QUALITY
Site
A, Las Vegas Airport
8, Basic School
C, Clark Cc. Health Oept.
D, Hend-rson Post Office
E, Las Vegas Fire Dept. 1
F, Las Vegas Fire Dept. 2
G, Nell is Air Force Base
H, North Las Vegas
I, Sahara Hotel
J, Las Vegas Stedium
Kt Sunrise Power Plant
L, Las Vegas ^sh.c
Correlation
Slope
Intercept
Grids
33
43
15
29
91
101
28
29
40
92
102
93
94
138
117
79
80
50
73
85
!
Uncalibrated
model
Impact
pg/m3
6.8
43.8
9.6
20.7
47.3
33.5
14.4
20.7
87.0
70.0
43.6
42.3
27.0
8.4
32.0
39.7
43.0
25.4
33.0
168.8
Allocation
factors
1/2
1/2
2/3
1/3
5/8
3/8
1/4
1/4
1/2
1/4
3/4
1/2
1/2
1/1
1/1
1/2
1/2
1/1
1/8
7/8
Grid
contribution
pg/m3
3
22
6
7
29
13
4
5
43
17
33
21
14
8
32
20
21
25
4
148
Model ed
air quality
at site
pg/m3
25
13
42
52
50
35
8
32
41
25
152
Measured
annual
geometric
meana
ug/m3 .
69
64
91"
92
82
105
51
105
117
64
92
32
0.689
1.012
51 .320
a Calculated using October 1974 through September 1975 data.
Site K was not included in the correlation because of an apparent sampling
site bias.
c Site L was not included in the correlation because It Is located outside of
the gridded study area.
-176-
-------
Table 2. CORRELATION OF RENO ANNUAL GEOMETRIC MEANS WITH
MODELED AIR QUALITY
Site
A, Washoe County Health Dept.
B, Cal-Neva Club
C. Jessie Beck School
0, Reno Airport
E, Nevada Fish and Game
F. Mamie Towles School
6. Greenbrae School
H. Sparks Nugget3
J, Verdi School
L, Sewer Plant
Correlation
Slope
Interceot
Grids
41
42
40
41
26
32
33
28
34
35
41
48
39
46
51
43
44
50
16
37
24
45
Unca libra ted
model
impact
ug/m3
81.4
115.0
50.6
81.4
18.1
35.3
62.3
26.1
40.1
19.6
81.4
51.8
32.9
40.5
39.0
39.1
61.3
38.6
10.5
13.1
19.2
22.6
Allocation
factors
7/9
2/9
1/3
2/3
2/9
6/9
1/9
2/9
6/9
1/9
1/3
2/3
1/2
1/2
1/1
5/9
1/9
3/9
1/2
1/2
1/2
V2
Grid
contribution
ug/m3
63
26
17
54
-4
23
7
6
27
2
27
35
17
20
39
21
7
13
5
7
10
11
Modeled
air quality
at site
pg/m3
89
71
34
35
62
37
39
41
12
21
Measured
annual
geonetric
aeanb
W/«3
77
74
44
65
57
32
SB
90
18
50
0.793
0.617
35.360
* Site H was not Included in the correlation because of an apparent sampling
site bias.
b Calculated using October 1974 thru September 1975 data.
-177-
-------
cation of background, also failed to improve the relation-
ship. This value of 51.7 vg/m3 did not favorably agree with
background levels measured by the local agency. An attempt
to force-fit a line through the suggested background value
of 35 yg/m would result in a calibration which would over-
predict the large concentrations and underpredict the low
ones.
Based on the investigation described above, it was
decided to use the slope of the original correlation, shown
in Table 1, and the background value (intercept) of 35 vg/m
suggested by the Clark County District Health Department.
All subsequent model runs predict air quality concentrations
according to the following expression:
predicted air quality = (1.012)(modeled air quality) + 35.
This equation produces predicted concentrations which are on
the order of 16 yg/m lower than those which would result if
the correlation intercept were used.
The slope and intercept were the parameters used to cali-
brate the model in Reno for comparison to actual data (Table
2). Subsequent model runs predict air quality concentrations
according to the following expression:
predicted air quality = (0.617) (modeled air quality) + 25.4
The value 25.4 may also be looked_upon as a background
concentration which is unaccounted for by any other source
of emissions. This value is considered to be an acceptable
background value by the Washoe County District Health
Department. It is also in good agreement with the intercept
from the emission density versus air quality correlation
presented in the Task A report.
-178-
-------
CONTROL STRATEGIES
Several control possibilities were discussed with Clark
County District Health Department, Air Pollution Control
Division personnel. These strategies were designed to
affect the major source categories in the study area.
Initially, the following strategies were considered:
Street cleaning,
Stationary source control,
Stabilization of railroad yards,
Control of construction emissions.
Reduction of emissions from cleared areas, and
Reduction of unpaved road emissions.
All of these strategies, except control of construction
emissions, were then evaluated using the Hanna-Gifford
relationship. The current regulations on construction are
considered by the agency as adequate; their limiting factor
being enforceability. The others were evaluated by first
determining the emission reduction efficiencies that are
expected for each strategy. A weighing factor was then
applied to better estimate the actual impact on air quality
that each strategy will have. (This is explained in the
following section.) The resulting emissions were then input
to the Hanna-Gifford model to predict the resultant air
quality.
As in Las Vegas, several control possibilities were dis-
cussed with Washoe County District Health Department, Air
Pollution Control Division personnel. These strategies were
designed to affect the major source categories in the study
area. Initially, the following strategies were considered.
-179-
-------
Street cleaning,
Stationary source control,
Control of construction emissions.
Reduction of unpaved road emissions, and
Reduction in emissions from controlled burning.
In discussions with the local agency, two of these
strategies, stationary source control and reduction in
emissions from controlled burning, were eliminated from
further consideration. This was because the stationary
source of principal concern would not be operating by 1985
and reduction in emissions from controlled burning is not
considered feasible. The other strategies were evaluated by
first determining the emission reduction efficiencies that
are expected for each strategy. A weighing factor was then
applied to better estimate the actual impact on air quality
that each strategy will have. The resulting emissions were
then input to the Hanna-Gifford model to predict the re-
sultant air quality.
DEVELOPMENT OF A WEIGHING FACTOR
In attempt to estimate more closely the impact on air
quality that will be expected from emission reductions,
development of a weighing factor was investigated in the
Task F Report, Particulate Characterization Study. This
factor was developed because of the large disparity between
the percent mineral observed on hi-volume sampler filters
and the percent mineral emissions in the microinventories
surrounding the hi-vol sites. A much higher percent of the
emissions were mineral than were collected on the filters.
Therefore, a reduction in mineral emissions can not be
expected to result in an equal reduction in filter loading.
From the microinventory of selected sites in Las Vegas
and Reno, all emissions were identified and classified as
-180-
-------
being from mineral, tire wear, or combustion sources. Also,
high volume samples (millipore and glass fiber), which were
taken at these sites, were analyzed under Task F for the
percent by weight of material from the same source clas-
sifications. The mineral percentage by weight of both the
emissions and the filter loadings are shown in Table 3.
It has been observed that a high percentage of all
micJroinventory emissions at the selected sites are from
mineral sources. One would expect a similar percentage of
minerals to appear on the sample filters. This, however,
was not the case. On the average, the sample filters showed
a percentage of mineral impact which was only 60 percent of
the percentage of mineral emissions. Therefore, reductions
in mineral emissions and reductions in ambient air quality
are not expected to occur in a one-to-one ratio.
Since the control strategies are applied to mineral
emissions, only 60 percent of any reduction in emissions due
to the control strategies will be observed in the ambient
air quality as measured by the high-volume sampler. There-
fore the weighing factor will be applied as follows:
Control strategy reduction, percent
times
Weighing factor, 0.60
equals
Effective percent reduction
The effective percent reduction is then applied to each
applicable emission category to generate the emissions that
are input to the Hanna-Gifford model for strategy evalua-
tion. This is as follows:
Affected category emissions (by grid)
times
effective percent reduction
equals
reduction in category emissions (by grid)
-181-
-------
Table 3. FUGITIVE DUST WEIGHING FACTOR
03
co
I
Overall Average Reduction Ratio 0.64b
Site
Henderson P.O.
Sahara
Sunrise
Average
Microinventory
emissions,
percent mineral
96
92
99
Glass fiber
filters,
percent mineral
70
51
75
Reduction ratio,
Glass fiber
Microinventory
0.73
0.55
0.76
0.68
Millipore filters,*
percent mineral
61
42
70
Reduction ratio,'
Hillipore
Microinventory
0.64
0.46
0.71
0.60
*Ratio of percent mineralcollected on the hi-vol filter divided by the percent mineral
emissions in the microinventory at the hi-vol site.
For this study, the reduction ratio (weighing factor) was rounded to 0.60, as agreed to
by the Clark County District Health Department.
Site
Cal-Neva Club
Sparks-Nugget
Waahoe Health Dept.
Average
Microinventory
emissions,
percent mineral
61
65
77
Glass fiber
filters,
percent mineral
38
43
41
Reduction ratio*
Glass fiber
Microinventory
0.63
0.67
0.53
0.61
Millipore filters,
percent mineral
33
32
51
Reduction ratio,
Millipore
Microinventory
0.55
0.50
0.66
0.57
Overall Average Reduction Ratio 0.59
°Ratio of percent mineral collected on the hi-vol filter divided by the percent mineral
emissions in the microinventory at the hi-vol site.
bPor this study, the reduction ratio (weighing factor) was rounded to 0.60, as agreed to by
the Wasnoe County District Health Department.
-------
Total category emissions (by grid)
minus
reduction in category emissions (by grid)
equals
emissions input to control strategy evaluation.
CONTROL STRATEGIES RESULTS
Figure 1 graphically displays reductions in TSP an-
nualized air quality achieved by applying all control
strategies to the 1985 uncontrolled emissions estimated for
Las Vegas.
According to the projected 1985 air quality, the area
which will have difficulty meeting the primary National
Ambient Air Quality Standards (NAAQS) of 75 yg/m is an
eight grid area (79, 80, 91, 92, 93, 101, 102, 103), concen-
trated in the central city. In these grids, the NAAQS are
anticipated to be exceeded by up to 40 yg/m in 1985. Even
after the control strategies are applied, grids 92, 93, 102
and 103 are still expected to exceed the national standards
by as much as 20 yg/m .
Other grids (40, 52, 85, 86), which are also shown to
exceed the NAAQS, do so because of the impact of emissions
from large local industrial sources. While control of these
sources are not projected to reduce the local air quality to
the level of the national standards, it will make a signifi-
cant improvement.
The air quality in grid 62 is also shown to exceed
the standards, primarily due to the emissions from cleared
areas. However, the control strategy of soil stabilization
did not include this area. Therefore, the air quality in
this locality is expected to remain above the standards.
-183-
-------
4020 ka
4010 kn
4000 Ion
3990 kn
3980 km
ISO In
660 kn
670 kn
680 hi
KEY
A LAS VEGAS AIRPORT
B BASIC SCHOOL
C CLARK COUNTY HEALTH DEPT.
D HENDERSON POST OFFICE
E LAS VEGAS FIRE OEPT. NO. 1
F LAS VEGAS FIRE OEPT. HO. 2
6 NELLIS AIR FORCE BASE
H NORTH LAS VEGAS
I SAHARA HOTEL
J LAS VEGAS STADIUM
,t SUNRISE POWER PLANT
L US VEGAS HASH
SCALE
KILOMETERS
10
MILES
? I ? T ! 7
LEGEND
XX
\
690 k»
REDUCTIONIN
TSP (u9/«3)
GRID NUMBER
Figure 1. Total change in 1985 Las Vegas air quality
due to all control strategies.
-184-
-------
The same situation applies to grid 43 since the control
strategy affecting emissions from unpaved roads, which is
the main source in this grid, does not include this area.
Finally, the area in and around grid 48 is also shown
by the model to slightly exceed the NAAQS. However, the
major source of emissions in this grid is a power plant.
Since the Hanna-Gifford model tends to over-predict sources
such as this (emissions from a relatively tall stack) actual
air quality in this grid is not expected to exceed the
national standard in 1985.
Figure 2 shows, for Reno, the projected 1985 TSP an-
nualized air quality resulting from the application of all
control strategies.
According to the projections, only grid 41 in Reno in
1985 will be exceeding the primary National Ambient Air
Quality Standards (NAAQS) of 75 yg/m . The source category
of normal paved streets accounts for 59 percent of the
emissions in grid 41. It is anticipated that the air quality
standards can be met in this grid with the application of
the proposed control strategy of street cleaning. This will
lower the air quality from an estimated 78 ug/m to approx-
3
imately 73 yg/m .
The other strategies reduce the projected emissions,
but only in grids that are already anticipated to be meeting
the primary NAAQS. Therefore, the only strategy that need
be considered for implementation is street cleaning.
-185-
-------
LEGEND
UT
WASIOE COUOTI HEALTH MM
CAl-HCVA ttUS
JISME ttcr. SCHOOL
HUAOA fill
KAMIE TO-JIES SCHOOL/
ir>*ii« N'.CUII
LAC TAllOt (MI iHOW)
vr«oi SCHOOL .
H»OS.O»IH CUE OEIT. (HOT SHOUM)
SE'JCR CUUil
242.5 a EAST;
100.000 ft. UST
7M.4IS n. KMTH
43S4.4 IM NORTH
4JM.2 tm mam
«;.«o ft.
4353.7 bi HORTK
660.000 ft. WSTK
100.009 ft. CAST
«1.J ta EAST;
Fig'ure 2. Total change in 1985 Reno air quality due
to all control strategies.
-186-
-------
ASBESTOS WASTE EMISSION CONTROL
By
Mary K. Stinson
U.S. Environmental Protection Agency
lERL-Ci
Edison, New Jersey 08817
Colin F. Harwood
Pall Corporation
30 Sea Cliff Avenue
Glen Cove, New York 11542
and
Paul Ase
IIT Research Institute
10 West 35th Street
Chicago, Illinois 60616
-187-
-------
ASBESTOS WASTE EMISSION CONTROL
Mary K. Stinson*
Colin F. Harwood**
Paul Ase***
ABSTRACT
Asbestos products manufacturing waste dumps can be
a significant source of asbestos emissions and consequently
a health threat. A strategy is revealed which evaluates
control concepts applicable to a model typical plant. As
a result of the evaluation, a curve can be drawn which gives
the least-cost control methodology to be applied to achieve
a given level of emission control. Field tests designed to
demonstrate the effectiveness of,the selected control systems
are described. Practical considerations on the field testing
of emission control systems are discussed.
* U.S. Environmental Protection Agency, lERL-Ci, Edison,
New Jersey 08817
** Pall Corporation, 30 Sea Cliff Avenue, Glen Cove, New
York 11542
*** IIT Research Institute, 10 West 35th Street, Chicago,
Illinois 60616
-188-
-------
INTRODUCTION
Adverse health effects associated with asbestos have
caused it to be listed as a hazardous substance by the Federal
EPA. Asbestos present in the ambient air may pose a health
threat to the general population who are not occupationally
exposed to asbestos. One source of asbestos in the ambient
air is the waste dumps located throughout the United States
where asbestos waste has been allowed to accumulate over the
years.
This paper reveals the strategy used in a study to in-
vestigate those methods which might be employed to reduce
the fugitive emissions from asbestos waste dumps. Although
the study was directed specifically towards asbestos waste
dumps, the strategy used to approach the problem can be used
in other industries where the waste can give rise to fugitive
emission problems.
In general terms the strategy was to divide the problem
into several tasks as:
9 Identify sources of emissions
«> Cost effectiveness evaluation of control
options
» Field test the most promising control options
»Provide conclusions and recommendations.
This paper discusses the emissions problem of asbestos
waste from the fabrication of asbestos cement pipes.
-189-
-------
FUGITIVE EMISSION SOURCES
The asbestos industry is not a small industry. World
consumption currently stands at about 5 million tons per
year. Approximately 1 million tons are used annually in the
United States. Asbestos cement products account for 70 per-
cent of the total United States usage. Products made from
asbestos cement include pipe, wall siding, roof shingles,
wallboard, and insulation products. It is estimated that
5-10 percent of the product material is dumped as scrap, of
which 10 percent is fine dust and 90 percent coarse scrap
from trimmings and breakage and from products which have
failed quality assurance testing.
The quantity of waste material from the asbestos cement
industry may be readily estimated. Of the 560,000 tons of
asbestos used in asbestos cement products annually, 7.5 per-
cent is scrapped. Since asbestos products contain an average
only 25 percent asbestos, then the total quantity of asbestos
cement products disposed of as scrap per annum is about
168,000 tons.
Over the years the scrap products have grown into sub-
stantial waste piles. Dumping of the scrap is usually
carried out with little or no attempt to control the fugitive
emissions associated with the various stages of the waste
transfer and dumping operations.
Fugitive emissions will result from four basic processes;
• Dumping of fines, reject pipe, and aggregates
-190-
-------
« Crushing and smoothing of the reject pipe on
the active pile
• Weathering of the active pile
o Weathering of the inactive pile
In a typical plant fines collected at several baghouse
locations are periodically emptied over the course of a day.
The fines are taken by truck and emptied onto the waste dump.
Reject pipe along with other large waste aggregates are also
collected several times a day and taken to the dump to form
a separate pile.
On a monthly basis, waste pipe is broken into fragments
by the action of a bulldozer (which weighs about 25 metric tons)
passing over the pipe. In this manner, a bed of fragments is
created. Fine waste material is then spread over the pipe
fragments to fill the voids and to form a relatively level
surface.
The active pile is that portion of the dump where fresh
waste material is added and is periodically crushed and leveled.
The active pile is the dump in use. It is subjected to vehicular
traffic and on an area basis results in a high emission factor.
An inactive pile is that portion of the dump to which new
wastes are no longer added and is not subject to any disturbance.
The emission rate is a function of the size and age of the in-
active pile.
COST EFFECTIVENESS EVALUATION OF CONTROL STRATEGIES
Environmental engineering estimates were made to assess
the relative magnitude of the emissions from the primary
-191-
-------
sources identified above. The results indicate that while
crushing and leveling has a very high emission rate, the
short time duration of the operation reduces its relative
emission rate to only 6 percent of the total. Estimates of
the four sources are shown diagramatically in Figure 1. Field
measurements taken subsequent to the estimation showed the
values to be reasonable and realistic.
Emission control technologies have been developed in
other industries for the control of fugitive emissions and
the stabilization of waste dumps. Some of the more feasible
control methods were studied for application with asbestos
waste. They are listed in Table 1.
Table 1
WASTE EMISSION CONTROL OPTIONS
Source
Duration
Control Option
Waste dumping While active Slurry, bag, agglomerate and
crush in-plant, spray-during
dumping.
Crushing and
leveling
Active pile
While active Crush in-plant, spray and
chemical treatment-during
operation.
<1 year
Inactive pile >1 year
Chemical stabilization, spray,
foam, and landfilling -
monthly.
Physical cover, vegetative
cover-annually.
A cost effectiveness analysis of these methods was
used to select control options for field testing. The
cost of controlling emissions from the waste dump activi-
-192-
-------
22%
FINES
DUMPING
6%
CRUSHING
8 LEVELING
60%
ACTIVE
PILE
12%.
INACTIVE
PILE
FIGURE I
PRIMARY DUMP ASBESTOS EMISSION SOURCES
AND THEIR RELATIVE EMISSION RATES
-193-
-------
ties were estimated by invoking a model plant with a daily
production of 250 metric tons, a daily disposal rate of
13.2 metric tons of aggregate waste, and 0.9 metric tons of
baghouse fines. The techniques available to reduce emissions
vary considerably in annual costs, as well as in the emission
reductions achieved. The eight basic control options shown
in Figure 2 were evaluated. These eight control options
were considered individually and in combination to determine
the lowest total cost methods for achieving the greatest
overall emission reductions. The results of the cost analysis
are summarized in Figure 3, and the least cost combinations
of the control options are shown in Figure 4.
Based on this analysis, four emission control options
were selected for field testing. Emission from the transfer
of baghouse fines can be reduced by (1) bagging, or (2)
slurrying the waste at the plant before dumping. Active pile
emissions can be reduced with a chemical stabilizer, and
inactive pile emissions can be reduced by a soil-vegetation
cover. To prevent the inactive pile emissions from increas-
ing each year with an annual increase in size of the pile,
a permanent cover is mandatory.
It was estimated that bagging would reduce dumping emis-
sions by 100 percent, while the use of a chemical binder on
the active pile would reduce its emissions by 90 percent, and
a soil-vegetation cover would reduce the long-term emissions
from the inactive dump by 100 percent. Application of the
three control options in combination would reduce the emis-
-194-
-------
Individual
Control Option
No Controls
A. Water spraying at waste
dump
B. Agglomeration of fines
before dumping
C. Water slurrying of fines
before dumping
D. Bagging of fines before
dumping
E. Sprinkling water on
active pile
F. Landfilling on active
ptle
G. Chemical stabilization
of active pile
H. Soil-vegetation cover
on inactive pile
Total
Annual
Cost, Controlled Dump Emissions
dollars % of Uncontrolled Emissions
0
0
100
Kx.i';>S':.vi waste dumping
active pile
crushing and leveling
inactive pile
1 Based on plant production rate of 75,000 metric tons per year
(250 metric tons per day).
Figure 2
INDIVIDUAL CONTROL MEASURES EFFECTIVENESS
AND THEIR APPLICATION COSTS
-195-
-------
Control Method1
No Control
H
A + H
E + H
B + H
F + H
C + H
D + H
G + H
B + E.+ H
C + E + H
D + E + H
B + F + H
C + F + H
D 4- F + H
B + G + H
C + G + H
D + G + H
Total2
Annual
Cost,
dollars
0
3,380
6,780
6,950
16,380
12,080
9,180
13,880
7,350
19,950
12,750
17,450
25,080
17,880
22,580
20,350
13,150
17,850
Controlled Dump Emissions,
% of Uncontrolled Emissions
waste dumping
active pile
100
crushing and leveling
inactive pile
1 Letter code refers to individual options listed in Figure 2.
2 Based on plant production rate of 75,000 metric tons per year
(250 metric tons per day).
Figure 3
COMBINATIONS OF EMISSION CONTROL OPTIONS — THEIR EFFECTIVENESS
AND THEIR APPLICATION COSTS
-196-
-------
30,000
25,000
20,000
tj 15,000
o
tfl
3
10,000
5,000
I
I
I
20 40 60 80
Asbestos Emission Reduction, Percent
100
Figure 4
LEAST COST COMBINATIONS OF EMISSION CONTROL OPTIONS
(For description of individual control options,
see Figure 2)
-197-
-------
sions by 91 percent at a total annual cost of $17,850 for
the typical model plant. Substitution of slurrying for
bagging in the dumping would reduce the combined control
to 87 percent at a much reduced annual cost of $13,150.
FIELD TEST OF OPTIONS
The separate contributions of each of the primary sources
were measured in field tests and the effectiveness of the
controls selected to mitigate their emissions was studied.
Following the recommendations of the cost effectiveness
analysis, field tests were conducted with emission controls on
only three of the four primary sources. Slurrying and
bagging were tested for fines dumping; chemical stabilization
was tested for the active pile; and a soil-vegetation cover-
age was tested for the inactive pile. Aggregate crushing
is difficult to control at the dump and expensive to conduct
as a plant operation. It is, however, a minor emission source
overall. This is because, although obvious high emissions
are created, the operation is conducted for only a short
period of time -- once every one to three months. For this
reason, aggregate crushing and leveling were monitored in the
field tests without emission controls.
The tests were conducted at the Johns-Manville asbestos
cement pipe plant located at Denison, Texas. The on-site
location offered easy haulage of waste materials and cover
soil, reasonable site security, and the availability of labor
to maintain the piles and perform periodic tests.
-198-
-------
FUGITIVE EMISSIONS MITIGATION
Field samples were collected and emissions monitored
from the following transfer activities:
» dumping of fresh, dry fines
o dumping of fines as a slurry
o dumping of fines in plastic bags
« crushing of waste pipe
* leveling of waste
• dump background
The emissions were collected with high volume particle sam-
plers and particle counters. The high volume samplers, in an
upwind-downwind array, collected emission samples which were
later examined for asbestos and analyzed for elemental com-
position, mass concentration, and fiber concentration. The
particle counters gave real time data of the emission cloud
concentration from the dump activities and the mitigation in
the emissions by use of controls.
Table 2 gives some typical asbestos concentrations found
near the dump while various activities were taking place.
Data obtained from the transmission electron microscope oper-
ating at a magnification of 20,00OX includes all fibers
greater than 0.06 urn in length, while from the optical micro-
scope, only those fibers greater than 5.0 ^im in length were
measured. Atomic absorption was used to estimate the mass
concentration of asbestos which is given in units of nano-
grams per cubic centimeter.
-199-
-------
Table 2
TYPICAL WASTE DUMP AMBIENT AIR ASBESTOS LEVELS MEASURED BY
ELECTRON MICROSCOPE, OPTICAL MICROSCOPE,
AND ATOMIC ABSORPTION
Asbestos Concentration
Activity
No Activity
Pipe Crushing
Waste Leveling
Distance
m
At Dump
90
35
Electron
Microscope
Fibers/cm
1.2
22.3
10.9
Optical
Microscope
Fibers/cm
0.05
0.25
0.30
Atomic
Absorption
ng cm
0.003
0.032
0.046
The contribution from the waste emission sources
are shown in Table 3. These results were calculated
Table 3
FUGITIVE ASBESTOS WASTE EMISSION ESTIMATES BASED
ON FIELD MEASUREMENTS
Emission Rate during Annual Emission
Emission
Source
Active Pile
Dumping
Inactive Pile
Crushing 8 Leveling
Total
Active Emission
kg/yr
1,100
1,300-6,300
230
500-6,300
Rate
kg/yr
1,100
90-430
230
6-70
1,400-1,800
from atomic absorption analyses of air samples collected
at the dump. The greatest contribution to the asbestos
emission comes from the active pile. In decreasing order
and significance, the others are: waste dumping, inactive
pile (assuming annual burial) and the crushing and leveling
-200-
-------
operation. The relative contribution from each of the
sources is in good agreement with the original estimates.
Both the dumping and the crushing and leveling opera-
tions show much higher active emission rates than the active
pile does, but due to the occassional nature of these
activities, the annual emission rates are considerably lower.
For the crushing and leveling the emission rate was less than
5 percent, while dumping gave between 5 percent to 30 percent
of the annual waste dump asbestos emissions. Application
of the control measure of bagging and slurrying reduced
dust emissions by more than 95 percent for the dumping
operation.
STABILIZATION OF WASTE PILES
Two options were used to demonstrate the stability of
waste pile emissions controls. They were tested in the field
over a period of seven months. Chemical stabilization was
tested for the active pile. A permanent cover is not recom-
mended for active pile because it is still subject to frequent
additions, vehicular traffic, and movement of waste on the pile.
Inactive waste piles, on the other hand, are sections of the
dump which are filled and have been abandoned. For inactive
piles, a revegetated soil cover was tested.
Recent interest in mineral waste stabilization has been
the source of many new agents for chemical stabilization. A
candidate for field testing was chosen based on limited labora-
tory tests for physical strength and durability. A resin
-201-
-------
stabilizer was selected which ranked near the top in
laboratory tests. Convenience in mixing and ease of appli-
cation were important criteria in the final selection. The
stabilizer was applied on the test pile with a sprinkler at
the rate of 0.^5 1/m3.
In general, it was observed that the chemically treated
pile was stabilized very effectively. The stabilizer adhered
well to the waste particles and reached a steady concentration
level within the surface of the pile which was not removed
after seven months of weathering by heat, cold, wind and rain.
Surface moisture tests showed that the presence of the stabil-
izer did not affect the water retention in the top surface.
The ability to retain water is important since moisture is
very effective in reducing surface emissions.
The inactive test pile was covered with a 30 cm thick layer
of sandy loam from a nearby area. The soil was fertilized
and broadcast seeded with a mixture of species similar to
those native to the area. A straw mulch was spread on the
2
pile surface at a rate of 0.45 kg/m and lightly crimped in
to keep it from being blown away. The straw provided good,
inexpensive protection for the establishment of the vegetative
cover. Moisture for germination and establishment was pro-
vided with a sprinkler system during the initial six weeks.
It was demonstrated that, given a sufficient depth of
soil and adequate fertilization, a stable, vegetative cover
can be developed on the highly alkaline, non-fertile asbestos
waste. A plant cover of 70 percent is regarded as a minimum
-202-
-------
to stabilize a waste pile against erosion by weather action.
In fact, a 95 percent cover was developed. The high pH and
electrical conductivity associated with the asbestos cement
waste can be tolerated if a sufficiently deep soil cover is
used and care is taken to avoid mixing of the waste with the
soil as it is deposited on the pile*
The work described in this paper represents a portion of
a study supported by the Environmental Protection Agency under
Contract Number 68-02-1872.
-203-
-------
FUGITIVE EMISSIONS FROM CHEMICAL FERTILIZER MINING
By
J. C. Ochsner and T. R. Blackwood
Monsanto Research Corporation
Dayton, Ohio
-205-
-------
ABSTRACT
FUGITIVE EMISSIONS FROM CHEMICAL FERTILIZER MINING
This paper summarizes the results of field testing programs to develop emis-
sion factors for dust generated in the storage, loading, and handling of
potash and phosphate rock. The samples collected were analyzed for composi-
tion and particle size distribution, and emission factors were developed both
for the total mass of emissions for the overall process and for unit opera-
tions. Visual observations were utilized to evaluate various control
measures which have been taken over the years by the industry. Data has been
collected from sampling programs other than those conducted by Monsanto
Research Corporation and the results compared and combined to give additional
perspective to the emissions and the extent of control obtainable. The major
obstacles in application of control technology lie in converting existing
facilities to accommodate the dust control equipment. Some of the problems
include door and hopper configuration, rail car height, length of duct runs
from dust area to control device, and the settling and compacting character-
istics of the dust to be handled. Several examples of the latest control
techniques which are effective are described.
-206-
-------
FUGITIVE EMISSIONS FROM CHEMICAL FERTILIZER MINING
J. C. Ochsner and T. R. Blackwood
Monsanto Research Corporation
Dayton, Ohio
Foreword
Monsanto Research Corporation is currently involved in a program aimed at
providing the preliminary data necessary for future planning of EPA action in
developing control technology for emissions from mining operations. Mining
emits air pollutants in a primarily nonpoint manner. These sources are gen-
erally large and diffuse in nature and are often comprised of several fugi-
tive emission points. Consequently, their source strength is quite difficult
to assess accurately.
The area covered by this paper is the mining and beneficiation of phosphate
rock and potash. These processes are responsible for most of the emissions
from the chemical fertilizer raw materials industry. The report addresses
the following questions:
1. What are the sources of air emissions?
2. Are they fugitive or stack emissions?
3. What is the magnitude of the emissions?
4. What control technology is currently in use?
5. What are the future trends in control technology?
Background
Vegetation requires several materials for adequate nutrition. The four major
plant food elements are nitrogen, phosphorus, potassium, and sulfur.1 Of
these only nitrogen, phosphorus, and potassium are needed by the plant in
relatively large quantities. The primary function of the fertilizer industry
is to supply these nutrients.2 Of these three, only the natural minerals of
phosphorus and potassium are of commercial importance. Manufactured nitrates,
-207-
-------
such as urea, are acceptable substitutes. Therefore, a discussion of the air
emissions produced by the mining and beneficiation of materials for the chem-
ical fertilizer industry needs to consider only two minerals. These are
phosphate rock and potash.
Phosphate Rock
Phosphate rock is the common term used for most commercial phosphate raw
materials. H least 90% of the world's annual phosphate rock supply is typi-
cally converted into phosphorus and phosphoric acid intermediates of which at
least 85% of the world supply is annually consumed in fertilizers. U.S. pro-
duction of marketable phosphate rock has grown from 18 million metric tons in
1960 to about 44 million metric tons in 1975, or about 6 percent per year.
Phosphate ore mining and processing is carried out in four different regions
of the United States. These areas and their contribution to the total output
are:
• Florida 78%
• Western States (Montana, Idaho, Wyoming, and Utah) , 12%
• North Carolina 5%
• Tennessee 5%
In Florida the deposits are concentrated primarily in a relatively small area:
a total of about 3h% of Florida's land.3 The phosphate that comes from this
limited area provides approximately one-third of the whole world's production.
Of the two available rocks, pebble and hard rock, pebble phosphate is the
more important. It has a P205 content around 30%. Hard rock has been of
little commercial value because of low phosphate concentration and/or high
mining and processing costs.
The actual ore is a sand/clay/phosphate matrix of about equal portions.
Table 1 gives analyses for P205, CaO, F, and U of Florida phosphate rock.1*
-208-
-------
TABLE 1. ANALYSIS OF PHOSPHATE ROCK: FLORIDA, 35*
Bone Valley formation
Composition
P205
CaO
F
U
Pebble
30.8
43.7
3.4
0.015
Concentrate
32.5
43.2
3.6
0.010
Hawthorne formation3
Pebble
23.5
40.0
2.7
0.009
Concentrate
29.9
44.8
3.3
0.007
aLies beneath the Bone Valley formation.
Most phosphate rock is mined by open-pit methods. Some locations in the West
mine by underground techniques; however, they account for only a minor part
of the total production. The common practice in Florida is to strip the over-
burden with electric-powered draglines and place it in windrows alongside the
cut.
No sampling for the dragline source in Florida has been done. However, some
estimates have been made on this fugitive emission from other source assess-
ment studies on mining.5 Dust emissions from surface mining of coal were
measured by MRC as 0.8 g/kg (0.17 Ib/ton) of overburden removed. Battelle
indicated that overburden removal was the largest emission source at strip
mines and estimated 0.5 k/kg (.1 Ib/ton).^ Engineering Research and Technol-
ogy, upon providing input on the air quality aspects of coal development in
northwest Colorado, proposed an emission factor of 0.024 g/kg (0.048 Ib/ton)
of overburden removal, including a correction for climatic conditions and
control measures (watering) at the mines.7
Overburden removal for phosphate rock mining is believed to be much less of a
fugitive dust .source than in coal mining because much less overburden mater-
ial is handled; the average overburden depths in Florida are about 6 m (20 ft)
versus up to 60 m (200 ft) for coal mining. Also, phosphate rock deposits in
Florida are generally mined in areas where the water table is near the sur-
face. Because the moisture content of the overburden is high, less dust is
produced.
-209-
-------
An emission factor for the overburden removal of phosphate rock definitely
lies toward the bottom of the range (0.048 to 0.10 Ib/ton of overburden)5
offered by PEDCo. PEDCo estimated particulate emissions from dragline opera-
tions at an open pit copper mine in Butte, Montana.8 The emission rate per
ton of ore mined was 0.004 g (0.008 Ib). This excavation area was noted to
be moist and nondusting. Although actual measurements were not taken at phos-
phate rock mines, it is felt that an emission factor of this order of magni-
tude is representative for the mining of the ore.
Since the stripping of overburden and mining of matrix is carried out under
somewhat "wet" conditions, industry does not consider this a source of air
emissions. Therefore, they have no controls on this operation and do not
expect a future need.
After the overburden is removed, the ore is placed into sluice pits where
hydraulic monitors break up the ore. It is subsequently pumped to the bene-
ficiation plant. Each company's beneficiation methods differ slightly and
are dependent on the characteristics of the matrix. In each case, however,
it is a wet beneficiation process and therefore is not a source of air
emissions.
The beneficiated rock is transferred to open storage piles. As this material
is also moist, no controls are used or expected in the future although some
emissions do result from the operations.
Typically, the moist material is conveyed by an underground belt operation to
a rotary dryer. The dust from the driers is handled by cyclone-wet scrubber
combinations. (One operation uses a rotary drier which has a cyclone scrub-
ber followed by twin wet scrubbers.) In this way, the emission becomes a
stack emission. Collection efficiencies of these control devices are greater
than 95%, and no further controls are expected in the future. The material
is sometimes ground at this point. However, the plant in our study did the
grinding under wet conditions. One plant which was visited does not use a
grinding operation at all.
-210-
-------
From the dryer, the material is transferred to storage silos. Control
devices on rock transfer junctions are typically scrubbers. The average col-
lection efficiency is over 95%. It is from the storage silos that the rock
is dropped into railroad hopper cars for transportation to fertilizer manufac-
turing centers. The loading facilities are typical of the phosphate industry.
Loading is accomplished manually from overhead storage by an operator stand-
ing at the roof level of the railroad car. Usually the drop areas have flexi-
ble rubber couplings which deliver the rock to within about 2 ft of the top
of the loading door. This helps direct the rock; however, the problem in the
loading is not dusting caused by the drop, but by the pickup of dust in the
displaced air from the railroad car. This is a dusty operation, and the
industry is working on the problem.
Companies have installed extensive ductwork systems which capture the dust
and exhaust it to a scrubber. However, as noted by one representative of the
industry, it is not always possible to get the operators to use the equipment
properly. It involves clamping enclosures onto the railcar doors tight
enough to capture the dust. One plant uses a mechanized system in which the
ends of eight loading ducts fit over the top of the eight railcar doors. The
displaced air is exhausted to a scrubber. In older facilities, this may not
be possible. At one location, the available headroom between the bottom of
the storage silo and the top of the rail cars is very confining, and cars are
typically not of uniform height or door configuration. Therefore, it was nec-
essary to design a system which would load the cars from an outside hopper
adjacent to the silos, allowing for more working space. In either case, it
appears that the future trend is to contain the fugitive emission, enabling
it to be treated as a stack emission.
The air emission sources of phosphate rock mining and processing are pictured
in Figure 1. The emission factor which corresponds to each source and the
discussion above are summarized in Table 2.
Potash
Agriculture is the major consumer of potash, accounting for about 95% of the
total consumption in the U.S. The percentage is even higher for the rest of
the world.
-211-
-------
Beneficiated
Phosphate Rock Ore
Dragline
Open Storage*
Drying
and
Storage
Beneficiation
Plant
Loading of Railroad Cars
*Emission factor not orovided.
Figure 1. Selected Phosphate Rock Sources
-------
TABLE 2. AIR .EMISSION SOURCES OF PHOSPHATE ROCK MINING AND BENEFICIATION
to
M
w
Dragline
Emission factor,
g/kg (Ib/ton)
Existing control
devices
Effectiveness of
existing con-
trol devices
Future trends of
control
devices
Stripping
of
overburden
.25
(.05)
None
None
Mining
of the
ore
.0004
(.0008)
None
None
Dryi ng
.071
(.142)
Cyclone and
wet scrubbers
>95%
None
Silo
storage
.105
(.142)
Scrubbers
>95%
None
Loading
.161
(.322)
Enclosure
a
Exhaust fugitive
dust to
control device
Total
.362
(.725)
aNot available.
-------
The United States' portion of production in North America has remained about
the same in volume (2.3 million metric tons K20 in 1975), but the share of
production has decreased from 100 percent in 1960 to about 30 percent in 1975,
as the vast resources of Saskatchewan were brought onstream.
Potash, like phosphate rock, is also produced in four different geographical
areas. However, four different processing methods are used. These areas,
their types of operation used, and their share of the total output are pre-
sented in Table 3.
TABLE 3. U.S. PRODUCTION AREAS OF POTASH9
Type of operation
Total
percentage
of output
New Mexico
Carlsbad
Utah
Great Salt Lake
Utah
Moab
California
Trona (Searles Lake)
Conventional mining, flotation. (One
company dissolves ore in brine,
(crystallizers)
Solar evaporation of natural or lake
brines, flotation
Underground solution brine, solar
evaporation, flotation
Steam evaporation of lake brine,
vacuum crystallizers
81
6
6
7
The area of interest for this report is in New Mexico due to its large share
of the output. Sylvinite ore, a mechanical mixture of sylvite (KC1) and
halite (NaCl), is a natural ore for potassium.10 This ore is found in commer-
cial quantities in the Permian Basin region near Carlsbad, New Mexico.
Table 4 shows the typical composition of potash ore in New Mexico.
Potash ores are usually processed in refineries adjacent to the mines.4 The
bedded deposits of New Mexico are all mined underground in a manner similar
to that used for coal. There are two processes employed in the Carlsbad area
refining facilities. One is used for recovery of potassium chloride from
sylvinite ore and the other for converting langbeinite ore to potassium
sulfate.
-214-
-------
TABLE 4. COMPOSITION OF POTASH4
Area Composition, 5
Carlsbad - New Mexico Polyhalite
KzSO,,
MgS04
CaSO^
Anhydrite
CaSOi+
NaCl
H20
Fe203
MgC03
Total mineral
Polyhalite
Halite
Anhydrite
Others
Eddy County - New Mexico Langbenite
K20
MgO
S03
Na20
Insoluble in water
f
j
21.93
15.15
34.29
8.17
12.91
4.53
2.29
0.73
75.9
12.9
8.2
3.0
22.37
19.15
57.44
0.48
0.05
Most of the potassium chloride is recovered from sylvinite ore by flotation
methods. The essentials of this process are to treat the ore with a hydro-
phobic material (e.g., an aliphatic amine) which will selectively coat one of
the constituents {the KC1) of the ore. Air is then bubbled through the
slurry, and the air bubbles attach themselves to the coated particles and
float them to the surface, while the uncoated particles sink. A process flow-
sheet is given in Figure 2. On a large scale and particularly with high-
analysis ores, the flotation process is much cheaper than one involving dis-
solution and crystallization. However, available resources of the high-grade,
-215-
-------
BRINE
ORE'
I
NJ
1
__ BRINE RECYCLE
* T
1 FLOTATION
WATER CHEMICALS
. 1 i'l
CRUSH
AND
GRIND
— »
DESLIME
AND
SEPARATE
— »
FLOTATION
— »
VENT
DEWATER
— ^
DRY
_^^
SCREEN
LEGEND:
ALTERNATE
ROUTES
»PRODUCT
BRINES
TO WASTE
OR
TO RECYCLE
SLIMES
TO
WASTE
T
TAILINGS
WASTE
AND
BRINE
Figure 2. Potassium Mining and Processing from Silvinite Ore
-------
low-clay ores that have been mined and concentrated during the past 40 years
are declining in grade, while the ore to be mined in the future has become
more difficult to beneficiate because of increased clay content which inter-
feres with flotation.
Research by industry and government has been in progress for several years to
fashion systems involving flotation and Teaching-crystallization processes
for treatment of the low-grade resources. One plant currently uses a
Teaching-crystallization process, while the remaining plants that process
sylvinite ore use flotation.
Langbeinite is a natural sulfate of potassium and magnesium, K2Mg2(SOi+)3,
usually intermixed with sodium chloride. Two facilities at Carlsbad process
langbeinite ore. This ore is mixed, crushed, and dissolved in water to which
potassium chloride is added. Partial evaporation of the solution produces
selective precipitation of K2S0lt which is recovered by centrifugation or fil-
tration from the brine liquor, dried, and sold. A simplified process flow-
sheet is given in Figure 3.
The air emission data for potash refinery activity near Carlsbad, New Mexico
is presented in Table 5. Some explanation of the table is necessary. The
seven companies (A-G), which operate in the vicinity of Carlsbad, New Mexico,
are listed. The two sets of emission factors, based on production and capac-
ity, are presented to equalize any production irregularities (such as E's cut-
back on production due to construction). A comments column is provided to
explain the deviation among the numbers. Except where designated, the emis-
sion factors represent the total particular emission of the refineries.
These emission factors account for operations such as rock transfer, grinding,
screening, drying, bagging, loading, and (in B and F's case) evaporation.
The major source of emissions is the drying operation, which is employed at
all locations. At present, the use of dry cyclones is the most evident form
of control technology. They are in use not only at drying locations, but at
other points of the process such as milling, screening, and bagging. A feel
for the amount of emission caused by drying can be gained by looking at
A and E's emission factors.
-217-
-------
00
I
POTASSIUM
CHLORIDE
WATER
DISSOLVER
I
BRINE LIQUOR RECYCLE
FILTRATION
LANGBEINITE ORE*
I
REACTOR
CM TDATinW 1
rlUi KAI lUW
WASTE MUDS
1
SOLID WASTE
CLARIFIER
WATER I r
VAPOR IT
EVAPORATOR
1
PARTIAL
EVAPORATION
MAGNESIUM
• CHLORIDE
CO-PRODUCT
POTASSIUM
SULFATE
PRODUCT
Figure 3. Langbeinite Mining and Processing
-------
TABLE 5. NEW MEXICO POTASH PARTICULATE EMISSION DATA
I
ro
Company
A
B
C
0
E
F
G
Annual l l
emission
rate,
g/s
(Ib/s)
22.58
(0.0497)
318.66
(0.7019)
10.86
(0.0239)
62.82
(0.1384)
26.8
(0.0590)
87.51
(0.1928)
91.97
(0.2026)
Average
K20
content, %
40
8
16
18
14
11
23
Ore processed,
1975
production,
kg/yr
(tons/yr)
1.655 x 109
(1,820,000)
3.625 x 109
(3,987,500)
1.932 x 109
(2,125,000)
1.944 x 109
(2,264,706)
0.468 x 109
(514,286)
1.653 x 109
(1,818,182)
2.202 x 109
(2,422,727)
Emission
factor
based on
production,
g/kg
(Ib/ton)
0.430
(0.860)
2.772
(5.544)
0.177
(0.354)
1.019
(2.038)
1.806
(3.612)
1.670
(3.340)
1.318
(2.636)
Ore processed,
current10
capacity,
kg/yr
(tons/yr)
2.121 x 109
(2,333,333)
4.318 x 109
(4,750,000)
1.989 x 109
(2,187,500
2.778 x 109
(3,055,555)
1.753 x 109
(1,928,571)
2.149 x 109
(2,363,636)
2.273 x 109
(2,391,304)
Emission
factor
based on
capacity,
9/kg
(Ib/ton)
0.336
(0.672)
2.327
(4.654)
0.172
(0.344)
0.713
(1.426)
0.482
(0.964)
1.284
(2.568)
1.334
(2.668)
Comments
Dryer emissions only
Includes a separate salt operation
Uses a leaching operation
_a
Dryer emissions only
Cutback due to construction
_a
a
aNot applicable.
Time basis: 31,536,000 seconds/yr
-------
Another major source of participate emission appears at the locations where
langbeinite ore is processed by an evaporation step. Ozark submerged-
combustion evaporators are in use at these plants. They consist of a tank to
hold the liquid, a burner and gas distributor that can be lowered into the
liquid, and a combustion-control system. One difficulty frequently
encountered in the use of submerged-combustion evaporators is a high entrain-
ment loss. Prior particle size measurements on a plant's Ozark evaporator
showed 100% of the particles to be less than 7 microns. These would all
remain suspended. A study is currently in progress to assess the application
of Monsanto's Brink® fine particle eliminators to the problem.
According to the New Mexico Environmental Improvement Agency, it would appear
that most all of the existing potash refineries will need to control their
present emissions. At this point, the agency is investigating different
types of control technology and their costs.11
Sampling and Analytical Methodology
A Gaussian plume equation was used to estimate the fugitive dust source
strength from ground level, ambient hi-vol air samplers. Since the disper-
sion equation gives a time-averaged description of the concentration field, a
number of receptors were used, as shown in Figure 4. The following rationale
applies: (1) the position of the sampler determines ambient dust levels, and
(2) wind direction and velocity were not constant and affect time-averaging
results. The positioning of samplers eliminated the need to correct for time
varying concentrations.
Sampler 1 was the primary source composition strength estimator and was
located as close to the source as possible. Sampler 3 was required for cor-
relations with downwind power law decay. Samplers 2 and 4 were required for
correlations with lateral dispersion estimates. Sampler 0 was used as a
blank to compensate for other dust sources upwind of the site under
investigation.
A portable mechanical meteorological station gave the remaining parameter,
wind speed. The azimuth was monitored throughout the sampling period. The
-220-
-------
Reference
Net Wind Direction
SOURCE
Angle
Downwind
Samples
Figure 4. Sampler Positions
-221-
-------
filter from Sampler 1 was analyzed by X-ray fluorescence techniques and com-
pared to source materials for composition. A particle count was also made
via optical microscopy to check for fibers and to give a different perspec-
tive to particle size distribution.
The sampling results for phosphate rock processing are summarized in Table 6.
In the comments column is a note regarding wind change. Excessive wind shift
was defined, through experience, as a change in wind direction greater than 20C
Conclusion
A comparison of the air emissions between the phosphate rock and potash indus-
tries shows that the total emission factors for potash are 2 to 4 times
greater than for phosphate rock. This difference may be explained by under-
standing that the pressure to control emissions has been somewhat greater on
the phosphate rock industry due to their large production and concentration
in one area. However, it appears that most potash refineries will have to
further control their emissions in the near future.
Both phosphate and potash emissions are relatively nontoxic; in fact, they
are generally considered as nuisance dusts. The danger is particle size.
This does not present a problem, however, in the phosphate industry and is
currently being dealt with in the potash industry.
Finally, a major difference between the two industries is the methods of con-
trol. The potash industry's primary control device is the dry cyclone. The
phosphate rock industry uses not only that control but also secondary stages
of control in the form of wet scrubbers. It is believed that, with the excep-
tion of the Ozark evaporators, the necessary control technology exists for
the potash industry.
-222-
-------
TABLE 6. PHOSPHATE ROCK PROCESSING
Material
Source
sample,
Run 3. % Run 5. % Run 6, %
Phosphorus
Sodium
Magnesium
Aluminum
Silicon
Sulfur
Potassium
Calcium
Titanium
Chlorine
Chromium
Manganese
Iron
Fluorine
32.6
0.40.
0.09
2.11
5.45
_a
0.14
41.4
1.10
_a
_a
0.11
2.47
3.82
25.5
_a
0.2
2.5
4.1
_a
0.08
60.0
0.16
_a
_a
_a
1.6
5.8
25.6
a
0.73
2.93
13.9
_a
0.07
49.8
0.15
_a
_a
_a
1.46
5.1
22.4
_a
0.75
3.19
5.3
_a
0.11
59.5
0.21
_a
_a
a
2.13
6.4
Minutes Total particulate Standard Confidence
run emission rate deviation limits
Run 1
Run 2
Run 3
Run 4
Run 5
Run 6
232
265
350
320
245
239
169.3 Ib/hr
_b
82.16 Ib/hr
250.2 Ib/hr
55.6 Ib/hr
126.6 Ib/hr
±67.6 Ib/hr
_b
±40.82 Ib/hr
±166.6 Ib/hr
±5.35 Ib/hr
±86.1 Ib/hr
±124.4
_b
±75.0
±306.1
±9.83
±158.2
Particle size
distribution
Emission
factor
Comments on run
Run
Run
Run
Run
Run
Run
1
2
3
4
r
6
0.
10. 11% less °'
than 7 y 0.
0.
0.
434
210
639
142
322
Ib/ton
_b
Ib/ton
Ib/ton
Ib/ton
Ib/ton
_c
Excessive wind
Storage
Total
Drying
_c
change
aNone detected. No good. Not applicable.
Note: Compositions were calculated after drying.
-223-
-------
BIBLIOGRAPHY
1. Edmiston, D. C. Fertilizer Minerals. Mining Congress Journal,
28(ll):25-29, 1976.
2. Kirk-Othmer Encyclopedia of Chemical Technology, Second Edition, Vol. 9.
New York, New York, 1966. 901 pp.
3. Phosphate Feeds You. Florida Phosphate Council, Inc., Lakeland, Florida.
14 pp.
4. Katari, V., et al. Trace Pollutant Emissions from the Processing of
Nonmetallic Ores. U.S. Environmental Protection Agency, Office of
Research and Development, NERC-RTP, Control Systems Laboratory, Research
Triangle Park, N.C. EPA 650/2-74-122. November 1974. 160 pp.
5. Energy Alternatives: A Comparative Analysis. University of Oklahoma,
Norman, Oklahoma. (Prepared for Council on Environmental Quality,
Energy Research and Development Administration. U.S. Environmental Pro-
tection Agency, Federal Energy Administration, Federal Power Commission,
U.S. Department of the Interior, and National Science Foundation,
May 1975.
6. Air Pollutant Emissions in the Northwest Colorado Coal Development Area.
Environmental Research and Technology, Westlake Village, California.
1975.
7. PEDCo - Environmental Specialists, Inc. Evaluation of Fugitive Dust
Emissions from Mining. U.S. Environmental Protection Agency, Cincinnati,
Ohio. Task No. 36. April 1976. 78 pp.
8. Emission Estimates for the Berkeley Pit Operations of Anaconda Company.
PEDCo - Environmental Specialists, Inc., Cincinnati, Ohio. (Prepared
for Montana Air Quality Bureau, Helena, Montana, September 1975.)
9. Johnson, A. B., et al. Beneficiation of High Clay Potash Ore by Flota-
tion. Bureau of Mines, Technical Progress Report No. 41, September 1971.
12 pp.
10. Keyes, W. F. Potash. In: Mineral Facts and Problems, Bulletin No. 667.
U.S. Department of the Interior, Washington, D.C., 1975. 15 pp.
11. Nicholson, B. R. Air Quality Estimates of the New Mexico Potash
Industry. Air Quality Division, State of New Mexico Environmental Pro-
tection Agency, Santa Fe, New Mexico, August 1976. 59 pp.
-224-
-------
FUGITIVE EMISSION CONTROL IN THE IRON & STEEL INDUSTRY
By
Arthur G. Nicola, Manager
Air Pollution Control Systems
Pennsylvania Engineering Corporation
Pi ttsburgh, Pennsylvani a
-225-
-------
FUGITIVE EMISSION CONTROL IN THE IRON & STEEL INDUSTRY
by
Arthur G. Nicola, Manager
Air Pollution Control Systems
Pennsylvania Engineering Corporation
Pittsburgh, Pennsylvania
Abstract
Fugitive Emissions related to the various steps in the Iron and Steel-
making process have been largely ignored or tolerated in the past mainly
due to light concentrations as well as the difficulty of collecting these
emissions. However, today with larger production units creating more
concentrated emissions, they can no longer be ignored. With present air
pollution control regulations, virtually no visible emissions are allowed
from the buildings.
Since air pollution is associated with practically all steps of Iron and
Steel Production, the related air pollution control equipment is an
important factor in all of these operations. A production process may
become obsolete if it is not capable of meeting today's stringent air
pollution control requirements.
Because of the emphasis by regulatory authorities today for complete
emission control during the Basic Oxygen Steelmaking Process, the limited
scope of this paper will concentrate on Fugitive Emission Control in this
area, but will also encompass the other main metallurgical production
processes.
-226-
-------
FUGITIVE EMISSION CONTROL IN THE IRON & STEEL INDUSTRY
INTRODUCTION
FUGITIVE EMISSION CONTROL DURING THE BASIC OXYGEN STEELMAKING PROCESS
Hot Metal Charging
Scrap Charging
Tapping
Slagging
Puffing During Oxygen Blow
HOT METAL RELADLING
FLUX HANDLING
ELECTRIC ARC FURNACE
Charging and Tapping Emission Control
Furnace Aisle Enclosure
CONCLUSION
By: Arthur G. Nieola, Manager
Air Pollution Control Systems
Pennsylvania Engineering Corp.
Pittsburgh, Pennsylvania
-227-
-------
Fig. 1 Hot Metal Charging for the Basic Oxygen Process.
Fig. 2 Schematic Arrangement and Design of EOF Furnace Enclosure.
Fig. 3 Schematic Arrangement Showing Components Effecting Fugitive
Emission Control with Furnace Enclosure.
Fig. 4 Tapping Emission Control with BOF Furnace Enclosure.
Fig. 5 Recent BOF Shop Installation with Fugitive Emission Control.
Fig. 6 Hot Metal Charging Emission Control with Furnace Enclosure.
Fig. 7 Schematic Arrangement - Ladle Hood for Reladling Emission
Control.
Fig. 8 Reladling Emission Control - Pouring through a Slot on Top
of the Ladle Hood.
Fig. 9 Reladling Emission Control - Pouring through a Vertical Slot
in the Ladle Hood.
Fig. 10 Schematic Arrangement - Electric Furnace Pollution Control
System with Enclosed Furnace Aisle.
Fig. 11 Furnace Enclosure for Emission Control on Electric Arc
Furnace.
-228-
-------
FUGITIVE EMISSION CONTROL IN THE IRON & STEEL INDUSTRY
INTRODUCTION
Fugitive Emissions related to the various steps in the Iron and Steel-
making process have been largely ignored or tolerated in the past mainly
due to light concentrations as well as the difficulty of collecting
these emissions. However, today with larger production units creating
more concentrated emissions, they can no longer be ignored. With
present air pollution control regulations, virtually no visible emissions
are allowed from the buildings.
Since air pollution is associated with practically all steps of Iron
and Steel Production, the related air pollution control equipment is
an important factor in all of these operations. A production process
may become obsolete if it is not capable of meeting today's stringent
air pollution control requirements.
Because of the emphasis by regulatory authorities today for complete
emission control during the Basic Oxygen Steelmaking Process, the
limited scope of this paper will concentrate on Fugitive Emission
Control in this area, but will also encompass the other main metallurgical
production processes.
Total Fugitive Emission Control can be broken down into two (2) main
components:
(1) Collection of the Fugitive Emissions.
(2) Gas Cleaning after the fumes have been collected.
With present technology, the alternatives available for collecting the
emissions generated during the main metallurgical production processes
are limited and have many disadvantages. The ideal solution is to
capture the emissions at their source. However, this is not always
possible.
-229-
-------
In new installations today where large volumes are involved, cleaning
of Fugitive Emissions is done mainly with Dry Bag Filter ("Baghouse").
However, when it is possible to capture the emissions at their source
with relatively low volumes or heat is a problem, Electrostatic
Precipitators or High Energy Scrubbers may be the best solution for
Gas Cleaning.
-230-
-------
FUGITIVE EMISSION CONTROL DURING THE BASIC OXYGEN STEELMAKING PROCESS
Overall, a number of sources exist which generate Fugitive Emissions
during the Basic Oxygen Process. Collection of these Fugitive Emissions
has become more of a concern than the primary Gas Cleaning itself. This
concern for Fugitive Emission Control is based on the difficulty of
efficient fume collection at the emission source, since once the fume
escapes into the building, it is almost impossible to control.
At the present time, twenty-seven (27) operating BOF/Q-BOP facilities
in the United States are in violation of current air pollution control
regulations for Fugitive Emissions and must bring these emissions under
control in the near future.
In an effort to comply with current Air Pollution Control Regulations in
the U.S., many systems of different design have been installed and are
in operation today with varying degrees of success. They include partial
or complete building evacuation by means of a canopy hood, local hoods
close to the emission source, partial or full enclosures and various
combinations of these systems. To successfully control Fugitive
Emissions with a minimum expenditure, they must be collected at the
source and not allowed to escape into the building.
Major sources of Fugitive Emissions in the Basic Oxygen Steelmaking
Process are:
1. Scrap Charging
2. Hot Metal Charging
3. Tapping and Ladle Alloy Additions
4. Slagging
5. Puffing During Oxygen Blow
6. Hot Metal Reladling
7. Flux Handling
-231-
-------
Other sources such as Ladle Transporting and Teeming also contribute
to the Fugitive Emissions, but are minor in nature since they are low
in volume and dissipate readily. They normally do not contribute to
monitor emissions in the EOF Shop.
HOT METAL CHARGING
The worst condition occurs during charging of Hot Metal into the furnace
which already contains scrap. The Hot Metal and the effect of the Hot
Metal on the scrap both contribute to emissions during this period.
Figure 1 shows the typical fumes generated when charging Hot Metal into
the furnace. Test results have shown that fumes generated at this time
are composed of approximately 35% Iron Oxide, 30% Kish and 35% others,
with particulate size less than 100 microns and an approximate emission
rate of 0.3 - 0.4 Ibs per ton of Hot Metal poured.
Since the prime factors controlling fume generation during this period
are the condition of scrap and the rate of Hot Metal poured, operating
practice is a big factor in the amount of fumes generated.
SCRAP CHARGING
Scrap Charging itself is not a primary source of pollution, however,
when the scrap contains a foreign substance such as dirt, paint or oil,
it becomes a major pollution source when the Hot Metal is added. The
quantities of pollution generated during Scrap Charging itself are minor
compared to the fumes generated during Hot Metal Charging and can be
minimized by scrap selection.
TAPPING
A dense fume normally results from the Tapping Operation itself and
with ladle additions such as Ferro Silicon and Ferro Manganese, the
magnitude of the problem increases. Fume composition here is dependent
-232-
-------
on the alloys employed but generally would consist of 75% Iron Oxide
with particle size less than 100 microns and an average emission rate
of 0.15 - 0.20 Ibs per ton.
SLAGGING
Emissions from the Furnace Slagging Operation can be a major problem
depending on the type of steel being produced which influences the
slag volume and emission rate at this time. The method of killing the
slag is also a factor which influences the emissions during the Slagging
Operation. Fumes are generated inside the furnace at this time and
continue with the Slag as it falls inside the Slag Pots below the
charging floor. These emissions normally have a particle size less
than 100 microns.
Slagging fumes are relatively cold and do not have the necessary thermal
energy to cause the fumes to rise into an Overhead Canopy Hood making
them difficult to capture.
PUFFING
Puffing during the Oxygen Blow occurs intermittently. When this occurs,
generally small quantities of fume are emitted around the mouth of the
furnace. This is not a major source of Fugitive Emissions, however, it
does contribute to the total emission problem in the shop.
With present technology, the alternatives available for collecting the
Oxygen Steelmaking Fugitive Emissions are:
1. Complete or Partial Building Evacuation.
2. Local Hoods (Fixed or Movable).
3. Furnace Enclosures.
Canopy Hoods for partial or complete building evacuation systems are
operating in a number of shops with varying degrees of success. In
-233-
-------
most cases, this method is normally considered as a last resort due to
the high capital expenditure and operating cost involved. These systems
usually are given consideration by maintenance and operating personnel
since they require minimum maintenance and do not restrict operating
practices. However, disadvantages are the high volume requirements
to achieve necessary collection efficiency due to the distance of the
hoods from the emission source and the adverse effects of cross drafts
in the shop which affect collection efficiency. Volume requirements
here would be in excess of 1,000,000 CFM for partial or complete
building evacuation.
The use of Local Hoods for collecting Charging Emissions is more
effective than the Canopy Hood located in the Building Trusses since
Local Hoods are in closer proximity to the emission source. However,
relatively high volumes are still required to effectively collect
these emissions in the manner. Although Local Hoods are 'more efficient
than Canopies, they are undesirable from the maintenance and operating
point of view.
The alternative of collecting Fugitive Emissions by means of a Furnace
Enclosure appears to be the most practical solution since it allows
collection of emissions at the source and prevents their escape into
the atmosphere. With properly designed Furnace Enclosures, it is
possible to effectively control Scrap Charging, Hot Metal Charging,
Furnace Tapping, Ladle Alloy Additions, Furnace Slagging and Puffing
Emissions with low volumes. At the present time, systems of this type
are doing an effective job of Fugitive Emission Control with volumes
of approximately 350,000 ACFM - 400,000 ACFM.
Figure 2 shows a typical schematic arrangement and design for the EOF
Furnace Enclosure where the enclosure essentially forms a gas tight
seal when the bi-parting doors are closed. Since the Furnace Enclosure
extends below the charging floor, the only openings are for the ladle
-234-
-------
car. However, these openings can be effectively reduced by the
addition of a vertical shield on the end of the ladle car, which
will also increase the efficiency of the Furnace Enclosure.
Design of the Furnace Enclosure is based on sound engineering principles
and actual testing, taking into consideration the volume of fumes
generated inside the furnace due to the reaction of pouring hot metal
into the furnace, as well as the velocity and temperature of the fumes
leaving the mouth of the furnace.
Since the furnace is charged at a known angle, it is possible to
calculate the horizontal and vertical components of the evolved
emissions, as shown in Figure 3 and determine the necessary volumes and
velocities required to contain these emissions inside the enclosure
and prevent their escape into the atmosphere.
Enclosures of this type were developed initially to control emissions
from the Q-BOP Process. However, the original enclosure designs have
been modified to effectively collect the Charging, Tapping and Slagging
Emissions generated during the Oxygen Steelmaking Process, in addition
to Puffing Emissions during the Oxygen Blowing Period.
Original enclosure designs did not include a capability for Charging
Emission Control, since a secondary hood was not included. To correct
this situation, a secondary hood outside of the enclosure was added
for collecting Charging Emissions. However, this proved to be in-
efficient due to the location of the secondary hood and insufficient
volume for efficient fume collection. Today's improved design in-
corporates the secondary hood inside the Furnace Enclosure with
sufficient volume for efficient charging emission control.
For controlling emissions during the blow and tapping period, the
enclosure doors are closed, essentially forming a gas tight enclosure
while the fumes are evacuated from the enclosure through the main or
-235-
-------
secondary hood. Figure 4 shows the effective collection of tapping
emissions at this time. During the charging, the bi-parting doors are
opened while charging scrap or pouring hot metal and the fumes are
collected through the secondary hood located inside the enclosure
directly above the furnace mouth.
With a properly designed Furnace Enclosure, it is possible to collect
secondary emissions from the Basic Oxygen Process with approximately
90% efficiency, provided the charging of hot metal into the furnace
is done at a controlled rate and the scrap is relatively clean.
One of the most recent EOF installations with effective Fugitive
Emission Control Facilities (Figure 5) incorporates individual Furnace
Enclosures over two (2) 350 metric ton vessels. The main Gas Cleaning
Systems, consisting of two (2) Baumco Suppressed Combustion Systems
rated at 295,000 ACFM at 164op, are also utilized for secondary emission
control and at the present time appear to be doing an effective job in
controlling the Fugitive Emissions associated with the EOF vessel itself
(Figure 6) .
The total system here has been designed to insure that sufficient
capacity is available at the secondary hood for Fugitive Emission
Control when required.
Practically all new BOF/Q-BOP vessels that have been installed in the
U.S. in the past two (2) years have included a partial or full Furnace
Enclosure for Fugitive Emission Control. At the present time, there
are five (5) operating BOF/Q-BOP Shops utilizing the Furnace Enclosure
for Fugitive Emission Control with three (3) additional shops under
construction. At the same time, Furnace Enclosures are being installed
in operating EOF Shops without interrupting normal melting operations.
Since the original enclosure designs had many deficiencies, these
systems are operating today with varying degrees of success. Present
-236-
-------
enclosure design incorporates a secondary hood inside the Furnace
Enclosure with sufficient volume for Fugitive Emission Control, capable
of effectively collecting the Oxygen Steelmaking Fugitive Emissions
under controlled conditions.
HOT METAL RELADLING
Hot Metal Reladling is another primary source of Fugitive Emissions in
the Oxygen Steelmaking Process. The fumes generated at this time
consist of 55% Iron Oxide less than 3 microns, 42% Graphite greater than
75 microns and 3% others with an approximate emission rate of 0.25 Ibs per
ton of hot metal. With today's technology, collection of reladling
emissions is not a problem since it is possible to utilize a variety
of canopy or local hoods in addition to close fitting ladle hoods
depending on the arrangement of the reladling facility.
The main factor affecting the collection of reladling emissions is the
distance of the hood from the ladle mouth as well as the rate of pouring
hot metal. In actual practice, since the handling time remains almost
constant regardless of the amount of hot metal handled, the volume
required to control reladling emissions with a canopy or local hood is
proportional to the volume of hot metal.
In contrast, utilization of a close fitting refractory lined ladle
hood for collecting the reladling emissions allows fume collection at
the source. The close fitting refractory lined ladle hood shown in
Figure 8 utilizes the "air seal" principle where the fumes are drawn
through a pouring slot in the hood and the gap between the hood and
the ladle with sufficient velocity to prevent fumes from escaping into
the atmosphere. Figure 9 shows a reladling fume control system with
the hot metal poured through a slot on top of the ladle hood.
-237-
-------
Actual efficiency of the ladle hood is demonstrated by the installation
shown in Figure 10 where the reladling emissions for a 350 metric ton
furnace are effecively collected with 125,000 ACFM. In this design,
the hot metal is poured through a vertical slot in the movable ladle
hood which serves two (2) reladling stations. The Gas Cleaning here
is accomplished by a Baumco High Energy Scrubber which is also utilized
for desulphurization emission control. In contrast, it is estimated
that canopy or local hoods would require volumes in excess of 300,000
ACFM to effectively collect these emissions.
FLUX HANDLING
Flux Handling also contributes to the total Fugitive Emissions in the
EOF Shop, but these emissions are minor in nature and easily controlled.
The most common solution to this problem is to enclose the belt con-
veyors in the monitors with exhaust ducts placed at intervals along the
belt and at each transfer point. A nominal volume of 500 - 600 CFM per
foot of belt width should be adequate to effectively collect the fumes
generated during Flux Handling. The fumes can then be exhausted to a
small package Baghouse or tied into an existing Gas Cleaning Source.
ELECTRIC ARC FURNACE - CHARGING & TAPPING EMISSION CONTROL
Development of the Electric Arc Melting Process over the years has
resulted in larger furnaces and increased emissions from the furnace.
With today's requirements for controlling all emissions and the
necessity for large volumes to collect the scrap charging and tapping
emissions, the Dry Bag Filter ("Baghouse") is primarily used for this
application. This allows the same system to be used for primary
emission control as well as secondary emission control.
In the past there has been little success in collecting these emissions.
However, recent developments in collecting Electric Arc Furnace charging
-238-
-------
and tapping emissions include enclosing the furnace aisle area, making
it possible to contain the fumes in the enclosed furnace aisle allowing
the Fugitive Emissions to be collected with smaller gas volumes than with
the conventional overhead canopy hood method.
Figure 10 shows an arrangement in which the furnace is installed in a
fully enclosed furnace aisle with one overhead crane traveling inside
for scrap charging and a transfer car with ladle for tapping. The only
openings in the enclosure are at yard elevation for the ladle and scrap
bucket transfer cars.
Recent developments include total enclosure of the arc furnace as shown
in Figure 11 in order to contain the charging fumes, allowing their
collection with a minimum volume. This also provides for sound control by
insulating the enclosure. In this case, tapping fumes are collected in a
close fitting hood, using a common gas cleaning system for both charging
and tapping emission control. The method makes direct evacuation or
side draft hoods for melting emission control unnecessary.
-239-
-------
CONCLUSION
Technology for Fugitive Emission Control has lagged behind the technology
developed for main Gas Cleaning Systems in the Iron and Steel Industry
because the Fugitive Emissions were considered minor in comparison to
the emissions generated during the actual melting periods, and the fact
that Regulatory Authorities were willing to tolerate these emissions while
concentrating on other more important areas. As a result, at the present
time, system applications for Fugitive Emission Control have not reached
the same degree of reliability.
This gap in technology is gradually closing due to current pressure by
the Regulatory Authorities to control all Fugitive Emissions. This is
evidenced by the evolution of the total Furnace Enclosure for Oxygen
Steelmaking and Electric Arc Furnaces, and development of the close
fitting ladle hood for controlling reladling emissions, both of which
are capable of effectively controlling Fugitive Emissions with minimum
volumes and cost.
Little has been done in the U.S. to develop new technology for Blast
Furnace Cast House Emission Control. However, we can expect more
activity in this area as the Regulatory Authorities increase pressure
for total control of Blast Furnace Emissions.
Fugitive Emission Control that will be acceptable to the Iron and Steel
Industry requires a technology that offers the most effective and
economical type of Fugitive Emission Control from the standpoint of
working environment which will allow these emissions to be collected
at their source and prevent their escape into the atmosphere.
-240-
-------
I
to
Figure 1: Rot Metal Charging for the Basic Oxygen Process,
-------
BUMPER
SECONDARY HOOD
HOT METAL CHARGING LADLE
FURNACE CHARGING DOORS
(Retractable)
SLAG POT
NS
NJ
I
X-X-X /£\ -jfV
*S \ it ~^v >•;
\ ' i
\ I i
\ 11 1
\
\ '
,.-.A.i-L___-.,
i ,-% ;
\ o /
\ /
1 1
^••--
^f
. — -^**
r
i
t
t
i
i
__A__
;' 5_>
Ci-r /
i \ /
\ \ /
» /
1 -1 &&
— • — ' — " — iys
n /
"*•- I
i
r
i
WATER COOLED HOOD
HOOD TRANSFER CAR
ADJUSTABLE SKIRT
TAPPING EMISSIONS DUCT
SEAL RING
FURNACE ENCLOSURE
OPERATING FLOOR
TEEMING LADLE
!T_T_I g
-v-^-rJ
SHOP AIR INDRAFT
DURING SLAGGING &
TAPPING
•TT>
FUGITIVE EMISSION CONTROL WITH FURNACE ENCLOSURE
Figure 2: Schematic Arrangement and Design of EOF Furnace Enclosure.
-------
I
N3
JN
(jO
I
FURNACE ENCLOSURE
Figure 3: Schematic Arrangement Showing Components Effecting Fugitive
Emission Control with Furnace Enclosure.
-------
K5
-p-
-(>-
I
Figure 4: Tapping Emission Control with EOF Furnace Enclosure.
-------
Ul
I
Figure 5: Recent EOF Shop Installation with Fugitive Emission Control.
-------
•n •
Figure 6: Hot Metal Charging Emission Control with Furnace Enclosure
-246-
-------
TORPEDO CAR
LADLE HOOD
IsJ
•P-
HOT METAL LADLE
RELADLING EMISSION CONTROL
Figure 7: Schematic Arrangement - Ladle Hood for Reladling Emission
Control.
-------
• -•-••....:.•:• •• I
• • • • -"••'"•:":- ' 1
1 :,• ^ •
I • 1
Figure 8: Reladling Emission Control - Pouring through a Slot on Top
of the Ladle Hood.
-248-
-------
I
to
Figure 9: Reladling Emission Control - Pouring through a Vertical Slot
in the Ladle Hood.
-------
,^:> '
:*' . '•'".:••• •;'; -. "• . ..'•,•<' .'• • . \v.
- - > '
.
.-
.
..,v,.-.-:.'.';
/.••'•••->vi.»; Ml
-u't>-
• •».-.;'. ' I • ••••<••• ]
J ( •
TT SIT A p I
" > ; • i ;..'v :-';•;:•/? v- - :;
o
i
. - -: • : •. • I
-«—**• I 1
,, ;.„ • | ;.,,,,; -
! ', '. ••_.:... , ,.
'. M fti •: H\
•fit - | I , 1
,*
Figure 10: Schematic Arrangement - Electric Furnace Pollution Control
System with Enclosed Furnace Aisle.
-------
Figure 11: Furnace Enclosure for Emission Control on Electric Arc
Furnace.
-------
Tuesday Afternoon - May 243 197?
SESSION IV: WATER
Chairman: Charles E. Billings, Ph.D., Consultant
Page
STORMWATER: ONE STEEL MILL'S EVALUATION 255
John W. Luton* and William B. Chadick, Armoo Steel
Corporation
CHARACTERIZATION OF COAL PILE DRAINAGE 269
Doye B. Cox and R. J. Ruane*, Tennessee Valley
Authority
MEASUREMENT OF NON-POINT SOURCES FROM A COAL-FIRED UTILITY AND 299
THE IMPACT ON RECEIVING WATERS
Gordon T. Brookman* and Willard A. Wade III, TRC -The
Research Corporation of New England
THE DEVELOPMENT OF A MATHEMATICAL MODEL TO SIMULATE INDUSTRIAL 337
NON-POINT SOURCE POLLUTION
James J. Binder* and Gordon T. Brookman, TEC - The
Research Corporation of New England
* indicates speaker
-253-
-------
STORMWATER: ONE STEEL MILL'S EVALUATION
By
John W. Luton, Associate Pollution Control Engineer
and
William B. Chadick, Senior Pollution Control Engineer
Armco Steel Corporation
Houston, Texas
-255-
-------
STORMWATER: ONE STEEL MILL'S EVALUATION
John W. Luton, Associate Pollution Control Engineer
William B. Chadick, Senior Pollution Control Engineer
Pursuant to effluent discharge permits, Armco Steel's Houston Works
has conducted an extensive stormwater survey. This survey has proven to be
the first report of its nature in the steel industry. The scope of the
survey deals not only with quantity, but also quality of direct stormwater
runoff.
Results of the survey tend to show that stormwater quality from the
Houston Works is generally of an inert, particulate composition. Little
evidence of a "first flush" effect or organic constituents was observed.
Generally, in all cases of well defined flow hydrographs, the maximum
constituent concentration corresponded to samples taken near peak flow
rates. It was also determined that for certain ranges of storm events the
average concentrations of runoff parameters was directly related to the
total volume of runoff. This paper covers the methodology and results of
the survey.
-256-
-------
STORMWATER: ONE STEEL MILL'S EVALUATION
Armco Steel Corporation operates a fully integrated steel mill in
Houston, Texas capable of producing approximately 1.8 million ingot tons
of steel per year. As a completely integrated steel mill, the Houston
Works starts with the raw materials of iron ore, coal and limestone and
produces finished products of plate and structural shapes. All treated
process wastewaters and stormwater discharge directly to the Houston Ship
Channel. The Houston Ship Channel being a dredged channel which provides
a deep water port to Houston, a city located fifty miles inland from the
Gulf of Mexico.
Pursuant to effluent discharge permits, the Houston Works has under-
taken an extensive stormwater survey. This survey has proven to be the
first report of its nature in the steel industry. The scope of the survey
deals not only with quantity, but also quality of direct stormwater runoff.
The most significant aspects revealed by the survey are the absence of a
"first flush" effect, the absence of significant quantities of organic
matter, and the establishment of ratios of those organics to carbon based
parameters (BODs/TOC).
After completing a literature search, it was determined that a
comprehensive stormwater survey had never been undertaken in the iron and
steel industry. This fact was confirmed by engineers currently under con-
tract to the United States Environmental Protection Agency (USEPA) now
performing such a study nationwide (Ref. 1).
-257-
-------
Armco undertook studies to determine drainage basin characteristics.
This included size, shape and activity of each basin. Then,hydrographs of
direct stormwater runoff were observed from one of the more typical basins.
These hydrographs were typical with those experienced in the literature
(Ref. 2). It was also determined that the runoff/rainfall ratios were with-
in the range predicted by empirical coefficients (Ref. 3).
Discrete grab samples were then taken sequentially throughout a direct
runoff event. These samples were analyzed for pH, Total Suspended Solids
(TSS), Chemical Oxygen Demand (COD), Oil and Grease (O&G) and Total Organic
Carbon (TOC). When concentrations of these parameters were plotted against
time (paramographs), it was noted that the concentrations plotted in the
general outline of the flow hydrograph. Figure 1 illustrates the hydrograph
and TSS and TOC paramographs for a 0.39 inch rainfall with an effective
duration of approximately fifteen minutes. This figure is representative
of all but one of twenty-two events sampled. The range of intensities and
durations of these events will be discussed later.
As seen in Figure 1, peak concentrations were found to align with the
peak flows. The concentrations, more or less tracked the flow hydrograph
outline. This has been termed a "flow dependent" effect as contrasted with
the "first flush" effect prevalent in the literature. "First flush" is
defined as: "The condition, often occurring in storm sewer discharges and
combined sewer overflows, in which a disproportionately high pollutional
load is carried in the first portion of the discharge or overflow" (Ref. 4).
This is significant both from the standpoint of design of stormwater treat-
ment facilities and from the impact of the stormwater on a receiving water-
-258-
-------
FIGURE I
TYPICAL HYDROGRAPH & PARAMOGRAPHS
FOR STORM DISCHARGES
400
TSS
MG/L
200
I0,000r
FLOW
GPM
5,000 -
TIME , MRS
-259-
-------
course. As the majority of pollutants in a "first flush" area occur before
peak flows are encountered, treatment facilities in these areas would be
considerably smaller than for the same basin exhibiting a flow dependent
effect. This, of course, is a matter of economics and the related cost/
benefit ratio.
From the standpoint of the impact of stormwater on a receiving stream,
a "first flush" effect may be significantly more deleterious than the impact
from an equivalent rainfall event from a flow dependent area. This is
attributed to the fact that concentrations in a flow dependent area are
proportional to total volume of flows or scouring velocities. It appears
that "first flush" areas contain matter that is considerably less dense and
smaller than flow dependent areas. Thus, once a minimum scouring velocity
is reached, the majority of the scourable material will be entrained in the
runoff flow. As flows increasingly exceed the minimum scouring velocity,
dilution of the effects of the high initial concentrations also increases.
However, storms that just exceed the minimum scouring velocity do not have
flow rates which substantially increase the receiving streams assimilative
and flushing capacities.
Armco's experience has shown relatively large particulate matter
comprises the majority of "pollutants" in storm runoff discharges from the
Houston Uorks. The samples that have been analyzed for soluble fractions
exhibit extremely low "pollutant" concentrations. Evidence of high
particulate, low organic, flow dependent areas have been observed in
Atlanta suburban regions, while urban and central business district samples
in Atlanta exhibited the typical "first flush" effect (Ref. 5). It appears
-260-
-------
that the flow dependent areas of suburban Atlanta and.the steel mill in
Houston contain relatively inert particulates. These particulates are
scoured in proportion to flow. No correlation is available, but it appears
that a minimum scouring velocity for a flow dependent area would be much
higher than that of a "first flush" area because of the nature of the scoured
matter in each basin.
A runoff event of the magnitude which produces a minimum scouring
velocity in a "first flush" area would contribute much higher concentrations
of organic pollutants than an equivalent event in a flow dependent area.
As the total volume of flow would be equal, the assimilative and flushing
capacities in the receiving stream would be the same. The "first flush"
effect would impart more of a negative impact than the flow dependent
effect. A flow dependent effect has the redeeming features of increasing
assimilative and flushing capacities with increased pollutantal load.
Armco also found significant statistical evidence that in the larger
basins (60.5 and 142.8 acres) studied, average concentrations varied
linearly with total flow. This was determined for TSS, TOC and COD. O&G
concentrations were found to exhibit no concentration pattern. This was
attributed to the fact that the highest concentration of O&G seen was 2.0
mg/1. The lack of significant quantities of O&G prohibited any statistical
correlations.
Figure 2 shows the relationship between average concentrations and
total flows for the 142.8 acre basin. The plots, as exhibited, are the
result of linear regression analysis applied to the available data. The
negative slope of TOC concentrations is attributed to the lack of
-261-
-------
FIGURE 2
AVERAGE CONCENTRATIONS VERSUS
TOTAL DISCHARGE
142.8 ACRE BASIN
80
COD
MG/L
40
40
TOC
MG/L
20
400r
TSS
MG/L
200-
0,2 0.4 0.6 0.8
FLOW, MILLION GALLONS
-262-
-------
appreciable quantities of organic material in the basin. This is confirmed
by two sets of 8005 samples which averaged less than 4 mg/1., It is assumed
that this limited quantity of scourable organic material is in fact diluted
by increasingly larger runoff events.
Figure 3 shows plots of average TSS, TOC and COD concentrations versus
total flows in the 60.5 acre basin. These plots are similar to those dis-
cussed in Figure 2. The projections represent the best linear fit for all
the given data. While the given lines for TSS and COD have correlation
coefficients greater than 99%, if the data from the lowest flow event were
removed, the correlation would be even more statistically precise. For the
TOC relation, removing the low point provides an increase from under 95%
to between 95% and 99%.
It is theorized that the low flow event did not exceed the minimum
scouring velocity discussed earlier. Therefore, concentrations from this
event would be lower than the predicted values, which can be seen in Figure
3. While the minimum scouring velocity is professed, there is not enough
data available to define its limits. It is anticipated that a specific
minimum scouring velocity would be required for each individual basin.
The plots from Figures 2 and 3 also indicate the site specificity of
data between basins. The two basins in question are similar in activity,
but the larger basin has a field determined rainfall/runoff coefficient
twice that of the smaller. The three parameters shown vary widely between
basins. Based on the Armco and Atlanta studies, transfer of data between
basins may not be a valid practice. This data indicates that even within a
specific industrial area each individual basin may exhibit totally different
-263-
-------
FIGURE 3
AVERAGE CONCENTRATIONS VERSUS
TOTAL DISCHARGE
60.5 ACRE BASIN
400
COD
MG/L
200
50
TOC
MG/L
25
2,400
TSS
MG/L
1,200
0.0
O.I 0.2 0.3 0.4
FLOW , MILLION GALLONS
-264-
0.5
-------
stormwater quality characteristics.
While no attempts were made to correlate the effects of the number of
antecedent dry days preceding each sampled event, the general findings are
interesting. Lager (Ref. 3) states that "higher concentrations of pollutants
are generally expected after prolonged dry periods". Kuhner and Shapiro
(Ref. 6) question whether this accumulation of material is linear or after
a long enough time frame a leveling off occurs. From the data at the 142.8
acre basin, there was only one event that was preceded by more than two
dry days. For TSS and TOC concentrations the actual value was almost twice
the expected value. There was slightly more variation in the number of
preceding antecedent dry days in the 60.5 acre basin. In general, those
events preceded by more than two dry days exhibited higher concentrations
than the predicted values. The available data is sufficient to provide
general quality relationships to flow. However, more data is required in
order to expand those relationships to include any effect of antecedent dry
days.
BODs is the parameter most widely used in the evaluation of non-process
wastewater effect on the oxygen level in a receiving stream. BODs is not
without faults; and has been criticized for its lack of reproducibility,
the time required for analysis, the necessity of seed acclimation and the
fact that a BODs represents one point on a rate curve (Ref. 7, 8, 9, 10).
TOC, on the other hand, is a very rapid, accurate measurement of which no
organic chemicals have been found that are resistent to the nature of the
procedure. If TOC is to be utilized as a parameter for waste characteriza-
tion and water quality monitoring, it must be correlated to the specific
-265-
-------
oxygen based parameter to be replaced, COD or BODs (Ref. 11). However,
there is no available correlations of TOC/BOD5 for stormwaters.
Armco has attempted this correlation for the 60.5 and 142.8 acre basins,
The data available is shown in Figure 4. Both linear projections represent
lines with statistical reliability greater than 99%. As with the data pre-
viously discussed, there is considerable difference in the relationships
between basins. Technology transfer would not be applicable in this case
either. However, even if a TOC/BODs relationship were developed, Ford (Ref.
12) warns that "TOC should not be considered a replacement for BODs or COD,
but rather an additional parameter for measuring organic pollution".
Armco feels it's efforts at characterizing stormwater quality at the
Houston Works have been very educational. Presently, it is felt that
rational attempts can be made at predicting the effects of a rainfall/runoff
event at the Houston Works. However, it must be remembered that each rain-
fall/runoff event is a unique occurrence. The variables entailed are
numerous; such as intensity, duration, antecedent dry days, and localized
non-routine activities within a drainage basin. While process wastewaters
may be fairly consistent, stormwater discharges vary dramatically between
events. Particularly important is the site specificity as it applies to
discharge quality. This precludes the inter-basin transfer of data which
is often relied on to make decisions concerning stormwater effect on
receiving streams. As costs for the treatment of stormwater can be con-
siderable, it is felt that careful consideration should be given to the
need for such treatment.
-266-
-------
FIGURE 4
TOC/BOD5 RELATIONSHIP
60
60.5 ACRE BASIN
40
TOC
MG/L
60
40
TOC
MG/L
142.8 ACRE BASIN
20
10
20 30
BOD5 ,M6/L
40 50
-267-
-------
REFERENCES
1. Brookman, Gordon T., The Research Corporation of New England.
Conversations with.
2. Hydrology for Engineers. Linsley, R.K., Kohler, M.A., and Paulhus, J.L.,
McGraw-Hill Book Co., N.Y., N.Y., 1958 p.193-213.
3. Design and Construction of Sanitary and Sjorm Sewers^. ASCE - Manuals and
Reports on Engineering Practice - #37, 345 East 47th St., N.Y.,
N.Y., 1970, p. 50-51.
4. Lager, J.A. et.al. "Urban Storm Water Management and Technology: An
Assessment", EPA-670/2-74-040, U.S.E.P.A., NERC - Cincinnati,
December, 1974, p.437.
5. Holbrook, R.F., Perez, A.I., Turner, B.G., and Miller, H.I., "Stormwater
Studies and Alternatives in Atlanta", Journal of the Environmental
Engineering Division - ASCE, Vol. 102, No. EE6, December, 1976,
p. 1263-1277.
6. Kuhner, J., and Shapiro, M., "Discussion of Urban Runoff Pollution
Control - State of the Art", Journal of the Environmental Engineering
Division, ASCE Vol. 102, No. EE1, February, 1976, p. 220-222.
7. Wastewater Engineering: Collection, Treatment and Disposal, Metcalf
and Eddy, McGraw-Hill Book Co., 1972, p. 24T^252T
8. Chemistry for Sanitary Engineers, Sawyer, C.A. and McCarty, P.L., McGraw
Hill Book Co., 1972, p.394-410.
9. Handbook for Monitoring Industrial Wastewater, U.S.E.P.A. Technology
Transfer, Nashville, Tennessee, 1973 (5-7 to 5-10).
10. Arm, M.L., "Monitoring With Carbon Analyzers", Environmental Science
and Technology, 8, 898 (1974).
11. Jones, R.H., "TOC: How Valid Is It?", Water and Wastes Eng., 9, 4, 32
(1972).
12. Ford, D.L., "Application of Total Organic Carbon Analyzer for
Industrial Waste Water Evaluation", Proc. 23rd Ind. Waste
Conference, Purdue University, West Lafayette, Ind., Ext. Ser.,
1, 132, 989 (1968).
-268-
-------
CHARACTERIZATION OF COAL PILE DRAINAGE
By
Doye B. Cox
R. J. Ruane
Special Projects Staff
Division of Environmental Planning
Tennessee Valley Authority
-269-
-------
ACKNOWLEDGEMENTS
This study was conducted as part of the project entitled 'Charac-
terization of Effluents from Coal-Fired Utility Boilers," and is supported
Wider Interagency Agreement Numbers EPA-IAG-D5-E-721 and TV-41967A between
TVA and EPA for energy related environmental research. Thanks are extended
to EPA project officers, Julian W. Jones and Ron A. Venezia, and TVA project
directors, H. B. Flora and B. G. McKinney. Appreciation Is also extended to
Randall L^ Snipes, F. G. Parker, J. E. Liner, Ralph D. Gillespie, and
Roger P. Betson for their aid in the project.
-270-
-------
ABSTRACT
Coal pile drainage is produced by rainfall percolation through
coal storage piles where it procures the oxidation products of metallic sul-
fides. To chemically characterize coal pile drainage, sampling and analysis
programs are being conducted at two TVA coal-fired steam-electric generating
plants. The drainage streams were found to contain high concentrations of
acidity to pH 7.0 (as CaCO.,), suspended solids, total dissolved solids, iron,
manganese and a variety of trace metals. The pH of this drainage remained in
the relatively narrow band of 2.3 to 3.1. Transfer of coal pile drainage to
an alkaline ash pond appears to provide adequate removal of the above. A
rainfall-runoff relationship was developed for prediction of total runoff
volumes for specific rainfall events.
-271-
-------
INTRODUCTION
Increases In energy use coupled with delays and restrictions on
nuclear plants and decreases in supplies of fuel oil and natural gas make
increases in coal production and utilization inevitable. Coal, whether it
is to be used directly for steam generation, for liquefaction, for gasifica-
tion or for other processes will have to be stored in huge quantities. The
volume to be stored, methods of handling, and the explosive nature of coal
dust all dictate outdoor storage. Outdoor storage presents a number of
interesting potential problems including blowing coal dust and an esthetically
displeasing appearance.
Another potential problem associated with open-air storage is rain-
fall runoff from the coal storage area. Long before major coal-fired steam-
electric plants were in exlstance, the Welsh were familiar with the acidic
character and distinguishing orange color of coal mine drainage. Naturalists
have for centuries been aware of the orange-yellow stains that many times
accompany coal outcrops. When similar drainages appeared emanating from coal
storage facilities at power plants they were largely ignored and passed off
as a localized and perhaps natural occurrence. However, with the advent of
the environmental movement, industrial discharges came under closer scrutiny.
In light of this, characterization of discharges from coal-fired power plants
was undertaken. As part of this overall study of potential discharges, a
program was designed to provide an adequate characterization of drainage from
coal storage piles.
Programs were established at two Tennessee Valley Authority (TVA)
coal-fired steam plants. Plant J has a rated capacity of 1700 megawatts
with a 90-day coal supply amounting to VL.26 x 10 cubic yards (^9.6 x
5 69
10 cu m) or 1.2 x 10 tons (1.1 x 10 kg). Plant E has a rated
-272-
-------
capacity of 1400 megawatts with a 90-day coal supply amounting to M..13 x 10
6 68
cubic yards (8.6 x 10 cu m) or 1.08 x 10 tons (9.88 x 10 kg).
Coal for Plant J is mined in eastern Tennessee and Kentucky and
transported by truck and rail to the plant site. It is stored prior to any
preparation. Coal for Plant E is mined in western Kentucky and is trans-
ported mainly by barge. This coal is also stored prior to any preparation.
A typical analysis of coal from both plants is presented in table 1.
Table 1
COAL ANALYSIS DRY BASIS
Plant J Plant E
Total Moisture
Volatile Matter
Ash
Fixed Carbon
Total Sulfur
Btu/lb
Ash Analysis
CaO
MgO
3.8
34.1
17.2
48.7
2.1
12270
1.4
1.1
4.2
37.7
15.0
47.3
3.9
Btu
12450
% of Ash
4.2
1.1
-273-
-------
BACKGROUND
To ensure uninterrupted generation of electricity, an outdoor coal
reserve is maintained at each power plant. This coal supply is available for
use in the event normal deliveries are delinquent, temporarily discontinued,
or inadequate to meet peak electricity demands. A 90-day coal supply is cus-
tomarily maintained to provide a sufficient safety factor. Factors which
preclude a large coal stockpile include the cost of land required for stor-
age, workmen and equipment needed to maintain the coal storage area, the cost
of the larger inventory, and oxidative degradation that occurs when coal is
stored for long periods of time. While the physical volume of coal storage
required varies with the plant consumption rate, coal piles are typically
25 to 40 feet (8 to 12 m) high and spread over an area 25 to 100 acres (10 to
40 hac). Normally 780 to 2340 cu yds (600 to 1800 cu m) of coal storage are
required for every megawatt of rated capacity.
Coal pile drainage results from percolation of rainfall through
stored coal. The water quality of the drainage is affected by the leaching of
oxidation products of metallic sulfides associated with the coal. The sul-
fide bearing minerals that predominate in coal are pyrite and marcasite, both
iron sulfide ores. Marcasite is unstable and degrades into pyrite. The
oxidation of pyrite results in the production of ferrous iron and acidity
(FWQA 1970):
(1) 2 FeS.(s) + 70. + 2H,0 »~ 2Fe+2 + 4H+ + 4SC-72
2 2. 2. 4
This ferrous iron then undergoes oxidation to the ferric state:
(2) 4Fe+2 + 02 + 4H+ *- 4Fe+3 + 2H+ + 20H~ (rate limiting step)
-274-
-------
Ferric iron then either hydrolyzes to form insoluble ferric hydroxide
producing more acidity:
(3) Fe+3 + 3H20 »- Fe(OH>3 (s) + 3H+
or can oxidize pyrite directly producing more ferrous iron and acidity.
(4) FeS2(s) + 14Fe+3 + 81^0 «-15Fe+2 + 2SO~2 + 16H+
The stoichiometry of this reaction reveals that for every mole of
ferrous iron oxidized in equation (2) there is a net increase of two moles of
hydrogen ion. This net increase in acidity provides hydrogen ions for further
oxidation of ferrous iron and subsequent acid production.
As the pH decreases below 5, certain acidophilic, chemoautotrophic
bacteria become active. These bacteria, Thiobacillus ferrooxldans, Ferro-
bacillus. ferrooxidans, Metallogenlum, and similar species are active at pH
2.0 to 4.5 and use CO- as their carbon source (Silverman 1967). They are
the main factor contributing to the oxidation of ferrous iron to the ferric
state, the rate limiting step in the oxidation sequence. Their presence is an
indication of rapid pyrite oxidation and is usually accompanied by waters low
in pH and high in iron, manganese, and total dissolved solids.
Factors that possibly affect production of acidity in coal piles
and the subsequent leaching of trace metals are (1) concentration and form
of pyritic sulfur in the coal, (2) size of the coal pile, (3) method of coal
preparation and cleaning prior to storage, (4) climate, including rainfall
and temperature, (5) concentration of CaCO. and other neutralizing substances
in the coal, (6) concentration and form of trace metals in the coal, and
(7) the residence time in the coal pile.
-275-
-------
Methodology
Plant J—In 1972 a system was Installed for collection of coal pile
drainage and the subsequent transfer of the drainage to an ash pond (Figure
1). Collection is accomplished by a series of maintained channels around the
coal pile which drain into a storage basin. A manually operated pump, an
associated piping system, and a secondary maintained channel are components
which transfer the drainage to the ash pond. The storage basin is designed to
contain the runoff from a moderately small storm at best. Due to this limited
capacity, the pump is activated manually at the start of almost all rainfall
events and actual detention time in the basin proper is small.
A sampling system (Figure 2) was designed so that pressure in the
line which runs from the pump to the ash pond forces a sample into the sample
collection barrels. The sample line is composed of tygon tubing with plastic
fittings. The sample barrels are plastic garbage cans with an approximate
volume of 150 1 each. Flowrate of coal pile drainage from the storage basin
to the ash pond is approximately 900 gpm (3400 1/m). This flowrate will be
refined as more data becomes available. Flow through the sample line was
adjusted to approximately 0.025 gpm (0.1 1/m). This arrangement supplies a
sample that is a composite of the total volume pumped to the ash pond. Due
to the acid nature of the waste and the desire to collect pH and acidity data
the sample was not preserved by acidification until the date of collection.
Collection was accomplished by first manually stirring the samples and then
collecting the samples from a line draining both barrels. Chemical analyses
were performed by the TVA Water Quality Laboratory using methods prescribed
in Standard Methods for the Examination of Water and Waatewater (APRS 1971),
and Manual of Methods for Chemical Analysis of Water and Wastes (EPA 1974).
A rain gage was placed adjacent to the coal pile so that a rainfall/
runoff relationship could be developed. This information will be used in design
-276-
-------
DRAINAGE AREA - 53.3 ACRES (21.6 hoc)
TO ASH
POND
COLLECTION
SUMP
FIGURE I
COAL PILE AND DRAINAGE COLLECTION SYSTEM
PLANT J
-------
TO ASH POND
SAMPLE
LINE
SAMPLE
BARRELS
COLLECTION
SUMP
FIGURE 2
SAMPLE COLLECTION SYSTEM
PLANT J
-278-
-------
of future storage basins and in estimation of evaporation and percolation
losses. Rainfall was compiled on a daily basis, tabulated and compared with
hours of pumping time. Pumping rate was determined by placing a temporary
weir in the drainage channel downstream from the pump. This installation was
subject to only minimal inflow from the immediate vicinity. In cases when
intermittant rain occurred over several consecutive days and the pump was in
almost continuous operation, the determination of individual rainfall and its
concurrent runoff was impossible. Instead, when consecutive days of rainfall
occurred, total rainfall and associated pumping over this entire period were
considered as one rainfall/runoff event.
Plant E—Drainage from the coal pile at this installation is collected
in no systematic manner, though plans for total collection and transfer to the
ash pond are under development. Instead the drainage emanates in three dis-
tinct directions (Figure 3). Drainages A and B unite at some distance down-
stream and flow into a holding pond. However, at this point there is significant
dilution of the coal drainage. Drainage C quickly spreads out onto a mud flat.
Due to the diverse nature of these discharges and the expense of installing
and maintaining even temporary flow gages, drainage volume at Plant E was not
gaged.
A modified ISCO model 780 automatic water sampler was placed onslte
at one of the drainways and a small sample pool constructed (Figure 4). This
sampler was equipped with a stage activation device so that the sampler initated
sampling with the rise of the storm hydrograph. Samples were collected at
hourly intervals and composited. Thus, these samples represent a simple com-
posite of each runoff event. Discreet samples were collected of a single storm
event on February 24, 1977. Total rainfall for this event was 2.10 in. (53 mm).
These samples were collected at 20-minute intervals and analyzed for pH, acidity,
-279-
-------
I
S3
00
o
I
FIGURE 3
COAL PILE AND ASSOCIATED DRAINAGE SYSTEM
PLANT E
B
-------
I
oo
M
I
SAMPLER
STAGE
ACTIVATOR
LINE v
SAMPLE
LINE—-x
STAFF GAGE
CONDUCTIVITY
CELL
FIGURE 4
SAMPLE COLLECTION SYSTEM
PLANT E
WEIR
-------
dissolved solids, suspended solids, sulfate, Iron, and manganese. Rainfall
was measured onslte so that runoff could be estimated. Loadings of pollutants
can be projected by application of this estimate to individual storm event
composites. Application of this simple composite—average flow method for
calculating loadings of pollutants has been demonstrated to be adequate by
Grizzard (1976).
Hydrology
In most cases rainfall either results as surface runoff, percolates
into the soil to become ground water, or is lost through evapotranspiration.
However, in coal piles, even that fraction of the rainfall that percolates
into the pile maybe subject to some evaporation, and no transpiration can
occur. Conventional estimates of evapotranspiration used in most hydrologi-
cal models are, therefore, questionable. Due to these problems and the
expense of installing large flumes for long term measurement of flows, detailed
hydrological models were not calibrated for use as a part of this study.
Instead a simple rainfall/runoff relationship was developed for use at TVA
facilities. Such a relationship could be used for estimating detention basin
design and calculation of acid loads to the ash pond because rainfall, tempera-
ture, wind velocity, humidity are similar throughout the Valley and ridge
geophysical province.
A rainfall/runoff regression analysis was performed for data col-
lected over a seven-month period at Plant J. Runoff was calculated by taking
hours of pumping time and multiplying by a known pumping rate. Rainfall was
monitored continuously onslte. A plot of the regression line and the 95 per-
cent confidence intervals of the mean are presented in Figure 5. This relation
-282-
-------
RUNOFF = 0.855 RAINFALL + 0.0082
i
PO
oo
U!
7 3
REGRESSION OF RAINFALL vs RUNOFF
-------
(Equation 5) can be used to predict the runoff In inches/acre for a given
storm event when total inches of rainfall are known.
(5) Runoff (in.) - 0.855 rainfall (in.) + 0.0082
Runoff can be converted to total runoff by applying a drainage area and
obtaining the appropriate volumetric term. Losses due to evaporation and
infiltration are about 14 percent. Losses decrease slightly at lower rain-
falls due to the inclusion of base flow in this relationship. It is, of
course, limited in application to coal piles of similar size. Additional
factors that could affect runoff Include amount of snowfall and soil permeability.
Acidity and pH
Both systems investigated exhibited highly acidic drainages. Acidity
was determined as "cold" acidity to pH 7.0 and expressed as CaCO_. Acidity
was quite variable in both cases (see table 2). However, pH was within a
rather tight band (2.3 to 3.1). This illustrates the point that acidity is
not a measure of hydrogen ion but rather a measure of available protons. Means
(arithmetic) are quite similar. In fact 21 of the 25 values fall between 2.6
and 3.0. Values of pH presented by Nichols (1974) exhibit a slightly broader
range of 2.1 and 3.0. Data presented by Anderson (1976) representing hourly
pH measurements over a 3-week period contained pH reading ranging from 2.2 to
5.8. A range of pH 2.4-3.0 was presented by Matsugu (1976) for 67 grab sam-
ples of coal pile leachate. For these same samples acidity varied from 10 to
120 meq/1. Thus, it appears that pH of coal pile drainage, at least for
eastern coal, is generally in the relatively narrow range of 2.2 to 3.1.
Another interesting observation involves comparison of pH values obtained at
Plant J with those obtained at Plant E. Even though coal supplied to Plant E
was high sulfur coal (>3%) and coal supplied to Plant J moderate in sulfur
-284-
-------
Table 2
MEANS AND RANGES OF PROJECT DATA
FROM PLANT J AND PLANT E
Plant
i
ho
oo
E
E
(Discrete Storm)
Range
Mean
N
Range
Mean
N
Range
Mean
N
£»
2.3-3.1
2.79
19
2.5-3.1
2.67
6
2.5-2.7
2.63
14
Acidity
(mg/1
CaCOq)
300-7100
3400
18
860-2100
1360
6
300-1400
710
14
Sulfate
(mg/1)
1800-9600
5160
18
1900-4000
2780
6
870-5500
2300
14
Dissolved
Solids
(mg/1}
2500-16000
7900
18
2900-5000
3600
6
1200-7500
2700
14
Suspended
Solids
(me/i)
8.0-2300
470
18
38-270
190
6
69-2500
650
14
Fe
(mg/1)
240-1800
940
19
280-480
380
6
62-380
150
14
Mn
(mg/1)
8.9-45
28.7
19
2.4-10.0
4.13
6
0.88-5.4
2.3
14
-------
content (1-3%) as classified by EPA (Oct. 1976) pH of the drainage was similar
for both plants. Carucclo (1976) has shown that total sulfur concentration Is
not directly related to acid formation from pyrltlc material. A second expla-
nation involves maintenance of an optimum pH range (2-4) of the autotrophic
bacteria responsible for pyrite oxidation as explained by Schnaitman (1969).
It is interesting to note that acidity is higher at Plant J than at Plant E
even though pH's are similar. In essence, any substance capable of donating
protons (most metals, many naturally occurring organic compounds) will be
measured as acidity.
Solids
Total suspended solids concentrations are of primary interest in
characterization of coal pile drainage. Elevated concentrations occur when
rainfall/runoff suspends coal fines in the pile. This is generally not a
problem during base flow conditions but occurs during runoff events at
levels up to 2300 mg/1.
Suspended solids concentrations at Plant J ranged from 8 to 2300 mg/1
with a mean of 470 mg/1. It should be noted, however, that these samples
were collected after the sample had passed through a collection sump where
some settling occurred. At Plant E where direct runoff was collected as a
single composite sample for each storm event, the mean and range of suspended
solids concentrations were somewhat lower. However, when the discrete-wlthin-
storm-event samples were examined the range and mean were similar to that
obtained at Plant J indicating that minimal settling occurred in the collec-
tion sump at that installation. Suspended solids values much higher than
this were presented by Matsugu (1976).
Total dissolved solids concentrations were somewhat higher at
Plant J than Plant E even when the discrete-storm-event samples were con-
sidered. An inspection of the data reveals that most of the total dissolved
-286-
-------
solids are sulfate salts. Hence, higher total dissolved solids concentrations
are a consequence of enhanced pyritic oxidation via equations 1 and 4. This
data complements the acidity and pH data in suggesting that enhanced pyritic
oxidation is occurring at Plant J.
Iron and Manganese
These metals are often discussed simultaneously due to their similar
behavior in water. Both are increasingly soluble with decreasing pH, exist in
both the reduced and oxidized state, and form coatings on particles that may
limit solubilities of other metals (Jenne 1968). Typically iron and manganese
concentrations in pyritic systems are quite high. Iron minerals are the sub-
strate necessary for acid production (Equations 1-4). As such, lower concen-
trations would be expected only where pyritic oxidation is repressed or where
pH is not depressed sufficiently to allow for iron solubility. Values for iron
presented by Nichols (1974) ranged from 0.17 to 93,000 ppm with a mean of
19,500 ppm. A somewhat narrower range of 10 to 5,300 ppm and lower mean of
1,150 was presented by Anderson (1976).
Iron concentrations at both plants are lower in range and mean than
encountered by these other investigators. Concentrations at Plant E ranged
from 280 to 480 mg/1 with a mean of 380 mg/1. This is considerably higher
than the discrete storm samples which ranged from 62 to 380 mg/1 with a mean
of 150 mg/1. Concentrations of iron at Plant J were higher with a range of
240 to 1800 mg/1 and a mean of 940 mg/1.
Manganese concentrations reported by Anderson (1976) ranged from
4.5 to 72.0 mg/1 with a mean of 17.1 mg/1. Somewhat lower concentrations
were presented by Matsugu (1976), ranging from 3.4 to 12 mg/1 with a mean of
6.9. Levels at Plant J (Table I) were comparable to those presented by
Anderson. Values for Plant E were somewhat lower.
-287-
-------
Trace Elements
There is a paucity of information available on trace element con-
centrations in coal pile drainages. Trace elements of environmental concern
in coal have been identified by EPA (October 1976) and are presented in table
3. These constituents, with the exception of Yttrium were analyzed in drain-
ages from both plants. In addition, several other elements including aluminum
and chromium were analyzed on several samples. Results of analyses for selec-
ted trace elements are presented in table 4. Several other trace elements
were also analyzed In drainages from both plants. Of these elements lead,
barium, and titanium were low or consistantly below the limits of detection.
Levels of antimony were above detection limits in several instances but were
not included because of a question of their significance to freshwater aquatic
life. When an element was below detection limits the detection limit was used
for statistical purposes. It is interesting to note that most means of trace
element concentrations at Plant J are 3 to 8 times higher than those at Plant E.
Table 3
TRACE ELEMENTS IN COAL
(EPA, October 1976)
Range (ug/g)
0.4-8
0.1-59
0.1-65
0.01-1.6
4-218
Element
Beryllium
Nickel
Copper
Zinc
Arsenic
Range (yg/g)
0-31
0.4-104
2-185
0-6000
0.5-106
Element
Selenium
Yttrium
Cadmium
Mercury
Lead
-288-
-------
Table 4
MEANS AND RANGES OF TRACE METAL DATA
FROM PLANT J AND PLANT E
1
NJ
00
1
Plant
J Range
Mean
w
N
E Range
Mean
N_
IT
J Range
Mean
N
»
E Range
Mean
N
N
Cu
0.43-1.4
0.86
0
19
0.01-0.46
0.23
0
6
Cr
<.005-.011
.007
11
17
<.005-.011
0.007
3
6
Zn
2.3-16
6.68
0
19
1.1-3.7
2.18
0
6
Hg
<.0002-.0025
.0004
12
20
0.003-.007
0.004
0
5
Cd
mg/1
<.001-<.001
<.001
19
19
<. 001-0. 003
0.002
2
6
As
mg/1
.005-0.6
0.17
0
19
0.006-0.046
0.02
0
4
Al
66.0-440
260
0
19
22.0-60.0
43.3
0
6
Se
<.001-.03
0.006
4
18
<.001-.001
0.001
3
4
Ni
0.74-4.5
2.59
0
19
0.24-0.46
0.33
0
6
Be
0.03-0.07
0.044
0
18
<. 01-0. 03
0.014
3
4
N = Number of samples below detection limits.
-------
Concentrations of copper are higher than criteria for Plant J
(EPA, July 1976) if discharged untreated. These levels are, however, lower
than those presented by Nichols (1974) or Anderson (1976). Concentrations
for Plant E are lower still and don't appear to be significant from a water
quality standpoint.
Levels of zinc are also high with respect to ambient quality.
The mean concentrations of 6.68 mg/1 and 2.18 mg/1 at Plant J and Plant E are
similar to the means of 5.9 presented by Nichols (1976) and 3.67 presented by
Anderson (1976), respectively. The public water supply criteria established
by EPA (July 1976) is 5 mg/1.
Cadmium concentrations are quite low in drainages from both plants.
At Plant J no samples exceeded detection limits. At Plant E four of the six
samples exhibited detectable concentration but none were above water quality
criteria (EPA, July 1976).
Aluminum is included as a toxic substance by the National Academy
of Science (1972) in developing proposed water quality criteria but eliminated
by EPA (July 1976) in their development of finalized criteria. Thus, the
significance of aluminum as a toxic substance is in question. Mean concen-
trations of 260 mg/1 and 43 mg/1 were found respectively at Plants J and E.
Concentrations of nickel are also above levels found in surface
water (National Academy of Science, 1972) but are significantly less than
aluminum.
Chromium concentrations were well below established criteria at both
plants and pose no threat to the aquatic community or man.
-290-
-------
Toxlcity of beryllium like several other metals is inversely related
to the hardness of the solute. Coal pile drainage is quite hard (the mean
calcium and magnesium concentrations for Plant J were 300 mg/1 and 230 mg/1,
respectively). Levels of beryllium are well below established criteria for
waters of this hardness (EPA, July 1976).
Mercury concentrations were an order of magnitude higher at Plant
E than at Plant J. Levels were found at both plants that exceed established
water quality criteria (EPA, July 1976).
Arsenic levels in drainage from Plant J ranged from 0.005 to 0.6
mg/1 with a mean of 0.17 mg/1. These values generally exceed established
criteria while those concentrations found at Plant E generally did not.
Selenium concentrations behaved similarly to those of arsenic in
that Plant J generally exceeded criteria while Plant E did not. This is
significant since selenium and arsenic exhibit antagonistic toxicities
(Levander 1976).
Mass Input vs. Output
Mass flows of several constituents were calculated at Plant J for
the month of June 1976. The total flow for this month was 13 MG (49 Ml).
4
Runoff from the coal pile during this period contained 50 tons (4.5 x 10 kg)
4 4
of iron, 17 tons (1.5 x 10 kg) of aluminum, 1.8 tons (1.6 x 10 kg) of man-
ganese, 335 Ibs (152 kg) of nickel, and 127 Ibs (58 kg) of copper along with
large quantities of sulfate, dissolved solids, suspended solids, and hardness.
In order to assess the contribution of rainfall to the coal pile
system, a rainfall sample was collected during a storm that occurred on June
29-30. This storm had a duration of 33 hours with a total precipitation
of 1.35 in. (3.43 cm). Total precipitation for June was 7.99 in. (20.3 cm).
-291-
-------
Concentrations of several constituents were analyzed in the rainfall samples
and loadings were extrapolated for the month of June. Calculated values were:
iron 13.5 Ib (6.1 kg), aluminum 19.4 Ib (8.8 kg), manganese 1.0 Ib (0.5 kg),
nickel 4.9 Ib (2.2 kg), and copper 6.7 Ib (3.0 kg). Thus, contaminants in
rainfall appeared to be insignificant.
Treatment
TVA plans to route all coal pile drainage to ash ponds where dilution/
neutralization/adsorption processes may take place. However, if the ash
sluice has an insufficient amount of base available for neutralization or the
pH of the ash sluice—coal pile drainage mixture is too low (<8.5-9.0)—precipi-
tation of many metal hydroxides will not occur (Sittig 1973). When pH is too
low for precipitation dilution and particle adsorption are the major processes
for reducing trace metals concentrations within the ash pond.
At Plant J, the ash pond effluent pH has exhibited a periodic char-
acter. The pH fluctuates from the 7.0-8.5 range in summer-fall to the 3.5-4.5
range in winter-spring. Total hardness, alkalinity, and total dissolved
solids also vary periodically. In the past, this phenomenon has been attri-
buted to changes in the raw makeup water for ash sluicing since water known to
contain lower alkalinity is used for this purpose during the spring. It was
felt that coal pile drainage may also play a role in establishing the seasonal
character of the ash pond effluent. However, there is a decrease in total
dissolved solids concentrations with decreasing pH in the ash pond effluent
which would indicate that this is not the case. Trace metals are monitored
monthly in TVA ash pond effluents for plants in Tennessee and quarterly in
other plants. Comparison of trace metal concentrations in ash pond effluent
at Plant J where coal pile drainage is routed to the ash pond with other ash
-292-
-------
ponds not receiving these Inputs reveals no obvious differences in mean efflu-
ent concentrations of trace metals. In addition, since mean effluent concen-
trations of trace metals are similar (see Table 5) at these plant even through
effluent pH's are dissimilar (some are alkaline, some neutral, some acid),
the adsorption/ dilution processes must equilibrate these concentrations. To
illustrate the importance of ash pond dilution in reducing the concentrations
of pollutants in coal pile drainage, yearly and daily flows of both waste
streams can be compared at Plant J. The total yearly ash sluice flow for
9 10
1976 was approximately 9.8 x 10 gal. (3.7 x 10 1). An estimation of
yearly coal pile drainage can be arrived at by applying the average annual
precipitation, 51 in. (129.5 cm), to equation 5 and using the drainage area
given in Figure 1. This produces an estimation of 6.3 x 10 gal. 2.3 x
g
10 1) or 0.64 percent of the average annual ash sluice flow. On a daily
basis the average ash sluice flow (calculated from 1976 data) is 2.6 x 10
8
gal. (1 x 10 1). An estimation of coal pile drainage at this plant is
best arrived at by assuming continuous operation of the coal pile drainage
pump for a 24-hour period. This estimate is 1.3 x 10 gal. (4.9 x 10 1)
or 5 percent of the average daily ash sluice flow. At a plant where no collec-
tion basin-pump system is operative, an estimation of coal pile drainage
volume could be arrived at by applying equation 5 using an appropriate design
storm. More information regarding these processes and the effects of pH on
trace metal concentrations in ash pond effluents is being developed by an
EPA/TVA concurrent project. Final determination of the role of coal pile
drainage in affecting ash pond effluents would require a detailed mass balance
study of the ash pond-coal drainage system.
In cases where dry ash disposal is employed, other treatment tech-
nologies may have to be used. Due to the similarities between coal pile
drainage and acid mine drainage many of the treatment methods used for the
-293-
-------
TABLE 5
ASH POND EFFLUENT DATA*
Lant
PH
Fe
Mn
Cu
Zn
Cd
Ni
Cr
Hg
As
Se
J
6.0
2.4
0.38
0.11
0.07
0.001
0.05
0.005
0.0003
0.041
0.004
E
11.2
0.16
0.01
0.08
0.05
0.001
<0.05
0.017
0.0002
0.028
0.007
C
7.2
3.9
0.19
0.06
0.14
0.004
0.06
0.008
0.005
0.018
0.007
A
(Fly Ash)
CONCENTRATION
(SU)
4.2
mg/l
2.2
0.49
0.33
1.4
0.038
0.08
0.072
0.0003
0.011
0.002
A
(Bottom Ash)
7.2
5.2
0.17
0.07
0.08
0.001
0.06
0.007
0.0005
0.007
0.002
D
8.5
0.32
0.02
0.03
0.03
0.001
0.06
<0.005
0.0002
0.034
0.07
^Average of quarterly data 1973-1975.
-294-
-------
latter drainage stream would be applicable. Methods most widely used are lime
or limestone neutralization. However, reverse osmosis (Rosehart 1973), and
sulfide precipitation (Ross 1973) have been investigated. Fly ash neutrali-
zation appears to be quite feasible when the ash is sufficiently alkaline.
Sediment samples were taken from the bottom of channels draining
the coal pile. These samples consisted of mud, coal fines, pyrite chips,
and an amorphous yellow precipitate. There were, in many places, green pre-
cipitates covering the small pebbles. These sediment samples were examined
for iron oxidizing bacteria. The genus Thlobaclllus was identified in all
samples. The amorphous precipitate appeared to be elemental sulfur, an
amorphous pyrite, or ferric chloride. This precipitate appeared on the sur-
face of gullies eroded into the sides of the coal pile. The precipitate only
appeared on the gully bottom and was not visible after removing approximately
one centimeter of the fines forming the gully bottom. If this was indeed
elemental sulfur, its source can be attributed to the bacterial oxidation of
pyrite where it has been identified as an intermediate in the oxidation of
sulfide to sulfate. Identification of iron oxidizers indicates the possi-
bility of inhibition of the responsible organism for drainage control. Investi-
gators working on control of acid mine drainage, a waste also mediated by
Thiobacillus, have reported dramatic decreases in acid production rates when
bacterial inhibitors were employed (Shearer 1968). One proven inhibitor is
ferrous iron in high concentrations. Application of an inhibitor for reducing
the rate of acid production could be accomplished by application along with
sprinkling for dust control. This method of control could be particularly
appropriate where dry ash disposal is employed.
-295-
-------
Conclusions
Coal pile drainage is a highly acidic waste stream containing high
concentrations of a wide variety of inorganic constituents. Further, the
acidity of this waste stream may not be governed by sulfur content of the
coal. However, transfer of this drainage to an ash pond where neutralization/
dilution/adsorption phenomena take place appears to provide adequate treatment.
-296-
-------
REFERENCES
1. American Public Health Association, American Water Works Association
and Water Pollution Control Federation, Standard Methods for the
Examination of Water and Wastewater, 13th Edition Published Jointly
by APHA, AWWA and WPCF, 1971.
2. Anderson, William C. and Mark P. Youngstrom, "Coal Pile Leachate
Quantity and Quality Characteristics," Presented at the Sixth Symposium
on Coal Mine Drainage Research, NCA/BCR Coal Conference and Expo III,
Louisville, Kentucky, October 19-21, 1976.
3. Caruccio, Frank T., Gwendolyn Geidel, , and John M. Sewell, "The
Character of Drainage as a Function of the Occurrence of Framboidal
Pyrite and Groundwater Quality in Eastern Kentucky," Presented at the
Sixth Symposium on Coal Mine Drainage Research, NCA/BCR Coal Conference
and Expo III, October 19-21, 1976.
4. Environmental Protection Agency, Manual of Methods for Chemical
Analysis of Water and Wastes. U.S. Environmental Protection Agency,
Publication No. EPA-625/6-74-003, Washington, DC, 1974.
5. Environmental Protection Agency, "Quality Criteria for Water," U.S.
Environmental Protection Agency, Publication No. EPA-440/9-76-023,
Washington, DC, July 1976.
6. Environmental Protection Agency, "EPA Program Conference Report.
Fuel Cleaning Program: Coal," U.S. Environmental Protection Agency,
Publication No. EPA-600/7-76-024, Washington, DC, October 1976.
7. Federal Water Quality Administration, "Oxygenation of Ferrous Iron,"
Federal Water Quality Administration, U.S. Government Printing Office,
Washington, DC, 1970.
8. Grizzard, T. J., C. W. Randall, and R. C. Hoen, "Data Collection for
Water Quality Modeling in the Occoquan Watershed of Virginia," Proceedings
of the Conference on Environmental Modeling and Simulation, Publication
No. EPA-600/9-76-016, July 1976.
9. Jenne, E. A., "Controls on Mn, Fe, Co, Ni, Cu, and Zn Concentrations in
Soils and Water: The Significant Role of Hydrous Fe and Mn Oxides,"
Trace Inorganics in Water, American Chemical Society, Washington, DC,
1968.
10. Levander, Orville, A., "Metabolic Interrelationships between Arsenic and
Selenium," Presented at the N.I.E.H.S. International Conference on
Environmental Arsenic, Fort Lauderdale, Florida, October 5-8, 1976.
11. Matsugu, R. S., Ontario Hydro Research Division Report, Submitted to
Mr. J. H. Waghorne, April 3, 1976.
-297-
-------
12. National Academy of Science, "Water Quality Criteria 1972," U.S. Environ-
mental Protection Agency, U.S. Government Printing Office, Washington, DC,
1973.
13. Nichols, C. R., "Development Document for Effluent Limitations Guide-
lines and New Source Performance Standards for the Steam Electric Power
Generating Point Source Category," U.S. Environmental Protection Agency,
Washington, DC, 1974.
14. Rosehart, R. G., "Mine Water Purification by Reverse Osmosis,"
Canadian Journal of Chemical Engineering, Volume 51, Number 6, December
1973.
15. Ross, L. W. , "Removal of Heavy Metals from Mine Drainage by Precipita-
tion," Environmental Protection Agency Technology Series Report
EPA-670/2-73-080, 1973.
16. Schnaitman, C. A., M. S. Korczynski, and D. G. Lundgren, "Kinetic Studies
of Iron Oxidation by Whole Cells of Ferrobacillus Ferrooxidans," Journal
of Bacteriology. Volume 99, August 1969.
17. Shearer, R. E., W. A. Everson, and J. W. Mausteller, "Reduction of Acid
Production in Coal Mines With Use of Viable Anti-Bacterial Agents,"
Presented at Second Symposium on Coal Mine Drainage Research, Mellon
Institute, 1968.
18. Silverman, Melvin P., "Mechanism of Bacterial Pyrite Oxidation," Journal
of Bacteriology. Volume 94, Number 4, October 1967.
19. Sittig, Marshall, Pollutant Removal Handbook, Noyes Data Corporation,
Park Ridge, New Jersey, 1973.
-298-
-------
MEASUREMENT OF NON-POINT SOURCES FROM A COAL-FIRED UTILITY
AND THE IMPACT ON RECEIVING WATERS
By
Gordon T. Brookman, P.E.
and
Willard A. Wade III, P.E.
TRC - The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
-299-
-------
ABSTRACT
Non-point source water pollution has become a growing concern in the
United States because in many areas non-point source pollution overwhelms
and negates the reductions achieved through control of point sources.
TRC - THE RESEARCH CORPORATION of New England was retained by the Indus-
trial Environmental Research Laboratory of the U.S. Environmental Protec-
tion Agency to perform, as part of a contract to develop fugitive emissions
management techniques, industrial runoff assessments of the coal-fired
utility and iron and steel industries.
This paper describes the coal-fired utility measurement program and
field survey results. The survey was performed at two utility plants in
Pennsylvania. One plant is near the headwaters of the Allegheny River in
Warren, Pennsylvania. The second plant is on the Delaware River in
Portland, Pennsylvania near the Pocono Mountains. The survey included
measurement of overland runoff from non-point sources and river sampling
upstream and downstream of the plant site. The non-point sources sampled
were stormwater runoff and leachate from coal storage piles (an infinite
source) and stormwater runoff from impervious areas such as parking lots
and roofs which were covered with dust fallout from coal handling and ash
handling operations (a finite source).
The data survey indicated that non-point source pollution from the
utilities had little impact on the two rivers compared to the impact from
sources upstream of each site. Although the river mass loadings were
higher (due to increased flow), the data showed no statistical difference
in mean concentrations of upstream versus downstream pollutant levels in
-300-
-------
either dry or wet weather conditions. Runoff at Warren showed dry weather
leachate from the coal pile had a greater mass loading than storm event
mass loading. The site location at Warren and the rain event at Portland,
contributed to the variability of river data.
The field survey results indicate that for a program to be cost-
effective, the sampling must be supplemented with modeling. Thus, before
regulations can be generated for industrial stormwater runoff, more prob-
lem definition (sampling and modeling) must be performed.
-301-
-------
INTRODUCTION
As most industries and many municipalities are meeting the point
source standards of the interim goal of 1977, the effect of non-point
source pollution on water quality is gaining more attention. The
National Commission on Water Quality reported that "non-point pollutant
sources are significant to the Commission's study because they may in
some instances overwhelm and negate the reductions achieved through
point source effluent limitations"1. Based on these findings, the
Commission recommended to Congress that "control or treatment measures
shall be applied to agricultural and non-point discharges when these
measures are cost-effective and will significantly help in achieving
water quality standards"2.
In January 1976, TRC - THE RESEARCH CORPORATION of New England was
retained by the Industrial Environmental Research Laboratory of the U.S.
Environmental Protection Agency to evaluate waterborne fugitive emissions
(non-point sources) in relationship to industrial activities. The objec-
tives of the program included:
o An evaluation of industrial sources which might contribute
to non-point source pollution.
o An assessment of present day sampling techniques for non-
point sources.
o A review of existing mathematical models for predicting
non-point source pollution.
o An evaluation (including a field program) of non-point
source pollution from the coal-fired utility and iron
and steel industries.
o The adaptation of a mathematical model for predicting
non-point source pollution from an industrial site.
-302-
-------
This paper presents the highlights of the coal-fired utility measure-
ment program. The main objective of this program was to quantify and
qualify runoff from two coal burning utilities and to measure the effect
of runoff on the receiving waters.
INDUSTRY SELECTION
Why should the runoff from the coal-fired utility industry be mea-
sured?
Electrical energy is generated from fossil and nuclear fuels at
approximately 1,000 sites in the United States. At these sites, coal
provides approximately 54% of the total heat input for electricity gen-
eration. In 1974, this amounted to a coal usage rate of 328 million
metric tons per year. Increasing demands for energy self-sufficiency
are likely to push coal usage up to 454 million metric tons per year by
1990. Subsequently, in 1990 coal storage, typically a 100-day supply,
will increase from the current 100 million tons to 138 million tons.
Land use for coal storage at electric facilities will increase to 81
million square meters from an approximate 1974 total of 58 million
square meters. Stormwater runoff from coal storage piles will also in-
crease 38% to an estimated yearly total of 100,000-140,000 cubic meters
per year3*1*.
The effect of stormwater runoff on the receiving bodies will become
more pronounced as water quality improves through regulation of point
sources. These projections are the basis for selecting the coal-fired
utility industry for a sampling program.
-303-
-------
TEST PLAN
TRC developed a test plan to quantify the pollutants associated with
storrawater runoff and to assess their effect on the receiving waters.
The field study was designed to determine:
1. Background conditions in the receiving water prior to a storm
event.
2. Volume of and pollutant concentrations in stormwater runoff
as a function of time for the storm event.
3. Post-storm effect of the runoff on the receiving water.
The following additional information was required to apply to the
predictive model development:
1. Air temperature and humidity
2. Dustfall accumulation
3. River flow rates
4. Surface permeability
To quantify the effect of stormwater runoff on the receiving body, a
sampling station would be installed upstream of potential plant site
effects and a second station would be installed downstream. Theoretically,
comparison of the data taken at these two sampling sites would show the
effect of runoff on the river water quality. The sampling sites were to
be placed at locations where representative samples could be taken with
respect to the river cross-section and depth. To allow adequate mixing
of the runoff in the river, placement of the downstream site was most
critical.
-304-
-------
To characterize the stormwater runoff, it would be necessary to in-
tercept some amount prior to its entering the river. As each runoff basin
was identified, sampling plugs would be deployed to catch representative
samples during a storm event. The flow rate of runoff would be quantified
to establish the time variable pollutant load on the river.
Composite samples would be collected from the river sampling stations
and runoff sites every ten minutes for the first 90 minutes of a storm
event. Half-hour samples would be collected after the first 90 minutes.
Each sample would be composited from five grab samples taken in a propor-
tionally smaller time period. Thus, five two-minute grab samples would
be composited to give a two liter ten minute sample, etc. During dry
periods before and after storms, samples would be collected hourly from
the upstream and downstream sites.
The pollutants analyzed in this program are listed in Table 1.
TABLE 1
PARAMETERS MEASURED IN UTILITY PROGRAM
pH Acidity/Alkalinity
Dissolved Oxygen Sulfate
Total Suspended Solids Iron
Total Dissolved Solids Aluminum
Temperature Manganese
-305-
-------
SITE DESCRIPTIONS
Two coal-fired steam electric generating facilities in Pennsylvania
were chosen for the field study to identify and quantify runoff character-
istics. These utilities were selected because they were located on rivers
which had good water quality. Specific characteristics'of each site are
shown in Table 2.
The Warren Station of the Pennsylvania Electric Company in Warren,
Pennsylvania is a small generating plant (84 Mw) and is used primarily as
a peaking facility. It is located on the Allegheny River below the Kinzua
Dam. This dam regulates the river flow at approximately 56 cubic meters/
sec (cms) with an average velocity of 0.3 to 0.6 meters/sec. Bituminous
coal is delivered by truck to the station on a daily schedule from mines
in Clarion County, Pennsylvania.
Figure 1 shows the site layout for the Warren Station. Coal pile
runoff is channeled to a drain pipe by a drainage ditch that parallels
unused railroad tracks next to an access road for the coal trucks. (The
road and tracks lie between drainage ditch and river on Figure 1.) The
drain pipe continually drains small quantities of leachate during dry
periods and substantial quantities of runoff during rainfall events. All
runoff from the coal pile must pass through the drain pipe for discharge
to the river. The paved access road is used by coal trucks to enter and
leave the coal unloading area. The road is covered with coal dust and
earthen materials, although the pavement is still visible through the
accumulation. The material is washed off during rainfall events. The
water drains across the road through a rockstrewn area of rubble approx-
imately 12 meters wide to the river bank. There are several distinctly
-306-
-------
TABLE 2
CHARACTERISTICS OF THE TWO SAMPLING SITES USED IN THE SURVEY
Utility
Plant
Location
Capacity
Mw output, net
Coal
Usage (metric tons/yr)
Source
Storage, metric tons
Sulfur %
Iron %
Manganese %
Aluminum %
Pennsylvania Electric Co.
Warren Station
Warren, PA
84
315,000
est. 1974
Clarion Co. , PA
27,200
1.84
0.35
0.003
0.56
Metropolitan Edison
Portland Station
Portland, PA
410
840,000
est. 1974
PA & W. VA
172,000
1.47
0.38
0.004
0.37
-307-
-------
Figure 1: Site Layout with Sampling Locations-
Warren Station of Pennsylvania Electric Co.,
Warren., PA
-308-
-------
visible areas where this road dirt and coal dust are carried to the river.
Vegetation is nonexistent in these drainage areas. Surface drains on
paved areas around the plant discharge into the main cooling water dis-
charge canal. The surface drains are not effective in collecting storm-
water due to the irregular pavement surface.
During the sampling period, the Allegheny River had a flow rate of
56 cms at the sampling sites. The river water was generally close to air
temperature and contained noticeable suspended material such as silt and
detritus. Plant cooling water was discharged at a rate of 3.6 cubic
meters per second into the discharge canal. A delta which split the dis-
charge in two segments had previously formed at the mouth of the canal.
The Portland Station of Metropolitan Edison Company is located in
Portland, Pennsylvania on the Delaware River. This 410 Mw station is
much larger than the national average of 150 Mw and is used as a baseload
station. Bituminous coal is delivered by railroad car from Pennsylvania
and West Virginia mines.
Figure 2 shows the site layout for the Portland Station. A substan-
tial portion of the stormwater runoff from the coal pile is intercepted
by the ash settling pond and never flows into the river. One sector of
the runoff does go to a surface drain and is discharged with parking lot
and road runoff into the river. Fly ash is kept on the north side of
the plant. There was little fly ash stored during the sampling period.
Stormwater runoff from this site washes directly into the river.
During the sampling period, the Delaware River had a flow rate of
300 cms and an average velocity of .3 m/sec. The water was generally
very clear and much colder than prevailing air temperatures. The river
-309-
-------
© Sampling locations
I I
Figure 2: Site Layout with Sampling Locations-Portland
Station of Metropolitan Edison Co., Portland, PA
-310-
-------
height and turbidity changed rapidly with rainfall activity in its water-
shed. The plant cooling water was discharged near mid-river from a sub-
surface discharge tunnel at an average rate of 5.4 cubic meters per
second.
WARREN SAMPLING PROGRAM
The test plan was implemented at Warren without major difficulties.
An upstream site was established approximately 152 meters upstream of the
cooling water intake, 7.6 meters from the river's edge. This location
was well upstream of the runoff area from the access road. An air-filled
buoy was used to suspend the pH/DO and temperature sensors and the sample
line at about half-depth, 1.2 meters above the bottom. Dissolved oxygen
(DO), pH and temperature were measured with a Model ODBC Aqua Monitor.
These data were recorded on a strip chart recorder during rainfall periods.
River samples were collected with an ISCO Model 16800L Sequential
Sampler. It was programmed to collect 200 ml grab samples every minute
to provide a two liter composite every ten minutes during the first 90
minutes of a rainfall period. From the 90th minute to the storm's end,
the sampler was programmed to collect 70 ml every minute to give a two
liter sample every half hour. During dry periods, the sampler collected
70 ml every two minutes to give an hourly composite of two liters. The
sequentializer was designed to provide backflushing before and after each
sampling sequence to preserve the time integrity of the sample. A river
depth profile was also made at this location as part of the predictive
model development.
-311-
-------
The downstream site was secured approximately 152 meters downstream
from the cooling water discharge-river interface, 46 meters from the
river's edge. Although the turbulent main flow of the river appeared to
be on the opposite side, a rudimentary mixing test indicated the sampling
location was adequate. An inflatable raft was used to suspend the sensor
probe and sample line at two meters, approximately mid-depth. The ISCO
Sequential Sampler was mounted on the raft since the distance to shore
was too far to run a sample line and maintain sample time integrity- An
Orbisphere Dissolved Oxygen meter and chart recorder housed onshore in a
tent were used to measure dissolved oxygen and temperature. River samples
were taken in the same time sequence as the upstream samples. A record-
ing rain gauge was installed at this site to record the rainfall rate
during storm events. A river depth profile was also conducted at this
location.
The mixing test was also performed to aid in the model development.
One dozen oranges were set adrift just above the upstream site and their
journey was timed past several landmarks and past the downstream site.
After traversing the rapids (approximately 75% of the distance from the
upstream site to the downstream site), the oranges dispersed across the
river fairly evenly. It was concluded that the downstream site was ade-
quately placed to obtain representative data.
The runoff drainage areas were well-delineated by the appearance of
vegetation between the road and the river. It was difficult to set the
sampling plugs vertically into the rocky, rubble-strewn surface, so they
were installed in the ground at a slight horizontal angle with the screen-
ed opening facing uphill. Approximately 50 plugs were distributed in the
-312-
-------
three main drainage areas and were kept covered with tape until a rainfall
event started. A compositing bucket was placed under the coal pile drain
pipe and an ISCO Sequential Sampler was used to collect runoff samples
from this bucket. In operation, the sampler intake was continually plug-
ged with push-along solids, and the sampler was replaced by manual
sampling.
PORTLAND SAMPLING PROGRAM
The program test plan remained basically unchanged for the Portland
survey, but modifications in implementing it were necessary to reflect
the differences between the two sites.
At Portland a small area of land north of the plant is used for fly
ash storage during winter months. Runoff from this area drains under the
plant fence and into the river. The upstream station was placed just up-
stream from this location approximately nine meters from the shore. An
air-inflatable raft was anchored at the upstream site to hold the sensors
and sample lines at mid-depth, approximately three meters below the sur-
face. The swift current, local turbulence and rocky bottom created great
difficulty when personnel attempted to anchor the raft to its station.
Broken and slipped moorings hindered the data-gathering effort at this
site throughout the program.
An ISCO Sequential Sampler was used to collect river samples in ex-
actly the same fashion as at the Warren site. In place of the ODBC Aqua
Monitor, an Orbisphere dissolved oxygen and temperature meter was used
for those parameters. A battery-powered chart recorder was used to re-
cord these data during a storm event.
-313-
-------
The lower station was established approximately 30 meters downstream
from the cooling water discharge tunnel, approximately 230 meters from
the upstream site. An identical instrumentation and sampling arrangement
as at the upstream site was used.
The coal pile runoff drained to both the ash pond and the storm
sewer. A portion of the runoff to the storm drain was intercepted for
sampling. Initially, an ISCO Sequential Sampler was installed in the
storm drain, but large coal particles continually plugged the sample in-
take line and pump. To solve this problem, sampling plugs were placed
in an array around the storm sewer inlet. Approximately 25 sampling plugs
were also deployed in the drainage basin of the fly ash storage area. As
with the sequential samplers, samples were collected every ten minutes
for the first 90 minutes, and every half hour for the duration of the
storm event.
During the one storm event sampled at Portland, river conditions
hampered sampling, since the only access to the river sites was by boat.
Runoff flow measurements were unsuccessful due to effluent turbidity
which masked the dyes and the complexity of storm drains which delayed
and trapped velocity markers. A second storm could not be sampled at
this site due to a prolonged dry spell, followed by the beginning of cold
weather and freezing conditions.
RESULTS OF FIELD SURVEYS
Despite the less than desirable amount of storm activity at Warren
and Portland, enough data were collected to show some interesting effects.
From the analyses of the coal pile runoff and receiving waters during dry
-314-
-------
and wet weather, some general characterizations can be made.
The laboratory analyses of the field data during dry and wet periods
at all sampling stations show a broad range of values. These ranges of
values were substantial enough to mask any apparent relationships between
sites and sampling locations. Several statistical summaries have been
prepared for selected pollutants during dry and wet periods at the two
sampling locations in the receiving body. These summaries included arith-
metic means, standard deviation, coefficients of variation and best esti-
mates of variance to provide the statistical basis for comparisons of re-
sults. These tests were performed within 95% confidence limits for the
selected pollutants of interest.
The calculations were done using two-tailed tests with non-detected
values distributed proportionally between 0 and the limit of detection
for the specific pollutant. It was assumed that the distribution of the
data was approximately similar to a normal distribution. A plot of
several sets of data on normal probability coordinates confirmed this
assumption. No attempt was made to evaluate the runoff data statistically.
WARREN RESULTS
Table.3 shows the range of pollutant concentrations at the various
sampling locations at the Warren Station. The only significant observa-
tion is that the pollutants in tKe coal pile discharge pipe are more con-
centrated during dry weather (leachate) than wet (runoff), as would be
expected.
The downstream pH values do appear lower under both wet and dry
sampling conditions. More data are necessary to establish a cause and
-315-
-------
TABLE 3
RANGE OF POLLUTANT CONCENTRATION AT THE SAMPLING LOCATIONS
AT WARREN STATION OF PENNSYLVANIA ELECTRIC CO., WARREN, PA
AUGUST"- SEPTEMBER, 1976
Pollutant
Total Suspended
Solids
Total Dissolved
Solids
Iron
Aluminum
Manganese
Sulfate
local Alkalinity
@ CaCOs
Total Acidity
@ CaCOa
pH
RANGE OF POLLUTANT CONCENTRATIONS ,- mg/1
Upstream
Dry
1-21
100 - 170
.14 - .40
N.D.1
.013 - .090
11 - 20
38 - 48
-
6.77 - 7.80
Wet
2-5
60 - 130
.09 - .17
N.D.1
.025 - .040
12 - 17
38 - 42
-
6.60 - 6.76
Downstream
Dry
1-11
80 - 180
.06 - .34
N.D.1
N.D.2- .040
11 - 22
36 - 45
-
6.77 - 7.60
Wet
2-12
-
.09 - 1.03
N.D.1- 26.6
.030 - .060
12 - 24
40 - 41
-
6.36 - 6.87
Coal Pile Discharge Pipe
Dry
12 - 19000
2300 - 21700
160 - 23500
20 - 1800
2 - 100
90 - 57000
-
200 - 38000
1.48 - 3.37
Wet
1700 - 13000
2300 - 115000
700 - 1400 .
70 - 100
9-15
1600 - 2700
-
1900 - 2900
2.35 - 3.36
JNone detected, <0.2 mg/1
2None detected, < 0.012 mg/1
-------
effect relationship between runoff and pH behavior in the river.
Table 4 presents the mean concentrations with 95% confidence limits
for selected pollutants in the Allegheny River. These data show the ex-
treme variability in the measurements made upstream and downstream.
Generally, the upstream and downstream sites appear to have similar pollu-
tant concentrations but a more detailed analysis was made using Student's
't1 and the 'F1 distribution tests.
Table 5 shows the results of the 't' and 'F' tests for comparisons
of data from the upstream and downstream sites during dry and wet periods.
There is no statistical difference between mean pollutant concentrations
at the upstream and downstream sites during dry weather. The sample var-
iances for TSS and Fe during the dry period at both sites were statis-
tically different.
Table 6 shows the characteristics of the coal pile leachate during
the dry weather sampling. The site layout with the drainage ditch and
pipe facilitated the collection of leachate samples. These data show
that the leachate is concentrated and extremely acidic. The leachate
flow rate was very low and no effect on the river was detected. The
total suspended and dissolved solids concentrations seem to be dependent
upon the length of time since the previous rain. As this time increases,
the concentrations decrease. The color of the leachate remained amber
during the dry period.
Table 7 shows the rainfall event data collected at Warren, Pennsyl-
vania. Other than suspended and dissolved solids, the pollutant concen-
trations were fairly consistent between the two storms. Since the first
storm was a short-term, moderately intense cloudburst, it was not possible
-317-
-------
TABLE 4
MEAN POLLUTANT CONCENTRATIONS WITH 95%
CONFIDENCE LIMITS IN THE ALLEGHENY RIVER AT
WARREN STATION OF PENNSYLVANIA ELECTRIC CO., WARREN PA
AUGUST - SEPTEMBER, 1976
Pollutant
TSS
S04
Fe
Mn
Alk
POLLUTANT CONCENTRATION, mg/1
Upstream
Dry
8.11 ± 2.26
13.89 ± 0.84
0.23 ± 0.02
0.028 ± 0.005
41.65 ± 0.85
Wet
7.25 ± 3.18
15.09 ± 0.94
0.12 ± 0.03
0.032 ± 0.003
40.33 ± 0.94
Downstream
Dry
4.13 ± 2.04
13.83 ± 1.45
0.21 ± 0.09
0.023 ± 0.005
39.33 ± 0.89
Wet
5.50 ± 2.71
16.65 ± 2.25
0.39 ± 0.27
0.043 ± 0.012
40.30 ± 0.34
95% confidence limits = x±t
V,.025 In
-318-
-------
TABLE 5
COMPARISONS OF MEAN VALUES & VARIANCES WITHIN 95% CONFIDENCE LIMITS
AT UPSTREAM & DOWNSTREAM SITES DURING DRY & WET SAMPLING PERIODS
WARREN, PENNSYLVANIA
AUGUST - SEPTEMBER, 1976
Pollutant
TSS
SO,,
Fe
Kn
Alk
TSS
SOi,
Fe
Mn
Alk
TSS
SOt,
5e
Mn
Alk
TSS
SOi,
Fe
Mn
Alk
Degrees of
Freedom
50
44
52
51
65
16
19
17
18
17
45
37
43
44
50
21
26
26
25
32
t Test
7.47
3.06
0.138
0.015
2.59
8.45
4.57
0.495
0.019
1.86
9.26
2.91
0.086
0.017
3.81
6.71
4.96
0.42
0.0202
2.80
Difference
Between
Means
UPSTREAM
3.98
0.06
0.02
0.005
2.32
UPSTREAM
1.75
1.56
0.27
0.011
0.03
UPSTREAM
0.86
1.20
0.11
0.004
1.32
DOWNSTRE
1.37
2.82
0.18
0.0200
0.97
Is Difference
Between Iteans
Significant?
DRY - DOWN
No
No
No
No
No
WET - DOWN
No
No
No
No
No
W F T - UPST
Xo
No
Yes
No
No
AM WET - DO
Xo
No
No
Marginal
No
Critical 'f for
95Z Confidence 'F' Ratio
STREAM DRY
2.69
2.53
2.416
2.422
2.173
STREAM WET
4.82
3.96
4.36
4.72
4.10
REAM DRY
2.50
2.57
2.52
2.42
2.44
WNSTREAM DRY
3.38
2.98
3.01
3.10
2.73
3.38
0.55
0.12
2.00
1.72
1.88
0.196
0.011
O.W
6.52
0.43
0.41
0.33
0.10
0.20
0.77
1.16
3.62
2.00
0.05
Is Difference
Between Variances
Significant?
Yes
No
Yes
No
No
No
Yes
Yes
Yes
Yes
No
No
No
Yes
Yes
No
No
Yes
No
Yes
Upstream > Downstream
Upstream < Downstream
Upstream < Downstream
Upstream < Downstream
Upstream < Downstream
Upstream > Downstream
Wet < Dry
Wet < Dry
Wet > Dry
Wet < Dry
-------
TABLE 6
CHARACTERISTICS OF COAL PILE LEACHATE-DRY WEATHER
AT WARREN STATION OF PENNSYLVANIA ELECTRIC CO., WARREN, PA
AUGUST - SEPTEMBER, 1976
Date
8/25/76
8/27/76
9/16/76
Hours Since
Last Rain
250
17
505
Pollutant Concentration, mg/1
TSS
200
18,700
12
TDS
40,000
82,600
21,700
S04
57,000
45,000
25,000
Fe
23,500
14,000
9,700
Al
1,800
1,400
1,100
Mn
100
70
70
Acidity
18,000
27,000
37,600
pH
2.4
2.1'
1.5
Discharge Flow Rate
1pm (gpm)
1.5 (.39)
1.5 (.39)
1.4 (.39)
O
I
-------
TABLE 7
CHARACTERISTICS OF RAINFALL EVENTS
AT WARREN STATION OF PENNSYLVANIA ELECTRIC CO., WARREN, PA
AUGUST - SEPTEMBER, 1976
Site
Warren
Storm
1
8/26/76
2
9/17/76
Elapsed Time
min
20
430
Total
Precipitation
mm (in)
2.8 (0.11)
8.9 (0.35)
POLLUTANT CONCENTRATION, mg/1
TSS
0.9
1.5
TDS
1679.
11.
SOtt
14.9
12.4
Fe
21.5
19.8
Mn
0.275
0.304
Al
3.18
2.71
pH
3.90
4.15
-------
to collect samples quickly enough to show any immediate effect of the run-
off on the river. From this storm it was verified that the plug locations
for sampling the surface runoff were adequate. The second storm lasted
much longer and provided the bulk of data for evaluation and conclusions.
Table 8 presents the characteristics of the coal pile runoff and the
surface runoff from the access road during the second storm event. At
the start of the storm, the "first flush" effect with its higher pollutant
concentrations can be seen. These values generally declined through the
rainfall period. Some perturbations do appear since the rain did not fall
at a constant rate throughout the day. All measured pollutant values are
lower during rain than during dry periods. When a comparison of the data
in Tables 6 and 8 is made, it appears that water stored in the coal pile
solubilizes various impurities in the coal and leaks out very slowly.
Rainfall washes out the stored water within the pile, thus greatly diluting
the impurities.
A comparison of the coal pile runoff with the dry weather leachate,
indicated that the rate of mass loadings of all pollutants on the river,
except suspended solids, is greater during the dry period. A closer ex-
amination of this behavior is warranted.
The surface runoff shows much similarity with the coal pile runoff
•-,
as the contaminant levels decreased through the storm's duration. Black
granular material was observed on the access road and runoff surfaces
during the field work, and the data indicate the presence of similar con-
taminants as in the coal pile runoff. The coal pile and surface runoff
responded very quickly to rainfall intensity. The ground around the coal
pile and the surface runoff areas had a very low porosity, practically
-322-
-------
TABLE 8
CHARACTERISTICS OF COAL PILE & ACCESS ROAD RUNOFF
DURING SECOND STORM EVENT AT
WARREN STATION OF PENNSYLVANIA ELECTRIC CO., WARREN, PA
17 SEPTEMBER 1976
TIME
1000 - 1015 - Rain Start
1015 - 1030
1030 - 1045
1045 - 1100
1100 - 1115
1115 - 1130
1130 - 1200
1200 - 1230
1230 - 1300
1300 - 1330
1330 - 1500 - Rain End
POLLUTANT (mg/1)
COAL PILE RUNOFF
TSS
9800
4200
6400
11400
5000
1700
1400
1600
1700
1700
23000
TDS
4600
3300
2400
2400
2500
3700
3800
3100
3000
-
500
SO 4
2300
2300
1600
1800
2100
2100
2700
1700
1000
-
200
FE
900
-
700
1400
700
500
-
300
200
-
-
AL
100
-
90
70
80
-
-
-
-
-
-
MN
40
-
10
10
10
-
-
-
-
-
2
ACIDITY
3200
. 2600
3100
2000
2200
2900
-
-
-
-
500
DISCHARGE
FLOW RATE
1pm (gpm)
22 (5.8)
20 (5.3)
20 (5.3)
17 (4.5)
AVERAGE SURFACE RUNOFF
TSS
11200
1400
4900
4400
3700
3000
1100
3100
1700
1500
2300
TDS
2800
900
900
600
500
400
600
700
700
-
1400
SO,
1000
600
500
900
400
200
600
500
-
-
1000
FE
100
200
200
-
300
200
-
200
-
-
400
AL
100
40
30
40
50
-
-
40
-
-
70
MN
10
5
4
5
5
-
-
7
-
-
6
ACIDITY
1700
600
600
500
300
100
500
500
-
-
1000
-------
zero. Within minutes after the rain stopped, the runoff declined to zero
and the coal pile discharge returned to its prior appearance and flow
rate.
The 't? and 'F1 tests presented in Table 5 show no statistically
significant effect of runoff on the river. However, in the case of sul-
fate, iron, manganese and alkalinity, the sample variances were signifi-
cantly different. In the cases where differences were noted, except for
total alkalinity, the upstream sample variance was lower than downstream.
This difference is partly related to the sampling locations. Although
both locations were as representative of the river's cross-section as
could be determined, it is likely that the downstream site contained a
greater number of anomalies. The river was very wide at this point with
a greater probability for peculiarities in flow patterns due to the delta
formation, rapids, and the large island just upstream of the site.
In a comparison of each river site during the wet and dry periods,
the data show only two statistically significant differences. At the up-
stream site, the data indicate a difference in the mean concentration of
iron. The dry period had much higher concentrations than the wet.period.
There was a marginal difference in manganese concentrations during wet and
dry periods at the downstream site. A comparison of these 'wet1 versus
'dry' variances with upstream versus downstream variances, indicates that
they are partly the result of differences in the characteristics of each
site as well as differences created by the rainfall events.
-324-
-------
PORTLAND RESULTS
Table 9 shows the range of values for each pollutant at the Portland
Station sampling sites. These ranges are similar to those measured at
the Warren Station sites. They commonly vary by up to an order of magni-
tude.
The pH values during the short sampling period at Portland appear to
cover a higher range downstream from the plant, contrary to pH values ob-
served at Warren.
Table 10 shows the 95% confidence limits for the upstream and down-
stream sites during dry and wet periods. As was true with the Warren
sampling data, most of the Portland data at each river sampling site seems
to be similar during both the 'dry' and 'wet' sampling periods. A com-
parison of Portland data with Warren data indicates that the Delaware
River at Portland has higher suspended solids, iron and manganese, but
lower alkalinity and similar sulfate concentrations.
Student's 't' and 'F' distribution tests of significance were per-
formed to establish any apparent relationships between sites and sampling
locations (see Table 11). As expected, the 't' and 'F1 tests on the dry
weather data show no significant differences between means or variances
at upstream and downstream sites. The sample variances at Portland were
noticeably greater than at Warren, due possibly to the smaller sample
size at Portland. The intrinsic characteristics of each river's behavior,
as well as the sampling techniques used, are also undefined contributors
to the sample variance.
The parameters associated with the single rain event at Portland are
presented in Table 12. The rain at Portland had noticeably lower con-
-325-
-------
TABLE 9
RANGE OF POLLUTANT CONCENTRATION AT THE SAMPLING LOCATIONS
AT PORTLAND STATION OF METROPOLITAN EDISON CO., PORTLAND, PA
OCTOBER 1976
Pollutant
Total Suspended
Solids
Total Dissolved
Solids
Iron
Aluminum
Manganese
Sulfate
Total Alkalinity
@ CaCOs
Total Acidity
(3 CaCO
PH
RANGE OF POLLUTANT CONCENTRATION, mg/1
Upstream
Dry
3-33
43 - 72
.18 - 2.0
N.oJ- .63
.03 - .14
10 - 18
12 - 25
-
6.2 - 6.8
Wet
10 - 20
62 - 89
.18 - .45
N.D.1
N.D?-. 03
9-22
16 - 19
-
6.5 - 6.8
Downstream
Dry
2-43
38 - 71
.18 - 1.4
N.D.1- 1.25
.01 - .18
5-12
12 - 21
-
6.3 - 7,2
Wet
4-11
46 - 67
.18 - .63
N.D.1
N.D.- .03
5-11
16 - 20
-
6.6 - 7.2
Coal Pile Runoff
Dry
-
-
-
-
-
-
-
-
-
Wet
220 - 3800
600 - 7500
18 - 400
2.75 - 88
3.75
380 - 6000
-
300 - 4600
2.35 - 3.10
Fly Ash Pile Runoff
Dry
-
-
-
-
-
-
-
-
-
Wet
840 -15200
730 - 2500
73 - 245
63 - 200
.03 - 25
100 - 1600
-
11 - 800
2.72 - 3.06
I
w
M
detected, <
2None detected, <
0.2 mg/1
0.012 mg/1
-------
TABLE 10
MEAN CONCENTRATIONS WITH 95% CONFIDENCE LIMITS
FOR SELECTED POLLUTANTS AT THE PORTLAND STATION
OF METROPOLITAN EDISON CO., PORTLAND, PA
OCTOBER 1976
Pollutant
TSS
S0i+
Fe
Mn
Alk
POLLUTANT CONCENTRATION, rag/1
Upstream
Dry
12.72 ± 4.86
12.86 ± 1.31
0.56 ± 0.22
0.051 ± 0.016
16.07 ± 1.82
Wet
13.54 ± 5.91
14.25 ± 6.12
0.30 ± 0.10
0.020 ± 0.010
17.60 ± 1.42
Downstream
Dry
11.66 ± 6.96
10.10 ± 1.10
0.56 ± 0.18
0.055 ± 0.020
15.59 ± 1.26
Wet
7.39 ± 2.20
8.15 ± 1.31
0.43 ± 0.21
0.016 ± 0.006
16.38 ± 0.43
-327-
-------
TABLE 11
COMPARISONS OF MEAN VALUES & VARIANCES WITHIN 95% CONFIDENCE LIMITS
AT UPSTREAM & DOWNSTREAM SITES DURING DRY & WET SAMPLING PERIODS
PORTLAND STATION
Ni
00
Pollutant
TSS
SOi,
Fe
Mn
Alk
TSS
SO
it
Fe
Mn
Alk
TSS
SO
Fe
Mn
Alk
TSS
SO
Fe
Mn
Alk
Degrees of
Freedom
27
35
37
36
29
11
16
16
16
11
16
20
22
21
17
22
31
31
31
23
t Test
17.48
3.32
0.55
0.05
4.20
9.82
7.34
0.631
0.044
2.22
17.40
7.02
0.791
0.058
6.52
20.29
3.38
0.568
0.055
3.79
Difference
Between
Means
U P S T R E
1.06
2.76
0
0.004
0.48
U P S T R E
6.15
6.10
0.13
0.004
1.22
U P S T R
0.82
1.39
0.26
0.031
1.53
D 0 W N S T R
4.27
1.95
0.13
0.039
0.79
Is Difference
Between Means Critical 'f for
Significant? 95% Confidence
AM DRY -
No
No
No
No
No
AM WET -
No
No
No
No
No
E A M WET
No
No
No
No
No
E A M WET
No
No
No
No
No
DOWNSTREAM DRY
2.96
2.62
2.53
2.55
2.86
DOWNSTREAM WET
5.52
4.12
4.04
4.04
5.52
- UPSTREAM DRY
4.12
3.73
3.44
3.50
4.00
- DOWNSTREAM DRY
3.29
2.72
2.73
2.73
3.22
'F' Ratio
0.378
1.18
1.21
0.50
1.65
3.26
5.16
0.091
1.00
4.81
0.350
3.76
0.052
0.100
0.131
0.041
0.861
0.688
0.05
0.045
Is Difference
Between Variances
Significant?
No
No
No
No
No
No
Yes Cpstrean > Downstream
Yes Upstream < Downstream
No
No
No
Marginal Dry < Wet
Yes Dry > Wet
Yes Dry < Wet
No
Yes Dry > Wet
No
No
Yes Dry > Wet
Yes Dry > Wet
-------
TABLE 12
CHARACTERISTICS OF THE RAINFALL EVENT AT
PORTLAND STATION OF METROPOLITAN EDISON CO., PORTLAND, PA
20 OCTOBER 1976
i
u>
NJ
\O
I
Site
Portland
Storrc
10/20/76
Elapsed Time..
win
810
Total
Precipitation
mm (in)
3.92 (1.55)
POLLUTANT CONCENTRATION, mg/1
TSS
2.
TDS
102
so,
9.7
Fe
0.18
i
Mn
ND1
Al
ND2
PH
5.70
*None Detected, < 0.012 mg/1
2None Detected, < 0.2 mg/1
-------
centrations of iron, manganese and aluminum than at Warren. Other para-
meters were similar between the two sites.
When compared with the Warren data, the coal pile runoff has sub-
stantially lower concentrations of pollutants (see Table 13). In part,
this is the result of the different sampling procedures required at each
site as determined by the site layout. At Warren, the entire runoff from
the coal pile was intercepted by a drainage ditch. At Portland, only a
small portion of the total runoff was captured from a coal pile that was
much farther from the sampling location. Collection of samples had to
be made near the surface drain since the terrain near the pile was uncer-
tain and the survey objective was to examine only the portion draining to
the river. It is also possible that the distance between the coal pile
and the surface drain allowed the soil to filter pollutants out of the
runoff.
Compared with Warren, the response of runoff flow at Portland was
much slower (i.e., there was a greater time lag) with respect to the rain-
fall intensity. The runoff did have sufficient force to transport quite
large (1-5 mm) particles. Plug sampling replaced automatic sampling after
the sequential samplers became inoperative from being jammed with these
particles. The plug collectors, even with screen covers, did collect some
of the push-along particles that the sequential sampler did not. This
could explain the change of pollutant concentrations at 1000 hours. The
large particles were removed from the sample within a few hours prior to
returning the samples for analysis, but their partial dissolution could
explain the increase in concentrations. There is also the possibility
that rainfall intensities, with their effect of washing out more of the
-330-
-------
TABLE 13
CHARACTERISTICS OF COAL PILE & FLY ASH AREA RUNOFF
DURING THE RAINFALL EVENT AT PORTLAND STATION OF
METROPOLITAN EDISON CO., PORTLAND, PA
20 OCTOBER 1976
TIME
0700 - 0730
0730 - 0800
0800 - 0830
0830 - 0900
0900 - 1000
1000 - 1100
1100 - 1200
1200 - 1300
1300 - 1400
1400 - 1500
1500 - 1700
17-00 - 1830
POLLUTANT (mg/1)
COAL PILE RUNOFF
TSS
240
300
350
-
230
280
-
-
1700
2200
2200
3800
TDS
-
-
500
600
600
3400
-
-
4200
7500
4800
4300
S04
-
-
500
500
400
2000
-
-
-
6000
-
2600
FE1
20
40
-
60
80
400
-
-
300
-
200
400
AL1
8
19
-
15
15
50
-
-
30
-
50
90
MN1
0.4
0.8
-
0.6
0.5
1.8
-
-
1.6
-
2.5
2.5
ACID-
ITY
290
300
-
370
300
2400
-
-
-
4600
-
2600
FLY ASH AREA RUNOFF
TSS
200
400
-
-
800
1200
6600
-
-
14000
2300
1200
TDS
200
400
-
-
1300
1700
2200
-
-
1800
1000
800
S04
100
200
-
-
-
1100
1400
-
-
1200
800
700
FE
-
-
-
-
100
200
-
-
200
150
-
AL
—
-
-
-
-
60
140
-
-
200
60
-
MN
—
-
-
-
--
2.3
1.3
-
-
0.2
1.2
-
ACID-
ITY
10
200
-
-
400
700
900
500
600
500
I
u>
UJ
1Concentrations as entering the river
-------
soluble material, could have caused this increase. The fly ash pile area
released much higher concentrations of suspended and dissolved solids into
the river than the coal pile runoff released. Acidity, sulfate, and metals
concentrations were lower. The flow from the coal pile and the fly ash
storage area could not be quantified with any success. If the study had
continued for another rain event, semi-permanent weirs would have been
installed to eliminate this problem.
As indicated in Table 11, the runoff from the coal pile and fly ash
area did not have any measurable effect on the river. Statistically,
there was no measured difference at either site during the wet and dry
sampling periods. These observations must be mitigated by the small
sample size as well as statistically significant differences in the sample
variances.
The sample variances at Portland, except for dry weather comparisons,
are statistically different for each of the compared sample sets. There
is no apparent consistency to these differences with respect to pollutant,
site, or sampling condition. It can be concluded that a rain event does
introduce an additional degree of variability to the data. The Portland
'dry1 data shows no difference in variance between samples taken at the
two sampling sites. This contrasts with the Warren 'dry' data which did
have some variances. The sampling location is another factor affecting
the Warren data but not the Portland data, where the river flow pattern
was less complex. The sample variances are similar at Warren and Portland
for each pollutant with few exceptions, despite the slight differences
in sample size. Total suspended solids and iron seem to have the greatest
degree of variation at both sites under the different sampling conditions.
-332-
-------
CONCLUSIONS
The following conclusions resulted from the field survey:
1. The pollutant concentrations in the river at both sites were
highly variable, often by an order of magnitude. These var-
iations were independent of river flow and weather conditions.
2. The mass loadings of pollutants in the Delaware River increased
substantially during and after the sampled storm event. This
was due primarily to an increased flow attributable to upstream
conditions and storm intensity. The mass loading of pollutants
in the Allegheny River remained essentially unchanged for both
sampled storm events since river flow was controlled by a dam
approximately six miles upstream and neither storm event was
substantial. Therefore, the pollutant concentrations in each
river at both upstream and downstream sampling stations were
not necessarily higher during storm conditions.
3. The data from these two sites generally show no statistical
difference in mean concentrations of upstream versus down-
stream pollutant levels in either dry or wet conditions.
4. The data show no statistical difference in sample variances
which are not consistently predictable with respect to pollu-
tant, site, and sampling period.
2
5. The main contributors to the change in S of the calculated
variance were site location and the storm event. The site
location was the major contributor "at Warren, while the rain
event was the major contributor at Portland. The sample
variances were generally consistent for each pollutant at
the Warren and Portland sites. The only exceptions were
total suspended solids and iron.
6. The storm data from Warren show a "first flush" effect from
the initial runoff of the access road which contained fugi-
tive fallout from the coal pile and coal trucks.
7. The pollutant concentrations of the leachates from the coal
pile at Warren were orders of magnitude higher than the
storm runoff pollutant concentrations. For a short dura-
tion, moderate intensity storm and a moderate duration low
intensity storm (the two events sampled at Warren), the
leachate mass loading was greater than the storm mass loading
because the leachate drained for several days. Thus, for
the two storms sampled at Warren, the pollutant loads on
the river from the utility were less during rain than during
dry weather with the exception of total suspended solids.
-333-
-------
8. The sample plugs worked effectively except for one problem;
soluble solids slipped through the screen prior to runoff
and leached into the sample, creating higher measured pollu-
tant concentrations.
The Warren and Portland station data do not show any coal pile run-
off impact on the river. It appears sample sizes may be too small (due
to lack of rain) to indicate a definitive conclusion of the runoff effects
at either Warren or Portland. The data certainly can be improved with a
larger data base (more rain events) and some improvement in sample var-
iances. This program points out the inherent problem of sampling runoff;
i.e., the need of representative storm events. A field program of this
nature can be costly if it does not rain. Even if it rains, a sampled
event must be related to other cases of rainfall duration and intensity
and to antecedent dry days before the storm. It is these problems which
make the use of a mathematical model to predict runoff and its impact on
receiving waters so valuable. Therefore, before regulations can be gen-
erated for industrial stormwater runoff, more problem definition (sampling
and modeling) must be performed.
-334-
-------
References
1. Staff Report, National Commission on Water Quality (Washington, D.C.,
April 1976).
2. Report to the Congress, National Commission on Water Quality
(Washington, D.C., April 1976).
3. Development Document for Effluent Guidelines and New Source Perfor-
mance Standards for the Steam Electric Power Generating Point Source
Category (Washington, B.C.: U.S. Environmental Protection Agency,
Effluent Guidelines Division, EPA-440/l-74-029-a, 1974).
4. 1975 Keystone Coal Industry Manual (New York, N.Y.: McGraw-Hill,
Inc., 1975).
-335-
-------
THE DEVELOPMENT OF A MATHEMATICAL MODEL
TO SIMULATE INDUSTRIAL NON-POINT SOURCE POLLUTION
By
James J. Binder, P.E.
and
Gordon T. Brookman, P.E.
TRC - The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
-337-
-------
ABSTRACT
Non-point source pollution is presently a significant concern to
those responsible for maintaining and improving water quality. Although
non-point source pollution has been studied in some detail for urban
environments, industrial non^point source pollution, a potentially
critical problem area, has not received jjuch attention..
\
The IKS.V Environmental Protection Agency, Industrial Environmental
Research Laboratory at Research Triangle Park, North Carolina, has re-
tained TRC - The Research Corporation of New England to identify non-
point source pollution from twelve industries, and to quantify and qualify
non-point source pollution from the coal-fired utility and iron and steel
industries. Recognizing that it is extremely costly to conduct extensive
field measurement programs on a site-by-site basis, the USEPA asked TRC
to develop and apply a mathematical model (in conjunction with a field
measurement program) capable of simulating both the quantity and quality
of industrial non-point source pollution and its impact on receiving
waters.
The TRC-developed model was applied at two coal-fired utility plants
to simulate non-point source pollution resulting from stormwater runoff
from coal storage areas and from plant areas covered with dust fallout
from coal handling and ash handling operations. When compared to field
measurements, the modeled results compared within a factor of 4 for both
the quantity and quality of stormwater runoff and its impact on the water
quality of the receiving waters. These results indicate that the model
-338-
-------
can be used with a minimum of field data to successfully simulate indus-
trial non-point source pollution and its impact on receiving waters for
the utility industry. In addition, due to inherent flexibility, the TRC
model can serve as a useful, cost-effective tool to evaluate non-point
source pollution from other industries.
A description of the TRC model development program and its results
are presented in this paper.
-339-
-------
1.0 INTRODUCTION
Non-point source pollution is presently a significant cause of con-
cern to those responsible for maintaining and improving water quality.
Although non-point source pollution has been studied in some detail for
urban environments, industrial non-point source pollution, a potentially
critical problem area, has not received much attention.
The U.S. Environmental Protection Agency, Industrial Environmental
Research Laboratory at Research Triangle Park, North Carolina, has re-
tained TRC - The Research Corporation of New England to identify non-
point source pollution from twelve industries and to quantify and qualify
non-point source pollution from two of these industries: coal-fired
utility and iron and steel. Recognizing that it is extremely costly to
conduct extensive field measurement programs on a site-by-site basis,
the USEPA asked TRC to develop and apply a mathematical model capable of
simulating both the quantity and quality of industrial non-point source
pollution and its impact on receiving waters. The TRC-developed model
was applied at two coal-fired utility stations and evaluated in conjunc-
tion with field measurements made at these stations.
A description of the TRC model development program, including model
selection, application, and evaluation, is presented herein.
-340-
-------
2.0 MODEL SELECTION
Prior to the work described in this paper, little had been done to
develop a mathematical model to quantify and qualify industrial non-point
source pollution and its impact on receiving waters for specific indus-
tries, with the exception of agriculture and mining. The objective of
this program was to develop such a mathematical model capable of quantify-
ing and qualifying non-point source industrial loadings associated with
stormwater runoff - the predominant mechanism for non-point source pollu-
tion - and the impact of such runoff on receiving waters. To increase
model utilization, the model was to be inherently flexible so that it
could be applied to various types of industry with only minor modifications.
To effectively satisfy the above objective, existing mathematical
models were reviewed and the model best able to meet the study objective
was chosen for development and adaptation.
Ten models were selected for review. A comparison of model capa-
bilities, application, complexity, cost, and availability is presented in
Table 1.
Of the ten models reviewed, the simplest, most flexible model requir-
ing the least amount of modification with the capability to quantify and
qualify stormwater runoff from industry and to determine the impact of
such runoff on receiving waters was the Short Stormwater Management Model
and Receiv II (SSWMM-RECEIV II).
The Short Stormwater Management Model (SSWMM) and Receiv II (RECEIV
II) are both modified versions of the EPA-SWMM model. SSWMM, developed
by the University City Science Center in 1976, is a simplified version of
-341-
-------
TABLE 1
A COMPARISON OF MOD#L CAPABILITIES, APPLICATION, COMPLEXITY, COST, AND AVAILABILITY
Model Identification
Name
EPA SWIM Release II
1
WRE Stormwater
Management Model 2
SSWMM/ 3
RECEIV II i*
Hydro coir.p Simula-
tion Program 5
Dorsch Consult 6
Corps, of Engi-
neers STORM 7
Battelle Waste —
Water Management
Model 8
Metcalf&Eddy Sim-
plified Stormwater
Management Model 9
EPA-Hydrocomp
ARM 10
Pyritic Systems: A
Mathematical Model
Date Released
1.974
L973
L976
L97Z
1974
L975
L975
L976
L976
L976
L972
Model Capabilities
Wastewater
Quantity
X
X
X
X
X
X
X
X
X
X
fN
4J
•H
rH
ra
3
O"
X
X
X
X
X
X
X
X
X
X
Dry Weather
Flows
X
X
X
X
X
X
X
X
X
Storm Runoff
X
X
X
X
X
X
X
X
X
X
Receiving Water
^
u
•H
U
C
n]
3
Cf
X
X
X
X
X
D
Quality
X
X
X
X
X
D
River
X
X
X
X
X
D
$
CO
X
X
X
X
X
D
Estuary
X
X
X
X
D
Quality Parameters
Temperature
•
X
Suspended
Solids
X
X
X
w
w
w
w
Total Dis-
solved solid?
X
i/>
§
w
X
X
X
X
X
w
w
w
§
u
w
D
W
Dissolved
OxvRen
X
X
X
Nitrogen
W
X
X
X
D
W
U
W
w
Phosphorous
W
X
X
X
D
W
W
W
W
X
o.
X
•o
C a
ca in
£3
r-l O
•H lJ
0 0
W
X
Pesticides
W
H2S04 - Only
Simulation ot
Single Storm
X
X
X
X
X
X
X
Simulation of
Multiple Storm
X
X
X
X
X
X
X
X
Potential
Model Application
Agriculture
X
X
X
X
X
X
X
X
X
60
c
•H
c
•H
S
X
X
X
X
X
X
X
X
X
Construction
X
X
X
X
X
X
X
X
>,
U)
IJ
y
c
bl
X
X
X
X
X
X
X
X
Other
Industries*
X
X
X
X
X
X
X
X
Program Modification I
Required 1
X
X
X
X
X
X
X
X
X
Relative Model ~1
Conn! p>H fv 1
H
U
M
H
H
M
M
L
M
M
Relative Model Cost
H
H
M
H
H
M
M
L
M
M
Computer Program
Available
X
P
P/X
P
P
X
X
X
X
X
*Cement, feedlota; inorganic cheoicals; fertilizer manufacturing; petroleum refining; iron and steel; non-ferrous metals; phosphate manufacturing;
timber.
- Key -
X _ Yes H ™ Complex/costly
W - Wastawater only M " Moderately complex/moderately costly
D - Currently being developed I- - Simple/low cost
P - Proprietary
-------
the runoff portion of the EPA-SWMM model, and RECEIV II, developed by the
Raytheon Company for the EPA in 1974, is a modified version of the receiv-
ing water portion of the EPA-SWMM model. When combined, SSWMM and RECEIV
II are capable of dynamically simulating both the quantity and quality of
stormwater runoff and the impact of such runoff on the quantity and quality
of receiving waters, including rivers, lakes, and estuaries. The user can
define, with certain restrictions, the quality parameters which he chooses
to simulate. Pollutant transport can be modeled by both overland flow and
sewer routing. Dry weather flows can also be simulated. The model is
primarily designed to simulate individual storm events but can be used to
model multiple storm periods.
Potential model application for both runoff and receiving water
simulation includes all industry categories identified in Table 1; however,
model development and modification is required for any such application.
-343-
-------
3.0 MODEL DEVELOPMENT
Model development is described in terms of the model development
tasks, a description of SSWMM-RECEIV II as developed by TRC, and model
utilization.
3.1 Model Development Tasks
To meet the program objectives, it was necessary to both modify
and interface SSWMM and RECEIV II. Specifically, TRC:
o Modified the stormwater runoff quality relationship in
SSWMM to make the model more suitable for industrial
application.
o Modified RECEIV II to increase the model's sensitivity
in the receiving water to a specific plant's point and
non-point discharges.
o Created a program to combine SSWMM and RECEIV II.
3.2 Description of SSWMM-RECEIV II
3.2.1 General Description
The SSWMM-RECEIV II model as developed by TRC12 consists of four
programs:
SSWMM (Short Stormwater Management Model Program)
LNKPRG (Link Program)
SETUP/QUANTITY (RECEIV II Quantity Program)
QUALITY (RECEIV II Quality Program)
SSWMM simulates both the quantity and quality of stormwater runoff.
LNKPRG interfaces SSWMM and RECEIV II (SETUP/QUANTITY and QUALITY).
RECEIV II SETUP/QUANTITY simulates hydraulics in the receiving water and
the impact of the stormwater runoff on these hydraulics. RECEIV II
QUALITY simulates water quality in the receiving water and the impact of
-344-
-------
the stormwater. runoff on the quality of the receiving water. A flowchart
for SSWMM-RECEIV II is presented in Figure 1.
SSWMM-RECEIV II is written in Fortran IV and was developed for
installation on a Univac 90/30 digital computer with a basic compiler
(equivalent to an IBM 370 Level G Compiler). The program requires 100K
bytes of core storage.
3.2.2 Input Information Requirements
Model input information requirements are fairly extensive and best
described in terms of the individual program requirements for SSWMM,
LNKPRG, SETUP/QUANTITY and QUALITY.
SSWMM Input
SSWMM input includes information such as physical descriptions of
user-selected simulation elements, storm activity, and pollutant generation
and washoff data.
As initial steps in preparing input information, the user must
divide the land area to be modeled into discrete spatial elements repre-
sentative of drainage patterns and land use characteristics, and must
choose the temporal framework for model computation. The discrete ele-
ments can be either subcatchments (drainage areas within a watershed with
overland flow) or gutters (drainage ditches, pipes, manholes, and inlets;
i.e., points of runoff entry to receiving waters). Information necessary
to establish this spatial framework is normally available from plant
engineering drawings. The temporal framework (computational timestep
length) for SSWMM should be chosen to reflect storm activity and the
user's needs.
-345-
-------
/
SSWMM
Input
>_
SSWMM
>
r
Vfc
r
SSWMM
Output
File
LIMKPRG
Output
File
r
SETUP
Input
SETUP
Output
File
QUAN
Input
QUAN
Output
File
SSWMM
Printout
QUAL
Input
>
r
QUAN
>
r
t
>
QUAN
Printout
>
r
QUAL
>
r
QUAL
Printout
Fig. 1
SSWMM - RECEIV II Flowchart
-346-
-------
Information necessary relative to storm activity includes rainfall
intensity, storm duration, and the number of days between storms. This
data may be obtained from a local National Weather Service meteorological
station or from rainfall data gathered by plant personnel. If neither
source is available, a rain gauge must be installed to measure rainfall
intensity at the plant.
Information model requirements for pollutant generation and washoff
data are critical. The amount of pollutant washed from the land surface
during a storm is, in part, related to the initial (pre-storm) mass of
pollutant on the land surface. The initial pollutant mass load is equal
to the dust and dirt accumulation rate multiplied by the area of the
watershed with that dust and dirt accumulation rate, the number of dry
days between storms, and the amount of a particular pollutant in the dust
and dirt. The dust and dirt accumulation rate and the amount of a par-
ticular pollutant in the dust and dirt can be determined by field measure-
ment and laboratory analysis, or from information available in the
literature. For most industrial sites, very little pollutant generation
data is available in the literature, and it is necessary to conduct a
field measurement and laboratory analysis program. The area of the
Watershed with the dust and dirt accumulation rate and the number of dry
days between storms are determined from the physical descriptions of the
simulation elements and from storm activity records.
-347-
-------
LNKPRG Input
LNKPRG input includes the information output file from SSWMM and
an input card deck. The card input consists of user-determined program
interface instructions to link SSWMM and RECEIV II (SETUP/QUANTITY,
QUALITY) and non-stqrm inputs to or withdrawals from the receiving water
(background receiving water flows and pollutant mass loads, industrial
process and cooling water flows and pollutant mass loads, etc.). Infor-
mation on non-storm inputs can be gathered from plant personnel and from
government groups such as the United States Geological Survey (USGS),
National Oceanic and Atmospheric Administration (NOAA), and the U.S. Army
Corps of Engineers.
SETUP/QUANTITY Input
Input requirements for the SETUP/QUANTITY portion of RECEIV II
include the information output file from LNKPRG and two input card decks.
The input card decks include geographical, hydraulic, and meteorological
data describing the receiving water.
As with SSWMM, in SETUP/QUANTITY the user must first divide, the
receiving water to be modeled into discrete elements representative of
the waterway's hydrology and characteristic uses (industrial withdrawals
and discharges, etc.), and choose, with certain restrictions, the tem-
poral framework for model computation. The discrete elements include
nodes or junctions (sections"of the receiving water with uniform hydraulic
and water quality properties) and channels linking the nodes. Informa-
tion necessary to establish this spatial framework is normally available
from National Ocean Survey bathymetric charts, USGS 7.5* topographic maps,
-348-
-------
and U.S. Army Corps of Engineers flood studies. More specific spatial
information might be available from plant personnel if they had conducted
any studies on the receiving water adjacent to the plant. The temporal
framework (computational timestep length) for SETUP/QUANTITY should be
chosen to reflect the user's needs, but must meet certain restrictive
requirements determined by the choice of the spatial framework.
In addition to this geographical and hydrological information,
meteorological information, including rates of rainfall and evaporation
which influence the volume and flow of water in the receiving water, can
be input to SETUP/QUANTITY. If required, this information is normally
available through the National Weather Service.
QUALITY Input
The QUALITY information requirements include the information output
file from SETUP/QUANTITY and a card deck input. The card deck input
includes information describing the initial pollutant concentrations in
the receiving water and pollutant reaction kinetics (reaction rates,
water temperatures, and temperature compensation coefficients). Values
for initial pollutant concentrations in the receiving water can be de-
termined from the USGS or NOAA, but more specific information might be
available from plant personnel. Information on pollutant reaction kinetics
is often available in the literature. If this information is not ade-
quate, a field measurement program may be conducted to determine reaction
kinetics; however, this is often costly.
-349-
-------
3.2.3 Output Information
Model results are printed for each of the programs (SSWMM, LNKPRG,
SETUP/QUANTITY, QUALITY) in the SSWMM-RECEIV II model.
Results from SSWMM include:
o Initial pollutant loads (mg) on each subcatchment prior to
the storm.
o Stormwater flow (cfs) and associated pollutant mass loads
(Ibs./min.) for each timestep.
o Total amount of rainfall (cu. ft.), total infiltration
(cu. ft.), total runoff (cu. ft.), total surface storage
(cu. ft.), and the percentage error computed for unac-
counted water.
o Total pollutant mass (Ibs.) washed from the land surface
during the storm.
LNKPRG results include the stormwater flows and pollutant mass
loads from SSWMM converted to a format acceptable to RECEIV II (SETUP/
QUANTITY, QUALITY).
Results from SETUP/QUANTITY include:
o Hydraulic head (m) or water level in the receiving water
at each node for each timestep.
o Water flow (m3/sec) and velocity (m/sec) in the receiving
water in each channel for each timestep.
Results from QUALITY include:
o Pollutant concentrations (mg/1) in the receiving water at
each node for each timestep.
o Daily maximum, minimum, and average pollutant concentra-
tions (mg/1) in the receiving water at each node.
-350-
-------
3.3 Model Utilization
SSWMM-RECEIV II can be used to effectively simulate industrial non-
point source pollution associated with stormwater runoff from material
storage piles and from areas of dust and dirt accumulation without per-
forming a detailed field measurement program. The model can also be used
to simulate the subsequent impact of this runoff on receiving waters
(rivers, lakes, or estuaries). Pollutants that can be modeled are user-
selected and include items such as solids, nutrients, and metals.
Typical model applications for new or existing plants might include:
o Defining industrial stormwater runoff flow and pollutant
concentrations.
o Defining the impact (flow and pollutant concentration
changes in the receiving water) resulting from the
stormwater runoff.
o Defining design criteria for stormwater treatment.
o Evaluating stormwater treatment alternatives.
In addition to the industrial uses cited above, SSWMM-RECEIV II can
be used with minor modification to simulate non-point source pollution
associated with stormwater runoff for urban and rural environments.
This information can be used in river basin planning or 208 planning
activities.
-351-
-------
As with any mathematical model, care must be taken to apply SSWMM-
RECEIV II correctly. The user must understand and work within the model
limitations. At this time SSWMM-RECEIV II:
o Cannot simulate stormwater percolation through or the
erosion of material storage piles, but can simulate
stormwater runoff from material storage piles.
o Has not been tested to simulate dynamic background
source flows and loadings in the receiving water.
o Must be used within temporal and spatial limits defined
in the model.
-352-
-------
4.0 MODEL APPLICATION
To test the validity of SSWMM-RECEIV II, the model was used to simu-
late storm-induced, non-point source pollution and the impact of such
pollution on receiving waters at two coal-fired utility plants. The
results were compared to field measured data. The two plants were the
Warren Generating Station (Pennsylvania Electric Company) on the Allegheny
River in Warren, Pennsylvania, and the Portland Generating Station (Metro-
politan Edison Company) on the Delaware River in Portland, Pennsylvania.
Table 2 summarizes the plant characteristics. The pollutants modeled
included total suspended solids, total dissolved silids, sulfates, total
iron, manganese, and aluminum.
The model test inputs and the comparison of the modeled and measured
field data are discussed in Sections 4.1 and 4.2.
4.1 Model Test Inputs
As discussed previously, model inputs to SSWMM-RECEIV II include land
and receiving water discretization information, pollutant generation
activity on land, storm activity, and background flow and pollutant load-
ings in the receiving waters.
As depicted in Figure 2, the land area at Warren was divided into
three watersheds based on drainage patterns and land use. These watersheds
are represented by Elements 1, 4, and 6. The discharge point of drainage
from each of the three watersheds to the river is classified as an inlet.
These inlets are labeled as Elements 3, 5, and 8. The first watershed
(Element 1) represents the coal pile drainage area. The runoff from this
area enters a culvert (Element 2) which empties into the Allegheny River
-353-
-------
TABLE 2
CHARACTERISTICS OF THE SAMPLING SITES USED IN THE SURVEY
1
Utility
Plant
Location
Capacity
Mw output, net
Coal Characteristics
Usage (metric tons/
yr)
Source
Storage, metric tons
Sulfur %
Iron %
Manganese %
Aluminum %
Coal Delivery
Pennsylvania Electric Co.
Warren Station
Warren , PA
84
315,000
est. 1974
Clarion Co. , PA
27,200
1.84
0.35
•
0.003
0.56
Truck
Metropolitan Edison
Portland Station
Portland, PA
410
840,000
est. 1974
PA & W. VA
172,000
1.47
0.38
0.004
0.37
Railroad Car
-354-
-------
r
•
L
Culvert to
drain coal pile
Cooling water
discharge canal
I
Legend
Element number
Watershed boundary (subcatchment)
Fig. 2
Discretization scheme schematic
of land area; Warren, PA.
-355-
-------
at Element 3. The second watershed (Element 4) describes the drainage
area between the plant, the coal pile, and the river. Runoff from this
area enters the Allegheny River at Element 5. The third watershed
(Element 6) includes the drainage area adjacent to the plant. Runoff
from this drainage area enters the cooling water discharge canal (Element
7) which empties into the Allegheny River at Element 8. Physical di-
mensions for each of these elements are listed in Table 3.
Based on the physical characteristics of the river and the location
of the stormwater runoff discharges to the river, the Allegheny River was
divided into ten nodes and eight channels, as illustrated in Figure 3.
Background river flow and pollutant loadings enter node 1. Stormwater
runoff discharges from the coal pile drainage area and the drainage area
between the plant, coal pile, and the river enter the river at node 4,
and the stormwater runoff discharge from the drainage area adjacent to the
plant enters the river at node 5. Cooling water for the power plant is
withdrawn at node 4 and is discharged at node 5. The river branches at
node 5 and is unified at node 9. Since the model requires the last node
in the discretization system to be immediately downstream from a dam,
even if one does exist, an artificial dam, i.e., a dam with no elevation
above the river bottom, was placed between nodes 9 and 10. Channels
connect each of the nodes except where the nodes are interrupted by the
dam. The physical characteristics for the node and channel discretization
scheme are listed in Table 4.
Based on drainage patterns and land use, the land area at Portland
was divided into four watersheds, labeled as Elements 1, 3, 5, and 8, in
-356-
-------
TABLE 3
DISCRETIZATION: LAND ELEMENTS
WARREN, PA
Element No.
1
2
3
4
5
6
7
8
Type
1
2
2
1
2
1
2
2
Width
(ft)
338.00
1.00
0.00
488.00
0.00
275.00
25.00
0.00
Slope
(ft/ft)
0.0180
0.1370
0.0000
0.1370
0.0000
0.0450
0.0220
0.0000
AR OR GL2
2.61
153.13
0.00
1.71
0.00
2.53
325.00
0.00
PI OR DF
100.00
0.95
0.00
100 . 00
0.00
100.00
23.75
0.00
Type 1 is a watershed
Type 2 is a pipe or manhole
2
AR = Area of watershed (AC); PI
GL = Pipe length (Ft) ; DF
Percent imperviousness (%)
.95 * Width
-357-
-------
N
06
mi
Plant
River mi. 185.43 • Field Station
River mi. 185.04
River mi. 184.69
Artificial Dam - elev. 0.0 [
River mi. 186.92
River flow - 68.66 m3/sec (2424 cfs)
185.83- Field Station 10
. 1 85.73 Cooling water withdrawal - 3.6 m3/sec { 1 27 cfs )
Coal pile storm and direct runoff input
mi. 185.63
Cooling water input - 3.6 m3/sec {127 cfs)
Branch withdrawal - 17.9 m3/sec (632 cfs)
Storm sewer input
River mi. 184.33
Branch Input -17.9 m3/sec (632 cfs)
UJ
UJ
LU
_l
J
<
0
10
River mi. 184.10
Legend
• Node (Junction)
1 Node (Junction) number
= Channel
[j| Channel number
Fig. 3
Discretization scheme schematic
of Allegheny River; Warren, PA.
-358-
-------
TABLE 4
DISCRETIZATION: RIVER ELEMENTS
WARREN, PA
Nodes (Junctions)
Junction
Number
1
2
3
4
5
6
7
8
9
10
Surface Area
(Sq M)
184072.00
111646.00
29350.00
23184.00
39078.00
68644.00
75948.00
80203.00
47265.00
47265.00
Channels Entering Junction
1
2
3
4
5
6
7
8
8
0
0
1
2
3
4
5
6
7
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
- o
0
0
0
0
0
0
0
Channels
Channel
Number
1
2
3
4
5
6
7
8
Length
(M)
1384.
370.
161.
161.
307.
460.
383.
690.
Width
(M)
133.
106.
121.
167.
167.
187.
172.
137.
Manning
Coefficient
0.033
0.033
0.033
0.033
0.033
0.033
0.033
0.033
Bottom Elev*
-1.7
-1.6
-1.4
-1.2
-1.2
-1.2
-1.0
-0.4
Junctions at Ends
1 2
2 3
3 4
4 5
5 6
6 7
7 8
8 9
* Measured positive downward from data plane.
-359-
-------
Figure 4. The four watersheds discharge to the Delaware River at two
inlet points represented as Elements 2 and 11. The first watershed
(Element 1) represents the ash handling and ash pile drainage area. The
runoff from this area enters the Delaware River at Element 2. The second
watershed (Element 3) is the drainage area for the plant substation, the
third watershed (Element 5) is the drainage area for the coal pile,
and the fourth watershed (Element 8) is the drainage area adjacent to the
plant. Stormwater runoff from these three watersheds enters a storm
sewer system (described by Elements 4, 6, 7, 9, and 10) which in turn
discharges to the Delaware River through Element 11. The physical di-
mensions for each of these elements are listed in Table 5.
Based on the physical characteristics of the Delaware River and the
location of the stormwater discharges to the river, the Delaware River
was divided into seven nodes and five channels, as illustrated in Figure 5.
Background river flow and pollutant concentrations enter node 1. Storm-
water runoff discharges from the ash handling and ash pile area enter the
river at node 3, and the stormwater runoff discharge from the substation,
coal pile, and the plant area enter the river at node 4. Cooling water
for the power plant is withdrawn at node 3 and is discharged at node 4.
As with the Warren discretization scheme, an artificial dam, i.e., a dam
with no elevation above the river bottom, was placed between nodes 6 and 7.
Channels connect each of the no.des except where the nodes are interrupted
by the dam. The physical characteristics for the node and channel
discretization scheme are listed in Table 6.
As described previously in this paper, the amount of pollutant washed
-360-
-------
Cooling water
\\, discharge
(j Element number
Watershed boundary
\
Fig. 4
Discretization scheme schematic
of land area; Portland, PA.
-361-
-------
TABLE 5
DISCRETIZATION: LAND ELEMENTS
PORTLAND, PA
Element No.
1
2
3
4
5
6
7
8
9
10
11
Type1
1
2
1
2
1
2
2
1
2
2
2
Width
(Ft)
808.00
0.00
1220.00
1.75
698.00
1.75
3.00
1320.00
1.75
3.00
0.00
Slope
(Ft/Ft)
0.0500
0.0000
0.0080
0.0100
0.0020
0.0169
0.0096
0.0500
0.0100
0.0096
0.0000
AR OR GL2
13.36
0.00
17.48
817.00
11.90
462.00
223.00
10.23
442.00
452.00
0.00
PI OR DF2
50.00
0.00
33.33
1.66
100.00
1.66
2.85
100.00
1.66
2.85
0.00
*Type 1 is a watershed
Type 2 -is a pipe or manhole
2AR = Area of watershed (AC); PI
GL = Pipe length (Ft)
; J)F
Percent imperviousness (%)
.95 * width
-362-
-------
N
Stormwater
drainage system
01 I River mi. 66.75
River flow - 297.2 m3/sec (10,500 cfs)
River mi. 66.50 - Field Station 110
River mi. 66.25 Cooling water withdrawal - 6.18 m3/sec (218 cfs)
Ash handling area direct runoff
River mi. 66.10 - Field Station 120
Cooling water discharge - 6.18 m3/sec (218 cfs)
Coal pile and parking lot runoff
Artificial Dam • elev. 0.0
River mi. 65.50
I
®6 /River mi. 65.0
Legend
© Node (Junction)
1 Node (Junction) number
^n Channel
| ) Channel number
River mi. 64.5
Fig. 5
Discretization scheme schematic
of Delaware River; Portland, PA.
-363-
-------
TABLE 6
DISCRETIZATION: RIVER ELEMENTS
PORTLAND, PA
Nodes (Junctions)
Junction
Number
1
2
3
4
5
6
7
Surface Area
(SQ M)
93861.00
92022.00
74722.00
130490.00
172638.00
71779.00
71779.00
Channels Entering Junction
1
2
3
4
5
5
0
0
1
2
3
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Channels
Channel
Number
1
2
3
4
5
Length
(M)
402.
402.
241.
966.
805.
Width
(M)
233.
224.
245.
209.
178.
Manning
Coefficient
0.033
0.033
0.033
0.033
0.033
Bottom Elev*
-2.8
-2.7
-3.4
-3.5
-1.6
Junctions at Ends
1 2
2 3
3 4
4 5
5 6
*Measured positive downward from data plane.
-364-
-------
from the land surface during a storm is, in part, related to the initial
(pre-storm) mass of pollutant on the land surface. The initial pollutant
mass load is equal to the dust and dirt accumulation rate multipled by
the area of the watershed with that dust and dirt accumulation rate, the
number of dry days between storms, and the amount of a particular pollu-
tant in the dust and dirt. The dust and dirt accumulation rate and the
amount of a particular pollutant in the dust and dirt used as inputs to
the model run were determined from dust field sampling programs for both
the Warren and Portland Generating Stations. The area of the watershed
with that dust and dirt accumulation rate and the number of dry days
between storms were determined from discretization schemes and meteoro-
*
logical records.
The results of the dust and dirt sampling and analyses programs are
summarized in Table 7. At the Warren Generating Station, one dust and
dirt accumulation rate (.003 Ibs/dry day - ft2) was representative of all
the watersheds. At the Portland Station, two dust and dirt accumulation
rates were developed — one for the coal pile, plant, and substation
watersheds (.003 Ibs/dry day - ft2) and one for the ash handling water-
shed (.00055 Ibs/dry day - ft2). The amount of pollutant in the dust and
dirt also varied between the Warren and Portland Generating Stations.
»
Total suspended solids and sulfate levels were slightly higher at Portland.
rtiile total iron, manganese, aluminum, and total dissolved solid levels
?ere substantially lower.
Storm activity input data consisted of field measurements of rainfall
Intensity. These results are summarized in Table 8. The storm of
-365-
-------
TABLE 7
SUMMARY RESULTS OF DUST ASTO DIRT
SAMPLING PROGRAM AS INPUT TO MODEL
Location
Warren
Generating
Station
Portland
Activity
Entire
Plant
Area
Coal Pile
and Adjacent
Area
Ash
Handling
Area
Dust and Dirt
Accumulating Rate
(Lbs/Dry Day-Ft2)
.0031
.00032
.000552
mg Pollutant Per' Gram
of Dust and Dirt
TSS
993.5
997.91
998.39
Sulfates
1.0
1.11
1.18
Total Fe
18.7
.016
.004
Manganese
.3
.003
.0013
Aluminum
3.4
.015
.0066
IDS
6.5
2.09
1.61
dust and dirt load increased at this accumulation rate without
reaching an ultimate level over a 15-day period.
2The dust and dirt load increased at this accumulation rate for
approximately four days after which the total dust and dirt load
remained at a constant level.
-------
TABLE 8
SUMMARY OF STORM ACTIVITY
I
bo
•-J
Location
Warren
Generating
Station
Warren
Generating
Station
Portland
Generating
Station
Date
8-26-76
9-17-76
10-20-76
Storm
Activity
Shower
Steady
Drizzle,
Intermittent
Showers-
Heavy at
Short In-
tervals
Steady
Drizzle,
Intermittent
Showers-
Heavy at
Short In-
tervals
Rainfall
Intensity
(Time)
1435-
1455
0800
0930
1050
1130
1230
1510
0800
0810
0910
1010
1110
1140
1155
to
2130
(in/hr)
0.33
0.00
0.04
0.05
0.11
0.07
0.03
0.00
0.06
0.10
0.05
0.09
0.16
0.16
0.12*
Total
Rainfall
(in)
0.11
0.33
1.55
Comments
Last Prior Storm
Occurred on 8-15-76
Last Prior Storm
Occurred on 10-13-76
*Average Intensity Based on Measured Value for Cumulative
Rainfall in that Time Period.
-------
September 17, 1976, at the Warren Generating Station, was used for model
calibration. This storm lasted approximately seven hours and was a steady
drizzle with intermittent, short but heavy, showers. The maximum rainfall
intensity was 0.11 inches/hour, and the total rainfall was 0.33 inches.
The storm of August 26, 1976, was used for model verification at Warren.
This storm was a short (20-minute) shower with a maximum rainfall intensity
of 0.33 inches/hour, and a total rainfall of 0.11 inches. The storm used
for the model run at Portland occurred on October 20, 1976. The storm
lasted approximately 14 hours and was a steady drizzle with intermittent,
short but heavy, showers. The maximum rainfall intensity was 0.16 inches/
hour, and the total rainfall was 1.55 inches.
Other input items of fundamental concern to SSWMM-RECEIV II are back-
ground flow and pollutant loadings in the receiving waters. Background
conditions are those flow and pollutant loadings which are not influenced
by the stormwater runoff of the particular site being studied.
In the case of the Warren and Portland Generating Stations, background
conditions were measured in the river upstream of the plant stormwater
discharges. Background flows and pollutant loadings were computed as a
three-day average of the river field measurements at the upstream station
for one day prior to the storm, the day of the storm, and one day after
the storm. The background conditions were'computed in this manner since a
model restriction requires that the background inputs to the model remain
unchanged for the model simulation period. A three-day simulation was
desired to allow computational stability and to monitor the storm-induced
flows and pollutant concentrations for up to one day after the storm.
-368-
-------
4.2 Model Test Results
The model results were compared to field measurements to test the
validity of the model where comparable information was available and at
time intervals where maximum runoff flows and pollutant loadings occurred
in the model. A comparative factor of 4 was considered to be adequate for
model development purposes.
For the storm of September 17, the initial model run at the Warren
Generating Station, the modeled stormwater runoff flow was 0.4 of the
measured flow, and the modeled stormwater runoff pollutant concentrations
were within a factor of 4; (i.e., the model concentration divided by the
measured concentration varied between .25 and 4.0) for total suspended
solids, total iron, manganese, and aluminum, but were greater than a
factor of 4 for sulfates and total dissolved solids. Modeled and measured
pollutant concentrations in the Allegheny River compared within a factor
of 3. The results are shown in Table 9.
In the calibration model run at the Warren Generating Station for
the storm of September 17, the impervious area water retention storage
depth was reduced from .062 inches to .001 inches to increase the modeled
percentage of the total rainfall that was runoff. This change was made
since the area was almost completely impervious. The percentage of
runoff, therefore, should be approximately equal to 100%. In the initial
model run, it was only 86%. In the calibration model run, the percentage
runoff was 99%. The impervious area water retention storage was main-
tained at .001 inches for the remaining model runs.
In the calibrated model run at Warren for the storm of September 17,
-369-
-------
TABLE 9
COMPARISON OF MODEL RESULTS TO FIELD DATA
STORM OF SEPTEMBER 17 (INITIAL RUN)
WARREN GENERATING STATION
SSWMM
Flow (CFS)
Pollutant
TSS
SOi,
Tot.Fe
Mn
Al
TDS
Model1
Inlet 3
.26
#/initi
6.635
.0325
.607
.010
.110
.211
mg/1
6819
33.4
623.8
10.3
113.
216.8
Meas*
Sta.
34
.68
mg/1
402-8.
2100.
700.
10.4
79.0
3135.
Factor
Model/
Meas
.4
1.7
.02
.9
1.0
1.4
.07
Model2
Inlet 5
,15
#/min
3.078
.0174
.325
.005
.059
.113
mg/1
3163
31.0
579
8.9
105.1
5.3
Meas^
Sta. 31
32,33
No Data
mg/1
2058 -
5410
225-475.
153-300
3.5-6.0
23.5-71.
357-618
Factor
Model/
Meas
-
1.5-.6
.1-.06
3.8-1.9
2.5-1.5
4.5-1.5
.01-. 00!
Model3
Inlet 8
.12
#/min
2.434
.0213
.398
.006
.072
.138
mg/1
5420
5.7
886.
13.4
160.3
307.3
Meas3
Sta.
42
No Data
mg/'l
2695.
3000.
1025.
16.0
111.0
5456.
Factor
Model/
Meas
-
2.0
.002
.9
.8
1.4
.06
1 Time » Model = 1130; Meas = 1110 - 1120
2 Time « Model - 1115; Meas = 1100 - 1115
3 Time « Model « 1030; Meas = 1020
RECEIV
Quality
Pollutant
TSS
S04
Tot.Fe
Mn
Al
TDS
Model Cone. "
(mg/1)
13.536
16.174
.355
.033
1.030
105.228
Junction
6
6
6
6
6
6
Background
(mg/1)
11.610
16.170
.190
.030
l.OOO7
105.198
Measurement"
Sta. 20
(mg/1)
4.5
16.0
.74
.06
l.OOO7
99.00
Factor
Model/Meas
3.0
1.0
.5
.6
1.0
1.1
**Quantity - Storm has no effect on flow characteristics
5Time = 1200
6Time = 1150 - 1210
71.00 = Non-Detectable Limit
-370-
-------
the modeled stormwater runoff flow and pollutant concentrations and the
modeled river pollutant concentrations compared to the field measure-
ments with approximately the same degree of accuracy as did the initial
model run for the storm of September 17. The results are shown in
Table 10.
For the storm of August 26, the verification model run at the
Warren Generating Station, the modeled stormwater pollutant concentrations
also compared to the field measurements with approximately the same degree
of accuracy as did the calibration model run for the storm of September
17. Modeled and measured pollutant concentrations in the Allegheny River
compared within a factor of 4.. The results are shown in Table 11.
For the storm of October 20, the model run at the Portland Generating
Station, modeled stormwater pollutant concentrations were different from
field measurements by greater than a factor of 4. Modeled and measured
pollutant concentrations in the Delaware River compared within a factor
of 5. Results are shown in Table 12.
The model was not calibrated at Portland because stormwater flow
field measurements were not available due to measurement difficulties,
and it is first necessary to calibrate flow in the model before any
other model adjustments are* warranted.
-371-
-------
TABLE 10
COMPARISON OF MODEL RESULTS TO FIELD DATA
STORM SEPTEMBER 17 (CALIBRATION RUN)
WARREN GENERATING STATION
SSWMM
Flow (CFS)
Pollutant
TSS
S04
Tot.Fe
Mn
Al
IDS
Model1
Inlet _3_.
.26
#/min
6.501
.0263
.493
.008
.090
.171
mg/1
6681
27.0
506. <
8.2
92.5
175.;
Meas1
Sta.
_M
.68
mg/1
4028.
2100
700
10.4
79.0
3135
Factor
Model/
Meas
.4
1.6
.01
.7
.8
1.2
.06
Model2
Inlet _5_
.15
iP/min
3.01
.0141
.263
.004
.048
.091
mg/1
5362.
25.1
468.5
7.1
85.5
162.1
Meas2
12,33=
No Data
mg/1
2058 -
22^75.
153-300
3.5-6.0
23.5-71.
357-618
Factor
Model/
Meas
-
2.6-1.0
.1-.05
3.1-1.6
2.0-1.2
3.6-1.2
.4-. 3
Model3
Inlet JL
.13
0/min
2.754
.0191
.358
.006
.065
.124
mg/1
5660
39.2
736.
12.3
133.6
255.
Meas3
Sta.
.4_2_
No Data
ELg/1
2695.
3000.
1025.
16.0
111.0
5456.
Factor
Model/
Meas
-
2.1
.01
.7
.8
.8
.05
1 Time = Model = 1130; Meas = 1110 - 1120
2 Tine *> Model = 1115; Meas = 1100 - 1115
3 Time - Model - 1030; Meas = 1020
Quality
Pollutant
TSS
S04
Tot.Fe
Mn
Al
IDS
Model Cone.5
(mg/1)
13.496
16.173
.324
.032
1.024
105.217
Junction
6
6
6
6
6
6
Background
(mg/1)
11.610
16.170
.190
.030
1.000 7
105.198
Measurement
Sta. 20
(mg/1)
4.5
16.0
.74
.06
l.OOO7
99.00
Factor
Model/Meas
3.0
1.0
.4
.5
1.0
1.1
**Quantity - Storm has no effect on flow characteristics
5Tirae = 1200
6Time = 1150 - 1210
71.00 = Non-Detectable Limit
-372-
-------
TABLE 11
COMPARISON OF MODEL RESULTS TO FIELD DATA
STORM AUGUST 26 (VERIFICATION)
WARREN GENERATING STATION
SSWMM
Flow (CFS)
Pollutant
TSS
SOi,
Tot.Fe
Mn
Al
IDS
Model1
Inlet _3_.
0/min
Dg/1
Meas1
Sta.
34_
io Data
fflg/1
So Data
Factor
Model/
Meas
Model2
Inlet =5_
.21
tf/min
1.939
.0143
.267
.004
.049
.093
mg/1
2467.
18.2
340.
5.1
62.3
118.
Meas2
Sta. 31
32.33
No Data
mg/1
1027 -
2965.
1750 -
108-700
9.6-45.
42.5-334
2405 -
8107.
Factor
Model/
Meas
-
2. 4-. 82
01-. oo;
3.1-.4J
.53-.!
1.5-.2
.05-.0]
Model3
Inlet &_
#/nin
Dg/1
Meas3
Sta.
42
So Data
ng/1
No Data
Factor
Model/
Meas
1 Time - No Data During Storm
2 Time = 1445
* Time = No Data During Storm.
Quality
RECEIV
Pollutant
TSS
SO^
Tot.Fe
Mn
Al
IDS
Model Cone.5
(og/D
7.839
12.481
.249
.030
1.003
114.502
Junction
6
6
6
6
6
6
Background
(mg/D
7.73
12.48
.23
.030
l.OOO7
114 ..5
Measurement .
Sta. 20
(mg/D
No Data
No Data
.06
.013
l.OOO7
No Data
Factor
Model/Meas
..
-
4.2
2.3
1.0
—
''Quantity - Storm has no effect on flow characteristics
5Time = 1600
6Time = 1550 - 1650
71.00 = Non-Detectable Limit
-373-
-------
TABLE 12
COMPARISON OF MODEL RESULTS TO FIELD DATA
STORM OCTOBER 20 (INITIAL RUN)
PORTLAND GENERATING STATION
SSWMM
Flow (CFS)
Pollutant
TSS
SO,,
Tot.Fe
Mn
Al
TDS
Model1
Inlet 2
.83
#/min
.543
.0037
•e.OOl
<.001
<.001
.005
mg/1
175.
1.2
<.32
<.32
<.32
1.6
Measure1
Station
131, 132
No Data
mg/1
2344-10932
1200-1600.
73-245
.03-2.5
125-138
1843-2504
Factor
Model/
Measure
-
.07-. 02
.001-. 0008
.004-. 001
10. -.13
.002-. 002
.0008-. 0006
Model2
Inlet 11
2.34
tf/min
1.015
.0058
<.001
<.001
<.001
.011
rag/1
116.
.7
<.l
<.l
<.l
1.2
Meas2
Sta.
134B
So Data
rag/1
280.
400.
385.
1.75
50.
698.
Factor
Model/
Me as
-
.4
.002
.0002
.06
.002
.002
1 Time
2 Time
1115
1045
Quality
Pollutant
TSS
SOi,
Tot.Fe
Mn
Al
TDS
Model Cone.1*
(mg/1)
16.070
14.611
.570
.050
1.000
67.455
Junction
4
4
4
4
4
4
Background
(mg/D
15.999
14.618
.570
.050
l.OOO6
67.486
Measurement^
Sta. 120
(mg/1)
7.1
9.0
.45
.01
l.OOO6
45.7
Factor
Model/Meas
2.3
1.6
1.3
5.0
1.0
1.5
' Quantity - Storm has little effect on flow characteristics in river
(Less than .05% increase in river flow at max. storm discharge).
* Time = 1200
5 Time - 1130 - 1200 and 1230 - 1300
6 1.000 = Non-Detectable Limit.
-374-
-------
5.0 EVALUATION OF MODEL DEVELOPMENT PROGRAM
The work completed in this study indicates that SSWMM-RECEIV II is
capable of predicting the quantity and quality of stormwater runoff and
its impact on receiving waters for specific industries, but model limita-
tions do exist.
SSWMM-RECEIV II is a versatile stormwater and receiving water model
suited for industrial application. It is inherently flexible so that
it is appoicable to many industries with only minor data input modifica-
tions.
The specific utility industry application described in this study
has demonstrated that, where adequate field data were available, SSWMM-
RECEIV II results compared favorably to field measurements. At the Warren
Generating Station, calibrated model results for stormwater runoff flow
and pollutant concentrations compared within a factor of 4 and river
pollutant concentrations compared within a factor of 3 to field measure-
ments. The model-field measurement comparative factor of 4 was maintained
for a second storm at Warren, indicating that the calibrated model could
predict the effects of different storm conditions with the same degree of
accuracy established in model calibration. In essence, the model was
verified, increasing model credibility.
Some difficulties were encountered in this model study. Modeled
stormwater runoff concentrations of total dissolved solids and sulfates
at the Warren Generating Station were different from the field-measured
values by greater than a factor of 4. In addition, due to measurement
difficulties, adequate field data were not available to ascertain the
-375-
-------
comparative validity of the model at the Portland Generating Station for
either stormwater runoff or the receiving water.
Inherent model limitations include the lack of capability to simu-
late storm erosion of infinite sources; i.e., material storage piles, and
to simulate stormwater percolation through material storage piles.
Although difficulties were encountered, and additional work is needed
to increase model credibility and usefulness, SSWMM-RECEIV II has been
demonstrated to be a valid stormwater runoff and receiving water model
suited to industrial application.
-376-
-------
REFERENCES
1. Huber, Wayne, Heaney, James, Medina, Maguil, Peltz, W., Sheikh, Hasan,
and Smith, George, University of Florida, "Storm Water Management
Model, User's Manual, Version II", Environmental Protection Agency,
Office of Research and Development, National Environmental Research
Center, Cincinnati, Ohio, EPA-670/2-75-017, March, 1975-
2. Discussion with Larry A. Roesner, Water Resources Engineers, Incor-
porated, Walnut Creek, California, April 5, 1976.
3. Hagarman, James A., "Short Stormwater Management Model Documentation
Report", (unpublished), University City Science Center, Philadelphia,
Pennsylvania, June, 1976.
4. Raytheon Company, "New England River Basins Modeling Project Final
Report, Vol. Ill - Documentation Report, Part I - RECEIV II Water
Quantity and Quality Model", Environmental Protection Agency, Office
of Water Programs, Washington, D.C., Contract No. 68-01-1890,
December, 1974.
5. Hydrocomp, Inc., "Hydrocomp Simulation Programming Operations Manual",
4th Ed., Palo Alto, California, January, 1976.
6. Geiger, W. F., "A New Method To Evaluate Urban Runoff Pollution and
its Effect on Receiving Waters", Dorsch Consult, Munich, Germany
(Toronto, Canada).
7. Hydrologic Engineering Center, U.S. Army Corps of Engineers, "Urban
Stormwater Runoff: STORM, Generalized Computer Program 723-S8-L2520",
(Draft), Davis, California, August, 1975.
8. Brandstetter, Albin, Engel, Roger, Arnett, Ronald, Wise, Stacey, and
Cearlock, Dennis, "Appendix A, User's Manual, Development of Hydraulic,
Water Quality, and Optimization Models for Wastewater Management",
Battelle Pacific Northwest Laboratories, Richland, Washington, March,
1976.
9. Lager, John A., Didriksson, Theodor, Otte, George B., Metcalf &
Eddy, Inc., "Development and Application of a Simplified Stormwater
Management Model", Environmental Protection Agency, Office of Research
and Development, Municipal Environmental Research Laboratory,
Cincinnati, Ohio, EPA-600/2-76-218, August, 1976.
10. Donigian, Anthony and Crawford, Norman, Hydrocomp, Inc., "Modeling
Pesticides and Nutrient's on Agricultural Lands", Environmental Pro-
tection Agency, Office of Research and Development, Environmental
Research Laboratory, Athens, Georgia, EPA-600/2-76-043, February, 1976.
-377-
-------
11. Morth, Arthur H., Smith, Edwin E., Shumate, Kenesaw S., Ohio State
University, "Pyritic Systems: A Mathematical Model", Environmental
Protection Agency, Office of Research and Monitoring, Washington,
D.C., EPA-R2-72-002, November, 1972.
12. Brookman, Gordon, Wade, Willard, and Binder, James, TRC - The
Research Corporation of New England, "Evaluation of Non-Point
Sources from Industry, Volume I, Technical Report, Sampling and
Modeling at a Coal-Fired Utility" (Draft Report), Environmental
Protection Agency, Industrial Environmental Research Laboratory,
Research Triangle Park, North Carolina, EPA 68-02-2133, February,
1977.
-378-
-------
Wednesday Morning - May 251 1977
SESSION V: INDUSTRY
Chairman: John E. Yocom, TRC/APCA
Page
CURRENT STATUS OF PROCESS FUGITIVE PARTICULATE EMISSION ESTIMATING 381
TECHNIQUES
John M. Zoller* and Thomas A. Janszen, PEDCo Environmental, Inc.,
and Gilbert H. Wood, EPA
AIR EMISSIONS IN IRON ORE MINING AND ENRICHMENT 457
G. V. Jorgenson*, J. P. Pilney, and E. E. Erickson,
Midwest Research Institute
DESIGN OF A STUDY TO MEASURE FUGITIVE EMISSIONS FROM PETROLEUM 479
REFINING
Donald D. Rosebrook, Radian Corporation
SOME AIR QUALITY AND ENERGY CONSERVATION CONSIDERATIONS FOR THE USE 509
OF EMULSIONS TO REPLACE ASPHALT CUTBACKS IN CERTAIN PAVING OPERATIONS
Francis M. Kirwan*, EPA, and Clarence Maday, Consultant
* Indicates speaker
-379-
-------
CURRENT STATUS OF PROCESS FUGITIVE
PARTICULATE EMISSION ESTIMATING TECHNIQUES
By
John M. Zoller, Thomas A. Janszen
PEDCo Environmental, Inc.
and
Gilbert H. Wood
U.S. Environmental Protection Agency
-381-
-------
CURRENT STATUS OF PROCESS FUGITIVE
PARTICULATE EMISSION ESTIMATING TECHNIQUES
John M. Zoller
PEDCo Environmental, Inc.
Gilbert H. Wood
U.S. Environmental Protection Agency
Thomas A. Janszen
PEDCo Environmental, Inc.
ABSTRACT
This presentation synopsizes the availability and
status of information useful for estimating the sources and
quantities of particulate matter classified as fugitive
process particulate emissions. As part of an EPA-sponsored
technical information effort, 24 major industrial classi-
fications were investigated. Fugitive emissions data
acquired through literature search, evaluation of pertinent
on-going EPA studies, interview of trade associations, and
on-site plant visitations has been tabularized. They have
been ranked according to a set of subjective criteria in an
effort to assign reliability indices. These indices are
presented by the authors in order to qualify both the cur-
rent status and possible utilization of these data for
"emission factor" purposes.
-382-
-------
INTRODUCTION
PEDCo Environmental, Inc. has recently completed a
project report for the U.S. Environmental Protecton Agency*
entitled "Technical Guidance for Control of Industrial
Process Fugitive Particulate Emissions." The document was
prepared as guidance for air pollution control agencies in
developing revisions, where necessary, to particulate matter
control strategies. A data base was assembled from litera-
ture sources, on-going U.S. Environmental Protection Agency
in-house or contractor project activities, air pollution
control agency records and files, interview of trade asso-
ciations, and visits to industrial facilities emitting
and/or controlling fugitive particulate emissions.
"Industrial Process Fugitive Particulate Emissions,"
often referred to hereafter as "IPFPE," result from either
one or both of the following categoricaj. groupings of pi -:~
ticulate emission sources - "fugitive emissions" and "fugi-
tive dust" originating within industrial facilities.
Fugitive dust emissions are generally related to natural or
man-associated dusts (particulate only) that become airborne
due to the forces of wind, man's activity, or both. Fugi-
tive emissions, on the other hand, include those particu-
lates that are emitted from industry-related operations and
Prepared under Contract No. 68-02-1375, Task Order No. 33.
Publication No. EPA-450/3-77-010. Gilbert H. Wood, EPA
Project Officer.
-383-
-------
which escape to the atmosphere through windows, doors,
vents, etc.; not through a primary exhaust system, such as a
stack, flue or control system. Therefore, for the purpose
of this investigation. Industrial Process Fugitive Particu-
late Emissions (IPFPE) were defined as ... "Particulate
matter which escapes from a defined process flow stream due
to leakage, materials charging/handling, inadequate opera-
tional control, lack of reasonably available control tech-
nology, transfer, or storage.
The document "Technical Guidance for Control of Indus-
trial Process Fugitive Particulate Emissions" contained
information on control technology for sources of IPFPE,
estimating the air quality impact of IPFPE, and integration
of IPFPE impacts into the State Implementation Planning
process. Another part of that document discussed sources of
industrial process fugitive particulate emissions, and the
topic of this paper is the development of that information.
SOURCES OF INDUSTRIAL PROCESS FUGITIVE PARTICULATE EMISSIONS
For each of 24 major industrial categories with poten-
tial IPFPE sources, the following information was given in
the guidance document:
0 Process description
0 Identification of IPFPE sources
0 Emission estimates and an example plant inventory
of IPFPE sources
-384-
-------
0 IPFPE emission characteristics
0 Control technology options for IPFPE sources
A difficult part of this task was to develop for all 24
selected industrial categories, a range of emission esti-
mates for each potential IPFPE source. This is a job that
would require years of testing and the expenditure of large
sums of money for definitive estimates. Since the deter-
mination of emission rates was but one part of the guidance
document, testing was not in the scope of work. Therefore
an extensive literature search was made to obtain available
fugitive emission estimates. Obviously, estimates could not
be found for all fugitive sources. For the sources where an
emission rate could not be found in the literature, best
engineering judgement was used to estimate the emissions.
This paper discusses development of the emission estimates
and an example plant inventory.
A range of emission estimates was presented where
adequate information was available. It was not the intent
of the study to determine a single factor (average or typi-
cal) for each IPFPE source, but rather to present a range of
values. However, when only one estimate was found in the
literature, that number was shown. For sources where emis-
sion factors were not available in the literature, estimates
were based on engineering judgment using emission factors
-385-
-------
for similar sources, observations during plant visits,
calculations from source reported or test data, or a percent
of the stack emissions.
A reliability rating was indicated for each fugitive
emission estimate. This was an indicator of the supportive
data used to develop the factor. The reliability ratings
used correspond to the rating system used in "Compilation of
Air Pollutant Emission Factors," EPA Publication Number
AP-42. The ratings in AP-42 are predominately applied to
stack (nonfugitive) emissions, and are defined as follows:
A - Excellent
B - Above average
C - Average
D - Below average
E - Poor
All fugitive emission rates in the guidance document were
determined to have ratings of C, D, or E. The criteria used
to determine these ratings were:
C - Supportable by multiple test data.
D - Supportable by limited test data and engineering
judgment.
E - Supportable by best engineering judgment (visual
observation, emission tests for similar sources,
etc.).
-386-
-------
Fugitive emission factors supportable by multiple test
data were given only a "C" (average) rating because of the
difficulty in accurately measuring fugitive emissions as
compared to measuring stack emissions. Also the plant to
plant variations in emission rates were expected to be
greater for fugitive sources than for nonfugitive sources.
Factors based on engineering judgment (e.g., visual observa-
tion or emission tests for similar sources) were given an
"E" (poor) rating. Factors with an "E" rating were at best
order of magnitude estimates; therefore, actual emission
rates at a given facility could differ significantly.
Fugitive emission factors based on limited test data and
engineering judgment were given a "D" (below average) rating.
A model plant IPFPE inventory was presented for each
industry, but was not intended to represent a "typical"
plant, only an example application of the uncontrolled
fugitive emission estimates. An average for any range of
factors presented was used in developing the inventory. For
a specific plant/ consideration of operating conditions must
be made when selecting the emission rate. Credit must be
given for controls or good operating practices that reduce
emissions. Also, poor maintenance or operating procedures
would likely result in an emission rate that is greater than
average.
-387-
-------
The Portland cement industry was selected to illustrate
the development of the fugitive emission factors and their
reliability ratings. Through this example, development of
the fugitive emission factors and reliability ratings for
the other industries (as shown in Appendix A) will be
better understood.
DEVELOPMENT OF FUGITIVE PARTICULATE EMISSION FACTORS FOR THE
PORTLAND CEMENT INDUSTRY
Figure 1 presents a process flow diagram for the
Portland cement industry showing 22 potential fugitive
particulate emission points. These emission points were
identified as potential fugitive sources as a result of an
extensive literature search and visits to several Portland
cement plants. Each source is identified and explained in
Table 1. Note that this table is reproduced as presented in
the guidance document. In footnotes to the table, section
numbers refer to sections in the guidance document and
references cited are listed at the end of this paper.
Emissions Calculated from Literature Sources
Some factors were based on data presented in the litera-
ture. For Source 1, raw material unloading, the factor for
coal was obtained from Reference 5. Since emission testing
was not done to develop this factor, it was assigned a
reliability rating of "E." The fugitive emission factor
range for unloading other raw materials (clay, sand, and
-388-
-------
TRUCK BARGE
I
lo
oo
RAW MATERIAL
UNLOADING
COAL, LIMESTONE, CLAT
GYPSUM, SAND, IRON ORE
LIMESTONE, CLAY
SAND, IRON ORE
— 1 mi. ,
DRY PROCESS OVERSIZED
FUEL FOR HEATING KILN
TO
TRUCK,-
BOX CAR
trd
TRUCK BARGE
Figure I. Process flow diagram for portland cement manufacturing showing
potential industrial process fugitive particulate emission points.
-------
Table 1. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PORTLAND CEMENT MANUFACTURING
i
OO
o
I
Source of IPFPE
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Raw material unloading (rail,
barge, truck) gypsum, iron
ore, clay, limestone, sand,
coal
Raw material charging to
primary crusher
Primary crusher
Transfer points and associated
conveying
Vibrating screen
Secondary crusher
Unloading outfall to storage
Raw material storage
Transfer to conveyor via
Clamshell
Raw grinding mill and feed/
discharge exhaust systems
Uncontrolled fugitive emission factor
Coal: 0.2 kg/Mg material*
(0.4 Ib/ton)
Other: 0.015-0.2 kg/Mg material3
(0.03-0.4 Ib/ton)
0.00015-0.02 kg/Mg
of rock charged^
(0.0003-0.04 Ibs/ton)
0.25 kg/Mg
of rock crushed
(0.5 Ibs/ton)
0.1-0.2 kg/Hq material handled*1
(0.2-0.4 Ib/ton)
0.74 kg/Mg screened6 '£
(1.5 Ibs/ton)
f
1.5-2.5 kg/Mg
raw material unloaded''
(3.0-5.0 Ib/ton)
h
h
0.05 kg/Mg
raw material milled5
(0.1 Ibs/ton)
Emission
factor
reliability
rating
E
E
D
C
E
C
-
E
-
_
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Coal unloaded"
79,175
(87,197)
Materials unloaded
161,240
(177,576)
Raw material0
591,421
(651,345)
Raw material crushed
591,415
(651,338)
591,415
(651,338)
591,267
(651,175)
-
590,824
(650,687)
-
Raw material
589,642
(649,387)
Uncontrolled
emissions
Mg/yr
(tons/yr)
16
(17)
17
(19)
6
(7)
148
(163)
89
(98)
443
(488)
f
1,182
(1,300)
h
h
29
(32)
-------
Table 1 (continued!. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PORTLAND CEMENT MANUFACTURING
Source of IPFPE
11. Raw blending
12. Blended material storage
13. Coal storage
Loading onto pile
Vehicular traffic
Loading out
Wind erosion
Uncontrolled fugitive emission factor
0.02 kg/Mg
raw material b.
(0.05 Ibs/ton)
(0.02) (KD (S/1.5)
(PE/100)*
/(O. 04) (Kl) (S/1.5)
\ (PE/100) 2
(0.065HK2) (S/1.5)
(PE/100) 2
/"(0.13) (K2) (S/1.5)
\ (PE/100) 2
(0.025)(K3)(S/1.S)
(PE/100) 2
f(0. 05) (K3) (S/1.5)
\^ (PE/100) ^
(0.055) (S/l. 5) D
(PE/100) * 90
Ao. 11) (S/l. 5) D
\ (PE/100) 2 90
tended9
-------
Table 1 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOB PORTLAND CEMENT MANUFACTURING
V£>
NO
Source of IPFPE
14.
15.
16.
17.
18.
19.
20.
21.
22.
Transfer of coal to grinding
mill
Leakage from coal grinding
mills
Unloading-clinker/gypsum
outfall to storage
Clinker/gypsum storage
Clinker/gypsum load-out
Finish grinding with leaks
from mill and from feed/
discharge exhaust systems
Cement silo vents
Cement loading
Cement packaging
Uncontrolled fugitive emission factor
Neg-0.1 kg/Mq transferred9
(0.2 Ibs/ton)
Negligible5
2.5-5.0 kg/Hg
of clinker and gypsum9' '
(5.0-10.0 Ibs/ton)
1
1
0.05 kg/Mg of cement9
(0.1 Ibs/ton)
Negligible9
0.118 kg/Mg
of cement loaded
(0.236 Ibs/ton)
Neg- 0.005 kg/Mg packed9
(0.01 Ibs/ton)
Emission
factor
reliability
rating
D
E
E
-
-
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/year
(tons/year)
coal transferred
79,173
(87,195)
_
clinker/gypsum
405,022
(446,059)
-
-
grinding feed^
403,627
(444,525)
-
cement loaded
375,443
(413,485)
cement packaged
28,164
(30,980)
Uncontrolled
emissions
(tons/yr)
8
(9)
__
1,519
(1,671)
1
1
20
(22)
-
44
(49)
negligible
-------
Table 1 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PORTLAND CEMENT MANUFACTURING
i
w
VO
U>
I
Source of IPFPE
Uncontrolled fugitive emission factor
Emission
factor
reliability
rating
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Uncontrolled
emissions
Mg/yr
(tons/yr)
Estimate based on data presented in Section 2.1.2. Range shown is for taconite and coal railcar unloading.
Reference 1.
0 Estimated based on crushed stone emission factors (Reference 2) . Note that approximately 80 percent of these emissions will
fall out on plant property.
Estimate based on data presented in Section 2.1.1.
Estimated based on crushed stone emission factors (Reference 2)
fall out on plant property.
Emissions for point 6 are included in emissions from point 5.
Engineering judgement based on visual observations during plant visits.
Emissions from point 9 are included in emissions from point 7.
For complete development of this emission factor, refer to Section 2.1.4
0.75, K2 = 0.5, and K3
Note that approximately 60 percent of these emissions will
For source 13 it was assumed that S
0.75.
4.0,
D = 90, PE - 100, and
Reference 3.
Emissions from points 12 are included in emissions from point 11.
Based on partially enclosed structure: open on both ends with roof.
Emissions from points 17 and 18 are included in emissions from point 16.
Reference 4. Based on tests of mechanical unloading of cement to hopper and subsequent transport of cement by enclosed
bucket elevator to elevated bins with a fabric sock over the bin vent.
88,930 Mg (97,940 tons) - clay
57,879 Mg (63,743 tons) - sand
14,431 Mg (15,893 tons) - iron ore
79,175 Mg (87,197 tons) - coal
Raw material charged, crushed, screened, and crushed (2nd)
- 88,930 Mg (96,751 tons) - clay
- 57,879 Mg (62,965 tons) - sand
- 14,431 Mg (15,698 tons) - iron ore
- 490,105 Mg (539,763 tons) - limestone
Includes 124 Mg (137 tons) of hydrophobe and grinding aid.
-------
iron ore) was assumed equal to the range of emissions for
taconite railcar unloading. The taconite factors were
derived from data presented in Reference 6. Since the
source of data in this reference was unknown an "E" rating
was assigned.
The factors for Source 2, raw material charging (truck
dump) to primary crusher, were obtained from Reference 1 and
were based on limited test data and engineering judgment.
As reported in that reference:
"Midwest Research Institute, in a sampling study of
aggregate handling operations, estimated that dumping
of crushed rock or gravel onto storage piles accounted
for about 12 percent of the total emissions of 0.33
Ib/ton from handling, or 0.04 Ib/ton. The truck dump-
ing operation was not sampled in isolation from the
other handling operations and the estimate of 12 per-
cent was partially subjective. This emission factor
for dumping of aggregate onto storage piles was re-
cently published in Supplement 5 of EPA's Compilation
of Air Pollutant Emission Factors.7
Monsanto Research determined an emission rate of
0-00034 Ib/ton for truck unloading at the hopper of a
primary crusher.^ The material being handled was
quarried granite with very little fine material present.
Since both engineering judgment and test data were used,
this range was assigned a "D" rating.
The factors for Source 4, transfer points and asso-
ciated conveying, were also obtained from Reference 1.
These factors represent total transfer and conveying opera-
tions at coal mines. Because the transfer and conveying
emissions for the Portland cement plant were assumed the
same as the coal plant factors, an "E" rating was assigned.
-394-
-------
Emission factors for the primary crusher, Source 3; the
vibrating screen, Source 5; and the secondary crusher;
Source 6, were taken from the EPA publication "Compilation
2
of Air Pollutant Emission Factors," AP-42. These factors
represent uncontrolled fugitive emissions for stone crushing
facilities and received a rating of "C" in AP-42. When
using these factors to estimate impact on air quality, note
that approximately 80 percent of the primary crushing emis-
sions and 60 percent of the screening and secondary crushing
emissions will settle out on plant property.
The coal storage pile factors, Source 13, were obtained
from Reference 3. These factors were a refinement of the
2
aggregate storage pile factor presented in AP-42. Note
that these formulas were preliminary and therefore subject
to further refinement and change when test results become
available. The storage pile emission factors were presented
for the unit operations of loading onto the pile, vehicular
traffic around the pile, loading out, and wind erosion.
Correction parameters were also incorporated into these
equations. Activity factors (K 0 ) were relative to the
i, /, j
operations performed by a front-end loader. Silt content
(S) and duration of material in storage (D) were also cor-
rection parameters. Because these factors were based on
both limited test data data and engineering judgment, a "D"
rating was assigned.
-395-
-------
Emissions for cement loading, Source 21, were obtained
from Reference 4. The emissions were based on tests of
mechanical unloading of cement to a hopper and subsequent
transport of cement by enclosed bucket elevator to elevated
bins with a fabric sock over the bin vent. This emission
rate was estimated to approximate truck, rail, or barge
loadout, and was given an "E" rating.
Emissions Calculated from Observations and Source Reported
Data
Emission rates for the remaining sources were deter-
mined by engineering judgment using observations made during
plant visits and a detailed emissions inventory. The emis-
sions inventory forms were received from a representative
Portland cement producing facility. Data were presented
showing controlled emissions from many potential fugitive
particulate emission sources. The emission rates were based
on stack tests of the control devices. Emission factors
were then calculated using the reported emissions, control
device efficiencies, and plant production. Consequently,
the emission factors derived from these forms were rated "D"
or "E", depending on the amount of alteration and engineer-
ing judgment that was necessary.
The first sources for which emissions were calculated
in this manner were clinker/gypsum outfall to storage,
storage, and load-out from storage, Sources 16, 17, and 18.
-396-
-------
At the plant for which emissions data were obtained, these
operations occurred in a totally enclosed building with a
vent to a baghouse. Knowing the annual emissions from the
baghouse, the amount of clinker handled per year, and the
efficiency of the baghouse, the uncontrolled emissions were
calculated to be 10 Ib/ton. From observations during plant
visits, it was judged that emissions from uncontrolled
clinker storage could be quite significant, but that an
average emission rate would be less than this value. There-
fore, a range of 5 to 10 Ib/ton was selected as the uncon-
trolled fugitive emission rate. This rate also assumes a
partially enclosed structure - open on both ends with a
roof. These emissions were assigned an "E" rating because
of the engineering judgment involved.
Emissions were next calculated for raw material outfall
to storage, storage, and transfer from storage to conveyor
(via clamshell), Sources 7, 8, and 9. Engineering judgment
alone was used for these sources since emissions data were
not available. It was observed during plant visits that the
operations involving raw materials were not as dusty as
those where clinker was handled. Therefore, it was esti-
mated that emissions from raw materials storage were 25
percent of the emissions from clinker storage. This results
in an emission factor of 1.25 to 2.5 Ib/ton. Since these
-397-
-------
factors were based on engineering judgment, they were given
an "E" rating.
Fugitive emissions from cement finish grinding, Source
19, occur through leaks from the feed and discharge exhaust
systems. Controlled emissions from the grinding mills were
calculated from the emissions inventory data to equal approx-
imately 0.1 Ib/ton. It was estimated that the same amount
also escapes from the hood pick-up systems at the feed and
discharge points and from leaks in the mill. Being an
engineering judgment, this factor was given an "E" rating.
Based on the emissions calculated for the finish mill
and visual observations during plant visits, fugitive emis-
sions from the raw grinding mill feed and discharge exhaust
pick-up systems, Source 10, were judged to be also 0.1
Ib/ton. This engineering estimate was given an "E" rating.
Raw material blending and storage, Sources 11 and 12,
were estimated to have one-half the amount of fugitive emis-
sions as that emitted by the raw grinding mill. This factor
was based on plant observations and was assigned a reli-
ability rating of "E".
Emissions due to transfer of coal from the storage pile
to the grinding mill, Source 14, were derived from observa-
tions at plant visits and data contained in Reference 1. An
emission rate for coal conveying and transfer at a coal mine
-398-
-------
is given in Reference 1 as 0.2 Ib/ton. However, during
plant visits it was observed that the emissions could be
negligible from this operation. Therefore, this range was
used for the factors. A reliability rating of "D" was
selected since both engineering judgment and an emission
factor specifically for coal conveying were used.
Leakage of coal from the grinding mills, Source 15, was
observed during plant visits to be negligible. This engi-
neering judgment was given an "E" rating.
Cement storage silos, Source 20, were found to be
almost exclusively controlled by fabric filters. Emissions,
therefore, were considered negligible from this source and
an "E" rating was assigned.
Based on the emission inventory data, fabric filter
controls for cement packaging, Source 22, were calculated to
emit 0.01 Ib/ton. It was estimated that emissions escaping
capture by the control system could be equal to or less than
this value. Therefore, the range of fugitive emissions was
from negligible to 0.01 Ib/ton, and was given an "E" rating.
Model Plant Inventory
The example plant inventory for Portland cement as
shown in Table 1 presents potential fugitive emission
quantities from the various uncontrolled sources within the
process. The inventory represents a plant which produces
-399-
-------
443,968 tons of Portland cement per year. The plant inven-
tory is not meant to display a typical plant, but merely an
example application of the uncontrolled fugitive emission
factors.
-400-
-------
REFERENCES
1. Evaluation of Fugitive Dust from Mining, Task 1 Report.
PEDCo Environmental, Inc., Cincinnati, Ohio. Prepared
for Industrial , Environmental Research Laboratory/REHD,
U.S. Environmental Protection Agency, Cincinnati, Ohio.
Contract No. 68-02-1321, Task No. 36, June 1976.
2. Compilation of Air Pollutant Emission Factors, AP-42.
U.S. Environmental Protection Agency, Office of Air and
Waste Management, Office of Air Quality Planning and
Standards. Research Triangle Park, North Carolina.
3. Open Dust Sources Around Iron and Steel Plants, Draft.
Midwest Research Institute. Prepared for U.S. Environ-
mental Protection Agency, Industrial Environmental Re-
search Laboratory. Contract No. 68-02-2120. Research
Triangle Park, North Carolina. November 2, 1976.
4. Personal communication to John M. Zoller, PEDCo Environ-
mental, Inc., from T.R. Blackwood, Monsanto Research
Corporation, Dayton Laboratory, 1515 Nicholas Road,
Dayton, Ohio. October 18, 1976.
5. Environmental Assessment of Coal Transportation. PEDCo
Environmental, Inc. Prepared for U.S. Environmental
Protection Agency. Contract No. 68-02-1321, Task No.
40, October 15, 1976. pp. 4-38 and 4-51.
6. Cross, F.L., Jr., and Forehand, G.D. Air Pollution
Emissions from Bulk Loading Facilities, Volume 6, En-
vironmental Nomograph Series. Technomic Publishing
Co., Inc., Westport, Connecticut, 1975. pp. 3-4.
7. Supplement No. 5 for Compilation of Air Pollutant Emis-
sion Factors. Second Edition. U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina.
April 1975.
8. Chalekode, P.K., and J.A. Peters, Assessment of Open
Sources. Monsanto Research Corporation, Dayton, Ohio.
(Presented at Third National Conference on Energy and
the Environment. College Corner, Ohio. October 1,
1975.) 9 p.
-401-
-------
APPENDIX A
IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS
This appendix contains the remaining fugitive emission
factor tables prepared for the guidance document. Note that
these tables are presented as prepared for the guidance
document. Table, section, and reference numbers refer to
that document. These tables are presented to allow rapid
dissemination of the information developed in that project.
Tables are presented for the following industries or
processes: coke manufacturing; iron production; steel pro-
duction; primary aluminum production; primary copper smelt-
ers; primary lead smelters; primary zinc production; second-
ary copper, brass/bronze production; foundries; minerals
extraction and beneficiation; grain elevators; lime produc-
tion; concrete batching; asphaltic concrete production; and
lumber and furniture industry.
-402-
-------
Table 2-9. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR COKE MANUFACTURING
o
OJ
Source of IPFPE
1 . Coal unloading
2. Coal storage
Loading onto pile
Vehicular traffic
Loading out
Hind erosion
3. Coal conveying and transfer
4. Coal pulverizing and
screening
Uncontrolled fugitive emission factor
0.2 kg/Mg coal unloaded3
(0.4 Ib/ton)
(0. 02) (Kl) (S/l. 5) kq/Mq material
(PE/100)^ loaded onto pileu
/(0.04) (Ki) (S/1.5) Ib/ton material "\
V (PE/lOO)^ loaded onto pile,/
(0.065) (K2) (S/l. 5) kq/Mq material
(PE/100H storedb
f(0.13) (K?) (S/l. 5) Ib/ton material^
V (PE/100)^ stored J
(0.025) (K3) (S/l. 5) kq/Mq material
(PE/100)2 loaded outb
f(0.05)
-------
Table 2-9 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR COKE MANUFACTURING
i
•e-
o
Source Of IPFPE
5. Coal charging
6. Coking (door leakage)
7. Pushing
8. Quenching
9. Coke handling
Uncontrolled fugitive emission factor
0.5-5.0 kg/Mg coal charged6' f'g/h
(1.0-10.0 Ib/ton)
0.20-0.45 kg/Mg coal chargede'E>1
(0. 40-0. 90 Ib/ton)
See Table 2-10
Clean water: 0.6 kg/Mg coke produced3
(1.2 Ib/ton)
Highly contaminated water:
1.0-3.0 kg/Mq coke produced3
(2.0-6.0 Ib/ton)
0.012-0.065 kg/Mg coke produced1*
(0.023-0.13 Ib/ton)
Emission
factor
reliability
rating
C
C
C
C
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Coal charged
1,542,500
(1,699,800)
Coal charged
1,542,500
(1,699,800)
Coal charged
1,542,500
(1,699,800)
Coke produced
1,000,000
(1,102,000)
Sized coke produced
1,000,000
(1,102,000)
Uncontrolled
emissions
Mg/yr
(tons/yr)
4,242
(4,674)
501
(552)
600
(660)
36
(42)
a Coal hopper car unloading emission factor as developed in Section 2.1.2.
For complete development of this emission factor refer to Section 2.1.4. For this example it was assumed that S = 4.0,
D = 90, PE = 100, KI = 0.75, Kj = 0.5, and K-j = 0.78. Reference 4.
Coal conveying and transfer emission factor as developed in Section 2.'.
Included in coal handling and transfer emission factor.
Reference 6.
Reference 7.
Reference 2.
Reference 8.
Reference 9.
Personal communication with Carl Edlund, U.S. Environmental Protection Agency, Division of Stationary Source Enforcement,
Washington, D.C., March 11, 1977. Emissions are for quench towers with baffles.
Coke conveying and transfer emission factor as developed in Section 2.1.1.
-------
Table 2-10. PUSHING EMISSION FACTORS
kg/Mg coal charged (Ib/ton coal charged)
o
Ul
I
Degree of
greenness
Green coke
Moderately
green
Cleanly pushed
coke
Sheda
0.32h - 0.51
(0.65 - 1.0)
0.28h
(0.57)
0.17f'g
(0.35)
Travelling hood
l.O1 - 1.8d
(2.1 - 3.5)
1.65d
(2.3)
0.75d
(1.5)
Direct-uncaptured
plume
1.5 - 2.0e
(3 - 4)
1.0°
(2.1)
0.196 - 0.26C
(0.38 - 0.52)
Includes most of travel emissions.
Does not include travel emissions.
Reference 11.
Reference 12.
Reference 13.
Reference 14.
Reference 15.
Reference 16 and 28.
Reference 17.
-------
Table 2-12. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR IRON PRODUCTION
i
4S
O
CTv
I
Source of IPFPE
1. Ship or railroad car
unloading
Iron ore - ship unloading
Iron ore - rail unloading
Limestone - ship unloading
Limestone - rail unloading
2. Iron ore storage
Loading onto pile
Vehicular traffic
Loading out
Wind erosion
Uncontrolled fugitive emission factor
Iron ore 0.01-0.015 kg/Mg unloaded3
(0.02-0.03 Ib/ton unloaded)
Limestone 0.1 kg/Mg unloaded
(0.2 Ib/ton unloaded)
(0.02) (Ki) (S/1.5) kg/Mg material .
(PE/100) ^ loaded onto pile
/(O. 04) (Ki) (S/1.5) Ib/ton material \
\ (PE/10012 loaded onto pilej
(0.065) (K2> (S/1.5) kg/Mg material
(PE/100) 2 stored1"
/"«>. 13) (K?) (S/1.5) Ib/ton material )
^ (PE/100) 2 stored /
(0.025) (K3) (S/1.5) kg/Mg material
(PE/100) 2 loaded outb
((0. 05) (K3) (S/1.5) Ib/ton material^
\ (PE/lOd)^ loaded out )
(0.055) (S/1.5) D kg/Mg material
(PE/100) ^ 90 storedb
f(0. 11) (S/1.5) D Ib/ton material)
V (PE/100) * 90 stored /
Emission
factor
reliability
rating
E
E
D
D
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Iron ore unloaded
1,680,300
(1,848,300)
Limestone unloaded
248,565
(273,422)
Iron ore loaded
1,680,300
(1,848,300)
Iron ore stored
1,680,300
(1,848,300)
Iron ore loaded out
1,680,300
(1,848,300)
Iron ore stored
1,680,300
(1,848,300)
Uncontrolled
emissions
Mg/yr
(tons/yr)
21
(23)
25
(27)
218
(240)
85
(94)
364
(400)
62
(68)
-------
Table 2-12 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR IRON PRODUCTION
i
j>
o
~-j
Source of IPFPE
3. Iron ore handling and-
transfer
4 . Limestone storage
5. Limestone handling and
transfer
6. Coke storage
7. Coke handling and transfer
8. Blast furnace flue dust
storage
9. Blast furnace flue dust
handling and transfer
10. Sinter machine windbox
discharge
11. Sinter machine discharge
and screens
12. Sinter cooler
Uncontrolled fugitive emission factor
1.0 kg/Mg ore handled0
(2.0 Ib/ton ore handled)
b
0.1 kg/Mg limestone handled*1
(0.2 Ib/ton limestone handled)
b
See Section 2.2.1e
Negligible1
0.15 kg/Mg flue dust9
(0.3 Ib/ton flue dust)
Negligible3'*1
0.28-1.22 kg/Mg sinter3 'h
(0.55-2.45 Ib/ton sinter)
0.16-0.4 kg/Mg sinter"'1
(0.32-0.8 Ib/ton sinter)
Emission
factor
reliability
rating
D
D
E
D
D
E
_
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Iron ore
1,680,300
(1,848,300)
Limestone
248,565
(273,422)
Limestone
248,565
(273,422)
Coke
586,613
(625,474)
_
Flue dust
49,713
(54,684)
_
Sinter
520,000
(570,000)
Sinter
520,000
(570,000)
Uncontrolled
emissions
Mg/yr
(tons/yr)
1,680
(1,848)
41
(45)
25
(27)
56
(63)
_
7
(8)
_
389
(427)
145
(160)
-------
Table 2-12 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR IRON PRODUCTION
-e-
o
00
Source of IPFPE
13. Sinter storage
14. Sinter handling and transfer
15. Blast furnace charging
16. Blast furnace upsets (slips)
17. Blast furnace tapping - iron
18. Blast furnace tapping - slag
19. Slag handling
20. Slag storage
21. Slag crushing
Uncontrolled fugitive emission factor
b
0.2 kg/Mg sinter^
(0.4 Ib/ton sinter)
Negligible
0.0019-0.019 kg/Mg iron produced
(0.0038-0.038 Ib/ton iron produced)
0.15-0.46 kg/Mg iron produced*'1
(0.3-0.92 Ib/ton iron produced)
Model values: . .
0.39-0.49 kg/Mg iron produced '
(0.78-0.92 Ib/ton iron produced)
1
0.01-0.05 kg/Mg slag"
(0.02-0.1 Ib/ton slag)
b
1.0 kg/Mg crushed0
(2.0 Ib/ton crushed)
Emission
factor
reliability
rating
D
E
E
E
E
D
-
C
B
A
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Sinter
44,000
(48,000)
Sinter
520,000
(570,000)
-
Iron
994,260
(1,093,686)
Iron
994,260
(1,093,686)
-
Slag
298,278
(328,106)
Slag
298,278
(328,106)
Slag
298,278
(328,106)
Uncontrolled
emissions
Mg/yr
(tons/yr)
7
(8)
104
(114)
-
10
(11)
423
(465)
1
9
(10)
49
(54)
298°
(328)
-------
Table 2-12 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR IRON PRODUCTION
o
VO
I
Source of IPFPE
Uncontrolled fugitive emission factor
Emission
factor
reliability
rating
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Uncontrolled
emissions
Mg/yr
(tons/yr)
Emission factor range for iron ore unloading (taconite pellets) derived from data presented in Section 2.1.2. Emission
factor for limestone unloading also derived from data presented in Section 2.1.2 assuming 50 percent of coal unloading
emission factor.
For complete development of this emission factor, refer to Section 2.1.4. Kor sources 2, 6, and 20 refer to section 2.1.4
for values for variables in formulas except for D which was assumed to equal 90. For source 4 it was assumed that S = 1.5,
and K,
1. Reference 1.
90, PE = 100, and K.
c ' ~"
Reference 2.
Engineering judgment, assumed 50% of coal handling emissions as reported in Reference 3.
e Emissions identified and included in Section 2.2.1. Additional emission factor of 0.055 Rg/Mg (0.11 Ib/tpn) pig iron
produced reported in Reference 3.
Blast furnace flue dust is normally handled in a closed system and so a negligible source.
g Engineering judgment, assumed equal to sand handling emissions as reported in Reference 3.
h Reference 4.
1 Reference 5.
* Engineering judgment, assumed equal to coal handling emissions as reported in Reference 3.
Reference 6.
1
m
n
° Estimated based on crushed stone emission factors for primary and secondary crushers in AP-42 (Reference 3).
approximately 65 percent of these emissions will settle out in the plant.
Reference 7. Emissions for source 18 included in emissions from source 17.
Emission for slag tapping included in iron tapping emission factor.
Engineering judgment, assumed equal to coke handling emission range as described in Section 2,
1.1.
Note that
-------
Table 2-14. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR STEEL PRODUCTION
o
i
Source of IPFPE
1.
2.
3.
4.
5.
Scrap steel unloading,
transfer and storage
Flux material unloading,
transfer, and storage
Molten pig iron transfer from
torpedos to charge ladles
(hot metal reladling)
Basic oxygen furnace - roof
monitor (total)
4a. Charging
4b. Leakage
4c. Tapping-steel
4d. Tapping-slag
Open hearth furnace - roof
monitor (total)
5a. Charging
Sb. Leakage
5c. Tapping-steel
5d. Tapping-slag
Uncontrolled fugitive emission factor
Negligible
Negligible
0.028-0.12 kg/Mg hot metala'b
-------
Table 2-14 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR STEEL PRODUCTION
Source of IPFPE
6. Electric arc furnace - roof
monitor (total)
6a. Charging
6b. Leakage
6c. Tapping-steel
6d. Tapping-slag
7. Ingot casting
8. Molten steel reladling
9. Scarfing
Uncontrolled fugitive emission factor
0.09-1.5 kg/Mg steelc'e'h>1
(0.18-3.0 Ib/ton steel)
Carbon steel: 1.5 kg/Mg steel
(3.0 Ib/ton)
Alloy steel: 0.75 kg/Mg steel
(1.5 Ib/ton)
j
j
j
j
0.014-0.06 kg/Mg steelk
(0.028-0.12 Ib/ton)
0.014-0.06 kg/Mg steelk
(0.028-0.12 Ib/ton steel)
0.0055 kg/Mg steel1
(0.011 Ib/ton steel)
Emission
factor
reliability
rating
C
E
E
C
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Steel produced
(350I carbon steel)
454,000
(500,000)
Steel produced
2,905,600
(3,200,000)
Steel produced
1,360,000
(1,500,000)
Steel produced
907,000
(1,000,000)
Uncontrolled
emissions
Mg/yr
(tons/yr)
511
(563)
108
(118)
50
(56)
5
(5)
Reference 2.
Reference 3.
c Reference 4.
Reference 5.
Reference 6.
DA^A^dnsiA "7
Emissions included with steel tapping emission factor.
Reference 8.
Emissions included with total open hearth building emission factor.
Emissions included with total electric furnace emission factor.
Engineering judgment, assumed to be 50 percent of the hot metal reladling emission factor because of the lower carbon
content of steel.
Reference 24.
-------
Table 2-16. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR PRIMARY ALUMINUM PRODUCTION
ro
I
Source of IPFPE
1. Material handling
2. Anode baking
3. Electrolytic reduction cell
3a. Prebaked
3b. VSS Soderberg
3c. HSS Soderberg
4. Refining
Uncontrolled fugitive emission factor
5.0 kq/Mq of aluminum produced3
(10 Ib/ton)
1.0-5.0 (1.5) kg/Mg of aluminum*3
producedc
(3.0 Ib/ton)
0.75-6.7 kg/Mg of aluminum produced0
(1.5-13.4 Ib/ton)
13.1 kg/Mg of aluminum produced0
(26.2 Ib/ton)
0.75-10.8 kg/Mg of aluminum produced0
(1.5-21.6 Ib/ton)
1.25 kg/Mg of aluminum produced
(2.5 Ib/ton)
Emission
factor
reliability
rating
E
D
D
D
D
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Aluminum produced
200,000
(220,400)
Aluminum produced
200,000
(220,400)
Aluminum produced
200,000
(220,400)
Aluminum produced
200,000
(220,000)
Uncontrolled
emiss ions
Mg/yr
(tons/yr )
1,000
(1,102)
300
(330)
745
(821)
250
(275)
a Includes unloading conveying, crushing, screening, mixing, and green anode or paste preparation. Individual emission rates
are not available. Most of these sources are enclosed/vented through a particulate control device.
b Reference 1 and 2.
c Reference 2 and 3. Includes charging, tapping, and anode replacement.
d No data available. Estimate based on the emission factor in Reference 1, assuming 95 percent hood capture efficiency.
-------
Table 2-21. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY COPPER SMELTERS
u>
Source of IPFPE
1. Unloading and handling of
ore concentrate
2. Ore concentrate storage
Loading onto pile
Vehicular traffic
Loading out
Wind erosion
Uncontrolled fugitive emission factor
5 kg/Mg material hi
(10 Ib/ton)
(0.02)(Ki) (S/1.5)
(PE/100) (S/1.5)
(PE/100)ndleda'b
kg/Mg material
loaded onto pilec
Ib/ton]
kg/Mg material
stored0
Ib/ton)
kg/Mg material
loaded outc
Ib/ton^
kg/Mg material
stored0
Ib/ton")
Emission
factor
reliability
rating
E
D
D
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Ore concentrate
305,140
(336,165)
Ore concentrate loaded
305,140
(336,165)
Ore concentrate stored
305,140
(336,165)
Ore concentrate loaded
out
305,140
(336,165)
Ore concentrate stored
305,140
(336,165)
Uncontrolled
emissions
Mg/yr
(tons/yr)
1,526
(1,681)
6
(7)
20
(22)
8
(9)
17
(18)
-------
Table 2-21 (continued). IDENTIFICATION AND QUANTIFICATION OF
POTENTIAL FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY COPPER SMELTERS
Source of IPFPE
3.
4.
5.
e.
7.
8.
9.
10.
Limestone flux unloading
and handling
Limestone flux storage
Roaster charging
Roaster leakage
Calcine 'transfer
Charging reverberatory
furnace
Tapping of reverberatory
Reverberatory furnace
leakage
Uncontrolled fugitive emission factor
b
c
11.5 kg/Mg copper produced 'e
(23 Ib/ton)
e
e
4.15-4.35 kg/Mg copper produced1 'g'h
(8.3-8.7 Ib/ton)
h
h
Emission
factor
reliability
rating
D
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Limestone flux
138,700
(152,570)
Copper produced
90,478
(99,645)
Copper produced
90,478
(99,645)
Uncontrolled
emissions
Mg/yr
(tons/yr)
b
23
(25)
1,040
(1,146)
e
e
385
(423)
9
g
-------
Table 2-21 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY COPPER SMELTERS
Ln
I
Source of IPFPE
11.
12.
13.
14.
15.
16.
17.
18.
19.
Slag tapping
Converter charging.
Converter leakage
Slag tapping from converter
Blister copper tapping
Blister copper transfer
Charging blister copper to
fire refining furnace
Copper tapping and casting
Slag tapping and handling
Uncontrolled fugitive emission factor
h
1.6-8.85 kg/Mg copper produced -1
(3.3-17.7 Ib/ton)
j
j
j
j
0.5-1.4 kg/Mg copper produced1'"1
(1.0-2.8 Ib/ton)
1.26 kg/Mg copper produced '
(2.52 Ib/ton)
k
Emission
factor
reliability
rating
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Copper produced
90,478
(99,645)
Copper produced
90,478
(99,645)
Copper produced
90,478
(99,645)
Uncontrol led
emissions
Mg/yr
(tons/yr )
g
545
(600)
i
i
i
i
86
(95)
114
(126)
j
Reference 1 also includes slag handling.
Emission from limestone unloading and handling included in emission factor for ore unloading and handling.
For complete development of this emission factor, refer to Section 2.1.4. The emission factor for source 4 is the same
source 2. For these examples it was assumed that S = 1.5, D = 90, PE = 100, and KI , K2, and K.. = 1. Reference 2.
Based on material balance using same percentage as estimated for SO, emissions from reference 3.
Emissions from roaster leakage and transfer are included in emission factor for roaster charging.
Reference 4.
Lower value of range is for plants with roaster, high value for plants without roaster.
Emissions from reverberatory tapping and leakage are included in emission factor for reverberatory charging.
Reference 5.
Emissions from converter leakage and tapping, and blister copper transfer are included with converter charging emission
factor.
Emissions from slag to tapping are included in casting building emissions.
Engineering judgment, assumed approximately equal to 25 percent of the reverberatory furnace fugitive emissions.
Reference 13.
-------
Table 2-24. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY LEAD SMELTERS
i
j>
h-1
ON
I
Source of IPFPE
1. Railroad car and truck
unloading
Limestone
Silica sand
Lead ore concentrate
Iron ore
Coke
2. Blast furnace flue dust
2a. Storage
2b. Handling and transfer
3. Limestone
3a. Storage
Loading onto pile
Uncontrolled fugitive emission factor
0.015-0.2 kg/Mg unloaded3
(0.03-0.4 Ib/ton unloaded)
0.015-0.2 kg/Mg unloaded3
(0.03-0.4 Ib/ton unloaded)
0.015-0.2 kg/Mg unloaded3
(0.03-0.4 Ib/ton unloaded)
0.015-0.2 kg/Mg unloaded3
(0.03-0.4 Ib/ton unloaded)
0.2 kg/Mg unloaded3
(0.4 Ib/ton unloaded)
Negligible15
Negligible13
(0.02)
-------
Table 2-24 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY LEAD SMELTERS
Source of IPFPE
Vehicular traffic
Loading out
Hind erosion
3b. Handling and transfer
4. Silica sand
4a. Storage
4b. Handling and transfer
Uncontrolled fugitive emission factor
(0.065) (K?) (S/1.5) kq/Mq material
(PE/100)Z Stored0
/1o. 04) (K2) (S/1.5) Ib/ton material^
V
-------
Table 2-24 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
-»
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY LEAD SMELTERS
00
Source of IPFPE
5. Lead ore concentrate
5a. Storage
5b. Handling and transfer
6. Iron ore
6a. Storage
6b. Handling and transfer
7. Coke
7 a . Storage
7b. Handling and transfer
Uncontrolled fugitive emission factor
c
0.82-2.5 kg/Mg handled1
(1.64-5.0 Ib/ton handled)
c
1.0 kg/Mg iron ore handled9
(2.0 Ib/ton iron ore handled)
c
0.06-0.1 kg/Mg coke handled11'1
(0.13-3.39 Ib/ton coke handled)
Emission
factor
reliability
rating
D
E
D
E
D
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Concentrate stored
317,782
(349,560)
Concentrate handled
317,782
(349,560)
Ore stored
45,000
(49,500)
Ore handled
45,000
(49,000)
Coke stored
94,494
(103,943)
Coke handled
94,494
(103,943)
Uncontrolled
emissions
Mg/yr
(tons/yr)
52
(57)
528
(580)
29
(32)
45
(49)
7
(8)
8
(9)
-------
Table 2-24 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY LEAD SMELTERS
vO
I
Source of IPFPE
8. Mixing and palletizing
9. Sinter machine
10. Sinter return handling
11. Sinter machine discharge and
screens
12. Sinter crushing
13. Sinter transfer to dump area
14. Sinter product dump area
Uncontrolled fugitive emission factor
0.57-1.70 kg/Mg lead product3
(1.13-3.39 Ib/ton lead product)
0.12-0.55 kg/Mg sinterk
(0.25-1.1 Ib/ton sinter)
2.25-6.75 kg/Mg sinter3
(4.5-13.5 Ib/ton sinter)
0.28-1.22 kg/Mg sinterk
(0.55-2.45 Ib/ton sinter)
1
0.05-0.15 kg/Mg sinter transfered3
(0.10-0.30 Ib/ton sinter transfered)
0.0025-0.0075 kg/Mg sinter dumped3
(0.005-0.015 Ib/ton sinter dumped)
Emission
factor
reliability
rating
D
E
D
E
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Lead produced
200,000
(220,000)
Sinter produced
349,979
(384,977)
Sinter produced
349,979
(384,977)
Sinter produced
349,979
(384,977)
Sinter transfered
349,979
(384,977)
Sinter dumped
349,979
(384,977)
Uncontrolled
emissions
Mg/yr
(tons/yr)
227
(247)
117
(130)
1,575
(1,732)
262
(289)
1
35
(38)
2
(2)
-------
Table 2-24 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY LEAD SMELTERS
N3
O
I
Source of IPFPE
15. Charge car or conveyor loading
and transfer of sinter
16. Blast furnace - monitor (total)
16a. Charging
16b. Blow condition
16c. Upset"
16d. Tapping
17. Lead pouring to ladle and transfer
18. Slag pouring
19. Slag cooling
20. Slag granulator and slag piling
21. Zinc fuming furnace vents
22. Dross kettle
Uncontrolled fugitive emission factor
0.13-0.38 kg/Mg charged0
(0.25-0.75 Ib/ton charged)
0.04-0.12 kg/Mg lead produced3'1"
(0.08-0.23 Ib/ton)
m
m
3.5-11.5 kq/Mg lead produced3'1"
(7.0-23.0 Ib/ton)
m
0.47 kg/Mg lead produced
(0.93 Ib/ton lead produced)
P
0.24 kg/Mg lead produced**
(0.47 Ib/ton lead produced)
Negligibler
1.15-3.45 kg/Mg lead3
(2.3-6.9 Ibs/ton lead)
0.12-0.36 kg/Mg lead3
(0.24-0.72 Ib/ton lead)
Emission
factor
reliability
rating
D
D
D
E
E
E
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Blast furnace charge
477,226
(524,949)
Lead produced
200,000
(220,000)
Lead produced
200,000
(220,000)
Lead produced
200,000
(220,000)
Slag crushed
200,000
(220,000)
Lead produced
200,000
(220,000)
Lead produced
200,000
(220,000)
Uncontrolled
emissions
Mg/yr
(tons/yr)
122
(131)
16
(17)
m
m
0
m
94
(102)
P
48
(52)
200
(220)
460
(506)
48
(53)
-------
Table 2-24 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY LEAD SMELTERS
Source of IPFPE
23. Reverberatory furnace leakage
24. Silver retort building
25. Lead casting
Uncontrolled fugitive emission factor
0.75-2.25 kg/Mg lead3
(1.5-4.5 Ib/ton lead)
0.45-1.35 kg/Mg lead'
(0,9-2.7 Ib/ton lead)
0.22-0.66 kg/Mg lead^
0.43-1.30 Ib/ton lead)
Emission
factor
reliability
rating
D
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Lead produced
200,000
(220,000)
Lead produced
200,000
(220,000)
Lead produced
200,000
(200,000)
Uncontrolled
emissions
Mg/yr
(tons/yr)
300
(330)
180
(198)
88
(96)
Engineering judgement based on data presented in Section 2.1.2; emission range derived using emission factors for unloading
of Taconite pellets and coal/hopper car; emission factor for coke derived from coal unloading emission factor only.
Engineering judgement, assumed enclosed handling and storage or direct recycle to system.
For complete development of this emission factor, refer to Section 2.1.4. The emission factor for sources 4a, 5a, 6a,
and 7a are the same as source 3a. For sources 3a, 4a, and 5a it was assumed that S = 1.5, D = 90, PE = 100, and K.,
and
1. Values for sources 6 and 7 can be found in Section 2.1.4. Reference 1.
Engineering judgment, assumed 50 percent of coal handling emissions as reported in Reference 2.
Engineering judgment based on aggregate storage pile emission factors in Reference 2.
Reference 3.
Reference 4.
Engineering judgment; calculated from emission factor (0.055 kg/Mg of iron) given in Reference 5.
Reference 6.
Reference 7.
Engineering judgment using steel sinter machine leakage emission factor given in Reference 8 and 9.
Emissions for sinter crushing included in emissions from sinter machine discharge and screens.
Emissions for charging, blow condition, and tapping included in total.
Emission factor for upset not considered part of normal operating conditions and is not included in emission factor
for the blast furnace roof monitor.
Emissions for blast furnace upset are not included in model plant inventory.
Emissions for slag pouring included in lead pouring to ladle and transfer emission.
Engineering judgment; estimated to be one half the magnitude of pouring and ladling operations (source number 17).
Granulated slag is wet and therefore most likely not a source of fugitive emissions. Reference 10.
-------
Table 2-27. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY ZINC PRODUCTION
to
I
Source of IPFPE
1. Railroad car or truck
unloading
la. Zinc ore concentrate
Ib. Sand
Ic. Coke
2. Zinc ore concentrate
2a. Storage
Loading onto pile
Vehicular traffic
Loading out
Uncontrolled fugitive emission factor
0.015-0.2 kg/Mg unloaded3
(0.03-0.4 Ib/ton unloaded)
0.015-0.2 kg/Mg unloaded8
(0.03-0.4 Ib/ton unloaded)
0.2 kg/Mg unloaded8
(0.4 Ib/ton unloaded)
(0.02) (Ki) (S/1.5) kg/Mg material .
(PE/IOOK loaded onto pile
Afl.04) (Ki) (S/1.5) Ib/ton material \
V. (PE/lOO)* loaded onto pile/
(0.065) (K?) (S/1.5) kg/Mg material
(PE/100M stored^
f(0.13) (K2) (S/1.5) Ib/ton material^
\ (PE/100)^ stored /
(0.025) (K3) (S/1.5) kg/Mg material
(PE/lOO)' loaded outb
((0. 05) (K3) (S/1.5) Ib/ton material^
V
-------
Table 2-27 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY ZINC PRODUCTION
1S3
OJ
Source of IPFPE
Wind erosion
2b. Handling and transfer
3. Sand
3a. Stored
3b. Handling and transfer
4. Coke
4a. Storage
4b. Handling and transfer
Uncontrolled fugitive emission factor
(0.055)(S/1.5) D kq/Mg material
(PE/100)^ 90 storedb
f(O.ll) (S/1.5) D Ib/ton material^
V^ (PE/100K 9~ff stored /
0.82-2.5 kg/Mg handled0
(1.64-5.0 Ib/ton handled)
b
0.15 kg/Mg handled3
(0.3 Ib/ton handled)
b
0.06-0.1 kg/Mg handled6' f
(0.12-0.2 Ib/ton handled)
Emission
factor
reliability
rating
D
E
D
D
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Ore concentrate
loaded out
195,447
(214,992)
Ore concentrate handled
195,447
(214,992)
Sand
1,613
(1,774)
Sand
1,613
(1,774)
Coke
5,734
(6,307)
Coke
5,734
(6,307)
Uncontrolled
emissions
Mg/yr
(tons/yr)
10
(11)
324
(356)
0.27
(0.29)
0.24
(0.26)
1
(D
0.46
(0.50)
-------
Table 2-27 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY ZINC PRODUCTION
i
.p-
N)
Source of IPFPE
5. Sinter machine windbox
discharge
6. Sinter machine discharge
and screens
7. Coke-sinter mixer
8. Retort furnace building
Ba. Retort furnace tapping
8b. Retort furnace residue
discharge and cooling
Be. Retort furnace upset
9. Zinc casting
Uncontrolled fugitive emission factor
0.12-0.55 kg/Mg sinter"5'9
(0.25-1.1 Ib/ton sinter)
0.28-1.22 kg/Mg sinter9'*1
(0.55-2.45 Ib/ton sinter)
i
1.0-2.0 kg/Mg of zinc^
(2.0-4.0 Ibs/ton of zinc)
k
0.25-1.0 kg/Mg of zinc1
(0.5-2.0 Ib/ton of zinc)
2.5-5 kg/Mg zincm
(5-10 Ibs/ton zinc)
1.26 kg/Mg zinc"
(2.52 Ib/ton zinc)
Emission
factor
reliability
rating
E
E
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter.
Mg/yr
(tons/year)
Sinter
70,788
(78,867)
Sinter
70,788
(78,867)
Zinc
68,100
(75,000)
Zinc
68,100
(75,000)
Zinc
68,100
(75,000)
Zinc
68,100
(75,000)
Uncontrolled
emissions
Mg/yr
(tons/yr)
24
(26)
53
(59)
9
61
(68)
i
43
(47)
86
(95)
-------
Table 2-27 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR PRIMARY ZINC PRODUCTION
i
.P-
NJ
IJ1
I
Source of IPFPE
Uncontrolled fugitive emission factor
Emission
factor
reliability
rating
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Uncontrolled
emissions
Mg/yr
(tons/yr)
Estimate based on data presented in Section 2.1.2; emission range derived using emission factors for unloading of Taconite
pellets and coal/hopper car; emission factor for coal derived from coal unloading factor only.
For complete development of this emission factor, refer to Section 2.1.4. The emission factor for sources 3a and 4a are the
same as source 2. For sources 2a and 3a it was assumed that S = 1.5, D = 90, PE = 100, and K, , K,, and K, = 1. Values for
source 4 can be found in Section 2.1.4. Reference 2.
Engineering judgment assuming that fugitive emission factors given for lead ore concentrate (Reference 3) to be similar to
those for zinc ore concentrate.
Reference 4.
Engineering judgment based on 50% of emission factor given for coal in Reference 4.
Reference 5.
Engineering judgment assuming that fugitive emission factors given for sintering machine in iron production (Reference 6)
is similar for .sintering in zinc production.
Reference 7.
Emissions from coke-sinter mixer included in emission factor for sinter machine discharge and screens (emission source 6).
Engineering judgment using emission factor from retort building in primary lead smelting (Reference 8).
Emissions from retort furnace tapping included in emission factor for retort building total (emission source 8).
Engineering judgment based on observation of retort residue discharging and cooling at a secondary zinc smelter which is
similar to primary zinc production.
Engineering judgment using emission factor for primary copper smelting given in Reference 9. Not considered part of
ordinary operations. Reference 10.
Engineering judgment assuming fugitive emissions from zinc casting equal to fugitive emission for copper casting given
in reference 11.
-------
Table 2-29. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR SECONDARY ALUMINUM PRODUCTION
i
-P-
ro
Source of IPFPE
1. Sweating furnace
la. Charging
Ib. Tapping
2. Crushing and screening
scrap metal
3. Chip (rotary) dryer
4. Smelting (reverberatory)
furnace
4a. Charging
4b. Tapping
5. Smelting (crucible) furnace
5a. Charging
5b. Tapping
6. Smelting (induction) furnace
6a. Charging
6b. Tapping
Uncontrolled fugitive emission factor
0.36 kg/Mg metal processed3
(0.72 Ib/ton metal processed)
b
b
Negligible0
0.36 kg/Mg metal dried*3
(0.72 Ib/ton metal dried)
0.11 kg/Mg metal processed3
(0.22 Ib/ton metal processed)
e
e
0.05 kg/Mg metal processed3
(0.09 Ib/ton metal processed)
f
f
0.05 kg/Mg metal processed^
(0.09 Ib/ton metal processed)
h
h
Emission
factor
reliability
rating
B
E
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Scrap processed
2,727
(3,000)
Metal dried
1,136
(1,250)
Metal processed
4,546
(5,000)
-
-
Uncontrolled
emissions
Mg/yr
(tons/yr)
1
(1)
Negligible
5
(6)
1
(1)
-
-
-------
Table 2-29 (continued), IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR SECONDARY ALUMINUM PRODUCTION
i
-o
fO
»vi
I
Source of IPFPE
7. Fluxing (chlorination)
8. Hot dross handling and
cooling
9. Pouring hot metal into molds
or crucible
Uncontrolled fugitive emission factor
25 kg/Mg chlorine used3
(50 Ib/ton chlorine used)
0.11 kg/Mg metal processed^
(0.22 Ib/ton metal processed)
Negligible15
Emission
factor
reliability
rating
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Chlorine used
455
(500)
Metal processed
4,546
(5,000)
-
Uncontrolled
emissions
Mg/yr
(tons/yr)
12
(13)
1
(1)
Negligible
Engineering judgement, assume 5% of uncontrolled stack emissions as reported in Reference 4, p. 7.8-1.
Emissions included with total sweating furnace emission factor.
Based on engineering judgement.
Assume uncontrolled fugitive emissions are equal to the emissions from the sweating furnace.
e Emissions included with total reverberatory furnace emission factor.
Emissions included with total crucible furnace emission factor.
" Assume uncontrolled fugitive emissions are equal to the emissions from the crucible furnace.
Emissions included with total induction furnace emission factor.
Assume that the amount of fluxing agent used is 10% of the weight of the metal processed in the reverberatory furnace
(Reference 1, p. 286).
3 Assume that the emissions are equal to the emissions from the reverberatory furnace.
i.
Emissions are negligible as reported in Reference 1, p. 285.
-------
Table 2-32. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR SECONDARY LEAD SMELTING
to
co
Source of IPFPE
1. Railroad car and truck
Coke
Limestone
Lead scrap
Iron scrap
2. Coke
2a Storage
Loading onto pile
Vehicular traffic
Loading out
Uncontrolled fugitive emission factor
0.2 kg/Mg unloaded3
(0.4 Ib/ton unloaded)
0.015-0.2 kg/Mg unloaded3
(0.03-0.4 Ib/ton unloaded)
Negligible
Negligible
(0.02) (Ki) (S/1.5) kq/Mq material .
(PE,?100)Z loaded onto pile
Ao. 04) (Ki) (S/1.5) Ib/ton material^
V (PE/100)2 onto pile )
(0.065) (K2) (S/1.5) kg/Mq material
(PE/100)* storedb
f(0. 13) (K7) (S/1.5) Ib/ton material)
V (PE/100)Z stored /
(0.025) (K3) (S/1.5) kq/Mq material
(PE/100)2 loaded outD
Ao. 05) (K3) (S/1.5) Ib/ton material^
\ (PE/100)^ loaded out /
Emission
factor
reliability
rating
E
E
E
E
D
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Coke unloaded
3,141
(3,455)
Limestone unloaded
1,713
(1,884)
Coke loaded
3,141
(3,455)
Coke stored
3,141
(3,455)
Coke loaded out
3,141
(3,455)
Uncontrolled
emissions
Mg/yr
(tons/yr)
0.63
(0.69)
0.18
(0.20)
negligible
negligible
negligible
-------
Table 2-32 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR SECONDARY LEAD SMELTING
ro
vo
I
Source of IPFPE
Hind erosion
2b Handling and transfer
3. Limestone
3a Stored
3b Handling and transfer
4 . Lead scrap
4a Storage
4b Handling and transfer
5. Iron scrap
5a Storage
Sb Handling and transfer
6. Lead and iron scrap burning
7. Sweating furnace (total)
*
Uncontrolled fugitive emission factor
(0.055) (S/l. 5) D kq/Mq material
(PE/1QO)^ M storedb
((0.11) (S/l. 5) D Ib/ton material}
V^ (PE/100)^ 90 stored 1
0.055-0.07 kg/Mg handledc'd
(0.11-0.13 Ib/ton handled)
b
0.1 kg/Mg limestone handled6
(0.2 Ib/ton limestone handled)
Negligible
Negligible
Negligible
Negligible
f
0.5-1.0 kg/Mg scrap burned
(1.0-2.0 Ib/ton scrap burned)
0.8-1.75 kg/Mg charged^
(1.6-3.5 Ib/ton charged)
Emission
factor
reliability
rating
D
D
D
E
E
£
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Coke stored
3,141
(3,455)
Coke handled
3,141
(3,455)
Lime stored
1,713
(1,884)
Limestone handled
1,713
(1,884)
-
-
-
-
Total scrap burned
151,437
(166,580)
Material charged
137,670
(151,437)
Uncontrolled
emissions
Mg/yr
(tons/yr)
negligible
0.20
(0.22)
negligible
negligible
-
-
-
-
114
(125)
176
(194)
-------
Table 2-32 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR SECONDARY LEAD SMELTING
LO
o
I
Source of IPFPE
7a Charging
7b Tapping
8. Reverberatory furnace (total)
8a Charging
Bb Tapping
9. Blast or cupola furnace (total)
9a Charging
9b Lead tapping to holding pot
9c Slag tapping
0. Tapping of holding pot
Uncontrolled fugitive emission factor
h
h
1.4-7.85 kg/Mg charged9
(2.8-15.7 Ib/ton charged)
i
i
6 kg/Mg charged^
(12 Ib/ton charged)
k
k
k
k
Emission
factor
rel lability
rating
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Total charge
135,000
(148,610)
Total charge
57,100
(62,810)
Uncontrolled
emissions
Mg/yr
(tons/yr)
h
h
625
(688)
i
i
343
(377)
k
k
k
k
-------
Table 2-32 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR SECONDARY LEAD SMELTING
Source of IPFPE
11. Pot (kettle) furnace (total)
lla Charging
lib Tapping
12. Casting
Uncontrolled fugitive emission factor
0.02 kg/Mg charged9
(0.04 Ib/ton charged)
1
1
0.44 kg/Mg lead cast™
(0.87 Ib/ton lead cast)
Emission
factor
reliability
rating
C
C
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Total charge
12,000
(13,200)
Lead cast
80,000
(88,000)
Uncontrol led
emissions
My/yr
(tons/yr )
0.24
(0.26)
1
1
35
(38)
Estimate based on data presented in Section 2.1.2; emission range derived using emission factors for unloading taconite
pellets and coal/hopper car; emission factor for coke derived only from coal unloading emission factor.
For complete development of this emission factor, refer to Section 2.1.4. The emission factor for source 3a is the same
for source 2a. For source 3a it was assumed that S = 1.5, D = 90, PE = 100, and K,, K,, and K, = 1. Values for source
2a can be found in Section 2.1.4. Reference 1.
Reference 2.
Reference 3.
Engineering judgment, assumed 50 percent of coal handling emission reported for coal in Reference 4.
Engineering judgment based on emission factor for zinc residual scrap (Reference 1) with 5 percent resulting in fugitive
emissions.
Engineering judgment based on lead sweating emission factor given in Reference 5 with 6 percent resulting in fugitive
emissions.
Fugitive emissions for charging and tapping included in emission for total sweating furnace operation (Source 7).
Fugitive emissions for charging and tapping included in emission for total reverberatory furnace operation (Source 8).
Engineering judgment based on emission factor given in Reference 6 with 5 percent resulting in fugitive emissions.
Fugitive emissions for charging, lead tapping to holding pot, slag tapping, and tapping of holding pot included in emission
factor for total blast furnace operation (Source 9).
Fugitive emissions for charging and tapping included in emissions for total pot furnace operations (Source 11).
Reference 7; fugitive emissions for primary lead casting assumed equal to fugitive emissions for secondary lead casting.
-------
to
to
I
Table 2-34. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR SECONDARY ZINC PRODUCTION
Source of IPFPE
1. Crushing/screening of residue
skimmings
2. Reverberatory sweat furnace
(total)
2a. charging
2b. Tapping
3. Kettle (Pot) sweat furnace
(total)
3a. Charging
3b. Tapping
4. Rotary sweat furnace (total)
4a. Charging
4b. Tapping
5. Muffle sweat furnace (total)
5a. Charging
5b. Tapping
Uncontrolled fugitive emission factor
0.5-3.8 kg/Mg crushed and screened3
(1.0-7.6 Ib/ton crushed and screened)
Neg.-0.63 kg/Mg product
(Neg.-1.3 Ib/ton product)
c
c
0.28 kg/Mg product
(0.56 Ib/ton product
d
d
0.28-0.63 kg/Mg product6
(0.56-1.26 Ib/ton product)
f
f
0.27-0.8 kg/Mg zinc scrap charged6
(0.54-1.6 Ib/ton zinc scrap charged)
g
g
Emission
factor
reliability
rating
£
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Residue skimmings
crushed
2,877
(3,165)
V
V
V
Metal produced
6,276
(6,904)
Residue skimmings
6,276
(6,904)
V
V
V
Uncontrolled
emiss ions
Mg/yr
(tons/yr)
6
(7)
V
V
V
2
(2)
d
d
3
(3)
f
f
V
V
V
-------
I
•p-
u>
OJ
Table 2-34 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR SECONDARY ZINC PRODUCTION
Source oC IPFPE
6. Electric resistance sweat
furnace (total)
6a. Charging
6b. Tapping
7. Hot metal transfer to melting
furnaces
8. Crucible melting furnace
(total)
8a. Charging
8b. Tapping
9. Kettle (pot) melting furnace
(total)
9a. Charging
9b. Tapping
10. Reverberatory melting furnace
(total)
lOa. Charging
lOb. Tapping
Uncontrolled fugitive emission factor
0.25 kg/Mg zinc scrap charged6
(0.5 Ib/ton zinc scrap charged)
h
h
i
0.0025 kg/Mg zinc produced''
(0.005 Ib/ton zinc produced)
k
k
0.0025 kg/Mg zinc product
(0.005 Ib/ton zinc produced)
1
1
0.0025 Kq/Mg zinc produced6
(0.005 Ib/ton zinc produced)
m
m
Emission
factor
reliability
rating
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
V
V
V
Zinc produced
5,230
(5,753)
V
V
V
Zinc produced
5,230
(5.753)
Uncontrolled
emissions
Mg/yr
(tons/yr )
V
V
V
i
0.01
(0.01)
k
k
V
V
V
0.01
(0.01)
m
m
-------
Table 2-34 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR SECONDARY ZINC PRODUCTION
Source of IPFPE
11. Electric induction melting
(total)
lla. Charging
lib. Tapping
12. Hot metal transfer to retort
or alloying
13. Distillation retort and
condenser (total)
13a. Charging distillation
retort
13b. Leakage between retort
condenser
13c. Upset in condenser
13d. Tapping
14. Muffle distillation furnace
and condenser
14a. Charging muffle dis-
tillation furnace
14b. Leakage between furnace
and condenser
Uncontrolled fugitive emission factor
0.0025 kg/Mg zinc produced
(0.005 Ib/ton zinc produced)
n
n
i
1.18 kg/Mg zinc produced
(2.36 Ib/ton zinc produced)
P
Negligible
2.5-5.0 kg/Mg zinc produced"
(5.0-10 Ib/ton zinc produced)
0.01-0.02 kg/Mg tappede'P
(0.02-0.04 Ib/ton tapped)
1.18 kg/Mg zinc produced0
(2.36 Ib/ton zinc produced)
r
Negligible
Emission
factor
reliability
rating
E
E
E
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
V
V
V
Zinc produced
5,475
(6,023)
-
-
P
Zinc produced
5,475
(6,023)
-
Uncontrolled
emissions
Mg/yr
(tons/yr)
V
V
V
i
6
(7)
P
-
w
P
(6)
(7)
r
-
-------
Table 2-34 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR SECONDARY ZINC PRODUCTION
10
Ln
Source of IPFPE
14c. Upiet in condenser
14d. Tapping
IS. Alloying
16. Casting
Uncontrolled fugitive emission factor
2.S-S.O kg/Hg zinc produced*
(5. 0-10.0 Ib/ton zinc produced)
0.01-0.02 kg/Hg zinc tapped* 'r
(0.02-0.04 Ib/ton tapped)
u
0.005-0.01 kg/Hg zinc cast"
(0.01-0.02 Ib/ton zinc cast)
Emission
factor
reliability
rating
E
E
E
Ho'del plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Cine cast
10,950
(12,045)
Uncontrolled
emissions
Mg/yr
(tons/yr)
w
r
u
o.oe
(0.09)
Reference 1.
Engineering Judgment based on emission factors given in references 1 and 2 assuming fugitive emissions to be equal to 5
percent of stack emissions. Ranges, when they appear, were derived from factors for clean and residual scrap.
Emission included in total for Source 2.
Emission included in total fox Source 3.
Engineering judgment based on, stack emission factor given in Reference 1 assuming fugitive emissions to be equal to 5 per-
cent of stack emissions.
Emissions included in total for Source 4.
Emissions included in total for Source 5.
Emissions included in total for Source 6.
Emissions included in emissions for individual furnace operations.
Engineering judgment assuming fugitive emissions from crucible melting furnace to be equal to fugitive emissions from
kettle (pot) melting furnace (Source 9).
Emissions included in total for Source 8.
Emissions included in total for Source 9.
Emissions included in total for Source 10.
Emissions included in total for Source 11.
Engineering judgment based on emission factor given in Reference 2, assuming fugitive emissions to be equal to 5 percent of
stack emissions.
Emissions included in total for Source 13.
Personal communication from J. P. Barnhart of H. J. Bullock, Inc. to Thomas Janszen. Personal estimation by J. P. Barnhart.
This is not considered part of normal operating conditions.
Emissions included in total for Source 14.
Engineering judgment assuming upset conditions for Source 13o to be equal to that for Source 14c.
Engineering judgment assuming tapping emissions for Source 13d equal to those for Source 14d.
Alloying often takes place with the sweating or melting operations (Reference 1), however, if performed separately, it is
an engineering judgment that fugitive emissions could range from negligible to as high as the emission factor given for
iron innoculation in Section 2.5 of this report.
Not Included in emission inventory.
Upset'conditions are not considered part of normal operating conditions and therefore are not included in the emission
inventory.
-------
Table 2-36. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR SECONDARY COPPER, BRASS/BRONZE PRODUCTION
Source of IPFPE
1. Sweating furnace (total)
la. Charging
Ib. Tapping
2. Drying (total)
2a. Charging
2b. Discharging
3. Insulation burning
4. Electric induction furnace
(total)
4a. Charging
4b. Tapping
5. Reverberator/ furnace (total)
5a. Charging
5b. Tapping
Uncontrolled fugitive emission factor
0.38 kg/Mg scrap charged3
(0.76 Ib/ton scrap charged)
b
b
6.85 kg/Mg scrap dried3
(13.7 Ib/ton scrap dried)
c
c
6 . 9 kg/Mg scrap burned
(13.8 Ib/ton scrap burned)
0.025-0.07 kg/Mg scrap charged6
(0.05-0.14 Ib/ton scrap charged)
f
f
1.33-3.92 kg/Mg charged5
(2.65-7.84 Ib/ton charged)
0.6-1.48 kg/Mg charged9'11
(1.2-2.95 Ib/ton charged)
0.01-0.025 kg/Mg chargedg'h
(0.02-0.05 Ib/ton charged)
Emission
factor
reliability
rating
E
E
E
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Material charged
19,789
(21,768)
n
n
n
Scrap burned
38,462
(42,308)
n
n
n
Material charged
12,500
(13,750)
Material charged
12,500
(13,750)
Material charged
12,500
(13,750)
Uncontrolled
emissions
Mg/yr
(tons/yr)
8
(9)
b
b
n
n
n
265
(292)
n
n
n
33
(36)
13h
(14)
0.2h
(0.2)
-------
Table 2-36 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR SECONDARY COPPER, BRASS/BRONZE PRODUCTION
i
.e-
u>
Source of IPFPE
6. Rotary furnace (total)
6a. Charging
6b. Tapping
7. Crucible furnace (total)
7a. Charging
7b. Tapping
8. Cupola (blast) furnace
(total)
8a. Charging
8b. Tapping
9. Casting
Uncontrolled fugitive emission factor
0.75-3.68 kg/Mg charged9
(1.5-7.35 Ib/ton charged)
0.3 kg/Mg charged9 jl
(0.59 Ib/ton charged)
0.015-0.045 kg/Mg charged9'1
(0.03-0.09 Ib/ton charged)
0.16-0.32 kg/Mg charged6
(0.32-0.64 Ib/ton charged)
j
j
0.5-1.75 kg/Mg charged*
(1.0-3.5 Ib/ton charged)
1
1
0.005-0.01 kg/Mg cast"1
(0.01-0.02 Ib/ton cast)
Emission
factor
reliability
rating
E
E
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Material charged
13,700
(15,070)
Material charged
13,700
(15,070)
Material charged
13,700
(15,070)
n
n
n
Material charged
15,714
(17,286)
Metal cast
30,000
(33,000)
Uncontrolled
emissions
Mg/yr
(tons/yr)
30
(33)
41
(4)
0.41
(0.4)
n
n
n
18
(19)
1
1
0.2
(0.2)
-------
Table 2-36 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR SECONDARY COPPER, BRASS/BRONZE PRODUCTION
i
-p-
00
Source of IPFPE
Uncontrolled fugitive emission factor
Emission
factor
reliability
rating
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Uncontrolled
emissions
Mg/yr
-------
OJ
VO
I
Table 2-39. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR FOUNDRIES
Source of IPFPE
1. Raw material receiving and
storage
la Unloading
Ib Storage
Loading onto pile
Vehicular traffic
Loading out
Wind erosion
Uncontrolled fugitive emission factor
Coke unloading 0.2
(0.4
<0.02)
-------
Table 2-39 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR FOUNDRIES
o
Source of IPFPE
2. Cupola furnace operation
(charging, tapping, etc.)
3. Crucible furnace operation
(charging, tapping, etc.)
4. Electric arc furnace operation
5. Open hearth furnace operation
6. Electric induction furnace
operation^
7. Pot furnace operation
8. Reverberatory furnace
operation
9. Ductile iron innoculation
Uncontrolled fugitive emission factor
0.05-1 kg/Mg iron"
(0.1-2 Ib/ton iron)
0.05-0.3 kg/Mg of metal processed"
(0.1-0.6 Ib/ton of metal processed)
2.5-5 kg/Mg of metal charged6
(5.0-10 Ib/ton of metal charged)
0.53-1.74 kg/Mg of steelf>1
(1.05-3.48 Ib/ton of steel)
0.05-0.45 kg/Mg of metal charged0'h
(0.1-0.9 Ib/ton of metal charged)
1.0 kg/Mg of metal charged6
(2.0 Ib/ton of metal charged)
0.75 kg/Mg of ironc
(1.5 Ib/ton of iron)
0.2 kg/Mg3
(0.4 Ib/ton)
4.15-4.35 kg/Mg of copperc
(8.3-8.7 Ib/ton of copper)
1.65-2.3 kg/Mg of ironf'9'1
(3.3-4.5 Ib/ton of iron)
Emission
factor
reliability
rating
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Iron castings
13,490
(14,840)
Iron castings
13,490
(14,840)
Uncontrolled
emissions
Mg/yr
(tons/yr)
26
<29)
-------
Table 2-39 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR FOUNDRIES
i
-P-
Source of IPFPE
10. Pouring molten metal into
molds
11. Casting shakeout
12. Cooling and cleaning castings
13. Finishing castings
14. Core sand and core binder
receiving and storage
15. Core sand and binder mixing
16. Core making
Uncontrolled fugitive emission factor
0.05-2.07 kg/Mg in gray iron
foundry iffc'f
(0.1-4.13 Ib/ton in gray iron foundry)
1.26 kg/Mg for copper1
(2.52 Ib/ton for copper)
0.47 kg/Mg for lead!
(0.93 Ib/ton for lead)
0.6-6.4 kg/Mg of ironf>™
(1.2-12.8 Ib/ton of iron)
cooling 0.08-0.4 kg/Mg iron
castingsf 'm
(0.16-O.B Ib/ton iron castings)
0.005 kg/Mg iron castings'"
(0.01 Ib/ton iron castings)
Sand unloading 0.015 kg/Mg sand"
(0.03 Ib/ton sand)
0.15 kg/Mg of sande
(0.3 Ib/ton of sand) .
0.38-4.12 kg/Mg of iron '
(0.75-8.24 Ib/ton of iron)
0.18 kg/Mg of cores6
(0.35 Ib/ton of cores)
Emission
factor
reliability
rating
D
E
E
E
D
E
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Iron castings
13,490
(14,840)
Iron castings
13,490
(14,840)
Iron castings
13,490
(14,840)
Iron castings
13,490
(14,840)
Sand
2,600
(2,860)
Iron castings
13,490
(14,840)
Cores0
2,698
(2,968)
Uncontrol led
emissions
Mg/yr
(tons/yr)
16
(17)
47
(52)
3
(4)
1
(1)
1
(1)
30
(33)
1
(1)
-------
Table 2-39 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR FOUNDRIES
NJ
I
Source of IPFPE
17. Core baking
IB. Mold sand and binder
receiving and storage
19. Sand preparation
20. Mold makeup
Uncontrolled fugitive emission factor
0.015-2.7 kg/Mg of coresm'p
(0.03-5.4 Ib/ton of cores)
Sand unloading 0.015 kg/Mg aandn
(0.03 Ib/ton sandT
0.67 kg/Mg iron castings
(1.3 Ib/ton iron castings)
0.02 kg/Mg iron castings'"
(0.04 Ib/tbn iron castings)
Emission
factor
reliability
rating
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Cores0
2,698
(2,968)
Sand
2,600
(2,860)
Iron castings
13,490
(14,840)
Iron castings
13,490
(14,840)
Uncontrolled
emissions
Mg/yr
(tons/yr)
4
(4)
1
(1)
9
(10)
1
(1)
Coke unloading emission factor baaed on the coal unloading emission factor presented in Section 2.1.2.
For complete development of the storage emission factor refer to Section 2.1.4. For source Ib it was assumed that
S » 1.0, D - 90, K = 0.85, K
0.40, and
0.93. Reference 5.
J
Reference 4. (Approximately 5% of uncontrolled emissions.)
Engineering judgment, assume !% of uncontrolled emissions as reported in Reference 4, p. 7.8-1 and 7.9-2.
Reference 6.
Reference 7.
Reference 8.
Reference 9.
Limited test data from Reference 10, indicate that emissions from electric induction furnace range from 0.15 to 0.3 kg/Mg
(0.3 to 0.6 Ib/ton) iron poured. These emissions include melting, pouring and innoculation.
Engineering judgment, assume 50% of uncontrolled emissions as reported .in Reference 4, p. 7.11-2.
* Reference 11.
Reference 13.
ra Reference 12.
n Sand unloading emission factor assumed to be of the same magnitude as the taconite pellets unloading emission factor as
presented in Section 2.1.2. Emissions from storage is estimated to be negligible since the sand is normally stored
indoors.
0 Assume that 20% of the weight of castings equals the weight of cores.
p Engineering judgment, assume all uncontrolled emissions as reported in Reference 14, are fugitive.
-------
Table 2-43. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR MATERIAL EXTRACTION AND BENEFICIATION
Source of IPFPE
1. Overburden removal
2. Drilling and blasting
3. Ore loading
4. Haul road truck
transport
5. Truck dumping
6. Primary crushing
7 . Transfer and Conveying
8. Secondary crushing/
Screening
9. Waste disposal
10. Storage
11. Land reclamation
Range of uncontrolled fugitive emission factors
0.0004a-0.225b Kg/Mg (0.0008-0.45 Ib/ton)
of ore
0.024c-0.05d Kg/Mg (0.048-0.10 Ib/ton) of
overburden
0.0005e-0.08f Kg/Mg (0.001-0.16 Ib/ton)
negligibleg-0.05 Kg/Mgh (0.10 Ib/ton)
of ore
0.233-0.62k Kg per vehicle-kilometer traveled
(0.8-2.2 Ib per vehicle mile traveled)
0.000171-0.02m Kg/Mg (0.00034-0.04 Ib/ton)
of ore
negligible-0.25 Kg/Mg° (0.5 Ib/ton)
of ore
negligible-0.75g Kg/Mg (1.5 Ib/ton)
of ore
0.022S-0.75t Kg/Mg (0.044-1.5 Ib/ton)
of ore
negligible-3.23v Mg/1000m2/yr
(14.4 ton/acre/yr)
0.0118W-0.2X Kg/Mg (0.0235-0.42 Ib/ton)
0.392X-1.48X Kg/1000m2/day (3.5-13.2 Ib/acre/da
use wind erosion equation
Emission
factor
reliability
rating
E
E
E
E-C1
depends
E
E
E
E
E
E,Dy
)
Emission factors by
Indus try^Kg/Mg (Ib/ton)
Coal
0.025b
(0.45)
0.024°
(0.048)
(0.05)h
(0.10)
on spee
0.01n
(0.02)
-------
Table 2-43 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR MATERIAL EXTRACTION AND BENEFICIATION
Estimate for open pit copper mining from comparison with emission rate for an active construction area.
Reference 1. pp. 25. Estimate for lignite mining from comparison with emission rates from similar fugitive dust
sources, such as construction and aggregate handling.
Reference 6. pp. 25. Estimate
aggregate handling and storage.
q
Reference 6. pp. 25. Estimate for western surface coal mining from comparison of published emission factor for
Reference 7. Estimate of total fugitive emissions at western surface coal mine, with overburden removal the predominant
contributor.
3. Estimate based on visual observation at open pit copper mine.
f Reference 8. Estimate for granite drilling of 0.0004 Kg/Mq (0.0008 Ib/ton) and 0.08 Kg/MG (0.16 Ib/ton) for granite blasting.
" Reference 8. Estimate for crushed granite plant.
Reference 1. pp. 35.
1 Reference 9. From sampling of crushed rock loading by front-end loaders.
•• Reference 10. Based on an uncontrolled "dry" rate of 1.04 Kg/VKT (3.7 Ib/VMT) and application of a climate correction of
0.44 (the fraction of days when the surface was not wet or frozen) and a control factor of 0.50 to account for watering of
the roads on dry days.
i* 11
Reference 1. Based on EPA's published emission factor for unpaved roads , 166 days per year with no rain or snow cover,
I and 50% control by watering during dry days.
•£"" 1
-P- Reference 8. For dumping granite into a primary crusher.
I m Reference 11. For dumping of crushed rock onto storage piles.
n Reference 1. Estimated by reducing the EPA published emission factor for unloading crushed rock to account for the larger
size of the coal and ore being handled and its higher moisutre content.
12. Includes both stack and fugitive emissions, 80% of which falls out on plant property.
p Reference 13.
^ References 1 and 14. Based on reported industry estimates of 0.075 Kg/MG (0.15 Ib/ton) with 90% control.
r Reference 15. Proportioned from a total fugitive emission factor of 0.22 Kg/Mg (0.44 Ib/ton) for western surface coal mines.
8. Includes both stack emissions and fugitive emission at a granite quarry.
Reference 12. Includes stack and fugitive emissions, 60% of which falls out on plant property.
u Reference 13.
v Estimates for dried tailings based on U.S. Dept. of Agriculture's wind erosion equation (see Section 2.1) and are a function
of regional climatic conditions, assuming no surface crusting.
Reference 16. For coal mines in the Southwest based on the average wind erosion rate.
x Reference 17.
" Reference 14. Based on sampling tests on coal storage piles; does not include loading and unloading emissions.
z Reference 14.
aa For further detail refer to Section 2.1.4.
-------
Table 2-46. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR COUNTRY AND TERMINAL ELEVATORS
Oi
i
Source of IPFPE
I. Terminal Elevators
1. Receiving
Truck unloading
Railcar unloading
Barge unloading
2. Transferring and conveying (total)
2a. Receiving elevator leg and
elevator head
2b. Garner and scale vents
2c. Distributor, trippers, and
spouting
2d. storage bin vents
2e. Turning
3. Screening and cleaning
4. Drying
Uncontrolled fugitive emission factor
0.16-1.75 kg/Mg unloadeda'b'c/a'e
(0.32-3.5 Ib/ton)
0.02-1.5 kg/Mg unloadeda'b/c'd>e
(0.04-3.0 Ib/ton)
0.04-1.75 kg/Mg unloadeda'b'c'd'e
(0.08-3.5 Ib/ton)
0.5-1.25 kg/Mg transferreda>b'd/e
(1.0-2.5 Ib/ton)
0.25 kg/Mg transferred6 ' f
(0.5 Ib/ton)
f
0.25 kg/Mg transferred6
(0.5 Ib/ton)
g
g
Emission
factor
reliability
rating
D
D
D
D
E
E
0.095-4.6 kg/Mg screened & cleaneda'b|d'e D
(0.19-9.2 Ib/ton)
0.095-4.0 kg/Mq drieda
-------
Table 2-46 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR COUNTRY AND TERMINAL ELEVATORS
Source of IPFPE
5. Shipping
Truck loading
Railcar loading
Barge or ship loading
II. Country Elevators
1. Receiving
Truck unloading
Railcar unloading
Barge unloading
2. Transferring and conveying (total)
which includes following:
2a. Receiving elevator leg and head
2b. Garner and scale vents
Uncontrolled fugitive emission factor
0.07-1.75 kg/Mg loadeda>b'°'d
(0.14-3.5 Ib/ton)
0.007-1.5 kg/Mg loadeda'b'C'd'e
(0.015-3.0 Ib/ton)
0.001-1.75 kg/Mg loaded3'*5' C(d/e
(0.002-3.5 Ib/ton)
0.16-4.0 kg/Mg unloadedb/c'd
(0.32-8.0 Ib/ton)
0.02-4.0 kg/Mg unloaded*3' °'d
(0.04-8.0 Ib/ton)
2.5-4.0 kg/Mg unloaded5 'd
(5.0-8.0 Ib/ton)
1.0-2.0 kg/Mg transferredb'd'h> i
(2.0-4.0 Ib/ton)
j
1.0 kg/Mg transferred '
(2.0 Ib/ton)
Emission
factor
reliability
rating
D
D
D
D
D
E
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
0
(0)
Grain shipped by
railcar
96,314
(106,208)
Grain shipped by
barge or ship
16,996
(18,742)
Grain received by
truck
16,635
(18,322)
0
(0)
0
(0)
Transferred &
conveyed
16,635
(18,332)
Uncontrolled^
emissions
Mg/yr
(tons/yr)
0
(0)
73
(BO)
15
(16)
35
(38)
0
(0)
0
(0)
25
(28)
-------
Table 2-46 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR COUNTRY AND TERMINAL ELEVATORS
Source of IPFPE
2c. Distributor, trippers and
spouting
2d. Storage bin vents
2e. Turning
1 3. Screening and cleaning
.p-
-j
1
4. Drying
5. Shipping
Truck loading
Railcar loading
Barge loading
Uncontrolled fugitive emission factor
1
1.0 kg/Mg stored11
(2.0 Ib/ton)
m
3.5-5.0 kg/Mg screened & cleaned ' rl
(7.0-10.0 Ib/ton)
0.095-4.0 kg/Mg drieda'b'c'd/e
(0.19-8.0 Ib/ton)
h r> H h
0.07-4.0 kg/Mg loaded '''
(0.14-8.0 Ib/ton)
0.007-4.0 kg/Mg loadedb'C'd/h
(0.015-8.0 Ib/ton)
0.001-4.0 kg/Mg loadedb'C 'd'h
(0.002-8.0 Ib/ton)
Emission
factor
reliability
rating
E
E
D
D
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Screened & cleaned
4,159
(4,575)
Dried
3,B37
(4,232)
Grain shipped
by truck
8,318
(9,161)
Grain shipped
by railcar
8,318
(9,161)
0
(0)
Uncontrolled
emissions
Mg/yr
(tons/yr)
m
20
(22)
a
(9)
17
(19)
17
(18)
0
(0)
-------
-P-
00
Table 2-46 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR COUNTRY AND TERMINAL ELEVATORS
Reference 5.
Reference 8.
Emissions for garner and scale included in receiving elevator emissions.
' Emissions for storage bin vents and turning included in total.
Reference 6.
1 Reference 9.
' Emissions for receiving elevator (country elevator) included in total.
This value is for scale vents only.
Emissions for distributor, trippers, and spouting included in total.
m This operation is not normally done in country elevators.
-------
Table 2-53. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR LIME PRODUCTION
i
-P-
-P-
Source of IPFPE
1. Limestone/dolomite charging
to primary crusher
2. Primary crushing
3. Transfer points and associated
conveying
4. Primary screening
5. Secondary crushing
6. Secondary screening
Uncontrolled fugitive emission factor
0.00015-0.02 kg/Mg of rock charged
(0.00030-0.04 Ib/ton)
0.25 kg/Mg of limestone crushed
(0.5 Ib/ton)
0.4 kg/Mg of quicklime produced
(0.8 Ib/ton)
0.75 kg/Mg of limestone 'crushed
(1.5 Ib/ton)
Emission
factor
reliability
rating
Model plant
fugitive emission inventory
Operating parameter
Mg/yr
(tons/year)
Limestone/dolomite
processed
100,000
(110,000)
Limestone/dolomite
processed
100,000
(110,000)
Quicklime produced
50,000
(55,000)
Limestone/dolomi te
processed
90,000
(99,000)
Uncontrolled
emissions
Mq/yr
(tons/yr)
1
(1)
25
(28)
20
(22)
68
(74)
-------
Table 2-53 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR LIME PRODUCTION
o
Source of IPFPE
7. Crushed limestone storage
Loading onto pile
Vehicular traffic
Loading out
Hind erosion
8. Quicklime screening
Uncontrolled fugitive emission factor
(0.02) (Kl) (S/1.5) kg/t]g matorial
(PE/100)2 loaded onto pile5
/ \
[(0.04) (Ki) (S/1.5) , , \
\ (PE/100) I
(0.065) (K2) (S/1.5) kg/ material
(PE/100)2 stored^
((0.13) (S/1.5) lb ,to j
(PE/100)2 /
(0.055) (S/1.5) (D natorial
(PE/100) storedf
/(O.ll) S/1.5) (D . , \
y (PE/100) ^ "° /
g
Emission
factor
reliability
rating
D
D
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Crushed limestone loade
40,000
(44,000)
Crushed limestone store
40,000
(44,000)
Crushed limestone loade
out
40,000
(44,000)
Crushed limestone store
40,000
(44,000)
Uncontrol led
emissions
Mg/yr
(tons/yr)
3
1
(1)
i
3
(3)
J
1
(1)
1
2
(2)
g
-------
Table 2-53 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR LIME PRODUCTION
I
Ui
M
Source of IPFPE
9. Quicklime and hydrated lime
crushing and pulverizing
with leaks from mill and
from feed/discharge exhaust
systems.
10. Lime product silo vents
11. Truck, rail, ship/barge
loading of quicklime and
hydrated lime
12. Packaging quicklime and
hydrated lime
Uncontrolled fugitive emission factor
0.05 kg/Mg of quicklime and hydrated
lime produced^
(0.1 Ib/ton)
0.118 kg/Mg of lime products loaded"1
(0.236 Ib/ton
Negligible - 0.005 kg/Mg of lime
products packaged*
(0.01 Ib/ton)
Emission
factor
reliability
rating
Model plant
fugitive emission inventory
Operating parameter
Mg/yr.
(tons/year)
Quicklime and hydrated
lime crushed and
pulverized
50,000
(55,000)
Lime products loaded
50,000
(55,000)
Lime products packaged
10,000
(11,000)
Uncontrolled
emissions
Mg/yr.
(tons/yr)
2
(3)
e
(6)
Negligible
Reference 1.
Reference 2 - AP-42. 80% and 60% of which falls out on plant property for points 2 and 5', respectively.
Engineering judgment, assumed approximately same as emission factor for dry phosphate rock as reported in Reference 3.
Emission from primary screening (point 4) included in emission factor for primary crushing (point 2) .
Emissions from secondary screening (point 6) included in emission factor for secondary crushing (point 5).
For complete development of this factor, refer to Section 2.1.4. For this example, it was assumed that S = 1 5
= =
90, PE = 100, and
and
1. Reference 4.
* Emission from quicklime screening (point 8) included in emission factor for quicklime crushing and pulverising (point 9).
^ "ngineering judgment based on controlled cement milling emissions reported by a cement manufacturing company.
1 Emissions from lime product silo vents (point 10) included in emission factor for lime loading (point 11).
j
Engineering judgment, assumed same as for loading of hydraulic cement obtained from Reference 5.
Engineering judgment based on observations and emission tests of controlled similar sources.
-------
Table 2-55. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR CONCRETE BATCHING
Source of IPFPE
1. Sand aggregate storage
Loading onto pile
Vehicular traffic
Loading out
Hind erosion
Uncontrolled
(0.02) (Ki)
fugitive emission factor
S/l . 5 )
(PE/100) loaded onto pile
/(0.04) (Ki) (S/1.5) \
\ (PE/100)
(0.065)
-------
Table 2-55 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL
FUGITIVE PARTICULATE EMISSION POINTS FOR CONCRETE BATCHING
i
•e-
01
Source of IPFPE
2. Transfer of sand and
aggregate to elevated bins
3. Cement unloading to elevated
storage silos
4. Weigh hopper loading of
cement, sand, and aggregate
5. Mixer loading of cement,
sand and aggregate (central
mix plant)
6. Loading of transit mix
(wet-batching) truck
7. Loading of dry-batch truck
Uncontrolled fugitive emission factor
0.02 kg/Mg of sand and aggregate
transferred0
(0.04 Ib/ton of sand and aggregate
transferred)
0.118 kg/Mg of cement unloaded*3
(0.236 Ib/ton of cement unloaded)
0.01 kg/Mg of cement, sand and
aggregate0
(0.02 Ib/ton of cement, sand and
aggregate)
0.02 kg/Mg of cement, sand and
aggregate0
(0.04 Ib/ton of cement, sand and
aggregate)
0.01 kg/Mg of cement, sand and
aggregate0
(0.02 Ib/ton of cement, sand and
aggregate)
0.02 kg/Mg of sand and aggregate0
(0.04 Ib/ton of sand and aggregate)
Emission
factor
reliability
rating
Model plant
fugitive emission inventory
_j-
Operating parameter,
Mg/yr
(tons/year)
Sand and aggregate
transferred
60,000
(66,000)
Cement transferred
to storage silo
7,000
(7,700)
Cement, sand, and
aggregate loaded
67,000
(73,700)
Cement, sand, and
aggregate loaded
67,000
(73,700)
Uncontrolled
emissions
Mg/yr
(tons/yr)
1.2
(1.3)
0.8
(0.9)
0.7
(0.8)
0.7
(0.8)
a AP-42 (Reference 2) reports total plant uncontrolled emission factor of 0.05 kg/Mg (0.1 Ib/ton) of concrete produced.
For complete development of this factor, refer to Section 2.1.4. For their example is was assumed that S = 1.5,
D = 90, PE = 100, and Kj, Kj, and Kj = 1. Reference 3.
Reference 4.
Reference 5. From testing of mechanical unloading to hopper and subsequent transport of cement by enclosed bucket
elevator to elevated bins with a fabric sock over the bin vent.
-------
Table 2-57. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR ASPHALTIC CONCRETE PRODUCTION
•e-
Source of IPFPE
1. Storage of coarse and fine
aggregate
Loading onto pile
Vehicular traffic
Loading out
Hind erosion
Uncontrolled fugitive emission factor
(0.02) (Ki) (S/1.5) , material
(PE/100) loaded onto pile
Afl.04) (Ki) (S/1.5) ib/to\
\ (PE/100) 1
(0.065) (K,) (S/1.5) , material
(PE/100) stored^
/(0.13) (K2) (S/1.5) ,b/tjl
\ (PE/100) 1
(0.025) (K3) (S/1.5) , material
(PE/100) loaded out3
/(0.05) (K,) (S/1.5) lb/ton\
\ (PE/100) J
(0.055) (S/1.5 .D nuLai-i.il
(PE/100) 3U stored^
/(O.ll) (S/1.5) (D lu/un\
\ (pE/ioo)2 9° y
Emission
factor
reliability
rating
D
D
D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Aggregate loaded
277,100
(300,000)
Aggregate stored
272,100
(300,000)
Aggregate loaded
out
272,100
(300,000)
Aggregate stored
272,100
(300,000)
Uncontrolled
emissions
Mg/yr
(tons/yr)
5
(6)
18
(20)
7
(8)
15
(16)
-------
Table 2-57 (continued). IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR ASPHALTIC CONCRETE PRODUCTION
01
Ui
I
Source of IPFPE
2.
3.
4.
5.
6.
Unloading coarse and fine
aggregate to storage bins
Cold aggregate elevator
Dried aggregate elevator
Screening hot aggregate
Hot aggregate elevator
(continuous mix plant)
Uncontrolled fugitive emission factor
Negligible - 0.05 kg/Mg of aggregate
(0.10 Ib/ton)
Negligible - 0.1 kg/Mg of aggregate*5''
(0.2 Ib/ton)
c
Negligible - 0.013 kg/Mg of aggregate'
(0.026 Ib/ton)
c
Emission
factor
reliability
rating
D
' D
D
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Aggregate
processed
272,100
(300,000)
Aggregate
processed
272,100
(300,000)
Aggregate
processed
272,100
(300,000)
Uncontrolled
emissions
Mg/yr
(tons/yr)
7
(8)
14e
(15)
c
2e
(2)
c
For complete development of this factor, refer to Section 2.1.4. For this example it was assumed that S - 1.5, D = 90,
PE = 100, and Kj, K2, and K3 = 1. Reference 1.
Reference 2.
c Emissions from points 4 and 6 are included in emissions from point 3.
Reference 3.
e Emissions from these points for many plants are negligible, since these operations are generally enclosed and exhausted
to the primary control system for the agaregate dryer (e.g., scrubber or fabric filter).
-------
Table 2-59. IDENTIFICATION AND QUANTIFICATION OF POTENTIAL FUGITIVE
PARTICULATE EMISSION POINTS FOR THE LUMBER AND FURNITURE INDUSTRY
i
•e-
Ln
Source of IPFPE
Sawmill
1. tog debarking
2. Sawing
3. Sawdust pile loading,
unloading, and storage
Furniture Manufacturing
4. Wood waste storage bin vent
5. Wood waste storage bin loadout
Uncontrolled fugitive emission factor
0.012 kg/Mg of logs debarked3
(0.024 Ib/ton of logs debarked)
0.18 kg/Mg of logs sawed3
(0.35 Ib/ton of logs sawed)
0.5 kg/Mg sawdust handled
(1.0 Ib/ton sawdust handled)
0.5 kg/Mg wood waste stored
(1.0 Ib/ton wood waste stored)
1.0 kg/Mg wood waste loaded out
(2.0 Ib/ton wood waste loaded out)
Emission
factor
reliability
rating
E
E
E
E
E
Model plant
fugitive emission inventory
Operating parameter,
Mg/yr
(tons/year)
Log£ debarked
740,000
(820,000)
Logs sawed
650,000
(720,000)
Sawdust handled
100,000
(110,000)
Wood waste stored
1,360
(1,500)
Wood waste loaded o
1,360
(1,500)
Uncontrolled
emissions
Mg/yr
(tons/yr)
9
(10)
117
(126)
50
(55)
1
(1)
It
1
(2)
3 Estimate based on material balance of the waste produced by the specific operation and
amount which becomes airborne.
Engineering judgement based on observations on plant visits. It is recognized that in
of a severe problem.
engineering judgement of the
some plants this may be more
-------
AIR EMISSIONS IN IRON ORE MINING AND ENRICHMENT
By
6. V. Jorgenson, J. P. Pilney and E. E. Erickson
Midwest Research Institute
North Star Division
10701 Red Circle Drive
Minnetonka, Minnesota 55343
-457-
-------
ABSTRACT
The iron ore industry has experienced radical change in the past
25 years and is growing rapidly today. There are a number of air emissions
from this industry. The mining activity and the large areas of barren
mines and dry tailings basins are the sources of fugitive emissions that
are very difficult and costly to control. Based on the data available from
tests at various sites, it is estimated that concentrating and pelletizing
cause on the order of 100,000 tons of particulate emissions in the U.S.
each year. Many of the sources contributing to these emissions already are
controlled with devices that average nearly 99 percent efficiency. Some
preliminary determinations of the characteristics of the particulate
emissions are presented.
-458-
-------
AIR EMISSIONS IN IRON ORE MINING AND ENRICHMENT*
G. V. Jorgenson, J. P. Pilney and E. E. Erickson
Midwest Research Institute
North Star Division
Minnetonka, Minnesota
INTRODUCTION
The iron ore industry has experienced a radical change in the past 25
years. Although iron ore beneficiation has been going on for more than
twice that long through the production of gravity concentrates, many natural
ores were still shipped in the 1950's. Most of the higher grade ores have
been mined and the vast reserves of lower grade ores (taconite) are now
being used. These ores are concentrated (usually about 2.5 or 3 to 1) and
shipped to the steel mills, mostly in the form of pellets. The importance
of beneficiation can be illustrated by looking at Minnesota's 1973 shipments
(more than 2/3 of U.S. production). That production consisted of 3 percent
direct shipping ores, 27 percent gravity concentrates and 70 percent
taconite concentrate pellets. Whereas the U.S. production of pellets in
1975 was approximately 60 million tons, it is expected to reach 90 million
••
tons in 1978 as several major expansion projects come on line.
Midwest Research Institute is conducting "A Study of Emissions from
Iron Ore Mining, Beneficiation, and Pelletizing" for the EPA's Industrial
Environmental Research Laboratory. The emissions sources associated with
"benef iciation and pellet production are confined sources that lend them-
selves in standard control technology. Most of the emissions from mining
*This work was supported under EPA Contract No. 68-02-2113.
-459-
-------
and from tailings disposal are fugitive emissions and, as such, do not lend
themselves to control other than through changing mining or handling
procedures.
The current program has not been completed; in fact, field testing is
being completed the end of this month. Therefore, we are unable to report
results of analyses of the sampling tests. Instead, this paper will give a
description of the current U.S. iron ore industry as it relates to the
air emissions. That description will include the mining, beneficiating and
pelletizing processes; the emissions sources from those processes; the
controls being used and something about the effectiveness of those controls.
That will leave many unanswered questions in the minds of fugitive emissions
experts. We hope to know more about these emissions when our current
program has been completed. We will also discuss the constituents that
could be expected to show up in the particulate. emissions from this industry.
Our current program should contribute to the overall knowledge in that area
also. The rapid expansion of the industry will probably bring about techno-
logical changes in the way all emissions are controlled, so this industry
will certainly remain of interest to those involved in the air pollution
business for some time.
-460-
-------
IRON MINING
Today's mining of iron ore is done almost entirely in open pit mines.
After the overburden has been removed, the ore is removed along different
levels, called benches. The operations that take place are drilling,
blasting, loading and hauling.
Blastholes of 12 to 18 inches in diameter are drilled to a depth that
is typically 35-50 feet deep, about 5 feet more than the bench height. The
most common drills are jet piercing drills and rotary drills. The burners
in the jet piercing drills are water cooled and the steam that is generated
when the water contacts the hot taconite expands and flushes the spalled
chips from the hole. The rotary drills use water flushing for dust
suppression and cooling. Consequently, the drilling operations do not
release many particulates.
The blastholes are loaded with explosives that are detonated to break
loose up to nearly 2 million long tons of ore at one time. This process
releases particulates, but it is not an important emission source. Most of
the material thrown into the air is so large.that it settles to the surface
rapidly. Also, blasting is intermittent, occurring only once every several
days, approximately once a week on the average. Meteorological conditions
are monitored to find the best times for blasting with minimum noise and
vibration. The pollutants from this source cannot be controlled, in the
sense of entrapment.
The loading of ore is a process that contributes significantly to
airborne pollutants. Large electric shovels (often 12 to 14 cubic yards)
pick up the broken ore and dump it into trucks or rail cars. In most mines
this is taking place at several sites, 24 hours per day. We feel that this
is the largest source of airborne particulates resulting from ongoing mining
-461-
-------
activity. For that reason we are setting up a fugitive emissions sampling
station at a shovel loading site as part of our current project; it will be
a modified ambient sampling array such as that shown in Figure 1. Glass
fiber filter media will be used to collect samples for mass loading measure-
ments , for trace metals analysis (As, Cu and Ni), and for particulate
sizing (Anderson). Membrane media will be used for collecting the samples
for asbestos analysis. Although it would not be possible to control the
dust that is raised as a result of filling the shovel, it is conceivable
that the dust resulting from dumping the shovel load into a truck or train
might be controlled.
Hauling the ore to the location of the primary crusher is a source of
fugitive emissions, especially when trucks are used. When temperatures
are above freezing, the mine roads are normally wetted down by water.
sprinkler trucks. This helps but does not eliminate the dust; the way
roads are continuously being built and abandoned control of road dust would
be extremely difficult. Finally, there is the dust that is raised on windy
days from the barren surface of the mine (typically 1 to 10 square miles).
This is another source that does not appear possible to eliminate or to
control except through early revegetation of areas that have been mined
out.
In summary, the amounts of particulate emissions in an open pit iron
ore mine are not known and are generally difficult to reduce. We feel that
a network of fugitive emissions sampling stations up- and down-wind from a
mine would be beneficial in getting some preliminary data on the quantity
and characteristics of pollutants that are carried outside of the mine.
-462-
-------
•30'
GJ
I
Top of Truck
Planked to
Give Work
Area
Hand Rail
Box Type Bed
2" Pipe Frame
Attached to
Truck Frame
Sampling Points for Mass Participate
(•) Sampling Points for Particulate Sizing
Figure ],. Sampling Station for Fugitive Emissions
-------
ORE BENEFICIATION
Approximately 40 percent of the iron ore that is mined in the U.S.
becomes shippable product for iron making. Thus, the beneficiation process
has to liberate 900 pounds of iron mineral grains from each long ton of
crude ore and dispose of 1340 pounds of gangue. The basic steps to
beneficiation are size separation, size reduction, size and density classi-
fication, minerals separation and dewatering. There are many variations
in the way these basic steps are arranged in different plants but the
process is basically the same. A typical beneficiation process flowsheet
with magnetic separation is shown in Figure 2. In a relatively small
percentage of U.S. ores iron is also separated by a flotation process.
The low grade iron ores being used today are very fine grained, so
beneficiation begins by size reduction. The first stage is a reduction to
approximately minus 6 inches in gyratory crushers. The size is further
reduced to 1/4 inch to 3/4 inch in cone crushers. The crushing is generally
dry with intermediate screening for size separation and recycle of the
large particles through the crushers again.
Pollutants from crushing operations are controlled by a variety of
scrubbers and by cyclones, rotoclones, and multiclones. Actual controlled
emissions are usually in the range 0.012 to 0.12 gr/SCF (grans/Standard Cubic
Foot), with 0.03 gr/SCF being typical. The gas flow rates are approximately
15 SCFM (Standard Cubic Feet per Minute) per LTPH (Long Tons Per Hour)
process weight rate, or 900 SCF per long ton of crushed ore.
There are essentially no data available on baghouse control of
crushing operations, and very few such systems exist. We are testing one of
those operations as part of our current program. We will use sampling method
EPA-5 (front one-half only) for collecting material for mass loading,
-464-
-------
CRUDE ORE
DUMP POCKET
PAN
FEEDERS
CRUSHER
STORAGE
& TRANSFER
CRUSHER
TRANSFER
i
•e-
Ln
I
PRIMARY
MILL
CLASSIFIER
TRANSFER
TAILINGS
THICKENER
HYDRO-
SEPARATOR ,
MAGNETIC
SEPARATOR
MAGNETIC
SEPARATOR
SECONDARY
MILL
TRANSFER
CLASSIFIER
TO TAILINGS BASIN
TO PELLET PLANT
Emission Points
Figure 2. Ore Beneficiation Flowchart
-------
asbestos, silica and trace metals (As, Cd, Co, Cr, Mo, Ni, Pb, V, and Zn)
analysis. Anderson or Brinks methods will be used for particle sizing.
After crushing, the fines are separated into iron bearing particles and
gangue, usually by magnetic separators, and the coarse material moves on to
grinding. Nearly all grinding is done wet, so nationally there is very
little air pollution from that process. However, the sizing by screens
and the ore handling and transfer while the ore is in the dry state
contributes significantly to emissions. The same types of control devices
are also in the same general range, being approximately 0.04 gr/SCF.
The outputs from the beneficiation process are the gangue that goes
to a tailings basin and the mineral concentrate that goes to the pelletizing
plant. Tailings basins (frequently many square miles in area) often
have large areas that dry out, especially along the beaches, and these can
become large sources of fugitive emissions on windy days. Much of the
solid material in these basins has been reduced to a very small size in the
grinding process, so it can become airborne quite easily. There is also
truck traffic on unpaved roads at the disposal site. Fugitive emissions
measurements have been conducted by MRI at one taconite plant tailings
basin. Modeling of the emissions from such a basin has been based on
that sampling and a climatological dispersion model. That modeling gives
a result of 25 tons per day as the expected fugitive emissions from a
large tailings disposal basin.
'Draft Environmental Impact Statement Technical Appendix; Reserve Mining
Company's Proposed On Land Tailings Disposal Plan; Appendix E, Midwest
Research Institute, "Air Quality" pp E177-E194; Prepared for State of
Minnesota Department of Natural Resources and Pollution Control Aeencv
October 1975.
-466-
-------
PELLETIZING
The pelletizing process consists of mixing the ore concentrate with a
binder (bentonite), making green pellets in a balling drum, indurating the
pellets, and cooling them. They are then stockpiled for shipping to the
steel mills. A typical pelletizing flowsheet is shown in Figure 3.
The concentrate has been dewatered before entering the pelletizing
plant, but is still wet enough to prevent dusting as it is handled and
transported to where the bentonite is added. The powdered bentonite is
added at a rate of 10 to 20 pounds per ton of concentrate. The bentonite
can become airborne when it is loaded into bins, unloaded, blended with the
concentrate, or transferred.
Bentonite emissions are usually controlled by scrubbers or bag
filters, and contribute very little to the total emissions.
The ore concentrate with bentonite added is too wet to cause dusting
as it moves to and through the balling drums and into the induration furnace
or kiln. Most of the large plants and the new plants use grate-kiln
systems. As the pellets tumble through the kiln, they begin to dust.
There is a large volume of counter-flow gas moving through the kiln (main
waste gas stream) , and this becomes heavily loaded with particulates.
Control of the waste gas streams at the different plants ranges from no
control to the use of a wide variety of control devices. Typical emissions
from controlled waste gas streams are 50 to 200 Ib/hr and an uncontrolled
stream will emit 4,000-5,000 Ib/hr. A large percentage of the material
from an uncontrolled source has a particle size so large that it settles
to the surface very close to the plant; a somewhat smaller fraction of
the emissions from the controlled waste gas streams settles out rapidly.
-467-
-------
CONCENTRATE
THICKENER
DISC
FILTERS
TRANSFER
CONCENTRATE
STORAGE
BENTONITE
STORAGE
oo
BLENDING
BALLING
DRUMS
FEEDERS
undersize
CHIP
REGRIND
TRAVELING
GRATE
ROTARY
KILN
COOLER
TRANSFER
PELLET DUMP
POCKET
Emission Points
Figure 3. Pelletizer Flowchart
-------
As the fired pellets emerge from a kiln they must be cooled before
they are stockpiled. They commonly pass through an annular cooler with
counter flow gas. There, the tumbling action abrades the pellets and
causes some emissions. Most annular coolers do not have emission control
devices and emit from 10 to 100 pounds per hour. The available information
on annular cooler emissions is quite limited, so we are testing one as a
part of our current program. Because many of the particles are being
released by the autogenous grinding of the pellets as they move through the
cooler, we expect that the average particle size might be quite small. If
the particles are small they are quite significant, even if the tonnate is
low, because they can remain airborne over greater distances.
Plants that use vertical shaft furnaces or straight furnaces rather
than rotary kilns for induration could be expected to produce fewer
emissions because there is not the tumbling action of a rotary kiln. That
also applies for grate type cooling compared to the annular coolers.
Pellets emerging from the cooler are transported to the pellet
loadout point. The emissions from loadout are usually controlled by equip-
ment such as rotoclones and contribute very little to the overall emissions
from a plant.
In addition to the emissions from the fixed sources, many plants have
significant fugitive emissions within the buildings. Large fans are used
for sweeping these to the outside or into a control device.
How much pollution does this add up to? In Table 1 we have shown
the ranges of emissions for ten sources, estimated to be the ten most
significant. The values are the low, median, and high value for all the
plants for which data were available (typically 3 to 8 plants) and the
units are in pounds per long ton of product. There are enough uncontrolled
-469-
-------
Table 1. Ranges of Emissions (Pounds/Long Ton of Product)
Source
Dump Pocket
Crushers
Ore Transfer
Fine Grinding/Milling
Bentonite Transfer
Bentonite Blending
Grate Feed
Grate Discharge
Waste Gas
Pellet Handling
CONTROLLED
Low
0.038
0.004
0.0006
0.360
0.001
0.002
0.007
0.002
0.584
0.0004
Median
0.147
0.013
0.003
0.387
0.0004
0.002
0.009
0.011
0.98
0.018
High
0.223
0.063
0.028
0.414
0.001
0.002
0.015
0.015
1.83
0.037
UNCONTROLLED
Low
0.252
0.143
0.017
72
0.011
0.159
0.035
0.088
16.68
0.021
Median
0.980
0.227
0.095
88
0.044
0.239
0.697
1.46
32.2
3.75
High
1.484
0.627
1.334
103
0.128
0.305
1.47
1.53
108
5.23
-470-
-------
sources and enough minor sources in the industry that we can assume the
median value, controlled sources, to be a minimum figure. If we obtain
the sum of median values and assume that value is applicable to U.S.
production, we obtain an estimate of 57,000 tons per year. We also esti-
mated the total U.S. emissions from pellet plants by adding the known
emissions from four plants that account for 25 percent of U.S. production
and multiplying the sum by four. These plants use technology that is
typical of that used for most of the U.S. production. Using that method
gives a sum of 112,000 tons of emissions per year. Most of the figures
used for these calculations are from 1974 and there have been some additions
of controls in old plants that would reduce those numbers. However, some
new plant capacity has also been added so we feel it is safe to conclude
that the emissions from the beneficiation and pelletizing of iron ore
is of the order of 100,000 TPY.
It is also interesting to look at the difference between the
median columns for controlled and for uncontrolled emissions in Table 1.
The sum of the median values for uncontrolled sources is 80 times
the sum for median level controlled sources. This shows that the controls
that are now being used on those ten sources have a control efficiency of
nearly 99 percent.
-471-
-------
POLLUTANT CHARACTERISTICS
What are the pollutants from this industry? As stated earlier, the
analyses of the samples from tests being conducted now have not been completed.
However, there are some clues concerning what the pollutants might contain.
In Table 2 is shown a typical chemical analysis of the tailings from a
beneficiation plant, and in Table 3 is shown chemical analyses of five
samples of igneous rocks from the iron range in Minnesota.
Arsenic compounds pose major problems in some foreign ores, but are
rare in Minnesota ores. We could find only one indication that they do
(2)
exist in Minnesota ores , and found no mention of arsenic with respect
to other domestic ores. A sample of material was picked up from the
floor of a baghouse hopper on a crushing operation outside of Minnesota
and chemically analyzed. The results are shown in Table 4. Note an
arsenic content of 5 parts per million in that sample. Because of the
indication that arsenic compounds may be present, we are hoping to deter-
mine how much is in these ores.
In summary, iron ore mining, beneficiating and pelletizing is a
massive industry that will soon mine approximately 350 million long tons
of iron ore per year to produce 90 million long tons of pellets and nearly
40 million long tons of other concentrates. The extent to which this
industry adds to air pollution outside the property lines of the mining
companies and the characteristics of those pollutants is not well known.
The present study being conducted for the EPA by MRI is a beginning. We
(2)
'Private communication with Dr. Roland Blake, U.S. BuMines, who stated
that Minnesota iron ores occasionally contain a few small crystallites
of Arsenopyrite.
-472-
-------
are hopeful that it will lead to sound recommendations concerning the best
opportunities for significantly reducing the emissions produced by iron
ore mining, beneficiation, and pelletizing processes.
-473-
-------
Table 2; Typical Chemical Analysis*
Element
Iron
Silicon
Aluminum
Calcium
Magnesium
Manganese
Titanium
Phosphorus
Sodium
Potassium
Sulfur
Lead
Zinc
Nickel
Copper
Chromium
Cadmium
Molybdenum
Vanadium
Cobalt
Carbon
Hydrogen
Oxygen
Percent
14.93
33.59
0.35
1.67
1.96
0.37
0.030
0.026
0.040
0.040
0.03
0.007
0.004
0.002
0.004
0.004
0.0003
<0.001
<0.001
0.002
0.11
0.10
46.64
Total 99.81
*Conference: Pollution of Lake Superior and Its Tributary
Basin, Minnesota-Wisconsin-Michigan. Proceeding, Executive
Session, Duluth, Minnesota, May 13,14,15, 1969, U.S.
Department of the Interior, Federal Water Pollution Control
Administration, Vol. 4, p. 1301.
-474-
-------
Table 3. Chemical Analyses of Some Igneous Rocks of Minnesota*
.Constituent
Fe 0
£» J
FeO
Si02
A100,
2 3
CaO
MgO
Ti02
MnO
Cr00_
2 3
V000
2 3
NiO
CoO
P00.
2 5
S
Na2°
K20
H20
co2
Total
6 12
35.14
(80.78)
32.40
2.02 13.81
2.68 4.00
trace 0.04
0.06
12.02 8.25
3.20
2.40 2.05
1.03
—
—
0.03 trace
0
—
—
—
—
100.00" 99.98
14
46.14
21.53
21.65
4.89
0.06
1.20
0
2.60
1.01
0.91
—
—
0.01
0
—
—
—
100.00
16
46.92
27.87
5.69
7.23
1.03
3.70
5.97
0.26
0.03
0.49
0.09
0.07
0.02
0.12
0.28
0.24
1.17
0.26
101.44
17
22.50
35.68
8.86
5.51
0.99
4.23
20.27
0.31
0.19
0.30
—
—
0.02
—
0.62
0.20
0.37
—
100.05
*A.P. Ruotsala and S.P. Tufford, "Chemical Analyses of Igneous Rocks",
Minnesota Geological 'Survey, Information Circular 2, (1965)
**See following page for sample identifications
-475-
-------
Sample Identifications for Table 3.
6. Magnetite ore.
Iron Lake, SE 1/4 sec 36, T. 65N., R. 3W., Cook Co.
[Minnesota Geological Survey Bulletin _6, p. 141 (1891)]
12. Magnetite ore (core at 75 ft depth)
Iron Lake, NE 1/4 NE 1/4 sec 1, T. 64N., R. 3W., Cook Co.
(Ibid., p. 106)
14. Magnetite ore (same core as No. 12, 65 ft depth)
Iron Lake, NE 1/4 NE 1/4 sec 1, T. 64N., R. 3W., Cook Co.
(Ibid, p. 106)
16. Magnetite band
Sec. 4, T. 63N., R. 12W., St. Louis Co.
[Minnesota Geological Survey Bulletin, 21, p. 74, (1926)]
17. Magnetite - Ilmenite
East of Benning Lake, sec 1, T. 64N., R. 2W., Cook Co.
Grout, The titaniferous magnetites of Minnesota,
IRRRC, St. Paul, p. 28 (1949-50)
-476-
-------
Table 4. Sample Analysis From Taconite Ore
in Crushing Operation
SAMPLE TYPES:
BATCH NO: 40
Soil
SAMPLE(S) COLLECTED BY: Client
DATE RECEIVED: 01-07-77
LAB NO:
SAMPLE SITE:
ANALYSIS:
54 54
TACONITE REPLICATE
DUST
#1 #2
Arsenic, mg/kg as As
Cadmium, mg/kg as Cd
Cobalt, mg/kg as Co
Total Chromium, mg/kg as Cr
Copper, mg/kg as Cu
Molybdenum, mg/kg as Mo
Nickel, mg/kg as Ni
Lead, mg/kg as Pb
Vanadium, mg/kg as V
Zinc, mg/kg as Zn
4
0.4
20
15
50
40
50
30
1000
23
6
0.4
20
15
50
40
40
20
1000
23
-477-
-------
ACKNOWLEDGEMENTS
The authors are indebted to Mr. Norman Plaks of EPA's Industrial
Engineering Research Laboratory for his support and to Mr. Bruce Harris
of the same laboratory for his assistance in selecting sources for sampling
and analyses to be conducted.
-478-
-------
DESIGN OF A STUDY TO MEASURE
FUGITIVE EMISSIONS FROM PETROLEUM REFINING
By
Donald D. Rosebrook
Radian Corporation
8500 Shoal Creek Boulevard
Austin, Texas 78766
-479-
-------
DESIGN OF A STUDY TO MEASURE
FUGITIVE EMISSION FROM PETROLEUM REFINING
by
Donald D. Rosebrook
Radian Corporation
8500 Shoal Creek Boulevard
Austin, Texas 78766
ABSTRACT
This paper describes the physical, chemical and
statistical considerations involved in designing a study to
sample and characterize fugitive emissions from petroleum
refining activities. The factors affecting the choice of
refineries and the individual samples are discussed in terms
of their eventual contribution to the data base. The concept
of choice and correlating variables is discussed and examples
of its application are shown. The final factorial design with
fractional replication is given.
-480-
-------
DESIGN OF A STUDY TO MEASURE
FUGITIVE EMISSIONS FROM PETROLEUM REFINING
by
Donald D. Rosebrook
Radian Corporation
8500 Shoal Creek Boulevard
Austin, Texas 78766
INTRODUCTION
Pollutants emitted into the air from a refinery fall
into one of two categories:
• controlled emissions or
• fugitive emissions.
The first type is released into the air at a controlled
rate from a point source such as a stack or a vent. The second
is released without control of rate or direction. Many types of
fugitive emissions cannot be measured by existing standard sampling
and analytical techniques. Therefore, development of reliable
measurement procedures is an essential prerequisite to the
development of strategies for the control of fugitive emissions.
This paper describes a program with dual objectives but will
concentrate on the fugitive emission aspect and within that
objective will concentrate on the design of the program.
-481r
-------
The dual objectives are:
• to determine emission factors, i.e.,
total hydrocarbons and nonmethane
hydrocarbons emitted from fugitive
sources within typical refineries,
and
• to assess the level and character
of emissions in terms of hydrocarbon
distribution and chemical function-
ality.
The decision to emphasize fugitive emissions in our
current study is the result of discussions with the API and the
EPA. The interest of the API stems from a need within the
refining industry to have access to updated emission factors.
The interest of EPA stems from its charge to develop adequate
control technology for potentially hazardous pollutants and from
a need for emission factors accurate and current enough to allow
trade-off changes to be made in existing refining operations.
The most cost-effective program should combine elements of both
fugitive and nonfugitive source sampling.
Previous studies (1958) have dealt with total hydrocarbon
emissions and with emissions of other criteria pollutants. This
-482-
-------
study will consider total and nonmethane hydrocarbons but will
ignore other criteria pollutants except from nonfugitive sources
where current data are not available. The present study will
also seek to identify the individual compounds being emitted
which may be potentially hazardous even at trace levels.
The makeup of a "refinery" is difficult to define since
the process units employed are dependent on the demands for the
refinery's products. Consideration is therefore given to a test
plan that is not definition limited.
APPROACH
The type and arrangement of process units within
refineries may vary considerably from one location to another;
however, the individual unit fittings and processes contained in
each refinery should show close similarities.
The unit fittings, ranging from valves, pumps, etc.,
to various seals, are available to industry in a smaller number
of varieties than the varieties which may result from the various
arrangements of the unit fittings to give unit processes. These
fittings largely control the amount of fugitive emissions poten-
tially available from a given unit process.
-483-
-------
Thus, the approach is to experimentally determine the
emissions from unit fittings. These data will allow calculation
of the emission factors as well as calculation of the fugitive
emission potential of individual unit processes.
In the case of operations involving fugitive emissions
other than those from unit fittings, special sampling approaches
have been devised to allow measurements or estimation of the
total emissions from those operations.
SOURCES OF HYDROCARBON EMISSIONS
There are several known sources of hydrocarbon emissions
in a refinery; they are listed in Table 1. Hydrocarbon emission
sources other than stacks or vents were considered to be fugitive
sources. Stacks or vents which can be identified as the principal
hydrocarbon emission points were considered to be process sources.
Some of the emission sources can qualify as either (or both).
Process emissions have received a great deal of study.
Control technology development has been concentrated on the
reduction of the emissions of hydrocarbons (and other criteria
pollutants) from these sources.
-484-
-------
TABLE 1. REFINERY HYDROCARBON EMISSION SOURCES
Process Emissions Sources:
• Compressor Engines
• Catalytic Cracker Regeneration Stacks
• Air Blowing
• Boilers and Process Heaters
• Flares and Slowdown Systems
• Vacuum Jets (Vented to the Atmosphere)
• Sulfur Recovery or Tailgas-Treating Unit Stacks
Fugitive Emissions Sources:
• Pipeline Valves
• Miscellaneous Joints
• Pressure-Relief Devices
• Pump and Compressor Seals
• Process Drains
• Blending Operations
• Cooling Towers
• Intermediate Storage Tanks*
• Blind Changing
• Maintenance Turnarounds (when vessels are vented
to the atmosphere)
• Wastewater Systems
• Barometric Condensers on Vacuum Jets
• Loading Operations*
• Sampling
• Decoking Operations
*Not within the scope of this work.
-485-
-------
Fugitive emissions, however, have received only
intermittant study since 1958. Fugitive emissions sources allow
hydrocarbons to escape to the atmosphere principally by accident,
inadequate maintenance, or poor planning. Some hydrocarbon
leakage normally occurs even in the absence of such conditions.
The sources are diverse both in their physical characteristics
and in the means by which they allow hydrocarbon escape.
The next step is to define the procedures by which
individual fugitive sources are to be selected for sampling.
Criteria were selected for choosing the refineries to be visited
and units to be sampled.
In evaluating all the possible variables which could
affect fugitive emissions from refineries, it was useful to
categorize these variables into choice parameters and correlating
parameters. A choice parameter was defined as a variable that
directly affects fugitive emissions in such a way that it should
be set up as a category in planning the number of samples taken.
Choice parameters are critical to selecting a statistically
accurate sample.
All other factors which could affect the level of
fugitive emissions will be used as correlating parameters. All
pertinent information on each source sampled will be recorded
-486-
-------
during field testing. Attempts will then be made to correlate
emissions to one, or a group of these parameters, yielding a
more convenient method to apply the restraints of this study.
CHOICE OF REFINERIES AND PROCESS UNITS
The first step in designating specific sources for
sampling was the choice of refineries. A rigorously accurate
sampling procedure would include most, if not all, of the 256
refineries operating in the U. S. Such a large sampling plan
would be virtually unmanageable as well as being cost-prohibitive.
Therefore, a number of representative refineries were selected
for sampling.
Refinery Location
Refinery location was an obvious variable. It has its
principal effect upon the hazardous components generated within
the refinery. It will influence the quality of the crude oil
processed, and the nature and relative quantities of the products
manufactured. The latter affects the severity of the operating
conditions used in the process units as well as the types of
units encountered. All of these considerations affect the by-
products that result from processing the crude, and will influence
the quantities and types of hazardous components produced.
-487-
-------
However, it is not believed that location will have a
direct effect upon rates of hydrocarbon emissions. Knowledge of
the specific chemical species involved is not necessary for -
classification of overall leak rates from fugitive sources.
Thus, this addresses the second of our objectives.
It was decided that location's effect on hazardous
species would be investigated by sampling refineries in four
different geographical regions. These are:
• East Coast,
• Gulf Coast,
• West Coast, and
• Middle U. S. (Midwest and Mid-Continent).
Refinery Age and Size
Two more principal parameters potentially useful in
the selection of refineries for sampling are age and size. These
variables affect such things as maintenance and degree of repair,
quality of equipment, and equipment design. Their use as inde-
pendent variables may not be entirely valid as closer examination
shows.
There is no doubt that age affects the characteristics
of refinery equipment and ultimately influences fugitive emissions
-488-
-------
There is some question, however, about the level at which age
becomes a significant variable. Age has three levels of complex-
ity:
• age of the entire refinery,
• ages of individual process units, and
• ages of individual pieces of equipment.
Age of the entire refinery is the most straightforward
variable to use in the selection of refineries for sampling. New
refineries have all new equipment and usually are more complex
than older ones. Older equipment designs will be found in old
refineries. Even when properly maintained, these could have
greater emissions potential than newer models. If two refineries
had identical maintenance budgets, more might be spent in the
older one just to keep the equipment operational.
Use of refinery age can be inconclusive, however.
Some of the equipment used in old refineries may be older, but
many newer pieces of hardware, and even new units can be the
result of turnarounds and expansions. Units having a variety
of age will be located at the older refinery sites. Specific
age groups and percentages of total capacity processed in old
units are totally site-specific. Sampling an old refinery will
not guarantee that all or any specific fraction of the equipment
studies will be as old as the refinery site itself. It will,
-489-
-------
however, yield data taken from equipment having a mixture of ages
which may tend to average somewhat the effect of age upon the
emissions.
The next level of decreasing complexity is the process
unit. Selecting process unit age as the criterion would eliminate
the problem of expansions since entire units are usually constructed
at one time. Turnarounds would remain a problem because any unit
more than three years old has probably undergone at least one
turnaround in which some of its hardware is replaced by newer
equipment. This again causes an averaging effect, although
newer units will have a smaller variety of age groups represented.
Equipment age is the most accurate variable to use in
studying the effect of age upon hydrocarbon emissions. Specific
operating lives of valves, pumps, etc., can be determined from
refiners and/or manufacturers. Age categories can be defined
for each equipment piece to be sampled, and the true effect of
age determined.
There are several problems with this method. Primarily
because determination of the ages of various equipment pieces will
be difficult at best. Turnaround and maintenance records for
each unit sampled would have to be checked to find the ages of
various pieces of hardware. This would also be true for a refiner
attempting to use the final results of the study.
-490-
-------
Refinery size will have an effect on such things as
the number and type of products manufactured, the number and type
of hazardous species formed, the types of units available for
sampling, the amount of effort put in on maintenance programs,
and the quality of equipment purchased. The most obvious break
in these factors is between very small and large refineries. In
order to use size as some indication of complexity, it was
decided that a realistic cut point is 50,000 bbl/day. A range
of refinery sizes above and below this size should be sampled
to prevent bias toward any individual size. Only those refineries
having sizes within 10% of the cut point should be avoided.
Becuase of the definition of size, size and refinery or
process age become dependent variables. There are not many new
refineries significantly smaller than 50,000 bbl/day. Also, it
would be difficult to find many new process units in refineries
that small. Although the logical solution is to pick hardware
age as the variable, implementation would be difficult if not
impossible. Therefore, it is recommended that refinery age and
size be broken into three categories:
• old/small,
• old/large, and
• new/large.
-491-
-------
Large and small have been defined. Old will mean any refinery
having its oldest operating unit more than 20 years old. New
will mean having units no older than 10 years.
Maintenance Practices
It is very obvious that maintenance is the key to
fugitive emissions and a precise characterization of refinery
maintenance practices could be a valuable parameter in correlating
fugitive emissions. Such a characterization, however, is very
difficult.
An important distinction must be made between corrective
and preventive maintenance. When a leak becomes bad enough to
be detected by operating personnel, most refiners will assign
maintenance forces to repair it without delay. It is this type
of corrective maintenance that is reflected in cost factors, but
it has very little bearing on total fugitive emissions. Over a
period of time, the subject fitting will develop a small leak
which degrades to the point where it is finally detected and
fixed. The time delay between detection and correction represents
only a small fraction of the time that the fitting has leaked.
All those months of slow leaking are responsible for far greater
total emissions. Only a preventive maintenance program wherein
periodic checks are made of all valves, flanges, pumps, compressors,
etc., can have a positive effect in decreasing fugitive emissions.
-492-
-------
The real problem, then, is finding out which refiners
practice preventive maintenance, and to what degree. This is
perhaps best done by a survey of refiners as to whether or not
any attempt at preventive maintenance is made, but it would still
be difficult to evaluate the effectiveness of those employing
such a program.
The conclusion of this analysis is that although
maintenance is very significant in determining the magnitude of
fugitive emissions, it is of no value as a choice parameter.
The only way to determine the effectiveness of maintenance at a
particular refinery would be to measure its fugitive emissions
and compare this to the industry norm. This conclusion is
supported by the fact that the Los Angeles County project (ST-427)
noted significant differences in maintenance at the refineries
they tested, and yet were unable to correlate their final results
to show the effect of the maintenance factor. All significant
parameters will be monitored during the field testing section in
an attempt to quantify the maintenance factor, but no use can be
made of it as a choice parameter. A survey method will be used
on the refineries tested to check possible correlations.
Process Units to be Sampled
Temperature and pressure have major effects upon
fugitive emissions from a refinery. They are available and useful
-493-
-------
variables and will be treated as choice parameters. In a refinery,
there can be found as many combinations of temperature and pressure
as one wishes to define. For the purposes of this sampling plan,
four pressure/temperature classifications were employed:
• high pressure/high temperature,
• low pressure/high temperature,
• high pressure/low temperature, and
• low pressure/low temperature.
These terms are defined as follows:
• pressure -
high - >150 psig,
low - <150 psig, and
• temperature -
high - >1QO°C, and
low - <100°C.
Several process units fall within each category of pressure/
temperature. It is doubtful that a single unit will have each
size, type and service that will be defined for each piece of
hardware. In most cases, several units will have to be sampled
to completely fill the required variable categories. In the
-494-
-------
event that all of the desired equipment categories do exist on
one process unit, samples will still be taken from as many other
units as possible. This will serve two purposes. Bias toward
a particular unit will be eliminated, generating data more
representative of the pressure/temperature class as a whole.
Also, differences between various units can be noted when the
data are analyzed. There may not be enough data to draw firm
conclusions about characteristic process emissions, but it will
give indications of the factors that may be significant.
It should be noted that the four pressure/temperature
classifications are simply guidelines by which units most likely
to have hardware in service at the desired conditions can be
identified for sampling. High pressure and temperature units
also employ valves, fittings, etc. that fall into the lower
pressure and temperature ranges. These need not be ignored, but
can be sampled, also, to fill appropriate choice variable cate-
gories .
Summary
Refineries will be chosen for sampling based on three
criteria:
• location,
• age, and
• size.
-495-
-------
Within each geographic location refineries from three age/size
categories will be chosen. One refinery will be sampled from
the classifications old/small and new/large. Two old/large
refineries in each geographic location should be sampled. This
will allow a greater chance for older equipment to be included
in the sampling scheme (since not all of the equipment in old
refineries is old). This sets the total number of refineries to
be sampled at 16.
BAGGABLE SOURCE VARIABLES
Variables affecting the fugitive emissions from baggable
sources can be classified into choice and correlating parameters,
as were the refinery selection variables in the preceding section.
The variables can then be further defined according to availability
and usefulness. Availability is determined based on the expected
level of difficulty encountered in obtaining the necessary data
in the field. Some information, such as pressure or temperature,
is readily available. Other facts, such as length of time since
a seal was replaced, would be very difficult to ascertain by the
sampling team. The final usefulness of the variable in the
computation of the fugitive emissions from a refinery was also
considered. Some very important variables were not categorized
for sampling because of their lack of ultimate usefulness. For
-496-
-------
example, using actual length of service for each equipment piece
would not yield useful information because it is doubtful that
refiners know how long each pump seal, valve, flange gasket, etc.
has been in service.
Prioritizing variables according to these criteria
allowed the most significant ones to be determined for each
baggable source. Decisions were then made concerning which
categories should be used to define the types and numbers of
fugitive sources to be sampled.
STATISTICAL APPROACH
The Value of a Statistical Approach
Seven variables have been selected as potentially
important for describing the leakage of baggable devices in
refineries. The variables include device type and size,
temperature and pressure of operation, gas or liquid service,
and refinery location, size and age. Even after dividing the
range of each variable into a minimal number of subcategories,
about 8,000 samples would be required to measure the leakage
from each combination of the variables just once. Obtaining a
statistically significant sampling for each combination of the
variables would require several times as many measurements.
-497-
-------
Fortunately, sampling plans utilizing fewer measurements
can be used to determine leakage rates within the desired accuracies,
By assuming that complicated interactions between the variables
are unimportant, the number of necessary measurements is reduced
considerably. Factorial experiment design procedures can be used
to select test combinations of the variables so that the effect
of each variable on the leakage rate can be determined. Analysis
of variance can be used to determine which variables significantly
affect the leakage rate. In our case, both the magnitude of
leakage rates and confidence intervals for the estimates will be
calculated. When appropriate, regression analysis will be used
to estimate trends. Tests of the validity of the assumption that
higher order interactions are negligible will indicate areas where
more measurements are required.
The value of statistical experimental design is that
it provides a systematic and orderly procedure for selecting a
specific set of measurements for a sampling program. The design
is based on assumption about the probability distributions of
errors, independence of effects of different variables, and the
insignificance of multivariable interactions. As the data begin
to come in, the validity of our assumptions will be tested and
adjustments made to the sampling plan.
Structured flexibility forms the tone of the sampling
plan. The structure assures that all preconcerned measurement
-498-
-------
and analysis requirements are efficiently covered. Flexibility
will be maintained within a procedural framework to apply what
is learned toward subsequent sampling and analysis.
The sampling plan structure consisted of outlining
detailed procedures before sampling begins. This included:
• specifying "choice" and "correlation"
variables (choice variables specify
sampling categories, while correlation
variables are others we wish to record
because they might be important);
• developing a form for recording sampling
and analysis results, and
• defining data analysis procedures and
outputs -
data digitizing,
graphic and tabular presentation
of raw data,
analysis of variance to select
important variables,
predicted leakage rates, with
confidence intervals,
tests for validity of assumption,
and
correlations of effect of variable
pairs.
-499-
-------
The flexibility to respond to preliminary results
includes altering choice variable categories for several reasons.
Inconsequential leakers may be dropped, new categories may be
added (based on preliminary results or engineering judgment) and
the most important "correlation" variables may be converted to
"choice" variables. Also, the data analysis procedures may be
improved as the sampling proceeds.
Factorial Experiment Design
Standard statistical techniques have been developed
for handling large numbers of variables, or "factors". The methods
are especially useful when only a few values of each variable need
be considered. Some important properties of the factorial design
for this program are described below.
• First, a list of the variables (factors)
which were expected to most strongly
influence the dependent variable was
prepared. In this case, the dependent
variable was the leakage rate for each
type of equipment.
• A number of levels, or ranges of values,
were selected for each factor. The
decision was based on a trade-off between
-500-
-------
precision of results (many levels)
and economy (few levels per variable).
The preliminary selections for inde-
pendent variables and their numbers
of levels were:
Choice of *
Refinery
(1) 4 geographic locations
(2) 3 age/size categories:
new/big
old/little
old/big
Operating
Conditions
(3) 4 temperature/pressure
combinations
(4) 2 types of service (gas/liquid)
(5) 3 sizes
Device
Types
(6) 8 device types:
block in-line valves
control in-line valves
open-end valves
flanges
pressure-relief devices
pump seals
compressor seals
unit drains
-501-
-------
• Each sample was identified by a set
of levels, one for each factor. For
example, a measurement could be
obtained for the third location,
largest size, old category, second
temperature/pressure category, etc.
• Factorial experiment design is. a
systematic procedure for selecting
a balanced set of experiments. All
levels for each variable are tested
an equal number of times, with
evenly distributed values for the
other variables. The result is
maximum efficiency, the same data
set is equally suitable for determining
the effect of all factors.
• Complete replication refers to a
sampling plan in which measurements
are performed for all possible com-
binations of variables. For the
preliminary variable selections,
about 8000 samples are required for
each complete replication.
-502-
-------
Fractional replication refers to
balanced subsets of the complete
design which can be analyzed for
the effect of each variable. The
ability to analyze for higher order
interactions between the variables
is lost.
Fractional replications are used to
obtain an overview of the results
after a limited number of samples
have been obtained. An estimate of
the effect of each factor on the
leakage rate is available. The
accuracy of the estimates may still
be low because of the small number
of samples.
A sampling plan which adapts according
to the early results will be the most
efficient. The proposed procedure
covers all possibilities by first
performing a fractional replication
and analyzing for the effects of all
factors on the leakage rate. At this
point, some types of devices or
-503-
-------
variables will be found to be adequately
characterized. This could be because
their contribution to total emitted
hydrocarbons has been found to be
small, or possibly because some factor
has a small effect on the amount of
emissions.
The subsequent sampling plan will be
designed to focus on those remaining
factors which are most important for
understanding the total hydrocarbon
emissions situation for a "typical"
process unit. The order of experiments
will be selected so that fractional
replications of variable sets will be
completed at regular intervals. The
results will be analyzed at these
times. The sampling plan will be
adapted at these decision points, so
that the project goal of identifying
the emission properties of equipment
types and typical process units is
most efficiently and thoroughly
accomplished.
-504-
-------
Number of Required Samples
The experimental design procedure which most efficiently
utilizes the data is adaptive in nature. Therefore, the number
of samples necessary to achieve a given hydrocarbon emissions
accuracy cannot be precisely determined. The procedures for
selecting samples and analyzing the results will now be described.
The initial choice variables for baggable sources can
be divided into refinery variables and device-type variables.
The refinery variables are:
• 4 geographical locations,
• 3 age/size categories,
new/big,
old/little, and
old/big.
Twice as many old/big refineries will be sampled relative to the
other two age/size categories. This is because old and large
refineries are common and they should contain a broad spectrum
of device types, ages, and maintenance conditions.
The device-type variables for baggable sources are:
-505-
-------
• 48 in-line valve categories:
2 gas/liquid,
2 block/control,
3 sizes,
4 temperature/pressure conditions,
• 2 open-end valve categories:
sampling/drain,
• 16 flange categories:
2 gas/liquid,
2 sizes,
4 temperature/pressure,
• 6 pressure-relief device categories:
2 gas/liquid,
3 temperature/pressure (low
temperature/low pressure excluded),
• 99 pump-seal categories:
3 sizes,
4 temperature/pressure,
3 shaft and packing classes,
3 Reid vapor-pressure classes
(for 3 temperature/pressure classes),
2 Reid vapor-pressure classes (for
the other temperature/pressure class),
-506-
-------
2 unit drain types, and
• compressor seals (all compressors will
be sampled).
For logistics reasons, it will be desirable to complete
the measurements at each refinery before moving to the next site.
A possible sampling plan for the first refinery is listed below:
No. of
Device Type Replications
In-Line Valves
Open-End Valves
Flanges
Pressure-Relief Devices
Pump Seals
Unit Drains
Compressor Seals
4
4
2
8
1*
4
1
Samples Per
Replication
48
2
16
6
189
2
(all)
Total
Samples
192
8
32
48
189
8
(all)
Total Samples =
477 + (all compressors)
*The replications for pump seals have been selected according to the RVP
category of the fluid being pumped. Thus:
1 replicate - RVP<1.5
2 replicates - 1.526
The number of replications per device type is selected according
to expected leakage rates and measurement accuracies.
For subsequent refineries, approximately the same number
of measurements will be obtained. The distribution of samples
-507-
-------
among the device types at each stage will depend on what has been
learned about:
* leakage rate for each device type,
• precision and accuracy of measurements,
prevalence of device types in refineries,
(this relates leakage to total emissions),
and
• relative toxicity of leakages.
The total number of baggable source samples obtained
will be about 8000 (500 each for 16 refineries).
ACKNOWLEDGEMENTS
This work is being supported by the U. S. EPA under
Contract No. 68-02-2147, Exhibit B. The contributions of Mr.
L. D. Zeagler, Mr. G. E. Harris, Mr. C. E. Riese and Dr. R. G.
Wetherold of Radian Corporation are graefully acknowledged.
-508-
-------
SOME AIR QUALITY AND ENERGY CONSERVATION CONSIDERATIONS
FOR THE USE OF EMULSIONS TO REPLACE ASPHALT CUTBACKS IN
CERTAIN PAVING OPERATIONS
By
Francis M. Kirwan
Environmental Protection Specialist
and
Clarence Maday
Consultant
Strategies and Air Standards Division
Office of Air Quality Planning and Standards
U. S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
-509-
-------
ABSTRACT
This paper reviews practices found in the use of liquified
asphalts for paving purposes. It examines (1) the solvent content
difference in asphalts liquified with petroleum distillates (cutback
asphalts) and asphalts liquified using water and an emulsion, and
(2) the amounts of hydrocarbons emitted when using cutback asphalts in
relation to national and state hydrocarbon emissions. The substitut-
ability of cutbacks and emulsions are discussed. Energy conservation
considerations are treated. The results of an eight-state telephone
survey of highway paving practices are presented.
-510-
-------
TABLE OF CONTENTS
Executive Summary
I. Purpose of Study
II. Asphalt Paving-General
III. Previous Efforts to Encourage Emulsified
Asphalt Use
IV. Air Quality Considerations
V. Energy Conservation Considerations
VI. Eight-State Survey of Paving Practices
and Economic Considerations
VII. Summary and Conclusions
References
Appendix A - Commonly Used Terms Associated with
Asphalt Paving Operations
Appendix B - Pros and Cons - Emulsified Asphalts
Appendix C - Liquified Asphalt Data: Sales,
Hydrocarbon Emissions Estimates - National
and State Summaries
Appendix D - Energy Conservation Considerations
Appendix E - Eight-State Survey of Paving
Practices and Economic Considerations
-511-
-------
Executive Summary
This paper examines the air quality and energy conservation aspects
of asphalt paving practices using liquified asphalt. There are two
basic types of liquified asphalt: (1) asphalts liquified with petroleum
distillates such as kerosene or heavy naphtha, called cutback asphalts,
and (2) asphalts liquified using water and an emulsifying agent, called
emulsified asphalts. Both types of liquified asphalt are "cured" through
the evaporation of the liquifying constituent. Cutbacks emit reactive
hydrocarbons during the curing process; emulsions emit almost no air
pollutants.
In 1975 cutbacks accounted for 2.3% of estimated national hydrocarbon
emissions. In some states the cutbacks accounted for more than 15% of
the state's estimated total hydrocarbon emissions. Some states, e.g.,
Wisconsin, Indiana, Illinois, Ohio, Pennsylvania, Virginia, and West
Virginias have significant air stagnation problems and require regulatory
control of hydrocarbon emissions to attain and maintain oxidant air
quality standards. These states also have had significant hydrocarbon
emissions attributable to paving with cutbacks. Since asphalt paving
operations occur predominantly during warm-weather months, when formation
of oxidants from photochemical synthesis; of hydrocarbon emissions is
most prevalent, the decreased use of cutback asphalt could provide major
assistance in oxidant attainment and maintenance strategies.
It is estimated that in 1975 more than 10 million barrels of petroleum
distillates were used nationally to liquify asphalt for paving purposes.
These distillates represent fuels which were evaporated to the atmosphere
-512-
-------
or were retained in the pavement. The total energy associated with
laying one gallon of cutback asphalt as pavement is about 50,200 Btu,
while the total energy associated with a gallon of emulsified asphalt is
about 2,830 Btu. For these reasons, the use of emulsified asphalt as a
replacement for asphalt cutback has energy benefits.
There are only three known instances when emulsions cannot be
substituted for cutbacks: (a) when long-life stockpiles are required,
(b) when ambient temperatures fall below about 50°F, and (c) possibly when
used as a penetrating prime coat. The price difference between the two
types of liquified asphalt was found to be not significant at this time.
Although there have been continuing and extensive efforts through
paving operation channels (highway engineering offices and the asphalt
paving industry) to promote the use of emulsions, no similar efforts
have been made to inform the general public and environmental groups of
the energy and environmental benefits accruing from the increased use
of emulsified asphalt.
-513-
-------
I. Purpose of Study
The purpose of this paper is to examine potential reductions in
hydrocarbon emissions which may be achieved through substituting one
kind of liquified asphalt for another in certain paving operations. The
paper reviews (1) the differences in asphalts liquified with petroleum
distillates (cutback asphalts), and (2) asphalts liquified using water
and an emulsifying agent (emulsified asphalts). Amounts of reactive
hydrocarbons emitted when using cutback asphalts are discussed, as well
as the substitutability of emulsified asphalts in place of cutback
asphalts. Energy conservation considerations are presented, and the
results of an eight-state telephone survey of highway paving practices
are summarized.
II. Asphalt Paving - General
Asphalt is a by-product of petroleum distillation (natural or man-
made) which man has put to use in many different ways. In ancient times
he used it in its natural form to caulk boats and ships, as mortar in
masonry construction, and as a cement for mending stone tools. Now we
use it for roofing, weatherproofing, floor tile, insulating materials,
molded electrical equipment, papers, shingles, coatings, and many other
applications. One of its better known uses is for pavements. Because
of its durability and weather resistant qualities we use it in many
different paving applications. These pavement uses can range from a
thin layer sprayed on a dirt road to keep down dust, to a heavy duty
pavement of thick layers of asphalt mixed with aggregate (crushed rock,
gravel, slag or sand) placed on a well prepared base and designed to
-514-
-------
carry heavy traffic. In between these two extremes, asphalt pavement
may be of a wide variety of thicknesses and strengths, depending on
the traffic it will have to carry./,x
Asphalt surfaces and pavements are composed of compacted aggregate
and asphalt. Aggregate materials are produced from rock quarries as
manufactured stone or obtained from natural gravel or soil deposits.
Metal ore refining processes produce artificial aggregates as a by-
product. The aggregate performs three functions. It transmits the
load from the surface to the base course, takes the abrasive wear of
traffic, and provides a nonskid surface. The asphalt binder holds the
aggregate together, preventing displacement and loss of aggregate, and
provides a waterproof cover for the base. Asphalts take the form of
asphalt cement (the residue of the distillation of crude oils), and
liquified asphalts, called asphalt cutbacks (asphalt cement thinned, or
"cut back" with volatile petroleum distillates such as naphtha,
kerosene etc.), and asphalt emulsions (nonflammable liquid, produced
by combining asphalt and water with an emulsifying agent such as soap).
Asphalt cement, which is semi-solid, must be heated to convert it to a
usable liquid. A slight amount of heat is often used with cutbacks to
facilitate spraying. Asphalt emulsions normally do not require heating.
Asphalt cutbacks and asphalt emulsions are produced in a wide variety of
types and grades related tp intended use, curing time and structural
design requirements. Some of the use.s are defined in Appendix A.
Emulsified asphalts are used widely in the construction and maintenance
of pavements ranging from high-traffic-volume highways and airports to
low-volume rural roads and city streets. Although emulsions have been
-515-
-------
available since 1903 and used extensively since the 1930s, recent
energy and environmental problems have focused attention on increased
use of these materials. The use of emulsions can reduce energy
requirements by reducing or eliminating petroleum distillates that
are used in liquified asphalts and by lowering heating requirements,
especially in heating aggregates to dry them. The elimination of petroleum
distillates also reduces air pollution by eliminating emissions of
hydrocarbons evaporated during the curing process.
Asphalt paving is a seasonal operation, with cold temperatures
and rainy weather severely limiting construction and maintenance
operations. Winter-time paving is limited to emergency repairs, usually
of pot-holes. Emulsified asphalts usually are not used when rain
is anticipated or when temperatures fall below 50°F. Generally
speaking, emulsified asphalt can substitute for cutbacks in almost any
application. Some believe that emulsions are not good for priming purposes,
others believe that proper soil preparation is the answer, and still
others question the very need for priming. Emulsified asphalt is not
recommended for long-term stockpiling (more than 3-4 weeks). The same
construction equipment used for cutbacks can be used for emulsions. A
moderate amount of training (one or two days) is recommended before
first using emulsions. This training is readily available from members
of the Asphalt Emulsion Manufacturers Association. Local policies which
encourage the use of cutbacks are the only known institutional constraints
that inhibit the use of emulsified asphalt. Appendix B presents more
detail on the advantages and limitations of emulsified asphalts.
-516-
-------
•III. Previous Efforts to Encourage Emulsified Asphalt Use
Some of the organizations concerned with energy problems affecting
the supply and use of asphalt road paving materials are: Department of
Transportation (DOT), Federal Highway Administration (FHWA), Federal
Energy Administration (FEA), U. S. Department of Agriculture (USDA),
Forest Service, Environmental Protection Agency (EPA), Transportation
Research Board (TRB), National Asphalt Pavement Association (NAPA), The
Asphalt Institute (AI), Asphalt Emulsion Manufacturers Association
(AEMA), American Society for Testing and Materials (ASTM), American Road
Builders Association (ARBA), American Association of State Highway and
Transportation Officials (AASHTO), and state and local highway agencies.
In December 1973 and again in January 1974, FHWA issued notices
concerning fuel conservation in federally funded highway construction
programs./2\ These notices encouraged state officials to minimize
the use of cutback asphalts by substituting emulsions and to reduce
mixing temperatures. They also provided guidelines on conserving fuel
and presented analyses which demonstrated the large quantity of
petroleum distillates which could be saved by substituting emulsified
asphalts for cutbacks. FEA and EPA studies resulted in the conclusion
that increasing fuel prices had already established a trend of
increased use of emulsions. To accelerate this trend, FEA contracted
with the National Research Council's Transportation Research Board
to produce a synthesis report/2\ on the use of asphalt emulsions for
pavements. This report was widely publicized by DOT and various trade
associations. FEA alone distributed 4,700 copies to city and county
-517-
-------
engineers in December 1975. In October 1975, EPA informed its regional
offices by letter of the advantages of emulsified asphalts over
cutbacks and advised the regional offices to encourage the use of
emulsions to save energy and reduce emissions of hydrocarbons.
Other agencies and organizations have been at work on the problem.
For example, NAPA and AI have published papers on energy conservation
in highway paving, AEMA has been making extensive efforts throughout its
membership to encourage the use of asphalt emulsions, and USDA Forest
Service has published a report on its experience in using asphalt
emulsions, as has the Navajo Area Bureau of Indian Affairs.
IV. Air Quality Considerations
The volatiles in cutback asphalts release hydrobarbons into the
atmosphere in amounts that vary according to the type of cutback.
Cutback asphalts fall into three broad categories: Slow Cure
(SC) (sometimes referred to as Road Oil), Medium Cure (MC) and Rapid
Cure (RC). Cutback content averages 35% diluents. /.\ SCs are a
fairly heavy crude oil in the Bunker C range. MCs are diluted with
a kerosene- type solvent. RCs are diluted with a heavy naphtha or a
gasoline-type solvent.
Table 1 is a summary of estimated hydrocarbon emissions resulting
from the use of cutback asphalts for paving purposes. The emission
calculations are based on the 35% volatiles contained in the cutbacks/^
and on the following evaporation amounts estimated by The Asphalt
Institute, g^: SC - 20%-30% evaporated (average: 25%), MC - 50%-60%
evaporated (average: 55%), and RC - 70% evaporated. Although no known
measurements of evaporation rates have been made, it is believed that
-518-
-------
most of the loss takes place early during paving operations: Continuing
amounts are lost to the atmosphere as time goes by but at an ever
decreasing rate, according to the Asphalt Institute/^.
Table 1
Summary of National Hydrocarbon Emission Estimates from
-the Use of Cutback Asphalt Paving Products
All Volatiles (Incl HC) HC Emissions
(Total tons/year) (tons/year)
1971 1,916,857 964,476
1972 1,830,724 933,920
1973 1,975,451 1,014,524
1974 1,613,454 817,753
1975 1,434,895 742,203
It is important to remember that paving operations are seasonal
and that the paving season occurs during the warm weather months when
formation of oxidants from photochemical synthesis of hydrocarbon emissions
is most prevalent. Attempting to arrive at specific answers to questions
about photochemical reactivity of the hydrocarbons emitted by cutbacks
is complicated by the fact that there are so many cutbacks of varying
chemical compositions. The situation is further complicated by the
variables of solar radiation, cloud cover, air mass stagnation, hydrocarbon
concentrations, and oxidant formation. However, cutbacks can be classified
as moderately to highly reactive as far as oxidant formation, is
concerned./5 &\
Emulsified asphalts, on the other hand,.consist of asphalt liquified
-519-
-------
with water containing an emulsifier. Emulsions are relatively
pollution-free with few volatiles to evaporate into the atmosphere.
FHWA has pointed out that there may be some distillates in some
formulations of emulsified asphalt.
Table 2 indicates the relationship of hydrocarbon emissions from
cutback asphalts used in paving, to national hydrocarbon emissions.
(Asphalt paving operations are not included as a source of HC emissions
in the national summary). Although the figures in Table 2 are estimated,
the conclusion of Table 2 is that cutback asphalt accounts for over 2%
of national hydrocarbon "emissions.
Table 2
Hydrocarbon Emissions from Cutback
Asphalt as a Percentage of National HC Emissions
Summary of Relationship of
National HC Cutback Asphalt
Emissions HC Emissions to
(10 tons/year) National HC Emissions
2.8%
2.7%
2.9%
2.4%
2.3%
Table 3 shows a breakdown of national hydrocarbon emissions
for mobile and stationary sources and displays the emissions from
cutbacks in context with the two other sources.
1971
1972
1973
1974
1975
33.3
34.1
34.0
32.9
30.9
-520-
-------
Table 3
U.S. Hydrocarbon Emissions by Category
(1CT tons/year)
MOBILE STATIONARY CUTBACK
SOURCES SOURCES SOURCES
19.6
20.1
20.3
20.4
19.2
1.0
.9
1.0
.8
.7
1971 13.7
1972 14.0
1973 13.7
1974 12.5
1975 11.7
Appendix C contains additional detailed information: Table C-l
summarizes annual national sales of cutback asphalts for paving purposes
from 1971 through 1975; Table C-2 displays by EPA Region and by State,
sales of cutback asphalts and sales of emulsified asphalts in 1975; and
Table C-3 displays by EPA Region and by State the statewide hydrocarbon
emissions from the use of cutback asphalts, State total hydrocarbon
emissions, and the percentage of State total emissions accounted for by
the cutback asphalts.
It is further noted that some states experience frequent air mass
stagnation and have oxidant air quality problems. Some of these states,
e.g., Wisconsin, Indiana, Illinois, Ohio, Pennsylvania, Virginia and West
Virginia, require regulatory control of HC emissions for attainment and
maintenance of oxidant ambient air quality standards. These states also
-521-
-------
have significant quantities of hydrocarbon emissions attributable to
paving with cutback asphalts.
V. Energy Conservation Considerations
In 1975, 10,249,250 barrels of petroleum diluents were used to liquify
asphalt for road paving operations. This amount of cutback is equivalent
to 464,906,000 gallons of gasoline, enough to fuel almost 558,000
automobiles for a single year in the United States. Rather than powering
automobiles, airplanes, or industry, however, energy in the form of
diluents was poured onto road surfaces, where some evaporated and some
remains. The energy impact of using cutback asphalts is just as striking
when viewed in terms of the energy expended per gallon of paving material.
The total energy associated with manufacturing, processing, and laying
one gallon of cutback asphalt is about 50,200 Btu. On the other hand,
analysis of emulsified asphalts shows that about 98% of the petroleum
diluents is replaced with water with the result that only about 2,830 Btu
is associated with each gallon of paving material. The components of
these energy costs are described in Appendix D.
VI. Eight-State Survey of Paving Practices and Economic Considerations
State highway maintenance divisions in eight states were contacted
for information, opinions, and experiences regarding the use of emulsified
asphalt paving materials. The states selected for this survey were the
larger users of asphalt. Since each state is responsible for some
fraction (which may differ for each state) of the roads within its
-522-
-------
boundaries, this survey addresses only those asphalt paving operations
for which the state is directly responsible.
In general, the survey showed that there has been an increased
use of emulsified asphalts. This increased use, which varies with each
state contacted, has been brought about primarily through fuel con-
servation measures and economic considerations. Relatively little
consideration is given to HC emission from paving operations. For example,
in Allegheny County, Pa., the Pennsylvania DOT uses emulsified asphalts
almost exclusively for county road paving operations because of conservation
and economics. In areas where such considerations do not exist, the
choice of emulsified asphalt or cutback asphalt depends largely upon
user preference as well as experience in specific materials and suppliers.
Individual responses ranged from general acceptance of emulsified
asphalts for paving to indifference and skepticism about emulsions.
Pennsylvania has changed from 30% emulsions/70% cutbacks in 1973 to
70% emulsions/30% cutbacks today. New York State uses 97% emulsions/
3% cutbacks.
New York and Pennsylvania have initiated training programs
to instruct their personnel and contractor personnel in the correct
use of emulsions.
Appendix D details further energy conservation considerations.
VII. Summary and Conclusions
The air quality and energy conservation aspects of the use of
liquified asphalt for paving operations have been analyzed to determine
the potentials for energy savings and reduced emissions. Cutback
-523-
-------
asphalts are liquified with hydrocarbon distillates such as kerosene or
naphtha; these reactive hydrocarbons are emitted during the curing process.
Emulsified asphalts use water and an emulsifying agent for liquifaction;
virtually no pollutants are emitted during the curing of emulsions.
Overall, more than 10,000,000 barrels of distillates are used annually
for paving purposes. Most of this is evaporated into the atmosphere; the
remainder is retained in the pavement. Use of emulsions would save almost
all of those 10,000,000 barrels of distillates for use as or conversion
to fuels.
In some states the curing of cutback asphalts accounted for a signi-
ficant amount of the state's total annual hydrocarbon emissions. This
problem is made more serious by the fact that asphalt paving operations
take place primarily during warm weather when oxidant formation from the
photochemical synthesis of hydrocarbon emissions is most likely. Reduced
use of cutback asphalts could decrease materially the oxidant problem in
these states.
It is anticipated that a minimal amount of cutback asphalt will
continue to be used at temperatures lower than 50°F and for dusty
surfaces. Also, some cutbacks will be used where portable plants are
not available, because the four week stockpile life of emulsions is a
problem. Other concerns with traffic control, rolling, aggregate selection
and application, emulsion selection, etc., (discussed in Appendix B), can
usually be met through good management.
Significant energy savings and air quality improvements can be
realized from the increased use of emulsified asphalts.
-524-
-------
REFERENCES
1. The Asphalt Institute, "Magic Carpet, The Story of Asphalt",
1975.
2. Transportation Research Board, National Research Council, National
Cooperative Highway Research Program Synthesis of Highway Practice
30, "Bituminous Emulsions for Highway Pavements", 1975.
3. U.S. Department of the Interior, Bureau of Mines, Mineral Industry
Surveys, "Sales of Asphalt in 1975", prepared July 19, 1976.
4. Special National Asphalt Pavement Association Report "Fuel Conser-
vation", by Charles R. Foster and Fred Kloiber.
5. Conversations with John Bachmann and Dave Patrick, OAQPS, EPA.
6. Conversations with Vyt Puzinauskas, Asphalt Chemist, The Asphalt
Institute.
7. Commonwealth of Pennsylvania, Department of Transportation, Bureau
of Materials, Testing and Research Informational Report, "Lets Get
Acquainted with Asphalt Emulsions", April 1974. Prithvi S. Kandhal,
P.E.
8. Conversation with Vaughn Marker, Chief Engineer, The Asphalt
Institute.
9. "National Air Pollution Emission Estimates, 1970-1975", undated,
MDAD, OAQPS.
10. "Asphal-t Emulsion Construction on the Navajo Reservation", W. R. Meier,
Bureau of Indian Affairs, Gallup, N.M., April, 1976.
11. EPA-450/2-76-007, May 1976, "1973 National Emissions Report",
National Emissions Data System (NEDS) of the Aerometric and Emissions
Reporting System (AEROS).
12. EPA-450/1-76-001, February 1976, "Monitoring and Air Quality Trends
Report, 1974".
13. 1974 Keystone Coal Industry Manual, Fuel Conversion Factors, p. 688.
14. Rural and Urban Roads, November 1976, article on page 26, "Water-
based Emulsions Help County Save Cash on Seal Coats".
15. Federal Register, Vol. 41, No. 233, Thursday, December 2, 1976,
page 52934.
-525-
-------
16. '"Asphalt Surface Treatments and Asphalt Penetration Macadam",
The Asphalt Institute, Second Edition, November 1969, Manual
Series No. 13 (MS-13)
17. "Energy Requirements for Roadway Pavements", The Asphalt Institute
(MISC-75-3) April 1975.
-526-
-------
APPENDIX A
COMMONLY USED TERMS ASSOCIATED WITH
ASPHALT PAVING OPERATIONS
-527-
-------
Appendix A
Commonly Used Terms Associated with
Asphalt Paving Operations
1. Crack Filler - Asphalt used to fill and seal cracks in existing
pavements.
2. Dust Palliative - A light application of liquified asphalt (cutback
or emulsified asphalt) for the express purpose of controlling
loose dust.
3. Maintenance Mix - A mixture of asphalt and mineral aggregate for
patching holes, depressions, and distressed areas in existing
pavements. These mixes are suitable for relatively small spot
application, hot or-at ambient temperature, using hand-laying
and compaction techniques. This includes mixes for immediate use
or for stockpiling for future use.
4. Penetration Macadam - Pavement construction using essentially
one-size coarse aggregate that is penetrated in place by an
application of asphalt binder. The asphalt application may be
followed by an application of finer aggregate to reduce the void
space.
5. Plant Mix, Cold Laid - A mixture of liquified asphalt (cutback or
emulsified asphalt) and mineral aggregate prepared in a central
bituminous mixing plant and spread and compacted at the job site
when the mixture is at or near ambient temperature.
6. Plant Mix, Hot Mix - Hot Laid - A mixture of asphalt and mineral
aggregate usually prepared in a conventional hot-mix plant or drum
mixer at a temperature of about 250°f and spread and compacted
at the job site at a temperature above 200°F.
-528-
-------
7. Prime Coat - An application of asphalt to an absorptive surface
to penetrate and bind the aggregate surface and promote adhesion
between it and the new superimposed construction.
8. Road Mix (Mix-in-Place) and Travel Plant Mix - A procedure by
which the asphalt and mineral aggregate are mixed on the job
site by means of travel mixers, motor graders, or other special
road-mixing equipment.
9. Surface Treatment - An asphalt surface that results from one
or more successive alternate applications of asphalt binder
and cover aggregate to a prepared consolidated gravel, crushed
stone, waterbound macadam, earth, stabilized soil, or similar
base. Multiple application of asphalt and mineral aggregates
may be used.
10. Seal Coat - An asphalt surface that results from one or more
successive alternate applications of asphalt binder and cover
aggregate to an existing paved surface.
11. Slurry Seal - A uniform application of a mixture of emulsified
asphalt, fine aggregate, mineral filler, and water to an existing
pavement. Single or multiple applications may be used.
12. Tack Coat - An application of asphalt applied to an existing surface
to provide a bond between new surfacing and existing surface and
to eliminate slippage planes where the new and existing surfaces
meet.
-529-
-------
APPENDIX B
PROS AND CONS - EMULSIFIED ASPHALTS
-531-
-------
Appendix B
Pros and Cons - Emulsified Asphalts
In any comparison of alternate methods for accomplishing a given
job there is usually a concern for the advantages and disadvantages of
one method relative to another. In the evaluation of emulsified asphalts
versus cutback asphalts, such concerns were raised and have been considered.
Reduced hydrocarbon emissions and fuel conservation have been cited as
advantages to be expected from the use of emulsified asphalts instead
of cutback asphalts. The reduction in emissions and the fuel savings
are significant. Cost comparisons are generally favorable to emulsified
asphalts but there are specific instances where the cost differential
is negligibly small. While the foregoing suggests that the large-scale
use of emulsified asphalts should be encouraged, there are a number
of reasons which prevent a complete switchover to emulsified asphalts.
Some of the reasons are intrinsic to the use of emulsions; the other
reasons or objections follow from local practices. There is little we
can do about the intrinsic reasons in the short term; research and
development, however, may remove these reasons in the long run. Reasons
or objections based upon local practices are usually amenable to change.
In the following, both kinds of reasons are considered.
Intrinsic reasons which stand in the way of universal acceptance
of emulsified asphalts include: 1) the time required for the water to
evaporate. This means that a newly paved road should be closed to traffic
-532-
-------
upwards of two hours to allow the asphalt to set up. Under unfavorable
conditions (high humidity, calm air, temperatures lower than 50°-60°F.
this could extend to a day or more. 2) the surface to be paved should be
swept clean of dust. This means also that the aggregate should be free of
dust which passes through a #200 sieve. Dust causes the emulsion to
break prematurely, that is, to accelerate the separation of the asphalt
and the water.
The second class of reasons for not using emulsified asphalts can be
regarded as non-intrinsic since there are ways of meeting these objections.
These include: 1) a stockpile life of only four weeks, (Cutback asphalts
have a long stockpile life). Use of portable mixing plants should
overcome this objection. 2) Possible washout in a heavy rain. Care should
be taken not to use emulsified asphalts when rain threatens. 3) "It's not
like buying an 'off-the-shelf item." Cooperation between the contractor,
aggregate supplier, and emulsion supplier should result in the proper
selection of an emulsified asphalt. 4) Miscellaneous reasons which can
be met through the implementation of a training program to instruct con-
tractors and operators in the proper use of emulsified asphalts.
In summary, emulsified asphalts cannot be used at temperatures lower
than 50° and they cannot be used on dusty surfaces. Where portable
plants are not available, the four week stockpile life is a problem.
Traffic control is more critical than for cutback asphalt paved roads.
Other concerns, such as rolling, aggregate selection and application,
emulsion selection, etc. can usually be met through good engineering and
construction practices.
-533-
-------
Some comments by FHWA follow:
"In some States, maintenance materials may be purchased on annual
contracts. This may create ideal conditions for the use of emulsions
since the available expertise of the emulsion manufacturer can be
utilized to insure a satisfactory product by adjusting the emulsion
formulation to suit the aggregates, designing the mixture and training
personnel. In the competitive bidding situation for construction projects
this may not be the case. Traditionally, State's material sections
have either designed the bituminous mixtures or verified the contractor's
submitted mix design for adequacy to the intended purpose. A number of
the States have had little or no experience in the design of asphalt
emulsion mixtures and currently there are no nationally recognized
standard procedures for the laboratory design of these mixtures as
there is for hot asphaltic concrete or cutback asphalt mixes. A large
research effort is necessary in this area."
"It can be safely said that all of the State highway agencies are
aware of and concerned with the needs for energy conservation and protection
of the environment. The slower than desired trend by some of these
agencies toward total substitution of emulsions for cutbacks can in part
be attributed to the following factors.
1. Lack of nationally accepted standard laboratory design procedures
for asphalt emulsion mixtures. A number of design methods have been
developed, mostly by members of the emulsion manufacturing industry.
None of these have been endorsed by the standardizing associations AASHTO
or ASTM. The availability of a recognized objective reproducible
-534-
-------
laboratory design procedure for use by the highway agencies is considered
a necessity in the effort to foster greater usage of asphalt emulsions.
2. Lack of reliable data on the structural worth of the asphalt
emulsion mix. The pavement designer needs this information in the
development of the structural section for a project. The general use of
emulsions in the past has not been conducive to obtaining this type of
data.
3. The tight money conditions that most State highway agencies
have experienced in the past 3 years. This has not enhanced the climate
for highway agencies to more or less experiment with a new product
particularly when there is no price advantage. In response to item 1
above, the FHWA's Office of Research and Development has underway an in-
house effort to develop a laboratory method of asphalt emulsion mix
design. However, the allocation of necessary manpower and financial
resources has been somewhat limited because of competing research
requirements. That is not to say that we believe the effort to be
unwarranted, just that strictly from a highway point of view there are
other, more urgent, research needs."
-535-
-------
APPENDIX C
LIQUIFIED ASPHALT DATA: SALES, HYDROCARBON EMISSIONS ESTIMATES
NATIONAL AND STATE SUMMARIES
-537-
-------
Appendix C
Liquified Asphalt Data: Sales, Hydrocarbon Emissions Estimates -
National and State Summaries
Table 4 displays a summary of national sales of cutback asphalt
for paving purposes. It is based on Bureau of Mines annual reports of
asphalt sales for the years shown. The Bureau of Mines data is further
broken down into the three categories of cutbacks based on an estimated
usage ratio of 65% Medium Cure (MC) and 35% Rapid Cure (RC)./8%
Table C-l
Summary of National Sales of Cutback Asphalts for Paving
(tons/year)
Year
1971
1972
1973
1974
19'75
Slow Cure
(Road Oils)
1,543,091
1,370,546
1,424,000
1,251,091
991,455
Medium
Cure
2,557,868
2,509,061
2,743,095
2,183,205
2,020,359
Rapid
Cure
1,375,775
1,351,033
1,477,051
1,175,572
1,087,886
Total
5,476,734
5,230,640
5,644,146
4,609,868
4,099,700
-538-
-------
Table C-2 Sales of Cutback Asphalt and Emulsified Asphalt in 1975,
by EPA Region and State
Cutback Asphalt sales Emulsified Asphalt
EPA Region (Includes Road Oils) Sales
and State _ (tons) _ (tons) _
Region I
Connecticut 20,355 2,340
Maine 23,702 16,742
Massachusetts 18,510 4,855
New Hampshire 22,982 12,994
Rhode Island 927 677
Vermont 321 299
Region II
New Jersey 26,528 55,357
New York 103,435 156,755
Region III
Delaware , ,fiQ 1,746
Maryland + D.C. 49 'i?? 62,122
Pennsylvania Jo nii 57,369
Virginia ll?ll 58.270
West Virginia 7Q7 31 '238
Region IV
45,138 51,829
46>387 26'753
50,003 56,318
7 QRQ 27
C » yby Ll '
Region V
312,041 31,042
129>783 17,815
39'305 24'441
297,417 182,896
Indiana 80'805 162'636
125'093 16'853
-539-
-------
Region VI
Arkansas 68,829 40,454
Louisiana 19,867 35,631
New Mexico 70,848 40,228
Oklahoma 390,836 7,372
Texas 189,648 131,079
Region VII
Iowa 98,242 17,496
Kansas 121,111 38,427
Missouri 470,686 13,102
Nebraska 21,928 8,519
Region VIII
Colorado 73,736 1.255
Montana 50,692 6,099
North Dakota 49,373 22>iJ?
South Dakota 41,152 5,771
Utah 27,397 17,006
Wyoming 15,705 1.931
Region IX
Arizona 152,121 100,987
California 236,685 136,802
Hawaii 5,220 - -
Nevada 47,183 6,828
Region X
Alaska 7,416 5,370
Idaho 60,426 21,008
Oregon 25,417 58,074
Washington 141,066 57,246
Total 4,099,700 2,143,877
-540-
-------
Table C-3
Cutback Asphalt Hydrocarbon Emissions Contribution
to State Total Hydrocarbon Emissions,,, ,-j
EPA Region
and State
Region I
Connecticut
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
Region II
New Jersey
New York
Region III
Delaware
Maryland & D.C.
Pennsylvania
Virginia
West Virginia
Region IV
Alabama
Florida
Georgia
Mississippi
Kentucky
North Carolina
South Carolina
Tennessee
Cutback Asphalt HC
Emissions Total* and %
of State HC Emissions
(tons) J
4,286
4,991
3,897
4,838
196
69
5,586
21,532
247
8,930
30,128
12,259
2,449
8,346
16,175
10,524
370
6,781
6,730
3,578
4,670
1.9%
3.8%
5.2%
1.9%
2%
3.6%
2.4%
2.1%
1.1%
1.4%
2.2%
1.9%
1.2%
1.1%
State Total HC
Emissions,
**
(tons)
11
221,736
125,390
520,930
88,366
85,283
41,372
639,325
1,116,330
64,813
427,337
811,992
483,501
113,711
729,076
,140,776
467,461
224,298
352,382
548,584
360,971
391,719
-541-
-------
Region V
Illinois 56,814 3% 1,831,465
Minnesota 23,627 5.8% 381,938
Michigan 7,157 <1% 824,898
Ohio 55,145 4.7% 1,117,174
Indiana 14,712 2.4% 603,771
Wisconsin 18,770 3.'2% 574,454
Region VI
Arkansas 12,726 6.3% 189,896
Louisiana 3,673 <1% 1,228,769
New Mexico 13,098 8.4% 143,116
Oklahoma 72,462 15.6% 391,672
Texas 35,281 1.6% 2,211,509
Region VII
Iowa 17,888 4.8% 353,844
Kansas 22,051 6.2% 336,756
Missouri 87,270 16.3% 448,299
Nebraska 8,023 2.1% 378,922
Region VIII
Colorado 12,857 5.4% 224,797
Montana 8,839 4.3% 197,518
North Dakota 9,127 11.3% 71,522
South Dakota 5,059 5.6% 85,161
Utah 4,778 4.4% 103,047
Wyoming 2,739 3.9% 67,654
Region IX
Arizona 23,577 8.9% 241,985
California 33,799 1.6% 2,115,039
Hawaii 771 <1% 94,405
Nevada 7,006 11.6% 53,429
-542-
-------
Region X
Alaska 1,095 <1% 109,370
Idaho 10,536 8.2% 118,532
Oregon 3,753 1.3% 290,495
Washington 20,823 5.8% 336,944
*Weighted average by state. See Table 1 narrative (Sect IV, p.5)
**State total hydrocarbon emissions estimates are based upon latest
available data on emission sources. All hydrocarbons are included.
Asphalt paving operations were not included as a source.
-543-
-------
APPENDIX D
ENERGY CONSERVATION CONSIDERATIONS
-545-
-------
Appendix D
Energy Conservation Considerations
The energy associated with cutback asphalt can be compared with
that associated with emulsified asphalt by considering their respective
compositions, the attendant process energy, and the energy required
for asphalt spray applications using an asphalt distributor. Since the
petroleum distillates content of cutback asphalt varies from 20% to 50%
of the total mixture, the energy associated with these asphalts spans a
wide range. In order to strike one kind of comparison, consider cutback
asphalt and emulsified asphalt, each with 65% asphalt. This is reasonable
because this percentage is generally recommended for emulsified asphalts
and the eight-state survey (Section VI and Appendix E) revealed that
the amount of cutback asphalt (typically containing 60% - 70% asphalt)
required for paving operations was about the same as for emulsified asphalts.
About 2500 Btu is required to process a gallon of asphalt for paving/,j\
and to this is added the energy content of the cutback which, for a medium
cure is about 135,000 Btu/gal. Thus, for one gallon of cutback asphalt
with 65% asphalt and 35% cutback the energy represented therein is 2500
+ .35 x 135,000 = 49,750 Btu. About 2050 Btu is required to obtain one
gallon of emulsified asphalt, independent of emulsifier energy content./-^
According to ARMAK (leading manufacturer of emulsifiers) there is about
7500 Btu/lb associated with the production of emulsifiers. At about
0.084 Ib emulsifier/gal (1%) an additional 630 Btu should be added to
the 2050 Btu for each gallon of emulsified asphalt giving an intermediate
total of 2680 Btu/gal.
-546-
-------
Finally, the energy requirements for asphalt spray applications
using an asphalt distributor amount to 444 Btu/gal for cutbacks and
144 Btu/gal for emulsions./,^ (An additional 300 Btu/gal is required to
heat the cutback asphalt). Hence, the total energy associated with
laying one gallon of cutback asphalt is about 50,200 Btu while the total
energy associated with a gallon of emulsified asphalt is about 2830 Btu.
Each ton of petroleum diluent used in cutbacks is equivalent to
about 300 gallons./2\ The tons of volatile diluents displayed in Table 1
convert to barrels (42 gals/bbl) as follows:
Table D-l
Petroleum Diluents Used to Liquify Asphalt
(Estimated Bbls/year)
1971 13,691,836
1972 13,076,600
1973 14,110,364
1974 11,524,671
1975 10,249,250
These petroleum diluents (fuels) were evaporated to the atmosphere
or were retained in the asphalt pavement. In terms of energy content, the
amount of gasoline equivalent to the 10,249,250 barrels of petroleum
diluents used in 1975 for cutback asphalts is:
Mn ?4Q ?RO bbl) x (42 aal/bbll x 135,000 Btu/gal diluent =
(10,249,250 bbi; x (U gai/DDi; x
125j000 Btu/gal gasoline
464,906,000 "gal gasoline
For an annual mileage of 10,000 at 12 mpg this gasoline equivalent
would fuel almost 558,000 automobiles for one year.
-547-
-------
APPENDIX E
EIGHT-STATE SURVEY OF PAVING PRACTICES AND ECONOMIC CONSIDERATIONS
-549-
-------
Appendix E
Eight-State Survey of Paving Practices
and Economic Considerations
Eight state highway maintenance divisions were contacted for their
opinions, facts, and experiences regarding the use of emulsified asphalt
paving materials. The states selected were the larger users of asphalt.
An initial contact with each state was made in order to establish an
overview of their operations. A second contact was made to obtain more
detailed information.
In discussions with the eight states it was found that there was an
insignificant difference (overall) in the costs of the two types of
asphalts at this time. Until three years ago the cutback asphalts were
less expensive but the increase in oil prices has changed this picture.
The local picture may be influenced by varying economic considerations.
However, the availability of light diluents from a nearby (and older)
refinery could make cutbacks very attractive. Newer refineries are able to
recycle naphtha and similar light stock to turn out a more profitable
product. When a nearby source of these light diluents is not available,
the additional freight costs to deliver the material usually tips the
scales in favor of emulsions.
A specific example is found in Waukesha County, Wisconsin.,^ The
switch to emulsions was made to conserve energy in 1973. In 1976 the
county used emulsions for resurfacing and realized about a 10% cost
saving in materials: 36.92
-------
Illinois
Primary roads are usually paved with portland cement concrete. Main-
tenance operations and paving of secondary roads are under local (city or
county) jurisdiction. The local agencies are being encouraged ("jawboned")
to use emulsified asphalts. The state "Specification Book" now lists
emulsions as a viable option. Cationic and High Float Emulsions (HFE) are
preferred (for definitions see reference 2). Stockpile life is no
problem. There is almost no cost differential. Most user problems are
expected to be resolved as more plants come into operation. Also, the
expanded use of "porta-pugs" (portable on-site mixing equipment) should
benefit pre-mix emulsions.
Paving with emulsions is not done after September 30. Under favorable
conditions (3 consecutive preceding days with maximum temperatures over
60 F in the shade and a night time minimum of 40 F or higher) emulsified
asphalts can be used until October 31. Use of emulsions is resumed in
the spring when these temperature ranges are obtained.
Missouri
Until 3 or 4 years ago cutback asphalt was used exclusively. Today,
emulsions are satisfactory for seal coats, priming, and in mixtures for
base construction. Emulsions are generally unsatisfactory for road
surfaces. The emulsified asphalt industry in Missouri is in its infancy
and has not been able to live up to its claims.
Problems include emulsions break-up and non-adherence to aggregate
resulting in rough surface. Procurement of emulsion asphalt is not like
buying an off-the-shelf item. The emulsifying agent and proportions must
-551-
-------
be tailored to the available aggregate the particular job. Their most
successful emulsions use 4-10% cutback.
Most of the paving work is done under contract.
Pennsylvania
The Department of Transportation has its own environmental section
which maintains liaison with the state Department of Environmental Resources.
The expanded use of emulsified asphalts was initiated as a fuel and
cost conservation measure about three years ago. DOT has continued to use
emulsions because they expect the cost of petroleum diluents to increase
further. Pennsylvania is directly responsible for and controls must of its
highway system (as do Virginia, North Carolina, and Texas), i.e., 45% of the
20,000 miles of roads in the state. The Pennsylvania DOT is under annual
contracts with asphalt suppliers and hence control the type of asphalt
to used on a given job.
About ten years ago emulsified asphalts were used for the first
time. Three years ago usage was 30% emulsions. Today, it is 70%
emulsions- No further increase is expected. The state is committed to
emulsions and has directed its districts accordingly. One-day training
programs have been set up for administrators as well as operators.
Hot-mix will be used for primary roads for some time because
suppliers have considerable investment in plants and equipment.
Special mixtures are required for emulsion asphalts which are to be
stockpiled. This operation demands tough quality control and requires
dry stone (which retards curing) and the right temperature. Further,
the mixture cannot exceed a depth of 8 feet. Pennsylvania also uses
Travel-mix plants to prepare the emulsion at the job site.
-552-
-------
• Emulsions are used for recycling pavements. Cutbacks are still used
for dust control.
Ohio
The increased use of emulsions is being encouraged. The state cannot,
however, mandate to the local governments. At present there is an
insignificant price differential but the cost of cutback asphalt is expected
to increase.
Oklahoma
Emulsified asphalts are used for seal coats and slurry seals - about
10-15% more emulsified asphalt is required than cutback asphalt for a
given surfacing requirement. Emulsions are not used for surface paving.
There is an insignificant price differential today. Emulsion costs are
expected to decrease but a trend to increased use of emulsions is not
foreseen.
Texas
The state is trying to increase use of emulsions as a result of
energy conservation efforts and financial conditions. Overall they find
an economic advantage in the use of emulsions. Emulsions are not used
for heavy traffic applications or where roads must be opened immediately
after re-surfacing. Stockpile life for emulsions is about one month.
There are 25 Districts in Texas and each District Engineer is
responsible for maintenance. Attitudes are important - some users have
been using emulsions with success for years while others are unalterably
opposed to it.
-^553-
-------
New York
In 1976, 97% of liquid asphalts used were emulsions; 3% were cutbacks.
New York state has an educational program to instruct its personnel in
the correct use of emulsions and recommends and stresses the advantages
of emulsions over cutbacks. In 1977 the use of cutbacks will be restricted
to pavement repairs only. The state knows that emulsions are actually
being used in paving operations by inspecting and testing material for
acceptance and adherence to specifications. They have experienced no
stockpile problems. They feel that four weeks is a good average for stock-
pile life and generally use the material within a four week period. Some
material has been stockpiled for longer periods with no problems. The
foregoing information pertains only to the use of emulsions by the New
York State Department of Transportation and does not include counties
and municipalities withagi the state.
California
California State Department of Transportation uses cutbacks as the
primary liquid asphalt. The emulsion that is used is mainly for seal
costs. They have no prejudice against emulsions per se, but they stockpile
for the whole winter, particularly for the remote areas in the northern
part of the state, and emulsified asphalt stockpiles just will not last
more than a few weeks before they set up and become unusable. The cost
differential between emulsions and cutbacks is negligible. Paving
operations are in compliance with environmental requirements.
Although not a part of the eight-state survey, a discussion with
the Southern California Air Pollution Control District, Metropolitan
Zone (formerly L.A. District) is believed pertinent. Cutbacks are
-554-
-------
forbidden to be used in that District under the Volatiles Rules, Rule 442
(Usage of Solvents) and Rule 443 (Labeling of Solvents). These rules
are formerly known as Rule 66.
^555--
-------
TECHNICAL REPORT DATA
(Please read Initructiaits on the reverse before completing)
1. REPORT NO.
EPA-600/7-77-148
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Second Symposium on Fugitive Emissions:
Measurement and Control (May 1977, Houston, Texas)
5. REPORT DATE
December 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
J. King, Compiler
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
TRC--The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
10. PROGRAM ELEMENT NO.
INE624
11. CONTRACT/GRANT NO.
68-02-2133
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings; 4-10/77
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES project officer StatnicK is no longer
contact D. Bruce Harris, Mail Drop 62, 919/541-2557.
with IERL-RTP; for details
16. ABSTRACT
The proceedings are a compilation of technical papers prepared for presen-
tation at the Second Symposium on Fugitive Emissions, May 23-25, 1977, Houston,
Texas. The papers discuss the scope and impact of fugitive emissions (non-point
sources) and present techniques which have been used the measure these emissions.
Fugitive emissions control technologies are also discussed.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Emission
Industrial Processes
Dust
Water
Measurement
Pollution Control
Stationary Sources
Fugitive Emissions
Non-point Sources
Particulates
13B
13H
11G
07B
14B
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
560
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73) -556-
------- |