EPA-600/2-76-246
September 1976
Environmental Protection Technology Series
SYMPOSIUM ON FUGITIVE EMISSIONS
MEASUREMENT AND CONTROL
(May 1976, Hartford, CT)
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five 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 five series are:
s
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products, constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-246
September 1976
SYMPOSIUM
ON FUGITIVE EMISSIONS
MEASUREMENT AND CONTROL
(May 1976, Hartford, CT)
E.M. Helming, Compiler
TRC, The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
Contract No. 68-02-2110
ROAP No. 21AUY-095
Program Element No. 1AB015
EPA Project Officer: Robert M. Statnick
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
FOREWORD
These proceedings for the symposium on "Fugitive Emissions:
Measurement and Control" constitute the final report sub-
mitted to the Industrial Environmental Research Laboratory
for Task Three of Environmental Protection Agency Contract
Number 68-02-2110. The symposium was held at the Sheraton-
Hartford Hotel, Hartford, Connecticut, May 17-19, 1976.
The objective of the symposium was to support the Environ-
mental Protection Agency's efforts to develop methods for
the measurement and control of industrial fugitive emissions.
Papers were presented that described the scope and regula-
tory aspects of fugitive emissions in general, reviewed the
impact of fugitive dusts on the atmosphere and indicated
the need for future measurement and control programs. Re-
cently developed measurement methods for air borne fugitive
emissions and the results of measurement programs utilizing
the methods in specific industrial applications were des-
cribed. Existing fugitive emissions control technologies
in typical industries were reviewed.
Dr. Robert M. Statnick of the Industrial Environmental Re-
Search Laboratory, Environmental Protection Agency, Research
Triangle Park, North Carolina, was the Project Officer and
General Chairman of the symposium.
Elizabeth M. Helming, Project Scientist at TRC - The Research
Corporation of New England, Wethersfield, Connecticut, was
the Symposium Coordinator and Compiler of the proceedings.
11
-------
TABLE OF CONTENTS
(* indicates speaker)
Page
17 May 1976
SESSION I: INTRODUCTION
FUGITIVE EMISSIONS PROBLEMS IN PERSPECTIVE 3
John E. Yocom, TRC
REGULATORY ASPECT OF FUGITIVE EMISSIONS 17
Gary D. McCutchen, EPA
SESSION II: MEASUREMENT
A GUIDELINE FOR THE MEASUREMENT OF AIR-BORNE FUGITIVE 33
EMISSIONS FROM INDUSTRIAL SOURCES
Henry J. Kolnsberg, TRC
COKE OVEN EMISSION MEASUREMENTS DURING PUSHING . . .51
Robert B. Jacko, Ph.D., Purdue University
PROBLEMS IN MEASURING FUGITIVE EMISSIONS FROM. 67
WASTE DISPOSAL PONDS
William R. King, Ph.D., FMC Corp.
CONTINUOUS ROOF MONITOR EMISSION TESTS 101
Abbas F. Souka, Ph.D., AIRCO Speer Carbon-Graphite
18 May 1976
SESSION III: IMPACT OF
FUGITIVE EMISSIONS
RELATIVE IMPACTS OF OPEN SOURCES OF EMISSIONS 123
Thomas R. Blackwood, Ph.D.* and
J. A. Peters, Monsanto Research Corp.
THE IMPACT OF FUGITIVE EMISSIONS OF FINE PARTICLES 143
Chatten Cowherd, Ph.D., Midwest Research Institute
iii
-------
TABLE OF CONTENTS (cont'd.)
Page
FACTORS INFLUENCING EMISSIONS FROM FUGITIVE . .159
DUST SOURCES
George A. Jutze and
Kenneth Axetell*, PEDCo Environmental
SESSION IV: CONTROL TECHNOLOGY
STUDY OF THE EFFECT OF ASBESTOS WASTE PILES ON AMBIENT AIR. . . .183
Colin F. Harwood, Ph.D.* and
Paul Ase, IIT Research Institute and
Mary Stinson, EPA
AN ASSESSMENT OF FUGITIVE EMISSIONS IN THE PRIMARY. ...... .203
ALUMINUM INDUSTRY
William D. Balgord, Ph.D., The Aluminum Association
MEASUREMENT OF IRON FOUNDRY FUGITIVE EMISSIONS .211
William D. Scott* and
Charles E. Bates, Ph.D., Southern Research Institute
CONTROL OF FUGITIVE EMISSIONS IN PETROLEUM REFINING 239
John H. Weiland, Texaco, Inc.
(Representing the American Petroleum Institute)
THE COST EFFECTIVE'NESS OF COKE OVEN CONTROL TECHNOLOGY 247
William A. Cote*, Robert E. Kenson, Ph.D. and
Norman E. Bowne, TRC
19 May 1976
SESSION V: FUTURE NEEDS FOR MEASURE-
MENT AND CONTROL TECHNOLOGY
FUTURE NEEDS FOR'MEASUREMENT AND CONTROL OF FUGITIVE DUST . . . ,267
Frederick A. Renninger, National Crushed Stone Assoc.
DETERMINING FUGITIVE EMISSIONS MEASUREMENTS NEEDS FOR AN 275
EMERGING INDUSTRY-ADVANCED FOSSIL FUEL UTILIZATION
Michael R. Guerin*, James E. Epler, Chuen-huei Ho, and
• . Bruce R. Clark, Oak Ridge National Laboratory
NONPOINT SOURCE WATER EMISSIONS: ENERGY AND INDUSTRY 305
PROCESSES
Robert M. Statnick, Ph.D.*, EPA and
Gordon T. Brookman, TRC
IV
-------
Session I:
INTRODUCTION
-1-
-------
Fugitive Emissions Problems in Perspective*
by J. E. Yocom
TRC - The Research Corporation of New England
Wethersfield, Connecticut
ABSTRACT
Primary emphasis on control of air and water pollutants has been
placed on point sources. Fugitive emissions, or those emitted from non-
point sources, are important contributors to environmental degredation in
many areas and in relation to many types of industrial operations. Because
of the potentially high cost of controlling such emissions it is important
that their significance be accurately assessed.
This paper discusses many types of fugitive emissions and the methods
for assessing them. Fugitive emissions are extremely site specific in
respect to their measurement and control, and this paper presents examples
of measurement programs that put fugitive emissions in proper perspective
in relation to other source categories and other environmental impacts.
For Presentation at a Symposium on "Fugitive Emissions - Measurement and
Control". Sponsored by the Environmental Protection Agency (IERL-RTP),
Sheraton-Hartford Hotel, Hartford, Conn., May 17-19, 1976.
-3-
-------
1.0 INTRODUCTION
The pollution control programs currently being developed and implemented
in this country have, for the most part, been directed at emission sources that
release pollutants into the environment from well defined points. These emis-
sion points are represented by stacks and ducts emitting air pollutants; and
pipes, culverts, and channels discharging water pollutants. Until recently
little attention has been given to fugitive emissions, or emissions that enter
the environment through other, ill-defined routes.
This brings up the question of terminology. It is easier to say what
fugitive emissions are not then what they are. The term "fugitive" is not
descriptive. It implies pollutants that have escaped and must be captured,
brought to justice, and incarcerated. But in considering fugitive air pollu-
tants that have already escaped from a roof monitor or an outdoor material
handling pile or fugitive water pollutants that have already entered a natural
stream or the ground water, it is out of the question technologically and eco-
nomically to attempt capturing and controlling them at this point in their
escape route. There are countless types of fugitive emissions and all of us
could provide many examples related to type of pollutant, mechanism of release,
and behavior in the environment. When we consider this class of emissions we
immediately recognize that each is extremely source and site specific and the
assessment and control of such emissions must recognize these fundamental
factors. While I do not advocate a change in the title of this conference,
nor will I stop using the term "fugitive emission", I believe that we could
come up with a more descriptive term for the type of pollutant emission we
will be discussing. At the moment I would vote for "non-point source emissions",
a term currently used by the water pollution control people.
The strategies for control efforts on point sources have been based upon
-------
a logical progession starting with quantifying the deleterious effects of
pollutants on human health and ecological systems, followed by prioritization
of sources to be controlled based upon the quantities of important pollutants
released. The programs for point source control are well advanced, largely
because such releases are already or have traditionally been released from
well defined conduits and uncontrolled emissions from these points often repre-
sent the largest potential emission from a process. Furthermore, such emissions
can be accurately measured using well established techniques.
In spite of this emphasis on point source control, EPA and the industrial
community have long recognized that fugitive emissions from certain types of
processes can be important contributors to environmental degredation. Many
obvious fugitive sources have already received considerable attention and con-
trol efforts have been implemented or initiated; for example, fugitive air
emissions from coke ovens, and non-point source emissions from abandoned coal
mines in the form of acid mine drainage. As point sources are controlled more
stringently, fugitive emissions (even those that are not now obvious) will be-
come increasingly important in determining environmental quality. The princi-
pal deterrents to the control of fugitive emissions have been essentially in
two areas:
1. Difficulties in measuring fugitive emission rates and
thereby assessing their impact on the environment.
2. The generally high costs of controlling such emissions
by consolidating them through process modification or
capturing them for removal in emission control systems.
2.0 IMPORTANCE OF FUGITIVE EMISSIONS
How important are fugitive emissions? Those living adjacent to an active
coal storage area on a windy day might tell you that the living conditions are
intolerable. Ecologists concerned about eutrophication of small lakes in rural
-------
areas will tell you that fugitive or non-point sources of nutrients are the
culprits. On the other hand, an industrialist faced with controlling fine
particulate emissions leaking from furnace buildings might tend to look at
these fugitive emissions as only an aesthetic problem. This is perhaps under-
standable since he might have already spent several million dollars in enclos-
ing furnaces and installing air pollution control systems and the leakage,
which he considers unavoidable based upon present control technology, is
thought to be a small part of the problem.
In view of the tendency for many fugitive emissions to take on an element
of emotionalism there is a need to develop some perspective about their rela-
tive importance. First of all, it is difficult to generalize. As stated
earlier, each fugitive emission problem is different and tends to be quite
site specific. One aspect that makes air fugitive emissions especially impor-
tant is the tendency for such emissions to be emitted over a broad frontal
area and at ground level. This means that fence line effects tend to be sig-
nificant over much of the plant's perimeter. Furthermore, except for large
particles which settle out rapidly as they move away from the source, concen-
trations of fugitive emissions do not fall as rapidly with distance as do
emissions from tall stacks. In general, ground level concentrations of pollu-
tants from tall stacks decrease with the square of the distance from the source
since such emissions are able to disperse both vertically and laterally. On
the other hand, ground level concentrations of pollutants emitted from low-
level fugitive sources covering a broad frontal area tend to be related to
the first power of distance from the source. This results from the inability
of the plume to disperse downward and the lack of significant dilution later-
ally because of the breadth of the plume.
In short, fugitive emissions as a class are important but their relative
importance in specific situations depends upon a variety of factors.
-6-
-------
3.0 SOURCES OF FUGITIVE EMISSIONS
While fugitive emissions are site specific, there are certain types of
situations that provide maximum potential for their release.
1. Open Operations
Such operations as mining, quarrying, outdoor materials
handling, and coke making produce fugitive air emissions
containing both particulates and gaseous pollutants.
Fugitive water emissions also emanate from such opera-
tions, for example, acid mine drainage from coal mining,
and surface runoff from all such outdoor operations.
2. Leaks and japills
Leaky piping, pump and valve glands, and loss of material
from spills are important sources of fugitive air and
water emissions. Gases and high vapor pressure liquids
that leak to the atmosphere are sources of air fugitive
emissions. Liquids leaking to the ground are capable of
finding their way into surface and ground waters.
3« Storage and Disposal of Materials and Wastes
Material storage piles (both worked and quiescent) and
dried sludge beds are sources of atmospheric particulate
matter. Sludge disposal beds are sources of contaminated
runoff and can also degrade ground water. Disposal ponds
containing volatile materials are sources of fugitive air
emissions.
4. Incompletely Controlled Point Sources
While a control device may adequately control the basic
process emissions from a source, certain operating con-
ditions, which may be planned or unplanned, create emis-
sions that the control system cannot adequately handle.
In the case of air pollutants an example would be an
electric steel furnace whose basic emissions are control-
led by hooding or furnace evacuation followed by a high
efficiency bag house, but during charging and tapping
the fumes are not captured and they escape through the
roof monitor. An example in the water area would be an
aeration or stabilization lagoon that releases fugitive
water pollutants during heavy rains when the capacity of
the system is exceeded.
5. Poor Housekeeping
Accumulations of materials that are likely to be
carried into the environment by the winds or rain water
are important sources of fugitive emissions. In render-
ing plants the accumulation of putrescible materials is
a significant source of fugitive odors.
-7-
-------
In some industrial processes examples can be found for many of these
categories of sources. Figure 1 shows a simplified schematic diagram of a
typical iron foundry. Even this relatively simple operation has many sources
of fugitive air emissions. As part of an EPA study, TRC identified a group
of industrial processes that have significant potential for fugitive air
emissions. These are shown in Table 1.
4.0 METHODS OF MEASUREMENT
A very simple method for categorizing fugitive emissions is by means
of the feasible methods of measuring them. In our work for EPA we have
evolved three basic approaches to the measurement of atmospheric fugitive
emissions. These will be discussed in greater detail by Hank Kolnsberg.
4.1 Quasi-Stack
In this method the fugitive emission is temporarily hooded or encapsulated
and a temporary duct or stack and fan is installed on the duct to permit sam-
pling by means of standard stack or duct sampling methods.
As an illustration of how the quasi-stack method is used to assess fugi-
tive emissions, we present Figure 2 from a TRC study. This test configuration
was set up in a ship building establishment to assess welding emissions. There
were three basic objectives of the study:
1. Determine indoor concentrations of important particulate
and gaseous releases.
2. Utilize the data to design a ventilation system.
3. Assess impact of fugitive and process emissions on the
surroundings.
Emission factors for each of 10 welding configurations were developed in
terms of weight rate of pollutant (particulates, metals, pollutant gases) per
pound of a welding rod. The reproducibility of results from duplicate samples
was excellent and since the plant maintained accurate records on amount and
-8-
-------
i
vo
i
Binder
SAND
PREPARATION
i.' Gas and
I" particulate
,f emissions
Dust
FINISHING
AND
SHIPPING
CASTING
SHAKEOUT
Core sand
and binder
Cores vC f-
COOLING AND
CLEANING
CORE
MAKING
Figure 1. Iron foundry process flow; sources of emissions.
-------
Table I. Classification of Industries With High
Potential for Fugitive Emissions
Category
Metallurgical
Industry/Pro cess
Primary Aluminum
Primary Copper
Steel Making
Iron & Steel Foundries
Coke Making
Major Fugitive Emissions
Fume,Fluorides, PNA*
Fume, S02, Dust
Fume, Kish, CO,Odors
Fume, Odors
Smoke & Fume, Hydrocarbon
Gases & Vapors, Odors,
CO, PNA
Fjiergy/Fuels
Coal Mining & Processing
Coal Gasification
Char-Oil-Gas
Shale Oil
Petroleum Refining
Oil Production
Dust
Hydrocarbons, Smoke &
Fumes, CO, PNA
Hydrocarbon Gases &
Vapors, CO, PNA
Fumes & Dust, Hydrocarbon
Gases & Vapors, CO, PNA
Hydrocarbon Gases &
Vapors, Odors, PNA
Hydrocarbon Gases &
Vapors, Odors
Chemical Products Plastics
Tire & Rubber
Hydrocarbon Gases &
Vapors, Odors
Hydrocarbon Gases &
Vapors, Odors
Rock Products
Phosphate Fertilizer
Lime
Sand & Gravel
Asphalt Batching
Dust, Fluorides, S02
Dust
Dust
Dust, Odors, PNA
Other
Agricultural Operations
*PNA - Polynuclear Aromatics
Dust
-10-
-------
Existing welding booth in
welder training school
Opacity
Meter
Light source-
,To
(.pump
Photocell
Particulate
filter
Inclined
draft
gage
To
blower
3 l/. ID smooth walled
aluminum tube,
10' long
Figure 2. Schematic diagram of test configuration for field
measurements of welding emissions.
-------
type of welding rod used, we were able to provide an accurate estimate
(probably ± 25%) of plant emissions. Furthermore, we were able to assess the
impact of changes in the quantity and mix of welding types. This approach,
using the quasi-stack method and computing emissions, was considered to be far
more accurate and cost-effective than sampling in the building openings. By
application of this method, we were able to place the impact of fugitive emis-
sions on the outdoor atmosphere in perspective as compared with the effect of
the emissions on the indoor environment.
Another advantage of the quasi-stack method is that it often can provide
data on the design of a control system, since sources of this type are most
often controlled by installing hoods, ducts, control systems and fans.
4.2 Roof Monitor
The term "roof monitor" denotes the generalized situation where there are
open sources inside of a building and the fugitive emissions are to be measured
at building openings, including the roof monitor, windows and doors, and venti-
lation openings. This method is applied when sources inside the building are
too numerous, too large, or too inaccessible to permit the use of the quasi-
stack method. This method tends to be less accurate than the quasi-stack
method since it is necessary to produce an accurate air and material balance
on all building openings. Since we will be hearing about an example of a
roof monitor measuring program and since Hank Kolnsberg will be giving further
details on our methods development program, I will not discuss this method
further.
4.3 Upwin d-Downwind
This term describes the general approach of sampling a fugitive emission
in the free atmosphere. The most common application is that of sampling simul-
taneously upwind and downwind of sources of fugitive emissions and using the
difference in sampling values as an indication of source contribution. This
-12-
-------
generalized method would also encompass sampling directly in free plumes.
We will hear more about this aspect from Bob Jacko.
The application of the upwind-downwind method is not straightforward and
carrying out such measurements and interpreting the data must be done with
great care. In general, a pair of simple measurements at ground level upwind
and downwind of a fugitive source will indicate at best only that there is a
source between the sampling sites. It will not provide even an estimate of
emission levels.
The sophisticated route is to utilize three-dimensional sampling and wind
flux arrays upwind and downwind of the source. A good cost-effective compro-
mise using ground level monitors can be used but it requires supplementary
data gathering efforts including
1. Accurate emission inventories for point sources.
2. Detailed knowledge of the chemical and physical
properties of point and fugitive emissions and
the collected pollutants.
3. Mathematical models that take into consideration
the decay, deposition or reaction of emitted
pollutants.
4. Tracer studies to calibrate the models.
Table II shows the results of a study that pieces together from the above
elements an assessment of the relative contribution of fugitive and point source
emissions of a mineral based operation in relation to background particulate
levels. Such an analysis puts the problem of plant fugitive emissions in per-
spective, and, in this case, the results of the study showed that control of
fugitive emissions was a more cost-effective method of air quality control than
further control of point sources.
-13-
-------
Table II. Distribution of Particulate Matter from Process Emissions
and Fugitive Dust at Several Downwind Sampling Points
Distance from Edge of Plant Property, mi
Process Contribution1 pg/m3
Plant Fugitive Dust2 vg/m3
Background, yg/n3
TOTAL yg/m
0
126 (37%)
170 (49%)
493(14%)
345
0.8
26 (18%)
40 (27%)
80'4(55%)
146
1.4
12 (10%)
23 (20%)
80 "(70%)
115
4.8
1.6 (2%)
2.4 (3%)
70 (95%)
74
'•Based on source inventory and diffusion modelling.
Calculated by difference.
3Based on upwind-downwind sampling at plant.
4Based on data for winds other than from direction of plant.
5Air quality data for downwind periods.
-------
5.0 CONCLUSIONS
Fugitive emissions are a complex and challenging category of sources that
have important impact on the environment. Their nature and relative importance
are exceedingly site specific and reliable methods of measuring such emissions
are the key to putting their relative importance into perspective. Control
measures are closely related to the processes generating the fugitive emissions
and are likewise site specific. Since fugitive emission control can be costly,
there is a great need for measurements for assessing the importance of the emis-
sions and the effectiveness of control systems.
-15-
-------
REGULATORY ASPECTS OF FUGITIVE EMISSIONS
Gary McCutchen
Emission Standards and Engineering Division
Environmental Protection Agency
Research Triangle Park, NC
May 17, 1976
In a June 1974 letter to Oregon's Department of Environmental
Quality concerning emission standards, the Asphalt Pavement Assoc-
iation of Oregon concluded with this statement: "If God wanted
perfectly clean air he would have made the earth out of asphalt
not dirt and he would not have created fire."
There are a few problems with this concept, but the writer
does have a point. Where there's fire, there's smoke; and where
there's dirt, there's dust. Windblown dust can comprise as much
as 90 percent of total ambient particulate concentrations, ^
agricultural open burning is by far the largest stationary source
of both particulate and hydrocarbon emissions out of 200 source
(2\
categories evaluated in a recent study, ' and industrial fugi-
tive emissions may in many cases exceed controlled stack emissions
from the process. What, then, are fugitive emissions, how much
of a problem do they represent, and what are EPA and other agencies
doing to regulate them?
To begin with, it is necessary to define what is meant by the
term fugitive emissions. Probably the most widely used definition
is that they consist of air pollution emissions which have not pass-
ed through a stack or duct. The distinction is a fine one, since
S0£ from a culm pile is considered fugitive, while S02 passing un-
controlled through an electrostatic precipitator is not. Li His
-17-
-------
and Young^ define two separate problems: industrial fugitive
emissions (gaseous and particulate emissions that result from
industrial related operations such as metallurgical furnaces
and materials transfer and storage and which escape to the atmos-
phere through windows, doors vents, etc.) and fugitive dust emis-
sions (dust storms and windblown particulate from unpaved roads,
tilled land, etc.). In discussing regulatory aspects, it is con-
venient to further define fugitive emissions on the basis of con-
trollability, as:
1. Confined - fugitive emissions which were initially contain-
ed in a duct, hood, building, or other facility, but which escape
to the atmosphere without passing through a control device or stack,
and
2. Confinable - fugitive emissions from storage piles, open
conveyor belts, material transfer operations and other sources, such
as coke ovens, which escape directly to the atmosphere, but for which
confinement or other control measures are possible.
Fugitive dust emissions constitute a third category, unconfinable,
since they originate from large tracts of tilled farm land, dust storms,
active volcanos, and other sources for which control measures are usually
unavailable, unjustifiable, or too costly. There is some overlap in
these categories. Paving dirt roads, for example, may be feasible in
one area and unjustifiable in another, so unpaved roads may or may not
be considered "unconfinable." In the context of this paper, the term
fugitive emissions has the same definition used by Lillis and Young
and refers to the first two of the three controllability categories:
--18-
-------
confined and con-finable emissions.
The Problem
Fugitive emissions constitute an increasingly important problem
in air pollution control strategy. Limited but persuasive data im-
plicate fugitive emissions as a major proportion of total emissions,
especially where stack emissions are reasonably well controlled.
Fugitive emissions of particulate during electric arc furnace charg-
ing, for example, are estimated to be from five to fifty times well-
controlled stack emissions from the furnace on a pounds per ton basis,
and tests at a lead smelter measured three times as much lead per
hour emitted from fugitive sources as from the controlled stack emis-
sions.<3>
Preliminary estimates indicate that 128 of the 247 Air Quality
Control Regions (AQCR's), 52 percent, are not expected to attain the
total suspended particulate (TSP) national ambient air quality stan-
dard (NAAQS) in 1976: ' Additional control of stack emissions will
be adequate for some, but not all, AQCR's. Assume that fugitive
emissions from a process equal two percent of total uncontrolled
emissions from that operation; if the captured emissions are treated
with a 98 percent efficient control device, then the amount of fugi-
tive emissions equals stack emissions. Also, the fugitive emissions
would likely have a greater local effect on ambient air pollutant
concentrations since they generally are emitted nearer ground level
with less vertical velocity than are stack emissions. Ambient air
measurements at a coke oven plant, for example, indicate that emissions
from this facility increase downwind TSP levels by 200 yg/m3 half a
-19-
-------
(8)
mile from the facility and by TOO yg/m3 a mile away.
The impact of fugitive emissions on air quality over a large
area has recently been estimated. Over 30 percent of all industrial-
(5)
ized urban monitoring sites analyzed in a recent study are in-
fluenced by particulate fugitive emissions; they average on an annual
basis 25 yg/m3 TSP higher than the other industrial sites. Fugitive
emissions in heavily industrialized cities are tentatively estimated
to increase annual citywide ambient TSP levels by roughly 10 v»g/m3.
These results are from 20 sites in five cities.
Many sources of hydrocarbon emissions are essentially fugitive
emission sources. The magnitude of these emissions is difficult to
determine, but the Office of Air Quality Planning and Standards (OAQPS)
has a task force studying the problem and establishing model regula-
tions. Oxidants, for which hydrocarbons are precursors, are the se-
cond most serious nonattainment problem (Table 1).
Regulation Under SIP
During the development of SIPs, states quantified emissions from
all sources using the best information available to them. Few fugi-
tive emission sources were specified in these plans, primarily due to
the magnitude of stack emissions from poorly controlled sources and
the lack of even rough emission estimates for most fugitive emission
sources. Even such general regulations as process weight curves are
/
usually applied only to stacks, not total emissions. This probably
is a direct reflection of the extreme costs, difficulties, and un-
certainties involved in testing most fugitive emission sources. No
widely accepted test methods exist, and testing wasn't even being
attempted on most of these sources until very recently.
-20-
-------
Many agencies did, however, utilize general regulations to address
the fugitive emissions problem. An analysis of all 50 state and 26 major
local gency regulations showed that 63 percent require that reasonable
precautions be taken to prevent fugitive emissions; 5 percent have nuisance
provisions; 14 percent limit property line concentration, particle size.,
or fallout' 11 percent regulate specific sources; and 7 percent have no
fugitive emission regulations/ ' Compliance is a matter of following
equipment specifications and approved operating procedures, meeting a
visible emission limitation, or meeting an ambient air property line fall-
out or concentration requirement (Table 2).
A year-long revision of SIPs, in areas where a need for revisions is
indicated, begins July 1976. The initial SIPs, as Figure 1 indicates, when
fully implemented will do a remarkable job of reducing process emissions.
Particulate emissions shown are those projected for 1985. The top line
represents uncontrolled emissions; the next delineates emissions with present
(1975) SIP requirements; the third represents a combination of 1975 SIPs
and best control of all post-1975 sources; the final line is a lower limit
representing zero emissions from all post-1975 capacity. This 1985 profile
shows that SIPs provide the major impact on particulate process emissions
(including combustion). Many of the SIP revisions in non-attainment areas,
therefore, are expected to concern fugitive emissions.
Several EPA projects are underway to assist states in this effort.
One is a report on the best estimates of both fugitive and stack emissions
from all iron and steel mill operations, including windblown dust from
storage piles. These factors are being formulated by a joint EPA-AISI
-21-
-------
task force formed in January 1976 to determine whether fugitive emissions
are, as has been claimed by industry, essentially "cosmetic." Finalized
factors are expected in mid-1976, but longer-range task force projects include
acceptable fugitive emission test methods, tests of specific fugitive
emission sources (at least partially through existing EPA-ORD contract
studies), and dispersion modeling a medium size well-controlled integrated
steel mill to determine ambient air impact.
Guidance on other sources is expected in an EPA report due January 1977
which will summarize all existing fugitive emission information for the states'
use. A similar study on fugitive dust is expected in July 1976. The reports
will include model regulations. The present control strategy recommended
by EPA is an equipment standard which uses visible emissions as an enforcement
tool.'1)
Regulation Under NSPS and NESHAPS
New Source Performance Standards (NSPS) are applicable to new or
modified stationary sources under Section 111 of the Clean Air Act (CAA),
and have been directed principally at stack emissions. Figure 2, which
is drawn by expanding the Figure 1 ordinate, permits an evaluation of the
impact of NSPS on total 1985 particulate emissions. As Figure 2 shows, the
impact - with the exception of the power plant standard - is hardly noticable
in terms of national emissions. This is at least partially due to the strin-
gent SIPs, which lessen the impact of an NSPS representing best control.
For other criteria pollutants, NSPS impact is much more dramatic,
and present policy emphasizes a need for hydrocarbon and nitrogen oxide
NSPS, a decision supported by growing concern over widespread non-attainment
-22-
-------
of the oxidant NAAQS, an anticipated rapid growth in nitrogen oxide
emissions, and a growing realization that both of these pollutants
represent difficult control technology/enforcement problems.
Increasing interest in and information on fugitive emissions
has refocused attention on particulate sources and will affect
NSPS priorities. It will also intensify efforts to establish
NSPS for total source emissions, both stack and fugitive.
Promulgated NSPS have usually included at least a visible emission
limitation applicable to both fugitive and stack emissions. A few
have actually established mass limitations, such as the fluoride emission
NSPS for primary aluminum production, which applies to both stack and
roof monitor emissions and specifies test methods for both emission
points. One serious limitation to NSPS for fugitive emissions is the
lack of specific authority to promulgate equipment standards. In certain
cases, such standards would be by far the most efficacious from the view-
point of both industry and control agencies.
Addressing fugitive emissions more directly than do NSPS, national
emission standards for hazardous air pollutants (NESHAPS) under Section
112 of the CAA require installation of control equipment with specific
performance characteristics and/or require implementation of certain
operating and maintenance practices. The asbestos NESHAPS, for example,
specifies control equipment parameters and mandates how asbestos tailings
piles are to be operated. The regulation also requires fencing (or
natural barriers) and even describes warning signs to be posted. The
mercury and beryllium NESHAPS utilize similar approaches.
-23-
-------
Enforcement
EPA's Division of Stationary Source Enforcement (DSSE) has identified
fugitive emissions as a major compliance issue, arid has initiated projects
to quantify total emissions from such difficult-to-measure sources as
coke oven battery coke-side leaks and pushing emissions^ and quench
tower emissions/8' These and other studies planned or underway in EPA's
Office of Research and Development and other groups will help quantify
the seriousness of fugitive emissions from many sources. Fugitive emission
sources in the 128 AQCRs where ambient TSP standards are not being attained
are of special concern to DSSE. Labeling these emissions as too difficult
to assess, insignificant, or cosmetic is insufficient; efforts to control
fugitive emissions will, in nearly all cases, be expected.
Until much less elaborate and costly test methodology is available,
opacity will remain the primary fugitive emission enforcement too!.
Opacity tests (such as the NSPS Method 9) have been upheld in court liti-
gation, give nearly instantaneous results, require only one or two qualified
observers, and eliminate the need for test equipment and laboratory analysis.
Use of visible emission standards is by far the most practical enforcement
approach in terms of minimizing both control agency and industry testing
resources.
It is surprising, then, that court action continues against SIP
visible emission standards and that a recently proposed CAA amendment^9'
would preclude EPA enforcement of such SIP regulations. The alternative,
for emission sources, is much worse: EPA under Section 114 of the CAA
can, among other things, require an owner or operator to sample emissions
-24-
-------
(as prescribed by EPA) and install, use, and maintain monitoring equipment
to determine whether a source is in violation of Section 110, 111, or
112 standards (or even to aid in developing such standards). Such require-
ments have to date been minimal only because visible emission regulations
provided an equivalent, less costly, alternative.
Summary
In conclusion, EPA's regulatory offices are becoming increasingly
concerned about fugitive emissions and their impact on ambient air pollutant
concentrations. Quantification and regulation of these emissions will
likely constitute a major EPA effort during the next five years. The most
practical approach appears to consist of equipment standards enforced
by visual observation; present sampling technology is too difficult and
costly as long as opacity regulations remain a viable option.
-25-
-------
TABLE 1. PRELIMINARY ASSESSMENT OF NUMBER AND PERCENT OF
AQCR's WHICH MAY NOT ATTAIN NAAQS IN 1976(4)
Pollutant Number Percent
•
TSP 128 52
S02 27 11
CO 48 19
Ox 65 26
NOX 13 5
-26-
-------
FIGURE 1. Projected 1985 Participate Emissions Under Various
Control Strategies
(2,6)
(O
(U
c
o
10
c
o
in
oo
CTi
c
o
•^
IA
fa
*3
o
(U
X)
o
150-
140
130
120
.110
100
90
80
70
60
50
40
30
20
10
0
/ / / State Standards Only
Best 19J_5 Technojpgy vQn_AlJ...Ppst-1975 So.urces, _
r -----•--«,-• ~*."~" ^s. ~~\ '"•v ~ . • T
Zero Emissions From All Post-1975 Sources
j_
8 10 12
Number of NSPS
14 16 1R 20
-27-
-------
FIGURE 2. Estimated Impact of NSPS Over a Ten-Year Period
on 1985 National Participate Process Emissions
(2,6)
(O
o>
t/i
c
o
c
o
s:
ur>
CO
01
c
o
in
l/l
01
+J
(O
*3
O
O.
C
O
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
State Standards Only
o--o-
Best 1975 Technoloqy On All Post-1975 Sources
\ N
\
\\v
Zero Emissions From All Post-1975 Sources
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Steam Generators (>250 MM Btu/hr)
Municipal Incinerators '
Portland Cement Plants
Asphalt Batch Plants
Petroleum Refineries - FCCU
Secondary Lead Smelters
Secondary Brass and Bronze
Iron and Steel, Basic Oxygen Furnaces
Sludae Incinerators
Primary Copper Smelters
Primary Zinc Smelters
Primary Lead Smelters
Coal Cleaning Plants
Iron and Steel: Electric Arc Furnaces
Ferroalloy Production
Kraft Pulp Mills
Coke Ovens
Grain Terminals
Boilers (Coal and Refuse)
Phosphate Rock Preparation
10 12 14 16
18 20
Number of Standards
-28-
-------
TABLE 2. PROPERTY LINE STANDARDS
(.5)
Jurisdiction
Standards
Hawai i
Kansas
Texas
Omaha
Missouri
Mississippi
Nevada
Illinois
Cleveland
150 yg/m3, above upwind concentration, 12-hour average
3.0 g/m2, fallout above upwind concentration, 14 day period
2.0 mg/m3, above background concentration, 60 minute average
100 yg/m3, 5-hour average
200 yg/m3, 3-hour average
400 yg/m3, 1-hour average
500 yg/m3, 60 minute average
80 yg/m3, 6 month geometric mean
200 yg/m3, 2-hour arithmetic mean
0.4 coh/1000 linear feet soiling index, 6 month geometric mean
1.0 coh/1000 linear feet soiling index, 8-hour arithmetic mean
>40 ym prohibited
5.25 g/m2, fallout above background
2 tons/mi2, 24-hour period
>40 ym prohibited
500 yg/m2, 60-minute average
-29-
-------
REFERENCES
1. Lillis, E.J., and D. Young, "EPA Looks at 'Fugitive Emissions,1" JAPCA,
25: 1015 (1975).
2. T.G. Hopper and W.A. Marrone, "Impact of New Source Performance Standards
on 1985 National Emissions from Stationary Sources." [Report prepared for
LPA under Contract 68-02-1382, by the Research Corporation (TRC) of New
England, 125 Silas Dean Highway, Wethersfield, Conn. 06109, Oct. 1975.]
This study includes 1400 pages of detailed referenced calculation sheets
yielding information on growth and obsolescence rates, emission levels, and
production t'rends and capacities.
3. Fugitive emission results were from PEDCo's Silver Valley /Bunker Hill Smelter
Environmental Investigation for EPA's Region X, (Febr. 1975). Stack test results
were from a telephone conversation with Mark Hooper, Region X, EPA.
4. Conversation with Henry C. Thomas, Control Programs Development Division,
Office of Air Quality Planning and Standards, Environmental Protection Agency,
May 6, 1976.
5. From information supplied by Tom G. Pace, Control Programs Operations Branch,
Control Programs Development Division, OAQPS, EPA. These data are from a study
being conducted for EPA by GCA Technology Division: National Assessment of the
Parti on ate Problem, Draft Final Report, April 1976.
6. ."Priorities and Procedures for the Development of Standards of Performance
for New Stationary Sources of Atmospheric Emissions," report prepared under
contract by Argonne National Laboratory, EPA Contract No. EPA-IAG-D4-0463,
Project No. 2,'April 1975.
7. Telephone conversation with Louis Paley, DSSE, EPA, May 5, 1976.
8. Conversation with Carl Edlund, DSSE, EPA, February 3, 1976.
9. Amendment No. 1597, proposed by Senator Randolph to S.3219, April 13, 1976.
-30-
-------
Session II:
MEASUREMENT
James A. Dorsey
Session Chairman
-31-
-------
A GUIDELINE FOR THE MEASUREMENT OF AIR-BORNE
FUGITIVE EMISSIONS FROM INDUSTRIAL SOURCES
Henry J. Kolnsberg
TRC - The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, CT 06109
Presented at the Symposium on
Fugitive Emissions: Measurement and Control
Sponsored by
Industrial Environmental Research Laboratory
Environmental Protection Agency
Research Triangle Park, North Carolina
May 1976
-33-
-------
ABSTRACT:
The paper presents a guide for the selection of the most
effective program for the measurement of air-borne fugitive
emissions from an industrial source.
The quasi-stack, roof monitor and upwind-downwind
techniques presently utilized for sampling a wide variety of
air-borne pollutants are described.
General criteria for the selection of the most effective
sampling program, relative to characteristics of the site,
process and emissions are discussed.
Baseline estimates of manpower, time and cost require-
ments for typical measurement programs for each technique are
provided.
-34-
-------
1.0 INTRODUCTION
The term "fugitive emissions" may be applied to any gaseous or
particulate pollutant entering the ambient atmosphere without first
passing through a stack, duct or other device designed to direct or
control its flow.
Industrial fugitive emission sources are, in general, complex com-
binations of a process or operation and a physical location or arrange-
ment, and are so varied as to make each source essentially unique. All
fugitive emission sources share one common characteristic in that stan-
dard measurement techniques such as stack sampling are useless for
determining the amount or nature of pollutant materials being transmitted
into the surrounding atmosphere.
This paper describes the three methods recognized as effective mea-
surement techniques for identifying and quantifying pollutants from almost
any industrial source. The methods, identified by their basic sampling
schemes, are the quasi-stack, roof monitor, and upwind-downwind techniques..
General criteria for consideration in the selection of the most effective
technique relative to characteristics of the process, its site, and its
emissions are presented. Baseline estimates of manpower requirements
and costs for the performance of typical measurement programs for each
method are included.
2.0 MEASUREMENT TECHNIQUES
Measurements of air borne industrial fugitive emissions may be made
at the source, before the pollutants begin to diffuse into the ambient
air; 4in the air immediately surrounding the source, where the diffusion
is limited to a relatively small volume of air; or in the ambient air,
-35-
-------
where the diffusion is essentially complete. The respective measurement
methods are the quasi-stack, roof monitor, and upwind-downwind sampling
techniques. Each is described in general terms below.
2.1 Quasi-Stack Sampling Technique
This technique captures the emissions at their source in a tempor-
arily installed hood and transmits them, by means of an exhaust blower,
through a duct of regular cross-sectional area where standard sampling
techniques are employed to measure the emission concentration and the
flow rate of the emission carrying air stream. (A simplified sampling
system is shown in Figure 2-1). The source strength of any pollutant
or combination of pollutants may then be determined as the product of
these two measured values.
The quasi-stack method is the most accurate of the air borne fugi-
tive emissions sampling methods in that it captures virtually all of the
emissions from a given source and, as a result of careful system design,
conveys them to their measurement locations with a minimum of dilution
by transport air. It is also the least applicable of the methods since
its use must necessarily be restricted to those sources of emissions that
can be physically and operationally isolated and are arranged to permit
the installation of the capture and measurement system in a manner that
will not interfere with normal plant operations or alter the character
of the emissions or their generating process.
2.2 Roof Monitor Sampling Technique
This technique is used to sample the emissions from processes or
operations taking place within buildings with a small number of openings
-36-
-------
Measurement
Hood
Exhaust
I
Y~7
/
Air flow
pitot
~| duct
T '
Particle
cfimnlar
T Gas
sampler g
a
I
Control
valve
Bypass
Blower
Source
Figure 2-1: Simplified Quasi-Stack Sampling System
-------
to the ambient atmosphere. Such a structure acts as a large hood,
confining the emissions to a finite volume of air before transmitting
them through one of its openings, such as a roof monitor, an exhaust
fan, or a door or window, to the outside air.
Samples are taken at the opening to determine the concentration of
the emissions in the transport air flowing to the ambient atmosphere,
and the flow rate through the opening is measured. (A simplified sampl-
ing arrangement is shown in Figure 2-2). The combined source strengths
of all the sources then producing emissions inside the enclosure is then
determined as the product of the measured concentration and flow.
The roof monitor method is not as accurate as the quasi-stack method
since a significant portion of the emissions may escape through other
openings before reaching the measurement point and since a much higher
degree of dilution with transport air occurs before measurement. It is
a generally more applicable method in that it can usually be utilized to
measure any indoor source, either for the composite of all processes
within a building, or for a specific combination of sources where processes
or operations may be selectively scheduled.
The roof monitor method requires instrumentation and trained personnel
capable of making measurements of usually very low air velocities through
a relatively large opening and mass balances of very small quantitites of
materials. It usually does not interfere with operations or schedules
since the required equipment and measurement activities are set up and
conducted away from the production floor.
2.3 Upwind-Downwind Sampling Technique
This technique is used to determine the emissions generation rate
-38-
-------
Togas
analyzers
Detail A
Gaseous emission ,
sample line
Hi-Vol
line
Detail B
Figure 2-2: Roof Monitor Sampling System
-------
of sources that cannot be effectively hooded for the quasi-stack tech-
nique and are not enclosed in a structure permitting the application of
the roof monitor technique. The emission concentration is determined
in samples taken from the ambient air approaching (upwind) and leaving
(downwind) the industrial site. The site contribution at the upwind
location is calculated as the upwind concentration minus the downwind
concentration. This contribution is then used in proved diffusion equa-
tions, along with measured wind speed and direction, to back-calculate
the source strength.
The upwind-downwind method is the least accurate of the three methods
described, owing to the very small portion of the emissions that are
captured for measurement and the extreme degree of dilution in the trans-
porting ambient air. It is the most universally applicable method, capa-
ble of identifying and quantifying emissions from sources indoors or out,
large or small, under any operating conditions or schedules. It is
sensitive to adverse weather conditions, changes in wind direction and
such other outside influences as wet or snow-covered ground, extended
dry periods and the like.
The upwind-downwind method can usually be utilized without even
considering its affect on plant operations or schedules since the entire
operation is so remote from the emissions site.
3.0 SAMPLING METHOD SELECTION
The selection of the most effective method for the measurement of
fugitive emissions at an industrial site is influenced by a number of
factors relative to the emissions, the process or operation involved,
and the source location and arrangement. The degree of influence of
-40-
-------
each of these factors in the large number of possible combinations of
factors at any given site is so variable that no simple selection pro-
cess can be developed to cover more than a few specific cases. Each
site must be considered as a unique situation and a selection made on
the basis of the criteria uniquely or most rigorously affecting that
site.
3.1 Selection Criteria
The general selection criteria described below are grouped into
three classifications common to all air borne fugitive emissions measure-
ment methods, providing representative examples of the influencing fac-
tors to be considered. Other factors will present themselves at any
specific site. Some degree of judgment will be required to determine the
relative importance any factor may carry.
SITE CRITERIA - factors influenced by geometry, physical layout
and location of the facility or source.
Source Isolability - can the emissions be measured separately
from other emissions? Can the source be enclosed?
Source Location - is the source indoors or out? Does the location
permit the installation of sampling equipment?
Meteorological Conditions - what are typical and critical situations?
Will wind or precipitation interfere with measurements? Will wet or snow-
covered ground alter emission rates or characteristics?
PROCESS CRITERIA - factors influenced by the nature and extent of
the process producing the emissions.
Number and Size of Sources - are emissions from a single location
or many scattered locations? Is a single source small enough to hood?
-41-
-------
Homogeneity of Emissions - are the emissions of the same character
throughout the process? Are reactive effects between emissions involved?
Process Continuity - will emissions be produced steadily for suffi-
cient time to obtain measurable samples? Must sampling be performed
over a number of process cycles?
Measurement Effects - are special procedures required to prevent
the making of measurements from altering the process or interfering
with production?
POLLUTANT CRITERIA - factors influenced by the nature or concen-
tration of the emissions.
Types of Emissions - are measurements of particulates, gases or
aerosols required? Are the emissions hazardous?
Emission Generation Rate - is the rate of emission release suffi-
cient to provide measurable samples in a reasonable time period?
Emission Dilution - will transport air reduce emission concentra-
tions below detectable limits before the sampling point? Are baseline
measurements of transport air required to identify possible masking
effects?
3.2 Criteria Application
The application of the general selection criteria described above
to each of the air borne fugitive emissions sampling methods is described
briefly and in general terms below. In practice, the application proce-
dure would be addressed to the measurement of specific emissions at a
specific site rather than to any measurement method.
Quasi-Stack Method
Effective use of the quasi-stack method requires that the source
-42-
-------
of emissions be isolable and that an enclosure can be installed capable
of capturing emissions without interference with plant operations. The
location of the source alone is not normally a factor. Meteorological
conditions usually need be considered only if they directly affect the
sampling.
The quasi-stack method is usually restricted to a single source
and must be limited to two or three small sources that can be effective-
ly enclosed to duct their total emissions to a single sampling point.
Cyclic processes should provide measurable pollutant quantities during
a single cycle to avoid sample dilution. The possible effects of the
measurement on the process or emissions is of special significance in
this method. In many cases, enclosing a portion of a process in order
to capture its emissions can alter that portion of the process by chang-
ing its temperature profile or affecting flow rates. Emissions may be
similarly altered by reaction with components of the ambient air drawn
into the sampling ducts. While these effects are not necessarily limit-
ing in the selection of the method, they must be considered in designing
the test program and could influence the method selection by increasing
complexity and costs.
The quasi-stack method is useful for virtually all types of emis-
sions. It will provide measurable samples in generally short sampling
times since it captures essentially all of the emissions. Dilution of
the pollutants of concern is of little consequence since it can usually
be controlled in the design of the sampling system.
Roof Monitor Method
Practical utilization of the roof monitor method demands that the
source of emissions be enclosed in a structure with a limited number of
-43-
-------
openings to the atmosphere. Measurements may usually be made only of
the total of all emissions sources within the structure. Meteorological
conditions normally need not be considered in selecting this method
unless they have a direct effect on the flow of emissions through the
enclosure opening.
The number of sources and the mixture of emissions is relatively
unimportant since the measurements usually include only the total emis-
sions. The processes involved may be discontinuous as long as a repre-
sentative combination of the typical or critical groupings may be in-
cluded in a sampling. Measurements will normally have no effect on the
processes or emissions.
The roof monitor method, usually dependent on or at least influ-
enced by gravity in the transmission of emissions, may not be useful
for the measurement of larger particulates which may settle within the
enclosure being sampled. Emission generation rates must be high enough
to provide pollutant concentrations of measurable magnitude after dilu-
tion in the enclosed volume of the structure.
Upwind-Downwind Method
The upwind-downwind method, generally utilized where neither of
the other methods may be successfully employed, is not influenced by
the number or location of the emission sources except as they influence
the locating of sampling devices. In most cases, only the total con-
tribution to the ambient atmosphere of all sources within a sampling
area may be measured. The method is strongly influenced by meteorologi-
cal conditions, requiring a wind consistent in direction and velocity
throughout the sampling period as well as conditions of temperature,
humidity, and ground moisture representative of normal ambient conditions.
-44-
-------
else identification and quantification of specific emission constituents
that a survey level system or similar reliable data indicates may be
present in concentrations of concern. They utilize the latest state-of-
the-art measurement instrumentation and procedures in carefully designed
sampling programs to provide data accuracy of +10 to +100% of actual
emissions, often over a range of process operating or ambient meteorologi-
cal conditions.
The differences in complexity of the two levels of sampling as
applied to the measurement of emissions from a foundry pouring opera-
tion are illustrated in Figure 4-1.
5.0 SAMPLING PROGRAMS COSTS AND TIME REQUIREMENTS
In order to prepare estimates of manpower, cost and time require-
ments that will permit valid comparisons among the three measurement
methods for air borne fugitive emissions, it is necessary to first de-
fine a factor or characteristic common to all methods as a basis for
comparison. The methodologies, techniques, and equipment requirements
are too diverse among the methods to provide that basis. The most
significant common factor among the methods is one that is within
control of the test program designer, the overall accuracy of the measure-
ments made. The accuracy is relatively easily manipulated over a generally
wide range by such means as instrument selection and test replication.
In preparing the estimates presented in Table 5-1 for overall accura-
cies of +500%, 200%, 100% and 50% for both survey and detailed measurement
systems in each of the three methods, the following assumptions were made:
o Site accessibility equal for all programs.
o Emission source accessibility and isolability as required
for each program.
-45-
-------
Figure 4-1: Survey and Detailed Systems for Foundry Mold-Pouring Measurements
Survey System
Capture
hood
Participate Measurement Devices
IKOR EPA CASCADE
IMPACTOR
HC and CO lira
Instruments
Detailed System
-46-
-------
TABLE 5-1
FUGITIVE EMISSIONS SAMPLING PROGRAMS
COST AND TIME REQUIREMENT ESTIMATES
Measurement Technique
Quasi-Stack
(Survey)
Quasi-Stack
(detailed)
Roof Monitor
(survey)
Roof Monitor
(detailed)
Upwind-downwind
(survey)
Upwind-downwind
(detailed)
Accuracy
Level-%
±500
±200
±100
± 50
±500
±200
±100
± 50
±500
±200
±100
± 50
Manpowers-
Man-hours
480
940
1,400
2,600
380
750
1,200
1,800
480
1,020
2,300
4,200
Field Study
Costs-$
1,000
2,000
4,000
7,000
800
1,700
3,400
6,200
1,100
2,600
5,000
10,000
Equipment
Costs-$
2,600
2,800
12,000
19,000
1,800
2,800
5,600
19,800
4,500
8,800
34,000
64,000
Total
Cost-$
18,000
33,000
58,000
105,000
14,000
27,000
45,000
80,000
20,000
42,000
100,000
200,000
Duration,
Weeks
12
16
24
36
12
16
20
28
12
16
26
40
-------
o No interference due to meteorological conditions.
o Emissions homogeneous and continuous for all programs.
o No process or emission alteration by measurement program.
o Particulates and gases measured in all programs.
The table presents comparitive estimates of total manpower requirements,
field study transportation and subsistence costs, equipment purchase
costs, total time duration from test planning through report preparation,
and total costs. Total costs include manpower at $30 per hour, field
study costs and equipment purchase costs. The total cost for each mea-
surement program is plotted versus % accuracy in Figure 5-1 for easier
comparison.
More detailed information relative to the selection, design and
application of the three air borne fugitive emissions measurement pro-
grams is contained in the Environmental Protection Agency Technical
Reports EPA-600/2-76-089a (b,c); entitled "Technical Manual for Measure-
ment of Fugitive Emissions: Upwind/Downwind (Roof Monitor, Quasi-Stack)
Sampling Method for Industrial Emissions."
-48-
-------
500-
VO
I
>• 200
u
u
100-
*-
z
LU
u 50-
Upwind-downwind
Roof monitor
Quasi-stack
25 50
COST - $ X 1000
100
200
Figure 5-1: Measurement "ropram Coats as a Function of Overall Accuracy
-------
COKE OVEN EMISSION MEASUREMENTS DURING PUSHING
Robert B. Jacko, Ph.D., Purdue University
I came here today to discuss the fugitive emission problem from coke oven
pushing. Purdue University's Civil Engineering Department has been involved
with a National Science Foundation funded program since 1972. The purpose of
the program was to characterize atmospheric emissions including trace metals
from a number of point sources. Some of those point sources included an open
hearth steel making furnace, a municipal refuse incinerator, a zinc smelter
coker, vertical retort and sinter plant, and a coal fired power plant. One of
the problems we identified back in '72 was the coke oven pushing emission
problem. At that time we didn't think about it in terms of a fugitive
emission -- indeed, the word was not yet coined. So we began to discuss various
sampling techniques; in order to quantify the coke oven pushing problem.
A mobile laboratory was constructed, the top of which contains necessary
hardware to enable us to sample at the coke ovens during pushing without the
constraints of a hood. In other words, sampling could be achieved in the
actual push plume in the free atmosphere with no restraints of a hood. A
description of the coke ovens is in the following text.
There were 65 ovens in the particular battery that was sampled. Refer to
Figure 1. Dimensions of the ovens are 18' wide, 12' high, and 40' in length.
Typical charge was about 15 tons of coal -- and obviously that varies,
however, this is a typical value. These ovens are approximately 20 years old.
Some of the problems that we encountered while designing the sampling
methodology was, first of all, should we position our sampling equipment off
the top of the battery and attempt to get into the plume that way. One of the
problems was that the Larry car is operated with large DC bus bars and they're
-51-
-------
located right in this area on the ovens. Therefore, we didn't want to position
any hardware off the side of the ovens which could possibly come in contact with
those DC bus bars. It was decided to go with a mobile approach, which provided
complete independence from the batteries. In other words, we don't have to
rely on, necessarily, perfect synchronization between the operators and our
sampling program -- we could operate independently. We wanted to be there when
they pushed the coke and we knew what the schedules were, so we worked it out
that way. But as many of you know, when you work in a large steel mill complex,
there are interdepartmental lines to work across — the research people work
with operations people and they have their own schedules to meet and you don't
want to interfere with those schedules any more than you have to. So we felt
that going with a mobile sampling arrangement such as this would provide us
with the flexibility that was needed.
The general arrangement of the coke ovens, quench car and mobile sampling
laboratory is seen in Figure 2. At the completion of the coking cycle (18-20 hrs),
the coke oven door is removed and the coke is pushed at a predetermined
constant rate by the ram into the quench car. Any remaining volatiles in the
carbonized coal ignited in the oxygen-rich atmosphere and incomplete combustion
results in the release of particulate matter to the atmosphere. The sampling
is initiated when the visible emissions start emanating from the coke quench
car. Blowers located on a boom approximately 30' above grade are then
activated. Samples would be withdrawn at as close to an isokinetic rate as
we could do, as we could set up, and we would draw a sample at about 35 CFM.
Now, we needed the high flow rate because, as you know, during pushing, a
typical push time is anywhere from 40 seconds to 50 seconds. The 35 CFM enabled
extraction of a relatively large sample volume over the push time so that the
sample would represent as closely as possible the contents of the actual plume.
The samples were taken isokinetically by monitoring a wind anemometer that we
-52-
-------
used on the end of the boom and plume temperature was also measured with two
thermocouples at the end of the boom.
Figure 3 shows the location of our two cameras which we use to characterize
the shape of the plume. The methodology we decided to use was to measure the
plume particulate concentration, velocity and cross-sectional area, then compute
the mass emission rate. That was the basic approach. The problem then becomes
one of trying to characterize the plume shape. At first, we thought it would
be just about impossible to do that on motion picture film, but we went ahead,
anyway, and looked at some of the results, and I think we're getting reasonable
values.
Figure 3 shows a quench car in a quasi-schematic form, with the development,
drawn in arbitrarily, of the push plume. We locate one camera, shown here as
"A", so that it views through a plane parallel with the front of the coke ovens;
and the other camera at location "B". We had two students, each manning a
camera -- one at "A" and one at "B". At the onset of pushing, we would commence
filming. 16 mm. motion picture films were taken from cameras that were
calibrated so that we knew that we were taking pictures at an exact frame rate,
so later on we could go back to cross check plume velocities against the
anemometer. In this fashion, we attempted to characterize what the shape of
the plume was through a plane taken at the sampling heads.
After analyzing many, many feet of film, a ratio of length A to B was
roughly 1.5 to 1.0. In other words, some of the data that I'll show you
assumes a circular plane cross section. That's an idealization, not necessarily
true. Analysis of the film shows that on the average, the plume shape, 30'
above grade, for both clean and green pushes, the aspect ratio, if you will —
was 1.5 to 1. We expected greater deviation than that and certainly there was
more deviation on individual pushes. But if you give me the liberty of taking
the mean value, we found it to be 1..5 to 1. We plan to try other cross sectional
-53-
-------
shapes and estimate the resulting effect on the mass emission rate. For
example, elliptic shape, possibly a hexagonal shape, and other shapes, but
we'll probably stay close to the circular or some modified circular shape.
So that was the idea — to characterize the concentration in the plume at that
same elevation and try to characterize the cross sectional area. Knowing the
flow velocity up through that plane, we computed the plume volumetric flow
rate and, ultimately, the particulate mass emission rate.
Figure 4 is a schematic, looking down on top of the entire sampling
apparatus. Note the 16' van truck, outfitted with the instrumentation inside
the truck. We have two 10-foot sections of off-the-shelf TV tripod tower
supporting the boom off the top of the truck. Since the truck sits about 10'
off grade, we are about 30' above grade to the boom. The boom is roughly 25'
long and is outfitted with high volume blowers. These blowers at free delivery
will develop about 60 to 80 CFM. Two blowers were located quite a distance
from the plume for environmental considerations. We didn't want to burn the
armatures on the motors. We ran aluminum tubes -- 2" aluminum tubes -- up
to three sampling heads, in which were mounted 8 x 10 glass fiber filters, the
same filters you use on the standard high volume samplers. At first, we
wanted to use a Stausscheibe pitot tube. However, the mean vertical velocity
of the plume was so low that we'd be operating down near its resolution points,
so we decided to use an anemometer mounted in a horizontal direction. We also
mounted two thermocouples on the end of the boom. So this was the general
arrangement of our sampling hardware. To measure the flow velocity of the
sample, we used orifices at the outlet of the blower and calibrated that orifice
with respect to the flow rates that we'd be using so that we could set up a
known flow.
Figure 5 is a schematic showing the instrumentation and data handling
system. We bring our sample in through a stainless steel sampling snout, which
-54-
-------
is about 2V long. The isokinetic sample, or the variation in the isokinetic
rate that we expected, was taken care of by welding an expanding cone on the
front side of the snout which could be shortened on site, just prior to
sampling to put us into an approximate isokinetic rate. The blowers were
variable speed to allow setting up isokinetic rates during sampling. The
snout came up inside the transition for the filter holder. We found on some
of our earlier samples that much of the material was extremely large, and after
we shut down the blowers after the push was completed, we lost material back
out the snout because of buildup. So we then modified the sampling snout to
bring it up inside the transition so that it would act like a quasi-settling
chamber and the big particles would not then fall back out the sampling snout.
This worked quite well. The sample then passes through an 8 x 10 glass fiber
filter, the same variety used in the high vol paths, and then via 2" diameter
aluminum tubes to the blowers. Each blower had an orifice plate installed at
its' outlet which was calibrated, and the resulting pressure was read out on a
magnehlic gauge in the truck. The anemometer was calibrated prior to use and
that, also, read out to a digital recording volt meter in the truck. By the
way, we wtre using an Ester!ine Angus D 2020 Digital Volt Meter to record all
sensor signals.
Comments at this point would be that we had expected problems due to
flame burning up our filter pads, and of the first 15 samples that we extracted
(my presentation deals with those first 15 -- we now have about 66 samples from
pushes at the coke oven, the same battery), we've only had flame problems on
two of the samples that we withdrew, which resulted in a fusion of the glass
pad. But the 2%' sampling snout appears to act as an adequate flame-arrester.
We had considered using other elaborate systems of flame arrest in the sampling
nozzle; but found that not to be necessary. Also, on the initial design of
the horizontal boom, we thought that it was going to be necessary to install
-55-
-------
water-cooling on the boom, but, again, had no major problem in that respect.
After our first few preliminary tests, we stripped off the water-cooling
arrangement that we had on the boom and on the structural members and we
haven't had.any problems due to temperature. However, the plume temperatures
were not as high as expected in the plume. Maybe the fact that we were
sampling a couple of days before Christmas last year and it was -18 degrees
Centigrade helped us.
The 16mm film record is a very important part of the data analysis, because
we have to correct the sample based on the time period that we were in the
plume, since we were not in the plume 100% of the time. The film was analyzed
with stopwatches to determine the net time that the sample heads were in the
visible portion of the plume. The total sample volume was then corrected for
the amount of time that we were not in the plume. We made the assumption that
when we were not in the plume and we couldn't see a visible emission, that the
concentration of particulates that we were sampling at that point in time would
be relatively low compared to the concentration of particulates when it was,
indeed, immersed in the visible portion of the plume. That's an idealization
and it has problems, but that's what we've done.
Now that we've taken the sample, we wish to retrieve the sample from the
hardware. The procedure here is to release the two guy wires from the back of
the truck that were connected up near the blower end, allowing the boom to
swing on a hinge plate arrangement that we designed. The sampling heads are
then lowered down into the vicinity of the top of the truck where the people
can get to the heads and retrieve the entire bottom half, including the glass
fiber filter. This assembly is taken into the laboratory environment within
the truck, where the filter is retrieved and the sampling head backwashed with
acetone.
How about summarizing, now, what we've done. We've sampled 15 coke oven
-56-
-------
pushes. This was on a 3-day period just before Christmas 1975 with relatively
low ambient temperatures of minus 18 C. We categorize clean and green pushes
by visual observation. We had 8 clean pushes and 7 green pushes. Refer to
Figure 6. Average plume temperatures in degrees centigrade were found to be
32 for the clean and 70 for the green pushes. Now, this may not be the true
temperature of the plume — it's the temperature that we measured with our two
radiation shielded thermocouples, located as best we could in the plume. The
plume does swing so this probably is an average of upper plume temperatures
and ambient temperatures. Average plume velocity was found in clean pushes
to be 5 meters per second; in green pushes, 4.4. These were taken with the
mounting of the anemometer in the horizontal plane.
Figure 7 contains the particulate concentrations and mass emission rates
from the 15 pushes that were sampled. Note that the clean and green push
particulate concentration is 1.5 and 2.4 grams/cubic meter respectively. The
mass emission rate of the green pushes was found to be 407 grams/sec as
compared to 147 for the clean pushes.
The average emission factors, based on a 50 sec push, were calculated as
1.6 Ib.-part./ton-coke for the 8 clean pushes and 4.5 Ib.-part./ton-coke for
the green pushes. In terms of tons of coal charged to the ovens, these values
are 1.0. and 3.0 for clean and green pushes respectively. The average is
2.0 Ib.-part./ton-coal and compares to 0.4 Ib.-part./ton-coal as published in
AP 42. Note the coefficient of variation (CV) on the mass emission rate. The
CV is defined as one standard deviation from the mean value, divided by the
mean value. As you can see, the variability is quite high. In clean pushes,
the standard deviation was as high as 74% of the mean value; in green, it was
92% of the mean value, and the overal value was 110.
It appears as though our concentration/ photographic technique, is perhaps
viable; that you can utilize it with some degree, hopefully, of representativeness
-57-
-------
of what is in that plume. Secondly, we found that the circular plume cross
section that we used for computation of a mass emission rate is at least a good
approximation, based on the film analysis and the aspect ratio of 1.5 to 1, and
this number of 1.5 to 1 comes from an observation of many, many frames of
motion picture film, literally thousands.
Conclusions
So what you've seen here is an attempt by us to quantify a fugitive emission
from a coke oven pushing operation. We designed the hardware and the sampling
methodology much before the time that fugitive emissions were discussed. Our
main objectives at that time were to quantify the flow of trace metal emissions
into the atmosphere, as well as looking at the total particulates. Work that
we have for the future includes not only the mass emission rate of the total
particulates, but we've modified our boom and we're going to instrument it for
hydrocarbons, for particle size distribution, using inertial impactors -- we'll
be mounting two inertial impactors at the boom so we can get Anderson
aerodynamic particle size distribution.
-58-
-------
COKE OVEN DATA
i
Ul
MANUFACTURER: Koppers
NUMBER IN BATTERY: 65
DIMENSIONS: Width 18 1/4", Height 12*, Length 40*
CAPACITY; 15 Tons Coal, Typical Charge
Figure 1 Coke Oven Description
-------
o
CHARGING
BY-: PRODUCT CAR
COLLECTION
MAINS
DOOR
MACHINE
COKE
PUSHER
CONTRO
CAB
RAM
PUSHING
MACHINE
SAMPLING
HEADS
BOOM BLOWERS
RECUPERATOR
QUENCH
CAR
WASTE
HEAT FLUE
Figure 2 Cross section of coke oven battery. Including sampling
vehicle
-------
Coke Oven
Push Plume
Camera
B
Quench Car
Figure 3 Location of Two 16mm Cameras
-------
BLOWERS (3)
KJ
TRUCK
QUENCH CAR
RAILS
SAMPLE
LINES (3)
/THERMOCOUPLES
'C2)
COKE GUIDE
ANEMOMETER
FILTER
HOUSINGS (3)
Flgurt 4 Top view of sampling vthlclt and cokt qutncti ear
-------
THERMOCOUPLES (2)
CENTRIFUGAL
BLOWERS (3)
ANEMOMETER
r
i
DIGITAL
RECORDING
VOLTMETER
/
THERMISTOR
MAGNEHELIC
PRESSURE
GAGES (3)
8"x 10"
GLASS
FIBER
FILTER
t
SAMPLING
NOZZLE
SAMPLING j
HE~ADS~(3)"
Figure 5 Schematic diagram of sampling heads and Instruments
used on coke oven sampling boom
-------
TOTAL NO. OF SAMPLES: 15
CLEAN PUSHES: 8
GREEN PUSHES: 7
AMBIENT TEMPERATURE: -18 C
AVERAGE PLUME TEMP., C
CLEAN PUSHES: 32
GREEN PUSHES: 70
OVERALL: 49
AVERAGE PLUME VELOCITY, M/SEC
CLEAN PUSHES: 5.0
GREEN PUSHES: 4.4
OVERALL: 4.7
Figure 6 General Summary of Coke Oven Push Sampling
-------
PARTICULATE RESULTS
GRAIN LOADING, GR/SCF G/M3
CLEAN PUSHES: 0.66 1.5
GREEN PUSHES: I.I 2.4
OVERALL: 0.83 1.9
MASS EMISSION RATE, G/SEC CV(%)'
CLEAN PUSHES: 147 74
GREEN PUSHES: 407 92
OVERALL: 268 110
*CV= COEFFICIENT OF VARIATION
= 100 x STD. DEV. /MEAN
Figure 7 Coke Oven Pushing Particulate Emissions
-------
PROBLEMS IN MEASURING FUGITIVE
EMISSIONS PROM WASTE DISPOSAL PONDS
William R. King
PMC Corporation
2000 Market Street
Philadelphia, Pennsylvania 19103
-67-
-------
PROBLEMS IN MEASURING GASEOUS FUGITIVE
EMISSIONS FROM WASTE DISPOSAL PONDS
William R. King
Abstract
This paper discusses a method of measuring the emission
rate of pollutants with finite vapor pressures from
contaminated-liquid storage ponds. The techniques described
were developed to quantify the emission
of fluorine compounds from wastewater disposal ponds connected
with wet-process phosphoric acid manufacturing plants.
However, they are applicable to any water soluble
chemical compound with a finite vapor pressure, such
as hydrogen sulfide, phenol or hydrocarbons up to their limit
of solubility in water.
heview of Available Methods
Pond emission measurement methods can be divided into
two general groups:
(1) Field measurements conducted at the actual emission
source or
(2) Laboratory measurements conducted on models.
Laboratory modeling usually utilizes a wind tunnel to
simulate the pond and the surrounding atmosphere. The
experimenter is concerned with evaluating Equations (1)
and (2).
-68-
-------
V
= Cfiu
where
NB= molar mass transfer rate per unit pond surface
of component B from the pond to the atmosphere
K = Overall gas-side mass transfer coefficient for
13
component B
p * = Partial pressure of B in equilibrium with
D
pond water
p = Partial pressure of B in atmosphere above
D
a pond
A = Pond surface area
B
C = Average concentration of B in atmosphere at
B
downwind edge of pond
u = Average wind speed or air velocity
In a wind-tunnel simulation, the average concentration -
velocity product—CD u—can be measured by placing a
B
mixing box after the wind tunnel and sampling the gas
stream coining out of the box. These measurements obviously
can be used in Equation (2) to produce good estimates
of the mass transfer rate. The wind tunnel data can
be further reduced through Equation (1) to a mass transfer
coefficient - Kg- This variable - Kfi- is the basic
parameter which must be translated from the
wind tunnel to the real world.
-69-
-------
Normally, the mass transfer coefficient is correlated
against a measure of the turbulance in the air stream
moving over the pond simulation in the wind
tunnel. This correlation is applied to real world
situations, if the pond could be considered a flat plant.
Equation (3) has been empirically demonstrated to describe
the mass transfer coefficient.
K X
c (XCu ) V5 ...
= .0365 ( ) (3)
CD (// }( CD )
aB ( .aB)
where
C = molar concentration of air
D = diffusivity of B in air
aB
// = air viscosity
-------
The characteristic length that is used to describe mass transfer
from a flat plate in a wind tunnel is well developed in the
literature. This is the distance the wind travels across the
plate. Unfortunately, the problem of characteristic length
is considerably less well defined in the atmosphere.
(No Paragraph)
Is it the pond length in the direction of wind or something
else? Likewise, the wind speed to be used is not very
well defined. In the wind tunnel the bulk average wind
speed can be measured easily and is normally used. In
the atmosphere the bulk average speed is not defined.
Based on this brief discussion, the reader can see the
difficulty in attempting to quantitatively translate wind
tunnel experiments into the real world. I don't think
the problem has been solved.
(No Paragraph)
Therefore, field measurements seem to be the only way
to quantify emissions from ponds.
Three methods to estimate emissions from ponds seem possible:
(1) In theory, direct calculation of an average pond
emission rate from measurements of wind speed and
pollutant concentration profiles in the vertical
direction should be possible. Equation (4) is the
defining equation.
NRA = C du (4)
D I
ground
-71-
-------
(2) Backing through a diffusion model from single point,
concentration measurements, Equation (5),
N = C_/f (A,X,Y,Z, 0",'ff) (5)
D ti ti •!•
where
X,Y,Z = distance
ff" .(T- vertical and horizontal dispersion coefficents
Z Y"
will also yield estimates of an average emission
rate.
(3) Finally, Equation (1) might be used to calculate emission
estimates, if field verified estimates of the mass
transfer coefficient were available.
Methods (1) and (2) directly measure some sort of average
emission rate. More than one pattern of emissions from
the diffuse source (pond) could satisfy either equation.
However, when coupled with a completely defined Equation
(1) (the third field measurement technique described above)
which rigorously defines the emission pattern, either method
can satisfactorily define diffuse source emissions.
-72-
-------
Methods Used In This Study
In this study a combination of methods (3) and a modified
method (2) were used to produce emission estimates and
to verify them. A correlation for the mass transfer
coefficient was developed from field data and used Equation
(1) to produce emission estimates. These estimates were
used in Equation (5) to predict pollutant ambient air
concentrations downwind of the emission source. The
predicted values were compared with field measurements
of the ambient air concentration of the pollutant. Good
agreement between the measured and predicted concentrations
demonstrated the accuracy of the emission estimate.
The major difference between this study and other efforts
to quantify fugitive emissions is that this study has
available a method to estimate rigorously the diffuse
source pattern of emissions. The emission estimates can
be varied with time and in space in an entirely logical
and rigorous manner.
Figure 1 is a work chart detailing the information development
necessary to apply this technique. Study of the work chart
will show that there are three major tasks. The first task
is to develop the emission estimate from the mass transfer
equation; the second is to measure ambient air concentrations
in the field, and the third is to relate the emission estimates
to the measured ambient air concentrations via diffusion
modeling.
-73-
-------
Calculation of Emission Rate
The right-hand side of Equation (1) is made up of the product
of two groups. The first of these is a driving force,PB*~PB»
which is primarily a function of pond water temperature and
the physical nature of the material being emitted by the pond.
The vapor pressure, p * can be measured in the laboratory;
B
the atmospheric partial pressure, pfi, is small for "normal"
sized ponds and usually can be assumed to be zero. Since this
driving force is specific to the pond and to the material being
emitted by the pond, it will not be covered in any detail in
this report. The second group is the mass transfer coefficient-Kn
D.
This term with straightforward modification, is applicable
to any pond. For this reason, its development is discussed
in this paper.
As stated previously, the overall average mass transfer
coefficient describing material transfer from a flat
plate into a flowing stream can be predicted from the
Colburn Analogy, Equation (3), when the stream-air mass-is
in turbulent flow.
\No Paragraph)
Since the physical parameters in Equation (3) do not
change significantly over the range of conditions found
in the atmosphere, they can be grouped with the constant.
The result is Equation (6).
K=au'8x"'2 (6)
-74-
-------
In 1950 and 1951 the U.S. government measured water evaporation
rates for Lake Hefner, the water supply reservoir
for Oklahoma City. The lake is roughly elliptical in
shape—about 2-1/2 by 1-1/2 miles. About 140 of the
500 separate daily, evaporation rate measurements made
during the government study were used to develop estimates
of the overall, average, mass-transfer coefficient.
In addition to the lake data, evaporation rates were
measured for four different types of evaporation pans:
1. The class A evaporation pan,
2. The Bureau of Plant Industries evaporation pan,
3. The Colorado evaporation pan and,
4. The screened evaporation pan.
Physical details of these pans are summarized in Table 1.
The measured evaporation rates were transformed into
24-hour-average mass transfer coefficients with Equation
(1). Least squares correlations were developed for the logarithmic
form of Equation (6)
0Kb = log10(aX"2) «. b Iog10ul6 (7)
where
u = wind speed 16 meters above pond surface
for Lake Hefner and each of the evaporation pans.
-75-
-------
The resultant correlations were tested statistically;
it was demonstrated that the correlations developed for
Lake Hefner tne Bureau of plan Industries pan and the
Colorado pan were not significantly different. On this
basis the three sets of data were combined and a new
correlation was developed from this combined data. The
statistics for the latter correlation are shown in Table 2.
The slope-b-of the best-fit-linear-least-squares correlation
for the screened pan was not significantly different from
the slope of the combined data correlation (or the 0.8
_. 2
slope); however, the intercept-ax —was significantly
different. This was not too surprising since the screened
pan is covered with a relatively small mesh screen that
reduces the air turbulence above the water at a given
wind speed. The expected effect of the screen is to
reduce the mass transfer coefficient at a given wind
speed; this was observed.
Both the slope and the intercept of the least squares
correlation for the class A pan are significantly different
from the combined data correlation. The class A pan
sits about 15" above the ground in contrast to the other
pans and the lake which sit flush with the ground. This
elevation makes the pan look more like a circular cylinder
than a flat plate. Wake separation can be expected to
occur in the air stream flowing over the surface of the
water. There has been no work done specifically to define
mass transfer from a cylinder head—the Class A pan's
configuration. However, measurements of mass transfer
-76-
-------
to fluids flowing past surfaces where wake separation
occurs(the curved surface of cylinders, rectangular bars,
streamlined cylinders and spheres) demonstrates a wind
speed dependency ranging from the 0.5 to 0.65 power.
The exponent predicted by the best-fit-least-squares
correlation for the Class A pan was 0.58—a reasonable
value based on the above discussion.
Since the correlation based on the pooled lake, Colorado
pan and BPI pan date shows a dependency upon wind speed
not significantly different from u,8-, the empirically
well-documented 0.8 velocity exponent was assumed correct.
The recommended predictor of the mass transfer coefficient
is
0.8 ,„,
K=.429u1<; Co)
•< 16
where
k= grams water per sq. meter-millimeter of mercury-hour
U =wind speed at 16 meters in meters per second
16
Examination of Equations (3) or (7) indicates that the
mass-transfer coefficient should be dependent on pond
diameter as well as wind speed. Statistically this difference
doesn't exist. But to test further for the length dependency,
Equations (3) and (8) were assumed to be the same equation
and they were solved simultaneously for length. With
this set of assumptions, length-X-is equal to 12,000
meters. if the further assumption is made that the equivalent
-77-
-------
length of the pans is 12,000 meters and the equivalent
length of the lake is 12,000 meters plus two miles, then
-0.2
the ratio of the empirical value of the pan's X
term to the lake's term should be 1.045. In fact, the
ratio of the average of the Colorado plus the BPI pan's
length-containing value to the lake's is 1.070. Perhaps
a length effect does exist, but the effect cannot be
separated statistically, and as demonstrated in this paragraph,
the value is unimportant in the estimation of mass transfer
from moderately sized bodies of water.
In the preceding paragraphs, a mass transfer coefficient
describing the evaporation of water into air has been
developed. A simple diffusivity adjustment must be made
to transform this into a mass-transfer coefficient describing
the evaporation of another material into air, Under assumptions
previously enumerated the overall, gas-side mass transfer
coefficient for water can be transferred into the overall,
gas-side mass transfer coefficient for another species
by adjusting the water coefficient by the diffusivity
ratio to the two species raised to the 2/3 power.
KB - K (D )2/3
" w( aB)
CDaW) (9)
where:
KB
= overall, gas-side, mass transfer coefficient
for species b
-78-
-------
D _ = diffusivity of B in air
ats
D = diffusivity of water in air
aW
Field Measurements
Two waste disposal ponds connected with wet-process-phosphoric-
acid manufacturing plants were studied. The digestion
of the phosphate rich minerals with sulfuric acids yield
product plus volitile fluorine compounds.
« (g,
These gaseous fluorine compounds are scrubbed from the
plant's air emissions and sent to wastewater ponds.
Over 85% of the fluorine entering the plant reports to
the ponds.
The ponds chosen for the ambient air studies are shown
in Figures 2 and 3. Pond 10 is a nominal 100-acre, roughly
rectangular, cooling pond. The land beyond the southeast
and southwest edges is grass covered for at least 300
meters; it is generally flat and at the same height as
the pond water. Pond banks are diked about two meters
above the water level except in the area around the two
sampling sites. Gypsum piles about 15 meters tall border
the northwest and part of the northeast sides of the
pond. Gypsum was not being accumulated on the northwest
pond during the test period; the gypsum disposal ponds
on top of the pile were dry. The northeast pile was
-79-
-------
used for gypsum disposal; 50 acres of ponds covered the
top of the pile, Pond 20 is a 160 acre-irregularly shaped
cooling pond constructed on an old mine site. By-product
gypsum was being stacked on the land mass in .the center
of the pond. The level of the pond was less than 1 meter
below the edge of the banks. Small weed trees, 2-3 meters
tall, cover the edge of the pond bank.
During the field work at pond 10, winds were usually southerly.
Overcast skies and squalls were common; 3(C) and 4(D) stability
classifications predominated. The wind speed measured three
meters above the ground varied from 1.5-6.3 meters per second.
At pond 20 the wind usually blew from the east although
it varied over the day. Sunny skies and 1(A) or 2(B)
stability classifications were common during actual sampling
period. However, rain squalls frequently occurred in
late afternoon and inversions were common in the early
morning. Wind speed at 3 meters varied from 0.9 to 3.5
meters per second.
The following data was obtained in the field:
1. One-hour-average, ambient-air fluorine concentrations.
2. Wind speed at three meters above the ground.
3. Wind direction at three meters.
-80-
-------
4. Clpud cover estimates, and air stablllty
estimate s«
5. Pond water inlet and outlet temperatures.
6. Pond water fluorine concentration.
In the study of pond 10,six upwind samples were taken
at station 3, Figure 2. Three of these samples showed
no ambient air fluorine content, and three showed an
average of about 0.8/^moles/m of fluorine not emitted
by the pond. The fluorine source was not identified;
however, assuming reasonable atmospheric dispersion,
it can be demonstrated that the maximum contribution
to the measured fluorine concentration at the downwind
sarolers was about 0. 10 /'cjmole/M of fluorine—about
5% of the total fluorine measured. This value is the
same magnitude as the limit of measurement. Upwind samples
from pond 2 were not obtained because of the inaccessibility
any upwind site. However, visual inspection of the area
revealed no obvious source of fluorine emissions east
of the pond closer than about 4 kilometers. Therefore,
the effect of other fluorine sources would be similar
to those experienced around pond 1; that is, their contribu-
tion to the measured concentration would be about the
same magnitude as the limit of measurement.
-81-
-------
Reconciliation of Data
In review, the paper has so far presented a method to
predict emissions from waste ponds. This method depends
only on the pollutant being emitted, the pond temperature
and the wind speed. The method was used to predict fluorine
emissions from wet-acid plant ponds. In addition, the
paper describes the measurement of ambient air fluorine
concentrations downwind of two wet-acid ponds. How can
these two sets of information be compared?
Examination of the diffusion model indicates two possible
ways of comparing the two data sets. A back-calculation
can be made from the field-measured concentrations through
the diffusion model to obtain estimates of the emissions
from the pond. These can be compared to the emission
rates calculated from the mass-transfer equation. Conversely,
the estimated pond emissions can be used in the diffusion
model to predict ambient air concentrations at the downwind
samplers. These values can be compared with the field
measured volumes. Neither method is entirely satisfactory.
Back-calculating emission estimates from ambient-air
field measurements requires that the fluorine emission
rate from the pond be distributed in some manner. As
you remember, at a given wind speed, the emission rate
depends upon the fluorine vapor pressure. This in turn
is a function of the pond water temperature and the tempera-
-82-
-------
ture variation over the pond. The logical way then is
to distribute emission rates over the pond surface
in proportion to the experimental vapor pressure. Unfortu-
nately the vapor pressure is also in the mass transfer
equation. Therefore, both the emission estimate developed
from the mass transfer equation and the estimate developed
from the measured ambient air concentration utilize the
measured vapor pressure of fluorine over pond water.
Because both are derived from vapor pressure measurements,
neither emission estimate is independent of the other.
Statistical comparisons of the two estimates are dangerous.
The statistics drawn from the two sets of data may compare
vapor pressure with vapor pressure.
This statistical problem can be avoided if ambient air
concentrations are compared. In this case, vapor pressure
is utilized only in the calculation of ambient air concentra-
tion from the diffusion model. In addition, this comparison
has an advantage in that all the estimated or measured
values—the rate of emission estimate and the diffusion
model—are lumped in the calculated concentration. The
field-measured concentration is a "pure" number which
has not been subjected to any manipulative calculations.
The disadvantage of comparing ambient air concentrations
is that the emission-rate-calculation method is indirectly
tested.
-83-
-------
A computer program was developed to calculate ambient
air concentration. This program divides the sample time
into 10 minute time periods. It divides the pond water
studied into a number of isothermal segments. A ten meter
wide strip perpendicular to the average wind direction
of this isothermal segment is approximately considered
to be a finite line source. The program calculates the
mass-transfer coefficient per unit length from the strip
and assumes that the entire emission comes from a single
line at the center of the strip. It calculates the time
weighted contribution of this finite line source at the
sampler and repeats the calculation for all strips within
an isothermal segment and for all segments of the pond.
Figures 4 and 5 are graphic comparisons of the calculated
and the actual field-measured ambient air concentrations.
(No Paragraph)
The figures also include the 45 , or perfect correlation,
line. Tables 3 and 4 summarize the statistical tests
which compare the measured and calculated corrections.
As these tables and Figures 4 and 5 show the hypothesis
that the least-squares, best-fit lines are not significantly
different from perfect correlation line ±s confirmed. Therefore,the
method of predicting ambient air fluorine concentrations
developed in this paper are confirmed by field measurement.
By extension, the method of predicting pond emissions
developed in thi-s paper has also been confirmed.
-84-
-------
As the reader knows, when making field measurements in
the atmosphere, every now and then a bad experiment occurs.
Sometimes these runs can be explained and sometimes they
can't. The field studies described in this paper were
no different. The experiments discarded in this study
are shown in Figures 6, 7, 8 and 9 along with a short
explanation on what may have gone wrong. Other than
the data shown in Figure 8 for pond 20, the discarded
rums could have been correlated with the main body of
data with no appreciable effect on the study's conclusion.
The same can not be said about the data shown in Figure
8; its inclusion would change the slope of the least squares
o o
correlation for pond 20 from +45 to -30 .
Conclusion
This paper proposes and demonstrates a method to estimate
emission from ponds containing water-soluble, volatile
atmospheric pollutants. The method has general applicability
To apply it to a specific pond and pollutant, the only
additional information that must be developed is the
pollutants vapor pressure curve.
R71629
BV24
-85-
-------
Acknowledgment
This paper is based on "Fluoride Emissions from
Phosphoric Acid Plant Gypsum Ponds"—EPA-650/2-74-095,
a report prepared for the Office of Research and
Development, U. S. EPA, Washington, D.C. EPA project
officer was B. N. Murthy, Control Systems Laboratory,
NERC, Research Triangle Park, N. C. 27711.
-86-
-------
Table 1 Pan structural details
Type
Diameter
(Ft.)
Depth
(Ft.)
Relation of Water
Surface to Ground
Surface
Class A
Screened
(% inch mesh screen
over top of pan)
Colorado
Bureau of Plant
Industries (BPI)
3 ft.
sq.
0.83
1.5
(est)15-18" above
ground
even with ground
even with ground
even with ground
-87-
-------
Table 2 Statistical Summary of Correlation Developed for
Combined Lake Hefner, FBI Pan and Screened Pan
Data Base
Iog10(ax~-2)
(Standard
Error )
-.3^87
(.02477)
b
(Standard
Error)
.7727
(.03^97)
Number of
Observations P
319 488.254
R2
.6063
QK = log10(ax--2) + b log1(J U16
O
K = g moles/hr m^ mm H«-
o
Ul6 = 16 meter velocity in meters/second
-88-
-------
Table 3 Least squares regressions of calculated ambient air
fluorine concentrations on measured concentrations
Pond
10
20
V V
a b
(Standard Error)
.2536 .9122
(.16825) (.06346)
.1043 .9888
Number of
Observations F
50 206.58
37 58.649
R2
.81145
.62626
(.15425) (.12911)
-89-
-------
Table i| Testing the coefficients of the least squares regressions
of the calculated fluorine concentration on the measured
fluorine concentration with the student's t test to
determine if av = 0 and bv = 1
Pond
10
20
Value of
t For
V
a
1.507
.67626
Significant3
No
No
Value of
t For
b
1.383
.0868
Significant3
No
No
aAt 95% confidence level
-90-
-------
Review and Correlate
Water Evaporation Data
Develop Wind Velocity
Profile Prediction
Review and Correlate Data
For the HF-H.O System
Design Equilibrium Still
Develop Preliminary
Emission Estimates
And Preliminary
Ambient Air Concentrations
Design Ambient Air
Sampling System
Develop Pond Simulation for
Computer
3
Predict Ambient Air
Fluorine Concentrations
Downwind of Ponds
Develop Pond to Air Mass
Transfer Coefficient For Water
Develop Pood to Air Mass
Transfer Coefficient For Fluorii
Test Analytical Hethods
a
Check Consistency of Still
on RF - Water System
Measure Vapor-Liquid Equilibrium of
Fluorine In Pond Water
Check Operation at Ambient Air Sampler
I
Heaaure Ambient Air Fluorine
Concentrations Downwind of Ponds
VariCy Emission Estimates by Comparing
Predicted and Measured Ambient Air Fluorine
Concentra t ions
Produce Emission Estimates
Figure l Work flow chart to produce an estimate of fluorine
compound emissions from wet process phosphoric acid
plant waste ponds
-91-
-------
Page 25
Study coordinate systen
ler 01
Wind Speed and
Direction
Recorder
Active Gypstim Pile
Figure 2 Pond 10 plot sketch
-92-
-------
V
•a
M
-------
o
LU
u
Measurements not used in
data analysis
I
I
2 3
MEASURED
Figure 4 Pond 10 measured versus calculated ambient air concen-
trations at samplers xlO~" g moles/or
-94-
-------
o
UJ
o
1.0
MEASURED
Figure 5 Fond 20 measured versus calculated ambient air
concentrations at samplers xlO~6 g moles/nr
-95-
-------
o
LU
§
O
o
Figure
3456
MEASURED
Pond 10 Sampler 01, high wind speed experiments
x 10~6 g moles/m3
Concentrations measured by sampler 01, the sampler nearest the pond.
when the wind velocity was over ^ meters/sec are considerably higher
than the predicted values. Since measurements made with sampler
02 during the same period did not show this trend and no
entrainment was noted, it is likely that the dispersion model
cannot handle the combination of high wind speeds and short
distances.
-96-
-------
6 8
MEASURED
12
Figure 7
Pond 10 wind flow over active gypsum pile xlO g moles/m
Measurements were made when the wind was blowing from the
active gypsum pile (wind direction 90° and over). The
measured values were considerably greater than the predicted
values. In this case, the piles caused extremely complex
wind patterns - including downwashes that the dispersion model
could not handle.
-97-
-------
o
LU
5
O
O
Figure 8
2345
MEASURED
Pond 20 Sampler 02 high fluorine measurements xlO
g moles/m
-6
On a number of days after 10:00 AM, the fluorine concentration
measured by sampler 02 increased dramatically. No similar
increase was noted in sampler 01. Since the phenomena was not
noted on the day after a plant shutdown, a tentative explanation
is advanced that some action by the plant in the gypsum
disposal area caused the high ambient air concentrations.
-98-
-------
o
UJ
Figure 9
1.0
MEASURED
Pond 20 strong solar radiation, low wind speed
experiments xlO"6 g moles/m3
A combination of low wind speeds (less than 1.3 meters/second)
and strong solar radiation (late morning and afternoon)
produced a greater degree of dispersion than could be
predicted by the dispersion model. Low wind speeds alone
(early and mid-morning) could be satisfactorily handled by the
model.
-99-
-------
CONTINUOUS ROOF MONITOR EMISSION TESTS
Abbas F. Souka, Ph. D. , Airco Speer Carbon-Graphite
Single story buildings in certain industries such as metal smelting, graphi-
tizing, silicon carbide manufacturing, glass plants, foundries, and many
other manufacturing activities require a substantial amount of ventilation
to maintain human comfort and safety. The most efficient and economical
method to achieve this purpose is the use of wall louvers and building monitors
capable of exhausting large amounts of heated gases at a very rapid rate. The
quantities of influent air and effluent gases are dependent on the building
design and location, the prevailing climatic conditions and on the activity with-
in the building. The two last factors are variable and greatly influence the
amounts of emissions from such structures.
In order to prove compliance with the air pollution regulations, most of the
above mentioned industries are required to sample the effluent gases as they
exit the building. At first, we at Airco Speer attempted to apply the standard
EPA method for particulate sampling of stack effluents only to realize that
such technique is not applicable and that an alternate approach is required.
First, I will enumerate the reasons why sampling using the standard EPA
method for stack sampling is not applicable to monitor sampling:
1. The velocity of the effluent gases as measured in the building
monitor fluctuates within a wide range. This can be readily seen by
inspection of figures (1) and (2). They show seven plots of instan-
taneous monitor exit velocity as a function of time. Four of these
plots were recorded on four consecutive days and the remaining
three were recorded on the same day. It can be seen that within
the course of one minute the exit velocity may fluctuate by more
than 600%
2. Roof monitor designs cause turbulence in the gas stream and
consequently do not fulfill the isokinetic sampling requirements that
sampling and velocity measurement points must be taken at locations
away from any disturbances.
3. The variation in activities within the building and the rapid changes
in the climatic conditions make it necessary to collect a large sample
of the effluent gas over an extended period of time in order to obtain
results that are representative of the actual contribution of the
manufacturing facility to the ambient air.
4. Pitot tubes are not suitable for measurements of low velocity-
heads.
-101-
-------
Sampling Procedure
The first published application of high volume samplers to sample emissions
from roof-monitors that we are aware of used suspended high volume samplers
which traversed the building monitor, figure (3). Samplers were positioned
along the length of the building; each to collect a sample representative of the
emission through a known area. Each sampler in each section was hung on
a cable so that it can be traversed back and forth across the width of the
monitor. The effluent velocity was measured using hot wire anemometers.
The procedure^ which has been applied in three separate tests by Airco Speer
for measurement of emissions from building monitors also included the use
of high volume samplers as described below:
The building monitor was divided into three equal sections. One sampling
station equipped with a high volume sampler was located at the center of
each section. Figure (4) shows the 3 sampling stations and fig. (5) shows
the high volume sampler affixed in its normal position with the filter facing
upwards. A goose neck attachment directs the flow into the filter. The
sampling velocity was maintained at about 100 fpm.; the normal sampling
rate used with high volume samplers. This rate is less than the isokinetic
velocity and gives emission values biased to values higher than the actual
values. Figure (6) shows the deviation between the observed concentration
to the actual concentration with departure from isokinetic sampling. ' It is
noted that the smaller the particle size the less the error in measurement
becomes. Figure (7) shows the sampling probes which consisted of three -
9" diameter nozzles each connected to a U-shaped 4" diameter pipe. Figure
(8) shows one of the high volume samplers mounted on a platform outside the
monitor. Particulates were collected on glass filter paper retentive to
particles larger than 0. 3 micron. Before initial and final weighing the filters
were dried at room temperature over silica-gel for 24 hours. Weighing was
made to the nearest 0. 1 mg with a balance having a precision of— .05 mg.
The volume of air sampled was determined by using calibrated rotameters.
Flow meters readings were generally obtained at least once an hour. In
order to obtain a sampling velocity between 100-120 fpm, a sampling rate
of about 1.4m /min. was necessary. By observing the rate of decrease in
flow meters readings an estimate as to the necessity of a filter change could
be made. Generally, the filters were changed when the flow meter indicated
a rate of 1. 3 m^/min.
The average velocity of the air as it exits the monitor was continuously recorded.
A Gill Propeller Electric Anemometer Model 27100 was mounted in the monitor
at each of the sampling stations to a depth of about 12 in. Figure (9) shows the
anemometer mounted in the monitor. The anemometer^ is an extremely sensitive
air speed measuring instrument employing a foamed polystyrene propeller.
The propeller rotates 0. 96/revolution for each foot of passing wind for all
wind speeds about 4 ft. /sec. Its threshold speed is 0. 8' /sec.
-102-
-------
The propeller drives a miniature D. C. tachometer generator. The
propeller anemometer will measure both forward and reverse flow. When
the propeller rotation reverses, the generator signal polarity reverses.
The propeller responds only to the component of the wind which is parallel
with its axis.
The anemometer was purged with filtered air in order to prevent dust or
corrosive gases from interfering with the proper function of the instrument.
Calibration of the anemometer was done by connecting the anemometer to a
synchronous drive with flexible coupling. At 1800 r.p.m. the D. C. generator
generates 500 m. v.
The signals from each anemometer, which are directly proportional to wind
speed,were recorded in the first two tests by a Leeds and Northrup
Speedomax W recorder which printed a reading from each anemometer every
75 seconds. This gave an extensive history of the monitor exit velocity.
Approximately 24,. 000 velocity readings were averaged. Interpretation of the •
data proved to be a tedious and time consuming operation and prompted us
to look for an alternative recording instrument. Subsequently, we purchased
a Martek Model EDP Environmental Data Printer, which was used instead
of the L.&N recorder and which greatly simplified the data logging.
The data printer can be operated in one of three modes; continuous, program
or manual mode. In the program mode record, period and length are
switch selected. Record period may be set at 4, 8, 16, 32, 64 or 128
minutes. Record length determines the duration of each recording and may
be set to either scan or 1/2 to 2 minutes position, recording will start by
printing the time and continue for the time selected. After the last active
channel has been recorded the instrument reverts to a standby condition and all
assemblies except the clock are turned off to reduce the power drain. In
the manual mode each channel is advanced by depressing the step switch.
The data collected by the data printer is in the form of a printed tape and a
recording on a magnetic tape cassette. The cassette can be read on a Data
Reader. A digital printout of the taped information can also be acquired
through external data processors.
Emission Data
The testing period for evaluating particulate emissions from our graphitizing
plant roof monitor was continuous for one week. Sampling was continuous
with the exception of time required to make filter changes and short intervals
to make minor equipment adjustment and repair.
The emission rate was calculated as indicated below:
1. Dust load/filter (mg/m3) = Weight of particulates collected (mg)/
total volume sampled.
-103-
-------
2. Average monitor velocity (ft/min). Velocity values were recorded
by means of LAN Speedomax type W recorder in the first two tests
and by a Martek Model EDP Environmental Data Printer in the third
test. These values were averaged over the time interval correspond-
ing to the filter paper used.
3. Station emission rate (#/hr) = Station fractional monitor area (ft^)
x average vel. (fpm) x dust load (mg/m ) x 2. 2 (#/kg) x 60 min/hr
x ID'6 (kg/mg)x .0283 (m3/ft3).
Figure (10) shows emission rate vs. time at one of the three sampling
stations. Studying of the bar chart shows periods of higher emissions and
other periods when emission rates are quite small. These fluctuations
reflect activity within the building and can serve as a guide to point out the
dusty operations so that corrective measures may be taken.
Top consideration has to be given to prevention of injury to the sampling
team. Installation of railings for high areas is a must. Sampling for
possible existence of harmful concentrations of gases must be performed
by the industrial hygienist. Sampling personnel must work in pairs. The
installation of a telephone for easy communication with the rest of the
plant is recommended.
Sampling Cost
The cost of performing this sampling technique can be divided into 3
portions:
1. Cost of Equipment
This is a one time expense which amounts to approximately
$4, 000.00 to cover the cost of 3 High Volume Samplers and
one Data Logger.
2. Cost of Sampling Station
This is a cost which is incurred every time a new monitor
is sampled. It covers the cost of setting up the shelters for
the high volume samplers as well as the Central station where
the recording equipment is situated. This cost will vary depend-
ing upon the monitor being sampled. An average figure of
$10,000 represents a realistic figure.
3, Cost of Conducting Test
This represents the cost of data collection, performing the
-104-
-------
necessary calculations and writing the final report. This cost
amounts to about $8, 000.
EPA Method for Sampling of Emissions from
Building Monitors
A sampling method for measuring emissions from primary aluminum
industry building monitors has been promulgated by the EPA (reference
methods 14 and ISA or 13B). In this method a manifold system and
connecting duct is permanently installed to draw an air sample from a
roof monitor to ground level, figure (11). The system includes eight nozzles
each having a diameter of 0.40 to 0. 5 meters spaced along the length of the
manifold which measures 8% of the monitor length with a minimum of 35
meters. The exhaust fan which is installed at the ground level draws a
portion of the exhaust at a rate equal to the isokinetic rate determined by
means of electric anemometers installed in the monitor. The monitor exit
velocity may vary during sampling as much as — 20% of the previous 24
hours average isokinetic velocity as recorded by the electric anemometers.
The sampling duration for fluorides has been set at a minimum of four to
eight hours.
Comparison of Monitor Sampling Methods
The first two sampling methods discussed in this presentation follow the
approach which uses high volume samplers to filter a large sample of
effluent gases over an extended time period in order to determine the
particulate concentration which fluctuates with the activity inside the build-
ing. This presentation also acknowledges that isokinetic sampling as defined
and applied to stack sampling is not applicable to monitor sampling.
The EPA methods 14, 13A or 13B promulgated for the Primary Aluminum
Industry can not necessarily be extended to other industries. The limitation
which require the average monitor velocity, misnomered isokinetic velocity,
on the day of sampling to be within i 20% of the average monitor velocity
recorded in the previous 24 hours is very restrictive. Disregarding the
changes in work activity within the buildings which, by itself, may result in
variation in the internal heat load to a degree producing change in the
monitor exit velocity approaching the — 20% allowable range, variations in
wind velocity by as little a 2 mph can produce as much as 20% change in
the monitor exit velocity.
Isokinetic Sampling
A method has been suggested2 in which an approximation of isokinetic
sampling may be realized. In this procedure, reproduced hereafter, two
probes are equipped with a Model 27100 Gill Electric Propeller Ane-
mometer or equivalent, figure (12). Each anemometer generates a direct
current proportional to the speed of rotation of the propeller. The electric
-105-
-------
signal produced by virtue of rotation of the comparator anemometer
represents the velocity of the effluent gases while that of the sampling
anemometer represents the sampling velocity. The net signal is fed into
a controller which actuates a motorized butterfly valve towards close or
open positions till the flows through the comparator and sampling probes
match. Alternatively this net signal may be fed to a control circuit
designed to adjust automatically the speed of rotation of the high volume
sampler till an isokinetic sampling rate is achieved.
When such a system is used for sampling emissions from building monitors,
a more representative sample may be obtained if a sampling chamber which
continuously travels from one end of the monitor to the other end is
substituted for the stationary sampling stations, figure (13).
A cknowledgment
This work would not have been possible without the participation and
valuable contribution of my colleagues at Airco Speer.
References
1. Electric Furnace Roof Monitor Emission: Testing Methodology
T. E. Kreichelt & Thomas G. Keller
64th Annual Meeting of APCA
Atlantic City
2. A New Approach to Roof Monitor Particulate Sampling
A. Souka, R. Marek & L. Gnan
APCA Journal, April 1975
Vol. 25, No. 4
3. H. Watson, Amer. Ind.
Hyg. Assoc. Quart. 15:21 (1954)
4. R. M. Young Co.
Gill Anemometer Brochure
5. Federal Register
Performance Standards for New Stationary Sources
January 26, 1976
-106-
-------
I
M
o
^-J
I
1 lb
•'iiTiJL
v*
"^H
7^
IN
\ /
V
V
A
-Jr
A;
-^-P4-J-4
Fig. (1) Variation of Monitor Exit Velocity With Time
-------
°°fl
&
4-3
A>
S
^f
t
to:.
,_O.
•f-
; I -<*t- .. i_ | — |. J 1__|.._.[.
-2-
=1=0
r\
^b
!VN
r
h/-^
i«.
1 ?-*
'•j..
£D
A
Fig. (2) Variation of Monitor Exit Velocity With Time
-------
I
I—•
o
I
HIGH
VOLUME
SAMPLER
F1GUREO) CROSS SECTION OF ROOF MONITOR SHOWING TEST EQUIPMENT (Ref. 1 )
-------
Figure 4
Sampling Stations
)
"
-110-
-------
Figure 5
Volume Sampler
CI
-111-
t
,
.!-.-~
-- -
VI
-------
Figure 7
Sampling Probe
"'
-112-
-------
Figure 9
Elec. Anemometer
'"
-113-
-------
.o 1-4
1.2
1.0
0.8
0.6
Limit for very
large particles
Perfect
^-* 31 n
•• ^
Alignment of sampling tube
parallel to wind direction
i i t lii i i
0.5
1.0
1.5
2.0
Ratio of wind speed to Inlet air speed (U */ U)
Figwre( 6)-—Diagram showing change in ratio of observed concentration Co true concentra-
tion with departure from isokinetic conditions. (Ref. 3)
-------
>**H.
l
Recorder
Ln
W»ITWI»<1
r..ct,ic./ 11
anemometer
9" nozzle
7
it
i
Sampler system.
4" U-tube
sample/
-------
lOr
o>
••* i
c
o
(1 'io
0-, .-.
UJ
H
nJU
lnhwjii
|h|-ji::^|:
.UL
n
4LiiiiJl.iI;
Luliiii
i
12M 12M 12M
Figure(ic» Emission rate vs. time interval, Central Station
12M 12M
Time
12M
12M
12M
-------
SAMPLE
MANIFOLD
W/8 NOZZLES
SAMPLE EXTRACTION
DUCT
35 on I.D.
ROOF MONITOR
/^ SAMPLE \
F~f
*&
£j
10 DUCT CIA:
EXHAUST MI8"MUM
3DUCTO.A.
MINIMUM
r~ . — I 1 — .y
-^
^r ^S^
NOZZLE
SAMPLE PORTS IN
VERTICAL Ol'CT
SECTION AS SHOWN
I—
7.5cmDIA.
POT ROOM
EXHAUST BLOUTER
Fig. (11-A) Roof Monitor Sampling System.
8.13 "^020 ?f!~~l!i7
1.0. I.D. '-°- '-D'
— 35-
I
TO SLOWER
0.35 0.025 DIA
1.0. CALIBRATION
HOLE
DIMENSIONS !N METERS
WOT TO SCALE
Fig. (1 1- B) Sampling Manifold and Nozzles.
-117-
-------
n
•
/
r
L&N multi
Pt.
Speedomax "
type W
recorder
50-0-50 m.v. "
^ ^
'•" ^
\
\/
n
, j
n
ii
i ! Mode
Gillane
j P^-To d
II or
|L sp
>. ~> ^
/
- /£
i o"y
-------
Traveling chamber
Partition-^
High volume-\ \
sampler
'"X
Monitor top
Figure d3) Traveling chamber*
-------
Session III:
IMPACT OF FUGITIVE EMISSIONS
Mary Stinson
Session Chairman
-121-
-------
RELATIVE IMPACTS OF OPEN SOURCES OF EMISSIONS
T. R. Blackwood and J. A. Peters
MONSANTO RESEARCH CORPORATION
DAYTON LABORATORY
1515 Nicholas Road
Dayton, Ohio 45407
-123-
-------
RELATIVE IMPACTS OF OPEN SOURCES OF EMISSIONS
T. R. Blackwood and J. A. Peters
Monsanto Research Corporation
Dayton, Ohio
Abstract
This paper describes approaches which can be taken for
comparing and assessing open sources based on the magnitude
and composition of the emissions. An open source is an indus-
try which emits air pollutants in a primarily non-point manner.
This includes sources of fugitive gases and dusts. These
sources can have a major impact on ambient air quality owing
to the persistence of the fine particle fraction and the dis-
perse geographical locations. Also described are source-
oriented sampling techniques which are used to determine the
emission rate while minimizing the cost of sampling activities.
A methodology for objective source comparison, called
source severity, is described which provides a consistent basis
for comparing the various emissions generated by a single source
as well as for comparing the environmental impact of diverse
sources.
Open sources which have been evaluated by MRC include
the following:
Coal Storage
Sand § Gravel Processing
Cotton Defoliation
Cotton Harvesting
Grain Harvesting
Grain Elevators
Crushed Stone Processing
Surface Coal Mining
Crushed Granite Processing
Crushed Quartzite Processing
Emission factors are presented for previously estimated and
never-before-sampled open sources. An example of the dif-
ferences found from previous estimates for crushed stone pro-
cessing is given.
-124-
-------
Introduction
An open source is an industry which emits air pollutants
in a primarily non-point manner. Open sources are generally
large, diffuse in nature, and are often comprised of several
fugitive emission points. Consequently, the source strength
of open sources is quite difficult to assess accurately since
emission rate is not a simple function of gas flow rate and
pollutant concentration.
Emissions of particulates and other pollutants from open
sources have drawn increased attention in recent years because
fugitive emission problems are frequently encountered. An
investigation of approximately 140 open sources is being con-
ducted under EPA Contract No. 68-02-1874 to provide a better
understanding of the magnitude and composition of emissions
for the purpose of determining needs for developing control
technology. An understanding of the extent of e -issions from
open sources is also necessary to provide a perspective on the
benefits to be derived from controlling point sources. If an
ambient air quality problem is mostly caused by fugitive and
open emissions, there is little to be gained in controlling a
few point sources.
Some typical open sources are dusty material transfer,
crushing, sizing, screening, landfill and excavation, blasting,
ship loading and unloading, unpaved roads, molten metal trans-
fer, beef cattle feedlots, tailings ponds, coal refuse fires,
agricultural tilling, agricultural chemical application and
volatilization, harvesting activities, storage piles, and
erosion of soil.
Consideration
Open sources can be easily observed adding emissions to
the atmosphere, but more often than not they are located in
sparsely populated and remote areas of the country. Thus,
why consider a source which, on the surface, may not affect
a large segment of the population? Unfortunately, in many
urban areas the ambient air quality standards for total sus-
pended particulates (TSP) are not being met. This can be shown
by looking at the yearly average TSP levels across the United
States. Although certain industrial areas are hotspots of
air quality degradation, the high annual TSP levels in the
Western U.S. are difficult to explain or rationalize without a
consideration of open sources of particulate emissions. Many
of these are naturally or quasi-naturally occurring. Also, a
mounting body of data indicates that particulate matter in-
jected into the atmosphere at one location can be deposited
-125-
-------
at locations up to several hundred miles away. The implica-
tion is obvious; in regions where this deposition occurs,
enforcement of strict pollution control procedures on a local
basis will have little impact upon air quality.
On a mass emission basis, open sources certainly do ac-
count for a disproportionate part of the impact on particulate
air quality standards, as illustrated in Table I. Unpaved
roads predominate U.S. mass emissions of particulates, but the
influence of other open sources such as agricultural tilling,
construction activities, grain handling, crushed stone pro-
cessing, grain harvesting, and wind erosion of dormant soil
are not to be discounted. This table lists the top 12 partic-
ulate emission sources from an overview matrix which includes
670 sources emitting criteria pollutants.
Measurement
Each open source has to be measured and evaluated on an
individual basis. This is important because the method of
evaluation depends o how the data and information are col-
lected on the source. Two techniques of field sampling are
briefly described below.
Suppose that the airborne drift is to be measured from
an airplane applying a pesticide to a fieldcrop. The field
can be divided into finite grid elements, and samplers are
placed at selected locations upwind and downwind of the site,
as shown in Figure 1. As the applicator moves back and forth
across the field, samplers at each location can be turned
on and off to correspond to the application zone represented
by the dot in the figure. Correlation with downwind power
law decay and crosswind variations can be accomplished without
relocating a sampler site.
In this instance it may be important, both from a safety
as well as operational standpoint, that control be maintained
at some location where the proper perspective and overview of
the application of a toxic material could be observed. This
is achieved by utilizing a series of sampling devices which
are controlled by radio frequency, as illustrated by the sche-
matic diagram in Figure 2. Each individual unit can be turned
on and off at the appropriate times through the remote control
point. We have used a pulse modulation multiplex system which
is capable of 360 simultaneous manipulations and operations.
The operation could also be converted to automatic or computer
control through the installation of proximity sensors located
throughout the field being evaluated. With this technique a
concentration for some time period is measured and then cor-
related back to a given application rate or area treated.
-126-
-------
Table I. Overview matrix for air pollution sources - July 1975.
Source
Unpaved .Roads
Coal-Fired Steam Electric
Utilities
Agricultural Tilling
Oil-Fired Industrial/
Commercial Boilers
Construction Activities
Grain Handling
Wind Erosion of Soil
from Dormant Land
Crushed Limestone
Coal-Fired Industrial/
Commercial Boilers
Cement Production
Steel Production
Grain Harvesting
Crushed Stone/Traprock
Lime Kilns
Crushed Granite
Mass of Emissions (1,000 kg/yr)
Particulates
99,990,000
6,059,000
5,500,000
3,633,000
3,414,000
2,935,000
2,139,000
2,034,000
1,185,000
887,000
526,000
452,000
395,000
312,000
299,000
Percent of Total
73.30
4.44
4.03
2.66
2.50
2.15
1.57
1.49
0.87
0.64
0.39
0.33
0.29
0.23
0.22
-127-
-------
WIND
SAMPLER
FIELD
CONTROL
Figure 1. Drift sampling approach
-128-
-------
I
K>
K3
I
SAMPLE IMPINGERS ON STAND
VALVE
DIFFERENTIAL
PRESSURE RECORDER
&4 1
TO ATMOSPHERE
PUMP
Y,
ANTENNA
_l —
_J—
_»—
[
1
1
J
n
&
i—| j
££— .
SOLENOID VALVES
r
RECEIVER & BATTERY
DIFFERENTIAL
PRESSURE RECORDER
TO ATMOSPHERE
ORIFICE
PUMP
AC GENERATOR
Figure 2. Airborne drift sampling apparatus
-------
Another example of open source sampling techniques for
continuously or semi-continuously emitting sources is to ob-
tain many real-time concentration readings within a plume,
preferably at several downwind distances along the centerline.
Figure 3 shows such a system which is fully portable and con-
sists of a portable anemometer, a beta attenuator dust monitor,
and a personnel cassette sampler. Accurate (±25%) concentra-
tion readings can be rapidly taken with this system and it
reduces field sampling costs by about 75% compared to Hi-volume
sampler arrays. Also, particle size separation between the
respirable fraction (<10 v) and TSP (<50 y) can be accomplished.
Source Severity
A methodology has been developed which provides a consis-
tent basis for comparing various emissions generated by a
single and for comparing the environmental hazard between dif-
ferent single sources. Called the source severity, this multi-
faceted approach can compare, in a relative fashion, an open
source with an elevated point source, an open source with
another open source, one pollutant emission with another pol-
lutant emission, the components of a pollutant with the whole
pollutant (e.g., free silica, Pb, Cd), one location of emis-
sions with another location, and the effects of emission param-
eter changes. As an assessment tool, the source severity
approach has proved to be extremely flexible and useful. A
discussion of its development and types of use follow.
The air pollution severity of a given source should in
some way be proportional to the degree of potential hazard it
imposes on a population in its environment. The relative
hazard, H, from a specific emission can be defined as being
directly proportional to the delivered dose, the probability
of dase delivery, and the number of people who receive it, and
inversely proportional to the toxicity of the material as
follows:
S a H a
NPijj
LD50
(1)
where S = source severity
H = relative hazard
N = number of persons
LD50 = lethal dose for 50% of the people exposed
P = probability of dose delivery
ijj = delivered dose = B • R' • /x(t)dt
B = average breathing rate
R' = lung retention factor
x(t) = concentration time history
-130-
-------
ANEMOMETER
WEATHER POLE
ANEMOMETER
HOUSING
CYCLONE SEPARATOR
RESPIRABLE DUST
MONITOR
SAMPLING PLATFORM
STOPWATCH
TRIPOD STAND
Figure 3. Sampling apparatus
-131-
-------
The source severity, S, has been defined as the ratio of
the dose of a pollutant delivered to a population relative to
some potentially hazardous dose. Since LD50 data are not avail-
able for humans as a measure of hazardous dose, another measure
was used. The potentially hazardous dose for a given pollutant
from a specific point source is defined instead as:
7
NBR' / TLV(t]Kdt (2)
where ty^ = potentially hazardous dose, g
N = population exposed to a specific source,
persons
B = average breathing rate, m3/s-person
R' = lung retention factor for the pollutant of
interest (dimensionless factor, 0
-------
where x(t) = the actual ground level concentration time
history of a pollutant of interest emitted
by a specific point source, g/m3
The value of x(t) is very difficult to obtain and was therefore
approximated by an average value, 7. The total actual dose
delivered for a specific pollutant from a specific source is
then:
B • R
(7)
Since our measure of source severity was defined as the ratio
of the two dosages, then:
c = lA = N • B • R' • T • y
b ^F N • B • R* • T • F (8)
or r. Y
s = f (9)
The above term for source severity results in a simple ratio
of the time-averaged concentration to a potentially hazardous
concentration from a single source, although it is based on
dosage. Ambient concentration measurements around a source
are commonly considered as the primary entity and dosage
derived from them. From the experimental point of view, the
inverse is true; mostly dosage is observed and, at the price
of involved assumptions, a concentration valid for some
smoothing or sampling time is reported.
In order to calculate the source severity, "x/F> and the
affected population, the Gaussian plume dispersion model is
employed, which relates the concentration of pollutant oc-
curring at ground level at any given distance from the source
of emissions to the rate and height of the emissions and the
local wind velocity. The ground level concentration is of
primary importance since it ;is the one to which the human
population is exposed. Also, the rate and height of emissions
are quantities which can be readily measured and/or compiled.
For elevated point sources the maximum ground level con-
centration, Xmax Cref. 1), is used to determine the numerator
of source severity:
20 o
Y = L_ (10)
*max
-133-
-------
where x x = maximum ground level concentration (3-min
averaging time), g/m3
IT = 3.14
e = 2.72
u = wind speed, m/s
h = emission height, m
a = vertical dispersion coefficient, m
£
a = horizontal dispersion coefficient, m
Q = emission rate, g/s
The above equation yields a value for a short-term averaging
time (^3 min) during which the Gaussian plume dispersion
equation is valid. For a continuously emitting source, the
average maximum concentration, Ymax> is a function of sampling
time, t, and it can be related to Xmax as follows (ref. 2):
xmax xmax \ t2 / (11)
where tj = 3 min
t2 = 24 hr
p = 0.17
We use national average meteorological conditions since they
prevail at most locations and it simplifies the comparison
process; thus, atmospheric stability is approximately neutral
(class C assumed), wind speed is 4.5 m/s, and a ^ a (ref. 2)
y z
or:
a
— Si1-0 (12)
y
The severity equation becomes:
S = = \ / - (13)
o
TT -.r _
TLV -7T • TM
or
(2) (105) Q o
s = - z = b-b M — (14)
ireuh2 a (TLV) (TLV) h2
-134-
-------
Since the criteria pollutants (participates SO NO
and HC) have established ambient air quality standards
propriate standard (in g/m3) is substituted for ^efo
hazard factor, F. Severity equations for each of the five
criteria-pollutants Ve alS° been derived, and the only data
rn
'
For most open sources, which are ground level sources
(h = 0) , Xmax occurs by definition at the nearest plant bound
ary or public access. Since this also occurs when y = 0, the
appropriate equation to calculate the "maximum" concentration
is (ref. 2):
x =
C15)
By letting D equal the distance to the occurrence of Xmax for
ground level sources (ref. 3):
ay = 0.209D0-903
az = 0.113D0-911
Thus, Xmax i-5 determined as follows:
= 42.56 Q
(16)
(17)
It will be noted that, under average meteorological conditions,
the xmax equations are identical with the algebraic substitu-
tion of:
h2 = 0.01737D1-81"
(19)
If a ground level source is a line or area source rather than
a point source, it is treated in a similar fashion to calculate
source severity.
Once source severities have been determined, one of the
uses is to compare locationally a source's impact on the popu-
lation. Affected population designates the number of persons
exposed to high concentrations, e.g., those for which ymdx/F
>1.0. This quantity is useful in characterization because a
given source may have a high severity, yet, if it is located
in a sparsely populated area it may have only a small effect
on human health.
-135-
-------
Another form of the plume dispersion equation is needed
to calculate the affected population since the population is
assumed to be distributed uniformly around the source. If the
wind directions are taken to 16 points and it is assumed that
the wind directions within each sector are distributed ran-
domly over a period of a month or season, it can be assumed
that the effluent is uniformly distributed in the horizontal
within the sector. The appropriate equation for the average
concentration as a function of distance, x(X), is (ref. 2):
Y(X) =
2.03 Q
a uX
exp
(20)
where
Q =
h =
X =
u =
a_ =
emission rate, g/s
effective emission height, m
downwind distance from source
4.5 m/s
0.113 x°-911 = vertical dispersion
coefficient m
m
To find the distances at which x(X)/F > 1.0, roots are deter
mined for the following equation:
-1.0
(21)
Since Equation 21 is a transcendental equation the roots are
found by an iterative technique using the computer.
For a specified emission from a typical source,
function of distance might look as follows:
x_
F
O.I _
as a
DI STANCE FROM SOURCE
-136-
-------
with Xi and X2 being the distances where S = 1 0 If those
distances are swept through 360°, an annulus is obtained con
taming the affected population:
The affected area in km2 is then computed as:
A = Tr(X22 - Xl2) (22)
where Xi and X2 are the two roots of Equation 21.
The population density, Dp, of the area of concern is
determined and the product A • Dp is designated as the affected
population and reported as number of persons.
Experimental Results
In order to apply source severity methodology in an assess-
ment of open source types, extensive field sampling was con-
ducted so that emission rates could be compiled. The simplest
way to take a "first look" at many open sources is to sample a
source thought to be representative of the whole and estimate
the emission factor. Tables II through V, present the emission
factors determined for open coal mining and storage, agricultural
activities, fugitive emission sources at grain elevators, and
crushed stone quarrying and processing, respectively. The em-
phasis was given to determining the fine particle, or respirable
(<7 ym), fraction where dust emissions were concerned. It is
this fraction which will impact human health and is likely to
disperse extensively beyond plant boundaries.
Comparison of Emission Factors with Previous Studies
Table VI compares the emission factors as determined by
downwind ambient sampling with those estimated in the Compila-
tion of Air Pollutant Emission Factors (ref. 4). As can be
seen, the emission factors for crushed stone processing as
determined in our work are two orders of magnitude lower than
previous estimates.
-137-
-------
Table II. Coal emission factors.
Source Type
Surface Coal Mining
Drilling
Coal Loading
Transport § Unloading
Blasting
Augering
Coal Storage
Emission Factor,
Ib/ton (resp)
0.025
3.2 x 10"3
3.4 x 10"3
9.7 x 10"3
8.3 x 10"3
1.6 x 10'" 3
0.013
Table III. Agricultural emission factors.
Source Type
Cotton Defoliation
(w/Arsenic Acid)
Cotton Harvesting
Picking
Stripper
Grain Harvesting
Emission Factor
12.2
2.63
24.6
2.34
Ib/ton
lb/mi2
lb/mi2
lb/mi2
(resp)
(resp)
(resp)
Table IV. Grain elevator emission factors.
Unit Operation
Emission Factor,
Ib/ton (resp)
Truck Unloading
Uncontrolled
Controlled
Truck Loading
Hopper Railcar Loading
Ship Loading
0.272
0.0028
0.0011
0.0031
0.0013
-138-
-------
Table V. Emission factors for aggregate unit operations.
u>
VO
I
Unit Operation
Drilling
Blasting
Loading at Quarry and
Vehicular Transport to Plant
Unloading and Primary Crushing
and Screening
Secondary Crushing and Screening
Conveying
Unpaved Road Traffic Between
Finished Stockpile and Nearest
Paved Highway
TOTAL
Fraction Respirable, by Weight
Emission Factor for
TSP in Ib/ton of Material
Processed Through Primary Crusher
Granite
-
0.16
_
-
0.044
-
0.048
0.25
61
Traprock
-
-
0.02
0.026
0.002
0.004
0.002
0.056
11%
Limestone
-
-
0.004
0.002
-
-
-
0.006
171
Quartzite
0.060
-
0.34
-
0.024
-
-
0.42
6%
-------
Table VI. Comparison of the emission factors at traprock plants
for AP-42 and MRC sampling.
Operation
Primary Crushing
Secondary Crushing
and Screening
Tertiary Crushing
and Screening
Fines Milling
Recrushing and
Screening
TOTAL
AP-42
Uncontrolled
Total
Emissions,
g/raetric ton
250( 0.5)a
7SO( 1.5)
3,000( 6.0)
3,000( 6.0)
2,SOO( 5.0)
9,500(19.0)
Percent
Settling
in Plant
80
60
40
25
50
41
Suspended
Emissions ,
g/metric ton
50( 0.1)
300( 0.6)
1,800( 3.6)
2,250( 4.5)
1,2SO( 2.5)
5,605(11.3)
MRC (Sampling at Traprock Plants)
Total
Particulate
Emissions,
g/metric ton
13 (0.026)
1 (0.002)
0.4(0.0008)
0.1(0.0002)
14.5(0.03)
Percent
Respirable
Emissions
10
55
18
16
15
Respirable
Emissions ,
g/metric ton
1.3(2.6xlO-3)
0.6(1.2xlO"3)
0.1(2.0x10-")
0.2(4.0x10-")
2.2(4.4xlO~3)
Note: Blanks indicate no data reported.
a' Numbers in parentheses are in English units, i.e., pounds per ton.
-------
There are several possible explanations for these differ-
ences. Our emission factors were determined by measuring am-
bient air concentrations around a source and then calculating
the emission rate using a dispersion equation. The samplers
were placed about 30 m to 40 m away from the source and thus
did not measure particles, that settled between the source and'
the samplers. Based on a particle count on high volume filters
the samplers did not collect particles more than 40 ym in diam-'
eter. Thus, the emission factor for total particulates in-
cludes only particulates less than 40 ym in diameter.
On the other hand, the previous emission factor estimates
were based on the results of sampling the dust loading at the
inlet of a baghouse used to control dust emissions from crushing
and screening operations at a crushed traprock plant, and on the
assumption that about 41% of the emitted particulates settle
within the plant. High air flows encountered in baghouse ducts
cause the entrainment of very large particles (up to about
350 ym in diameter) , so the emission factor for total partic-
ulates based on the dust loading at the baghouse inlet may be
an inflated value.
One of the purposes of determining an emission factor is to
estimate the ambient air concentrations due to a source. The
sampling methods outlined in this report allow a more realistic
estimate of source emissions than previously reported since
they are based on measurements of air concentrations around the
contributing source of interest. Based on this, the emission
factors in Tables TI through V will predict, using the same
dispersion methodology, real-life ambient air contributions of
sources.
Future plans call for a more extensive evaluation of the
major sources of emission, overall mass emissions, respirable
mass emissions and composition of some of these sources, and
a more complete analysis in order to best judge relative
severity and emission,burden.
Acknowledgment
Funds for conducting this research were provided by the
Industrial Environmental Research Laboratory of the Environ-
mental Protection Agency at Research Triangle Park> North
Carolina.
-141*
-------
References
1. Meteorology and Atomic Energy 1968, D. H. Slade (ed.)?
U.S. Atomic Energy Commission, (NTIS TID-24190), July
1968, 445 p.
2. D. B. Turner, Workbook of Atmospheric Dispersion Estimates,
U.S. Department of Health, Education, and Welfare, Cincinnati,
Public Health Service Publication No. 999-AP-26, May 1970,
65 p.
3, E. C. Eimutis and M. B. Konicek, Derivations of Continuous
Functions for the Lateral and Vertical Atmospheric Dispersion
Coefficients, Atmospheric Environment, 6^:859-863, March 1972.
4. Compilation of Air Pollutant Emission Factors, U. S. Environ-
mental Protection Agency, Research Triangle Park, Office of
Air Programs, Publication No. AP-42, 1972, pp. 8-19.
-142-
-------
THE IMPACT OF FUGITIVE EMISSIONS
OF FINE PARTICLES
by
Chatten Cowherd, Ph.D.
Midwest Research Institute
-143-
-------
THE IMPACT OF FUGITIVE EMISSIONS
OF FINE PARTICLES
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 contri-
bute 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. As shown in Figure 1, most of
the recognized adverse effects are attributable to fine particles, i.e.,
particles smaller than about 10 p,m in equivalent aerodynamic diameter,
which may be transported over distances of regional scale. By way of con-
trast, soiling phenomena which result from the rapid settling of coarse
particles, tends to be confined to localized areas in the vicinity of the
source.
This paper focuses on two topics critical to the assessment of the
impact of fugitive emissions of fine particles: (a) methods for fugitive
emissions quantification, and (b) analysis of the potential for atmospheric
transport of particulate matter.
Methods for Fugitive Emissions Quantification
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.
In large part, proven methods for quantifying fugitive emissions have
not been fully developed. Atypical quantification problems are presented
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.
-144-
-------
I
j
Atmospheric
Electricity
Atmospheric
Visibility
Condensation
Nuclei for Precipitation
Soiling Phenomena
(Horizontal Surfaces)
Upper Respiratory
Tract Deposition in Man
Peripheral Airways and
Alveolar Deposition-Man
Soiling Phenomena
(Verticol Surfaces)
Atmospheric
Chemistry (Gos-Solid)
W//////////
y///////////.
/ Main Aerosol Moss
Comprising//^//
////.
I
irr
10-3 10~2 10'1 10°
Particle Diameter -
10'
102
Figure 1 - Effects of Particulate Air Pollution versus Particle Size'
-------
MRI has employed two basic sampling strategies to quantify emissions
from open dust sources: exposure profiling and dispersion model calcula-
tion. Each of these techniques is discussed below.
Exposure profiling; Particulate emissions from an open source can be
determined directly by measuring the total passage of airborne particulate
matter (after subtraction of background) at some distance downwind of the
source. This method is feasible (a) for area sources smaller than about
0.1 km in diameter or (b) for line (elongated area) sources of width smaller
than 0.1 km.
The passage of airborne particulates can be obtained by spatial integra-
tion (over the effective cross-section of the plume) of distributed measure-
ments of exposure (mass/area). The exposure is the point value of the flux
(mass/area-time) of airborne particulates integrated over the time of measure-
ment. Mathematically stated, the total mass emission rate is given by
1 ff «(h.w)
tJJ.
where m = particulate catch by exposure sampler after subtraction of
background
a = effective intake area of sampler
t = sampling time
h = vertical distance coordinate
w = lateral distance coordinate
A = effective cross-sectional area of plume
In order to obtain an accurate measurement of airborne particulate ex-
posure, 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.
A variation in the exposure profiling method has recently been developed
by MRI for application to buoyant plumes emitted by pyrometallurgical processes^
This method utilizes a two-dimensional grid of thermocouples, horizontally
-146-
-------
distributed above the source, for temporal and spatial characterization of
the plume. Laboratory experiments have shown that plume temperature and con-
centration profiles are linearly related.
Dispersion model calculation^: Atmospheric dispersion models can also
be used to calculate rates of emission from area sources based on ground-
level measurement of upwind and downwind air quality under known meteorolog-
ical conditions. This method of emissions quantification is required when
the extent of vertical dispersion at the downwind edge of the source is too
large to allow for ground-based characterization of the vertical exposure
profile.
Atmospheric Transport of Fugitive Dust
This section presents an assessment of factors which determine the
drift distances of fugitive dust particles in the atmosphere. Drift dis-
tance is defined as the horizontal displacement from the point of parti-.
culate injection to the point of particulate removal by ground-level
deposition.
Factors to be considered in this assessment may be grouped into two
categories:
1. Meteorological factors - properties of the atmosphere which affect
contaminant advection and turbulent diffusion over surfaces of varying
roughness scales.
2. Source factors - height of injection and particulate properties
which affect gravitational settling and vertical mixing.
This assessment does not treat atmospheric washout of particulate matter.
Meteorological factors: Fugitive dust particles are typically injected
into the lower portion of the "surface layer" region of the atmosphere which
extends from ground level to a height of about 100 m. In this region the pro-
file of the wind and its turbulence characteristics are strongly dependent
on surface roughness properties.
For neutral atmospheric stability, the vertical profile of mean wind
speed, u(z) , in the surface layer is described by a logarithmic relationship:
u(z) .JL*(L\ (2)
-147-
-------
where u* = friction velocity
k = von Kantian1 s constant (0.4 for clear fluids)
z = surface roughness height
Neutral stability occurs with wind speed exceeding 12 mph or with over-
cast conditions regardless of wind speed.
The friction velocity, u^ , is related to the rate of momentum exchange
at the surface:
u* " T0/Pa) (3)
where TO = surface shear stress
pa = density of air
Within the surface layer, the vertical flux of momentum (and hence u.) is
known to be roughly constant and the eddy diffusivity is given by
c (z) - ku* z (4)
Aerodynamic roughness height, z0 , is related to the size, shape and
spatial density of the roughness elements. Based on similarity concepts
Lettau2/ has derived the following expression for evenly spaced elements:
(5)
where H = effective height of roughness elements
a = silhouette area normal to the wind
A » total ground area per element
1/2 • average drag coefficient.
Figure 2 gives roughness heights for various natural and man-made
roughness features.
-148-
-------
High Rise Buildings.
(30+Floors) I/
Suburban
Medium Buildings-
(Institutional) I/
u
X
O
UJ
X
«2 Suburban
Z
x
O
§
2/
Residential Dwellings
Wheat Field •*•/
Plowed Field
Zo (cm)
1000
—800-
—600-
—400-
—200-
—80.0-
—60.0-
—40.0-
-20.0-H
10.0
4/
Natural Snow— •
100
-8.0
-6.0
-4.0-
-2.0-
1.0
I—0.8-
-0.6-
U0.4-J
-0.2—J
0.1
Urban Area
Woodland Forest -2
I *^ /
r Grassland —'
Figure 2 - Roughness Heights for Various Surfaces
-149-
-------
Source factors; The primary source factors which affect the drift
distance of a fugitive dust particle are injection height, h , and particle
settling velocity, Vs , which may be approximated by the Stoke's relation-
ship:
V = 0.00301 p D2 (6)
s p
where V = terminal settling velocity (cm/sec)
s
3
p = density of particle (g/cm )
P
D = particle diameter (p-m)
Fugitive dust particles typically have a mineral composition with a density
of about 2.5 g/cm3.
Calculation of drift distance: In the past, most analyses of the
atmospheric dispersion of particles with appreciable settling tendencies
have focused on the distribution of settling rate, S(x) , expressed as:
S(x) = V C (x) (7)
s o
where C — the ground-level concentration of particulate with settling
velocity V
x = downwind distance from the source
Accordingly, an Eulerian approach to the problem has been taken.
However, analysis of particle drift with no net effect of atmospheric
turbulence, is most conveniently treated by a Lagrangian approach. This is
illustrated in the following section.
Case 1: Monodisperse particles, single injection height,
negligible turbulence effect.
Consider the case of a steady stream of monodisperse particles
released from a continuous crosswind line source at height h . It is assumed
that each particle during its lifetime in the atmosphere is subjected to a
balanced set of vertical turbulent velocity fluctuations with the result that
the particle does not deviate appreciably from the trajectory it would have
in the absence of turbulence.
-150-
-------
The vertical position, z , of the particle as a function of time
is given by
z (t) = h-V t
P s
Substitution of Eq. (8) into Eq. (2) gives the following expression for
the horizontal speed of the particle:
(9)
The particle drift distance, x , is given by:
(10)
where the upper limit of integration is the lifetime of the particle in
the atmosphere. Integration of Eq. (10) yields
(11)
To determine the effect of injection height and roughness height
on the drift distance of particles of given aerodynamic sizes, the wind
speed at z - 100 m was fixed at 6.9 m/s (15.4 mph) and friction velocities
were determined from Eq. (2). The results are shown in Table 1 for injec-
tion heights of 1, 3 and 10 m and for roughness heights spanning the range
given in Table 1. Figure 3 shows the variations of Xp for h = 3 m, measured
above z .
o
As expected, for particles of a given size, drift distance in-
creases with injection height and decreases with roughness height. The latter
effect is a direct result of the decrease in wind velocity near the surface
caused by obstacles to the flow.
Case 2; Monodisperse particles, single injection height, turbulent
atmosphere.
-151-
-------
Table 1. PARTICLE DRIFT DISTANCES CALCULATED FROM EQ. (11)
NO
I
Injection Roughness
height,^ height,
h zo
(m) (m)
1 0.01
0.05
0.10
0.50
3 0.01
0.05
0.10
0.50
1.00
10 0.01
0.05
0.10
0.50
1.00
Friction
velocity,
u*
(cm/ sec)
30.0
36.4
40.0
52.2
30.0
36.4
40.0
52.2
60.0
30.0
36.4
40.0
52.2
60.0
Drift distance, xp , by particle size
30 Rm
40.6 m
29.5
24.2
12.5
157.1 m
128.2
112.9
73.5 ,
56.4
655 m
582
541
423
363
20 urn
91.2 m
66.4
54.4
28.1
353 m
288
254
165
127
1,474 m
1,309
1,216
952
816
10 urn
366 m
266
218
113
1,418 m
1,157
1,019
663
509
5.92 km
5.25
4.88
3.82
3.28
5 um
1,460 m
1,060
871
450
5.66 km
4.62
4.07
2.65
2.03
23.6 km
21.0
19.5
15.3
13.1
1 urn
36.6 km
26.7
21.8
11.3
141.8 km
115.7
101.9
66.3
50.9
592 km
525
488
382
328
a/ Injection height measured above roughness height
-------
Ul
OJ
Injection Height (h) = 3 m above zo
——— Natural Snow (zo= 0.1 cm)
Plowed Field (z0= 1.0cm)
———— Grassland (zo = 3.0 cm)
——— Suburban Residential
Dwelling (ZQ- 5.0 cm)
——— Suburban Medium
Building (zo= 70.0 cm)
I 1 1 1^» i ' I
JO3 104
DRIFT DISTANCE (meters)
105
Figure 3 - Relationship Between Particle Size and Drift Distance
-------
The analysis presented under Case 1 assumed that all particles
generated from a particular fugitive dust source were deposited at the same
point downwind (x ). Clearly, however, particles subjected to a preponderance
of downward turbulent velocity fluctuations will settle from the atmosphere
at distances less than xp and particles propelled above the trajectory de-
fined above may drift far beyond xp . In other words, because of the random
nature of turbulent velocities, x^ approximates the distance at which half
of the particles have deposited on the surface.
The specific question addressed here has to do with the form of
the settling rate distribution. Recalling Eq. (7), this problem reduces to
finding the distribution of ground-level concentration by solving the ap-
propriate transport equations and accompanying boundary conditions.
The phenomena of quasi-steady advection and turbulent diffusion
from a continuous line source under the condition of uniform wind speed is
described by the following equation:
u ac = pu4_f z *cY .v dc
)
dx ' dzV dzl 's dT (12)
where C = particulate concentration
U = uniform speed of crosswind
p = turbulence parameter
The uniform wind speed, U , is assumed to have the value given by the Case
1 velocity profile at z = h. The quantity pUz becomes the coefficient of
eddy diffusivity.
Although Eq. (12) is not amenable to analytical solution for the
.case in point, it has been shown&/ that the distribution of ground-level
concentration has the following form:
-h/px
00(30 = A
where A « constant
-154-
-------
The function given in Eq. (13), and hence the settling rate, reaches
maximum at:
(14)
and then decays to zero as x-* <*> . Values for x^.^ are given in Table 2
based on values of p determined by comparing the two forms of the eddy
diffusivity, yielding
p = ku^/U (15)
The constant A in Eq. (13) may be evaluated by equating the
emission rate E to the integrated settling rate.
V * a <16>
o
With the transformation y = b/x where b = h/p , the above equation
becomes
(17)
b1
where TO*) is the gamma function.
Similarly it can be shown that the mass fraction K of particles
remaining suspended beyond some distance x is given by:
(18)
where the incomplete gamma function r(or,b/x) is defined as
F (",£)=/ e-yy«-*dy (19)
o
-155-
-------
Table 2. DISTANCES TO POINT OF MAXIMUM SETTLING,
CALCULATED FROM EQ. (14)
ON
I
Injection Roughness
height, height,
h zo
(m) (m)
1 0.01
0.05
0.10
0.50
3 0.01
0.05
0.10
0.50
- 1.00
10 0.01
0.05
0.10
0.50
1.00
Turbulence
parameter,
P
0.0347
0.0534
0.0695
0.2308
0.0281
0.0391
0.0470
0.0893
0.1456
0.0232
0.0302
0.0347
0.0534
0.0695
Friction
u*
(cm/sec)
30.0
36.4
40.0
52.2
30.0
36.4
40.0
52.2
60.0
30.0
36.4
40.0
52.2
60.0
Values of or and XQ,
30
or
0.564
0.465
0.423
0.324
0.564
0.465
0.423
0.324
0.282
0.564
0.465
0.423
0.324
0.282
Urn
xma,xj
18.4
12.8
10.1
3.27
68.3
52.4
44.9
25.4
16.1
276
226
203
141
112
20 urn
a ,Xmax,
0.251 23.0
0.207 15.5
0.188 12.1
0.144 3.79
0.251 85.3
0.207 63.6
0.188 53.7
0.144 29.4
0.125 18.3
0.251 345
0.207 274
0.188 243
0.144 164
0.125 128
ax (m) by particle size
10
or
0.0625
0.0515
0.0469
0.0359
0.0625
0.0515
0.0469
0.0359
0.0312
0.0625
0.0515
0.0469
0.0359
0.0312
Urn
Xmax
27.1
17.8
13.7
4.18
100.5
73.0
61.0
32.4
20.0
406
315
275
181
140
5
Of
0.0157
0.0129
0.0118
0.0090
0.0157
0.0129
0.0118
0.0090
0.0078
0.0157
0.0129
0.0118
0.0090
0.0078
Vim
"max
28.4
18.5
14.2
4.29
105.1
75.7
63.1
33.3
20.4
424
327
285
186
143
1 urn
or
0.00062
0.00052
0.00047
0.00036
0.00062
0.00052
0.00047
0.00036
0.00031
0.00062
0.00052
0.00047
0.00036
0.00031
"max
28.8
18.7
14.4
4.33
106.7
76.7
63.8
33.6
20.6
431
331
288
187
144
-------
The above analysis assumes that particles of all sizes are uni-
formly responsive to turbulent diffusion. More realistically, the time
constant of particle response to vertical velocity fluctuations increases with
increasing aerodynamic particle size.
In studies of the vertical flux of particulates over an agricul-
tural field undergoing wind erosion, Gillette et al.-£' have characterized
this phenomena in terms of the ratio Vg/u* . If settling velocity is small
compared to the root mean square velocity fluctuation, i.e., Vs/u*'< 0.1,
the particulate is dispersed as a gas. On the other hand for Vs/u* '** 1,
settling effects begin to predominate. Clearly, in the latter case, the
settling distribution is more strongly focused around the distance X .
P
Case 3; Polydisperse particles, distributed injection height,
turbulent atmosphere.
This case is treated by separately analyzing the dispersion of
particles within narrow size ranges and injection height ranges and by
superimposing the results. The analytical techniques to be used are those
described above.
-157-
-------
References
1. Corn, M., "Particle Size: Relationship to Collector Performance, Emis-
sion Standards and Ambient Air Quality," presented at Technical Session
41 of the Second International Air Pollution Conference of the Inter-
national Union of Air Pollution Prevention Associations, Washington,
D.C., December 10, 1970.
2. Lettau, H. H., "Physical and Meteorological Basis for Mathematical
Models of Urban Diffusion Processes," Chapter 2, Proceedings of
Symposium on Multiple-Source Urban Diffusion Models. U. S. Environ-
mental Protection Agency, Publication No. AP-86 (1970).
3. Davenport, A. G., "The Relationships of Wind Structure to Wind Loading,
in Wind Effects on Buildings and Structures," National Physical
Laboratory, Symposium 16, Her Majesty's Stationary Office, London
(1965).
4. Deacon, E. L., "Vertical Diffusion in the Lowest Layers of the
Atmosphere." Quarterly J. Royal Meteorological Society. 7Ji.:89 (1949).
5. Gillette, D. A., and P. A. Goodwin, "Microscale Transport of Sand-
Sized Soil Aggregates Eroded by Wind." J. of Geophysical Research.
79.(27):4080-4084, September 20, 1974.
6. Bosanquet, C. H., and J. L. Pearson, "The Spread of Smoke and Gases
from Chimneys," Trans. Faraday Soc.. 32_:1249-1264 (1936).
7. Gillette, D. A., and I. H. Blifford, Jr., "The Influence of Wind
Velocity on Size Distribution of Aerosols Generated by the Wind
Erosion of Soils." J. Geophysical Research. 79(27):4068-4075,
September 20, 1974.
-158-
-------
PEDCo - ENVIRONMENTAL
SUITE13 • ATKINSON SQUARE
CINCINNATI. OHIO -45246
513 / 77 1 -433O
FACTORS INFLUENCING EMISSIONS FROM FUGITIVE
DUST SOURCES
by
George A. Jutze
Kenneth Axetell
Presented at the
"SYMPOSIUM ON FUGITIVE EMISSIONS"
May 16-19, 1976
Hartford, Connecticut
Sponsored by: U.S. ENVIRONMENTAL PROTECTION AGENCY
BRANCH OFFICES
Suite 110. Crown Center Suite 104-A, Professional Village
Kansas City, Mo. 64108 Chapel Hill, N.C. 27514
-159-
-------
FACTORS INFLUENCING EMISSIONS FROM FUGITIVE
DUST SOURCES
by George A. Jutze and Kenneth Axetell
Emissions from fugitive dust sources are known to be
highly variable over time and geographic area and dependent
on a number of different parameters. While developing
emission factors for fugitive dust sources during the past
three years, PEDCo has attempted to identify the specific
parameters which influence emission rates most and to derive
general expressions describing the relationships between
these parameters and resulting emissions (correction func-
tions to the emission factors).
At least six different parameters have been found to
substantially affect fugitive dust emission rates:
0 soil type
0 windspeed
0 surface moisture
0 precipitation
0 vegetative cover
0 traffic activity across the surface.
This paper discusses the theoretical basis for the relation-
ship between each of these parameters and emission rate and
presents a quantitative method for estimating the effect of
each.
-160-
-------
SOIL TYPE
Soil type affects almost all fugitive dust sources,
since it is usually the native soil from the surface of the
source that becomes airborne as the suspended particulate
emission. Two different characteristics of soil are most
important in determining emission rates: (1) soil struc-
ture, or its resistance to breakdown into its component
particles; and (2) soil texture, or the size distribution of
the individual component particles. Such factors as cloddi-
ness (soil cohesion), surface crusting, bulk density, and
mechanical stability give a soil its "structure." Soil
texture provides an indication of the potential amount of
suspended particulate that can result if wind erosion or
repeated mechanical disturbance of the soil causes the
disintegration of the soil's natural structure.
The particle size distribution, or texture, of a soil
is an important factor in defining a soil's structure. For
example, soils containing a high proportion of silt form the
most stable clods and most compact crust. Since the two
soil characteristics which affect emission rate are somewhat
dependent on one another, soil scientists have generally
used soil texture as the single basis for comparing the
erodibilities and potential dust emission rates of different
soils.
The standard measure of dust losses from soil surfaces
is the erodibility index, I, in units of ton/acre/yr. This
index estimates the relative erodibility of different soil
surfaces, not the actual quantities erodible under field
conditions. The erodibility indexes for different soil
textures are shown in Table 1.
In previous fugitive dust emission factor development,
either the erodibility index or the percent silt was used as
the correction function to account for different soil
-161-
-------
Table 1. ERODIBILITY AND AVERAGE PERCENT SUSPENDED
PARTICULATE-SIZED PARTICLES BY SOIL TYPE
Soil texture
Erodibility index,
ton/acre/yra
Aggregates < 0.05 nun,
percent by weight*3
Sand
Loamy sand
Sandy loam
Clay
Silty clay
Loam
Sandy clay loam
Sandy clay
Silt loam
Clay loam
Silty clay loam
Silt
220
134
86
86
86
56
56
56
47
47
38
38
0.9
1.0
2.1
0.8
0.8
6.6
4.1C
1.0C
4.1
2.5°
4.1
0.8C
Source: D. G. Craig, J. W. Turelle. Guide for Wind Ero-
sion Control on Cropland in the Great Plains States.
U.S. Department of Agriculture, Soil Conservation Ser-
vice. July 1964. Table 1.
Source: W. S. Chepil. Soil Conditions that Influence
Wind Erosion. U.S. Department of Agriculture, Agri-
cultural Research Service and Kansas Agricultural
Experiment Station, Washington, D.C. Technical
Bulletin Number 1185. June 1958. Table 1.
Estimated from most similar soil types.
-162-
-------
234
types. ' ' However, both of these parameters have short-
comings. The erodibility index estimates the relative soil
loss for different soil types, but does not adequately
account for the greatly varying percentages of that total
soil loss which are in the suspended particulate size range
and thus available to become airborne. Example percentages
are also shown in Table 1. Most of the eroded soil is
suspended for a short distance and then resettles, or is
pushed along the ground. Movement of larger soil particles
near ground level is referred to as saltation or surface
creep.
The use of silt content as the correction factor is an
attempt to account for the amount of soil in the suspended
particulate size range, but it does not consider the struc-
tural resistance of silty soils to breakdown into the indi-
vidual particles.
Comparative particle size ranges for different soil
types and soil movements are shown in Figure 1. The par-
ticle size limits shown in the figure are approximate--they
depend on particle density, shape, and surface forces as
well as size. Loam, the fourth soil texture used in the
classification system, is a mixture of clay, silt, and sand.
It is recommended that the relative dusting potential
of different soil types be estimated from the product of
their erodibility index and average percent of suspended
particulate-sized particles (less than 0.05 mm):
Soil texture Correction weighting factor
Sand 1-98
Loamy sand 1.34
Sandy loam 1.81
Clay 0.69
Silty clay °-69
-163-
-------
20
100
840
suspended
settleable saltation, non-
surface erodible
creep
0.5
respirable
30
hi-volume sampler collection
74
clay
silt
sand
J_
0.1
Figure 1.
1.0
10
100
Particle diameter, microns
Particle size ranges for soil types and soil movements,
1000
-------
Soil texture Correction weighting factor
Loam 3-70
Sandy clay loam 2.30
Sandy clay 0.56
Silt loam 1.93
Clay loam 0.94
Silty clay loam 1.56
Silt 0.30
The normal crusting characteristics of the soils are
considered in their erodibility indexes. However, under
certain conditions natural crusting can be enhanced by
surface treatment with chemicals, compaction, or controlled
watering. A soil surface that is well crusted loses dust at
a rate about one-sixth that of the same soil in a non-
crusted state. If the surface is not completely crusted,
or if the crust is weathered or damaged by traffic, the
emission rate will be intermediate between an unprotected
and a well-crusted surface.
A study of particle size distributions in the atmos-
phere indicated that these distributions are quite similar
in all regions and appear to be independent of the prevail-
ing soil types in the region (except in the case of dust
storms). This conclusion is certainly not in conflict with
the assumption made in the above discussion that emission
rate from a soil surface is a function of the percent of
soil particles in the suspended particulate size range.
Essentially, the same sizes of particles appear to contri-
bute to atmospheric particulate concentrations regardless of
the size distribution in the parent soil.
Soils undergo seasonal changes in their erodibility as
a result of biological activities, alternate wetting and
drying, and freezing and thawing. These actions decrease
-165-
-------
cloddiness and mechanical stability of the soils in winter
in all areas where the soil is moistened at least occasion-
ally and increase cloddiness and mechanical stability during
the summer. The amount of seasi
is also a function of soil type.
the summer. The amount of seasonal change in erodibility
WINDSPEED
Windspeed affects emission rates primarily for the wind
erosion sources, e.g., cleared land and agriculture.
Several investigators have found that when windspeed is
greater than that required to barely move the soil, the rate
of soil movement is directly proportional to the friction
velocity cubed. The friction velocity at the surface is
proportional to windspeed measured at a height above the
ground, so the rate of soil movement is also proportional to
the ambient windspeed cubed after it attains some minimum or
threshold speed necessary to initiate movement.
One investigator reported threshold speeds to range
from 13 to 30 miles per hour (mph) at 1-foot height, depend-
ing on the history of the field. Other investigators have
found threshold velocities as low as 9 and 11 mph at about
8 9
the same height. '
The threshold velocity is influenced by the size of
soil particles on the surface, and is lowest for aggregates
of 0.10 to 0.15 mm in diameter. As shown in Figure 2, the
threshold velocity increases with either an increase or
decrease in the size of particles outside this range. The
high resistance of fine dust particles to erosion by wind
appears to be due partially to cohesion but mainly to the
fact that the particles are too small to protrude above a
laminar and viscous layer of air close to the surface of the
ground. Fine dust is lifted from the ground mainly by
impacts of larger grains, which are more erodible because
-166-
-------
0)
to
o
100
+> 80
•*H
o
O
rH
-------
they protrude farther into the fast moving, turbulent
currents of higher air.
Turbulence is as important as average forward velocity
in lifting and transporting the soil. Thus, maximum momen-
tary wind velocity and gustiness also are factors in deter-
mining the impact of windspeed on fugitive dust emission
rates. All surface wind velocities over about 2 mph create
some turbulence.
One research report recommended that the rate of soil
movement be estimated as being proportional to the windspeed
cubed times its duration above the threshold speed. In
practice, it is difficult to obtain measurements of the
duration above a given windspeed, such as 13 mph, because
climatological summaries only record one- or three-hour
averages. These averages are normally less than 13 mph, but
may contain significant periods of time when the windspeed
is greater than 13 mph. To avoid this problem, U.S. Depart-
ment of Agriculture (USDA) researchers have assumed that all
locations have a similar windspeed frequency distribution
and, therefore, that wind erosion potential in an area is
proportional to the cube of the average windspeed for any
specified period.
This assumption in turn creates a problem with lack of
standardization in height and exposure of recording anemom-
eters; i.e., an anemometer at 30 ft may record an average
windspeed one to two mph higher than the same instrument at
20 ft height. Since the erosion rate is related to the cube
of windspeed, this small difference causes a change of 30 to
100 percent in the resulting correction function.
The climatic factor, C, in USDA's wind erosion equation
is an available correction function that includes the aver-
age windspeed cubed term and also a term (Thornthwaite's
precipitation-evaporation index) to account for surface
moisture. Values for C have been compiled for the entire
-168-
-------
country with the exception of the four southwestern states
of California, Nevada, Arizona, and New Mexico. These C
values are shown in Figure 3.
Recognizing the limitations to accuracy of any correc-
tion function that utilizes average windspeed data, the
climatic factor is recommended as the correction function to
reflect emission differences due to- windspeed. Many avail-
able emission factors already employ this term.
SURFACE MOISTURE
The soil moisture content of a fugitive dust source
directly affects the emission rate from the source. For any
surface moisture content above air-dried, a distinct decrease
12
in erodibility is observed. Soil movement is reduced only
slightly by the first moisture, but decreases more rapidly
with additional moisture. At a moisture content approxi-
mately corresponding to the permanent wilting point for
vegetation, soil loss is completely arrested. For different
soil types, the percent water required to eliminate soil
12
loss is:
Soil type Moisture content, %
Dune sand 1.28
Sandy loam 3.89
Silt loam 11.21
Silty clay 20.71
Average surface moisture contents of soils are not
routinely measured. Instead, Thornthwaite's precipitation-
evaporation (P-E) index, calculated from the sum of 12
monthly ratios of measured precipitation to measured evapo-
transpiration, is used as an indicator.
-169-
-------
o
Figure 3. Climatic factors for use in the wind erosion equation,
Source: Armbrust, D. V. and N. P. Woodruff, 1968.
-------
Empirical data show that soil loss varies inversely as
the square of the P-E index. The P-E index value squared is
included in the C factor previously described in the WINDSPEED
section. Again, the C factor is recommended for use because
of its acceptance as a broadly applicable correction factor.
Because of large variations in monthly rainfall, the
monthly P-E values do not give meaningful monthly climatic
factors. Also, the P-E index is not precise enough to
evaluate soil moisture conditions for periods as short as
one month. The U.S. Department of Agriculture has not
developed any seasonal variation function for surface
moisture.
The correction factor for moisture content overlaps
somewhat with that for precipitation. The distinctions made
between the two are:
0 Only average soil moisture for the entire
year is considered in the surface mois-
ture function, and this value is used to
adjust the locally applicable emission
factor to reflect emission rates with
normal retained moisture.
0 The correction factor for precipitation
only applies to days with rainfall or
snow cover and accounts for the total
absence of dust emissions from very wet
soil surfaces.
While different soil types stop dusting with greatly
varying amounts of surface moisture (see data above), there
is no information to indicate that surface moisture affects
the particle size distribution of fugitive dust emitted from
a particular source. In other words, surface moisture does
-171-
-------
not sort an erodible soil by differentially retaining
certain fractions.
PRECIPITATION
As long as a soil surface is noticeably moist, it will
not dust, even with traffic or surface activity. '
EPA's recommended emission factors for emissions from
4
unpaved roads assume that no emissions occur on days when
there is measurable precipitation or snow cover. Analyses
O O T A T IX
performed by PEDCo ' ' and others on particulate concen-
trations near fugitive dust sources on days with rain and no
rain showed that hi-vol readings on rainy days are generally
about half of those on non-rain days. Since these hi-vol
measurements include some contribution from point sources,
conventional area sources, and influx background in addition
to fugitive dust, it appears that the combination of large
reductions in fugitive dust emissions and rainout (during
actual periods of rain) account for this 50 percent reduc-
tion in ambient concentrations. It can be shown that in
some cases fugitive dust still occurs on days with rain, and
may be even higher than normal due to strong winds that
sometime precede thunderstorms. However, on an annual or
seasonal basis, the current assumption embodied in EPA
emission factors of no emissions on days with rain seems
reasonable.
In most soils, crust formation due to rainfall has no
permanent effect in reducing fugitive dust emission rates.
In fact, small showers tend to smooth the soil surface,
loosen some surface particles, and accelerate rather than
alleviate soil movement by wind. Rains may also bring
additional fines to the surface in areas that have been
cleared or disturbed and cause a temporary increase in
dusting after the surface dries. A related natural phenome-
non, the seasonal deposition onto a lake bottom of fine
-172-
-------
particles from water erosion and subsequent wind erosion of
the dried lake bed, provides another example of the negative
secondary effects of rainfall on fugitive dust emission
rates.
Several attempts have been made to demonstrate that the
dust emission rate in an area increases with the number of
days since the last rain, as the surface becomes progres-
sively drier (see SURFACE MOISTURE). This relationship has
never been shown to be significant, possibly due to actions
such as those described above. Therefore, the assumption is
made that emission rate does not increase with time since
rain.
The proposed method for incorporating the correction
for precipitation is to count the number of days during the
period of interest with rain and snow cover (using a National
Weather Service climatological summary for a local station),
convert this count to a percentage of the total days, and
reduce the emission factor by this percentage. This method
is directly adaptable to estimating seasonal variations.
VEGETATIVE COVER
Ground cover primarily affects fugitive dust generated
by wind erosion. Vegetation reduces wind erosion losses in
two different ways. First, it absorbs some of the drag at
the surface and decreases the wind velocity locally.
Secondly, the roots of the plants act as a soil binder. The
retained moisture necessary to support shallow-rooted vege-
tation also reduces the rate of soil loss.
The initial ground cover on a field usually provides
more protection than additional increments of vegetation.
The relationship between the amount of vegetation and wind
erosion rate is best quantified in the USDA wind erosion
equation, which has been developed with over 30 years of
experimental data.
-173-
-------
The amount of vegetation is expressed in Ib/acre of
air-dried residue in the wind erosion equation. For ground
cover more than about 250 Ib/acre, additional vegetation
reduces soil loss approximately linearly. Depending upon
the potential for wind erosion in a particular geographic
area, ground cover of 1000 to 2000 Ib/acre are required to
essentially eliminate soil loss.
The amount of vegetation obviously changes seasonally.
Due to the wide range in seasonal changes for different
plants, it is difficult to quantify this factor for use as a
seasonal correction function. Deciduous trees and bushes
are reported to lose 20 to 40 percent of their effective-
ness when they are defoliated in the winter. This range may
provide a usable estimate of average seasonal variations due
to vegetation--a 20 percent increase from annual emission
rates during winter and a 20 percent reduction during summer
for sources with ground cover.
TRAFFIC ACTIVITY
Traffic movement over unpaved surfaces causes dust to
be thrown into the air at contact points between the vehicle
and the surface or pulled into the air by the vehicles'
wakes. Also, traffic may break protective surface crusts
and leave the soil more susceptible to subsequent wind
erosion. Thus, surface activity has a direct, immediate
effect on emission rates and an indirect, longer term
impact.
The fugitive dust emitted by traffic movement is
related to the amount of traffic or activity and its average
speed. For many source categories, such as road shoulders
and construction, it is difficult to find an activity param-
eter for which data are generally available. For sources
with automotive traffic (unpaved roads, unpaved parking
-174-
-------
lots, paved streets, and street sanding for snow control),
the emissions are assumed to be directly proportional to
vehicle miles of travel (VMT) even though there would be a
small amount of emissions from these sources with no traf-
fic. For agricultural tilling, the number of tilling oper-
ations per year on the fields is used as the measure of
activity. For aggregate storage operations, traffic levels
in the storage area are simply classified as active or
inactive.
2 3 17 18
Several studies ' ' ' have examined the effect of
average vehicle speed on emission rates from unpaved roads
and have variously proposed that the emission rate is
linearly dependent on, a function of the square of,17 or
exponentially related to vehicle speed. Since these
analyses were each based on data for only three or four
different speeds, it could be anticipated that many differ-
ent curves would approximately fit the experimental data
points. The important conclusion is that all these studies
show emission rates increase with higher vehicle speeds
throughout the normal operating range of the vehicles.
Within the range of accuracy of the basic emission factor
and the experiment designs it may be assumed that the actual
relationship and the resulting correction function should be
linear. The current EPA-recommended emission factor for
unpaved roads incorporates a linear correction for average
speed.
19
The only study identified which investigated the
effect of speed on emissions from paved streets concluded
that the relationship is also linear. However, the availa-
bility of material on the road surface for resuspension may
be a limiting factor on the emission rate from this source.
More data are needed to establish this correction factor.
A speed correction function for agricultural tilling
has also been reported.2'20 However, most farm implements
-175-
-------
are designed to operate over a narrow speed range and, as a
practical consideration, there are usually no means of
obtaining data on actual operating speeds when performing an
emission inventory. Similar situations exist for vehicular
traffic at construction sites and aggregate storage areas..
Therefore, it is recommended that speed correction functions
not be developed for any source categories other than paved
and unpaved roads.
For sources such as roads and construction sites that
have continual traffic, it has been determined that most of
the fugitive dust results directly from traffic movement.
For plowed fields and cleared areas, the total emissions are
due to a combination of surface traffic and wind erosion.
Total agricultural emissions may be estimated by calculating
the tilling and wind erosion components separately. A pro-
posed emission estimation procedure for cleared areas
utilizes the wind erosion equation, but it may be appro-
priate to add a specified percentage to the estimated emis-
sions if it is determined that there is surface traffic over
the cleared area. Correction values of five percent for
occasional traffic (1 veh/day) and 20 percent for regular
traffic (more than 5 veh/day) were calculated from the
previous estimates of the relative impacts of traffic and
wind erosion on unpaved roads.
Traffic activity over native surfaces tends to destroy
the surface crust or layer of pebble-sized particles ("desert
pavement") that normally shield the surface from further
wind erosion. The increase in emissions due to this traffic
is a function of the surface area with tire tracks and the
extent of damage to the natural protection. As mentioned
previously in the SOIL TYPE section, an uncrusted surface
loses soil at a rate about six times as great as a com-
pletely crusted surface. Therefore, a correction function
-176-
-------
for areas with off-road traffic could be calculated as
follows:
correction function =1+5 (area with daM*ged surface)
total area
^fraction of crusting lost.
in disturbed areas
Seasonal variations in a traffic correction function
are dependent primarily on seasonal variations in activity
(VMT, plowing operations, etc). The decision of whether to
apply seasonal corrections should be based on the availabil=
ity and quality of such data.
-177-
-------
REFERENCES
1. Chepil, W. S. Soil Conditions that Influence Wind
Erosion. U.S. Department of Agriculture, Washington,
D.C. Technical Bulletin Number 1185. 1958.
2. Development of Emission Factors for Fugitive Dust
Sources. U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. Publication Number EPA-
450/3-74-037. June 1974.
3. Investigation of Fugitive Dust—Sources, Emissions and
Control. U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. Publication Number EPA-
450/3-74-036. May 1973.
4. Compilation of Air Pollutant Emission Factors, Supple-
ment 5. U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. Publication Number AP-
42. April 1975.
5. Sehmel, G. A. Influence of Soil Erosion on the Airborne
Particle Size Distribution Function. (Presented at Air
Pollution Control Association annual meeting. Chicago,
Illinois. Paper Number 73-162. June 1973.)
6. Skidmore, E. L. and N. P. Woodruff. Wind Erosion
Forces in the United States and their Use in Predicting
Soil Loss. U.S. Department of Agriculture, Washington,
D.C. Agricultural Handbook Number 346. 1968.
7. Chepil, W. S. Dynamics of Wind Erosion: II. Initiation
of Soil Movement. Soil Sci. 6_0 (5) : 397-411, 1945.
8. Malina, F. J. Recent Developments in the Dynamics of
Wind Erosion. Amer. Geophys. Union Trans. pp. 262-
284, 1941.
9. Chepil, W. S. Dynamics of Wind Erosion: III. The
Transport Capacity of the Wind. Soil Sci. 60(6):475-
480, 1945.
10. Zingg, A. W. and W. S. Chepil. Aerodynamics of Wind
Erosion. Agr. Engr. 3_1 (6) : 279-284 , 1950.
11. Woodruff, N. P. and F. H. Siddoway. A Wind Erosion
Equation. Soil Sci. Soc. Amer. Proc. 29 (5);602-608,
1965.
-178-
-------
12. Chepil, W. S. Influence of Moisture on Erodibility of
Soil by Wind. Soil Sci. Soc. Amer. Proc. 20 (2)-288-
292, 1956. — ' '
13. Woodruff, N. P. and D. V. Armbrust. A Monthly Climatic
Factor for the Wind Erosion Equation. J. Soil Water
Conserv. pp. 103-104, May-June 1968.
14. Analysis of Probable Particulate Non-attainment in the
Kansas City AQCR. PEDCo-Environmental Specialists,
Inc., Cincinnati, Ohio. Prepared for U.S. Environmen-
tal Protection Agency, Kansas City, Missouri. February
1976.
15. Kosky, K. F. and M. P. Wanielista. Fugitive Particulate
from Highway Construction. (Presented at Air Pollution
Control Association annual meeting. Boston, Massachu-
setts. Paper Number 75-36.3. June 1975.)
16. Craig, D. G. and J. W. Turelle. Guide for Wind Erosion
Control on Cropland in the Great Plains States. U.S.
Department of Agriculture, Washington, D.C. 1964.
17. Roberts, J. W., et al. The Measurement, Cost, and
Control of Traffic Dust in Seattle's Duwamish Valley.
(Presented at the Air Pollution Control Association
Pacific Northwest Section annual meeting. Eugene,
Oregon. Paper Number AP-72-5. November 1972.)
18. Heinsohn, R. J. , C. Birnie, and T. A. Cuscino- Fugitive
Dust from Vehicles Using Unpaved Roads. (Presented at
the Third National Conference on Energy and the Environ-
ment. Oxford, Ohio. September 1975.)
19. Sehmel, G. A. Particle Resuspension from an Asphalt
Road Caused by Car and Truck Traffic. Atmospheric
Environment. 7:291-309, 1973.
20. Bocharov, A. P. and E. Yu. Terpilovsk. Study of the
Action of Machine-Tractor Units on the Upper Soil
Layer. (Translation by National Tillage Machinery
Laboratory, Auburn, Alabama.) Electrification and
Mechanization of Soviet Socialist Agriculture. 8:11-
14, 1970.
-179-
-------
Session IV:
CONTROL TECHNOLOGY
Norman Plaks
Session Chairman
-181-
-------
STUDY OF THE EFFECT OF ASBESTOS WASTE PILES ON AMBIENT AIR
Colin F. Harwood*
Mary Stinson**
Paul Ase*
ABSTRACT
The fabrication of asbestos products is a major industry
involving about one million tons per annum of asbestos.
Asbestos in the atmosphere is associated with increasing
incidences of cancer in the general populace. One source of
the atmospheric asbestos is fugitive emissions from the waste
piles located throughout the United States. Individual
sources of emissions from the steps involved in the formation
of waste piles are considered. Methods available to control
these fugitive emissions are discussed and estimates presented
on the relative costs of mitigation.
* IIT Research Institute, 10 West 35th Street, Chicago,
Illinois 60616
** Environmental Protection Agency, IERL, Edison, New Jersey
-183-
-------
STUDY OF THE EFFECT OF ASBESTOS WASTE PILES ON AMBIENT AIR
Introduction
Waste dumps from asbestos product manufacturing operations
are frequently located in high density population areas. Fugi-
tive asbestos emissions are created as the material is trans-
ferred to the dump and also as the surface of the pile is
eroded by weather action. At the present time, fugitive dust
control techniques are seldom used to mitigate the asbestos
emissions.
The lack of adequate emission control at asbestos waste
dumps is regrettable since there is growing evidence that low
levels of asbestos exposure may be harmful to health. Evidence
from the literature is presented which suggests that industrial
activity involving asbestos is leading increasing numbers of
mesothelioma cases. Mesothelioma, or cancer of pleura, is a
rare disease known to be caused by asbestos.
This paper discusses methods by means of which the fugi-
tive asbestos emissions may be mitigated. The technology
which has been developed and applied to other industries is
considered for application to the case of asbestos emissions.
Asbestos Waste Dumps
The asbestos industry is not a small industry; world
consumption approaches five million tons per annum, while in
the United States the amount is nearly one million tons per
annum (see Table 1 for 1973 figures). Chrysotile asbestos
accounts for approximately 96% of all the asbestos used in
the United States.
Asbestos cement products account for 70% of the total
United States usage. Products include asbestos cement pipe,
asbestos cement siding, asbestos cement shingles, asbestos
cement wallboard, and insulation products. It is estimated
that there is between 5-10% of the product material dumped
-184-
-------
Table 1
ASBESTOS PRODUCTION, 1973
(Minerals Yearbook)
Short Tons
World Production 4,598,000
U.S. Production 150,000
U.S. Consumption 876,000
Chrysotile 839,200
Crocidolite 18,000
Amosite 4,300
Anthophyllite 1,200
-185-
-------
as scrap every year. The scrap material is about 10% of
fine material collected from baghouses and 90% of coarse
scrap from cuttings and drillings and from products which
have failed quality assurance tests (Table 2). Thus, the
total annual scrap from asbestos cement operations alone
may be calculated. If 1,000,000 tons of asbestos are used
per annum, and of this 70% is for asbestos products, and
there is a 7.5% scrap rate, then the total waste asbestos
dumped is:
1,000,000 x — x — = 52,500 tons
100 100
Assuming that the average asbestos cement product contains
25% asbestos, then the total asbestos cement waste is:
52,500 x — = 210,000 tons
25
Thus, the nation faces an annual problem of disposing of
210,000 tons of waste material containing hazardous asbestos.
The Hazardous Nature of Asbestos
Recognition of the hazardous nature of asbestos is
relatively new. Although asbestos is associated with various
cancers of the stomach, colon, etc., there is no doubt of its
effect in producing two illnesses, asbestosis and mesothelioma.
Asbestosis is a non-malignant fibrosis of the lung and is
only found among asbestos workers with a relatively heavy
exposure level. Asbestos was first observed by Murray (1) in
1907. In 1930, Mereweather (2) gave a detailed description
of the disease, and this lead to the 1931 United Kingdom
regulations on asbestos usage. Similar regulations were
issued in the United States in 1938 following the studies by
Dressen, et al. (3).
Mesothelioma, or cancer of the lung, was not fully recog-
nized as a disease until 1960 when Wagner, et al. (4)
published their work on the South African mine areas. They
-186-
-------
Table 2
ASBESTOS CEMENT WASTE
Dust from baghouse collectors and sweepers 10%
Aggregates from breakage, cuttings, and
drillings 90%
Waste pH ^12.5
Waste composition:
Cured Portland cement 40 to 55%
Quartz silica sand 24 to 33%
Asbestos 15 to 35%
-187-
-------
cited 33 cases of mesothelioma and, of these, 17 were not
occupationally exposed. After 1960, the standard medical
textbooks, e.g., Willis (5), were changed to include meso-
thelioma; until this time mesothelioma had been described
as a metastasis from a primary site.
A number of studies have shown that asbestos is
ubiquitous to the ambient air [Selikoff and Nicholson,
1970 (6); Holt and Young, 1973 (7)]. Other researchers
have shown asbestos is commonly found in the lungs of urban
dwellers [Urn, 1971 (8)].
Case history studies in various countries have given
increasing evidence of mesothelioma occurring in people
non-occupationally exposed to asbestos. Newhouse and
Thompson (9) in 1965 reported in England 76 cases of meso-
thelioma; only 40% had occupational exposure. Lieben and
Pistawka (10) in 1967 working from Pennsylvania hospitals
V
discovered 42 mesothelioma cases; only 24% of these were
occupationally exposed. A further 24% could not be estab-
lished, while 52% were not occupationally exposed.
Borrow (11) in 1967 studied mesothelioma cases in the town
of Manville, New Jersey (population 15,000), the site of a
large asbestos products plant. Over a three-year period,
17 mesothelioma cases were found with two being non-
occupationally exposed. Bohlig (12) in 1969 reported his
study of the population surrounding a German asbestos plant.
He found 319 mesothelioma cases between 1958 and 1968. He
was able to follow the case histories of 119 and found only
46% of these were occupationally exposed to asbestos.
Much further research work is needed to answer the many
questions raised concerning asbestos exposure and health.
However, enough evidence is available to make it mandatory
that modern technology by applied to limit the emission of
asbestos to ambient air. Such techniques will be discussed
in the following pages.
-188-
-------
Utilization and Disposal Options
There are a number of options available for the utili-
zation of asbestos wastes, Table 3. The first option to be
considered is the re-use of the material. The industry
makes every attempt to re-use asbestos cement waste by
recycling it into the production line. However, there is
a limit on the amount that can be recycled set by the exacting
standards on the quality of the product. This is particularly
true for asbestos cement pipe where every section of pipe is
tested for its ability to withstand pressure and also its
ability to withstand flexing.
Attempts have been made to use the waste material in
building blocks. It has been shown that asbestos cement pro-
duct waste can be fabricated into bricks which meet the
standards of the ASTM C-90 and the ASTM C73-67 specifications.
To date, asbestos waste has not been used commercially to
produce these bricks because of the fear that a health hazard
might be associated with such a product.
Another suggestion for a use of the waste material is to
mix it with acid waste from other mining operations; the hope
being that the neutralized waste would be more suitable for
reclamation and plant growth.
Waste alteration by thermal or chemical decomposition
has also been suggested. Chrysotile asbestos when heated
above 600°F decomposes due to loss of water of crystallization
to give forsterite and talc, both of which are thought to be
harmless amorphous particles. Chemical leaching has been
found to remove the magnisia from the crystal structure of
chrysotile asbestos leaving behind the silica skeleton.
While both thermal and chemical decomposition' is possible,
it is economically impractical.
The only viable alternate to waste re-use or alteration
is dumping. The methods which may be applied to mitigate
the emissions from dumping operations'will now be considered.
-189-
-------
Table 3
UTILIZATION AND DISPOSAL OPTIONS
Waste Re-use -- recycle, cement blocks, steam
cured bricks, waste neutraliza-
tion with acid wastes
Waste Alteration -- thermal decomposition, chemical
leaching
Waste Dumping -- waste aggregates, waste fines
-190-
-------
The Dumping of Asbestos Wastes
The dumping of asbestos waste creates fugitive emis-
sions at a number of points. The flow sheet for the dumping
of waste materials is shown in Figure 1. The sources of
asbestos emissions are given in Table 4 along with an esti-
mate of the relative significance of each emission source.
It should be noted that in these figures are not emissions
rates but rather the relative magnitude of the total emis-
sions set out such that the importance of each source may be
assessed. Thus, even though the crushing and leveling is
observed to have a high emission rate, the operation is of
such a short time that its total contribution to the fugi-
tive emissions is relatively minor. Obviously then, the
major effort in reducing emissions should concentrate on
the dump rather than the dumping, crushing, and leveling
operations.
Emission Control Options
In Plant Options
The best method of controlling the dust is at the source,
before they ever become fugitive emissions. Fine waste can
be slurried, agglomerated, bagged, or at least wetted down.
Larger waste material or aggregates may be crushed, washed
free of fines, and kept wet. The use of various water-
additive agents will extend the effectiveness of wetting
techniques.
Dump Emission Control
Control of the emissions from waste dumps includes
physical, chemical, and vegetative.
Physical include the use of physical barriers including
straw, bark, or gravel applied directly to the surface of
the dump. Other physical methods involve the judicious
placement of wind breaks or banks of trees, placed in the
direction of the prevailing winds, which protect the dump
from wind erosion.
-191-
-------
Wrisle AGGREGATES
Segregate
Small
Aggregates
Waste FINES fro-,i
Collectors i
Large
Aggregates
^
\
>..!..
f \
/
Dump, Active Pile
Crush and Level
V S
Figure 1
ASBESTOS CEMENT WASTE DISPOSAL
-192-
-------
Table 4
WASTE DUMPING EMISSION SOURCES
Transfer Emissions -- dumping, crushing, and
leveling
Dump Emissions -- active pile weathering,
inactive pile weathering
Emission rate estimates:
Fines dumping 22%
Aggregate crushing and leveling 6%
Active pile weathering 60%
Inactive pile weathering 12%
-193-
-------
Chemical binders are finding increasing utility in
reducing erosion from soil banks. The selection of the
binder is important and the ideal binder would have the
following properties:
• Low application cost
• Water soluble
• High bondability to the particles under consideration
• Long life and stability
• Resistant to heat and cold
• Non-toxic
• Biodegradable
• No water pollution problems from drainage water
• Easy to clean from application devices
• Effective in low dilutions
Obviously no one binder material is able to score highest
in each of the above categories. For the purposes of the
present study, about 30 binders were considered (some of these
are listed in Table 5). Of these, five were considered in
greater detail. Field experiments were conducted using Coherex
which, although it was not the most efficient binder in terms
of stabilizing soil, it did have the best overall properties
including water solubility, biodegradability, and resistance
to leaching once applied.
Vegetative covers, if properly applied and maintained,
offer the surest, most permanent method of eliminating emis-
sions caused by surface erosion. A cover of 6 to 12 inches
of soil vegetated with fertilizer, seed, and mulch, and
regularly watered is generally found to be effective. It is
important to overcome the proplerms inherent in the waste
material such as the lack of plant nutrients and microbial
populations, and also the pH of the waste. For this reason,
plants cannot generally be grown directly on the waste, and
-194-
-------
Table 5
MATERIALS
Product
Amine D Acetate, SOS
Polyrad 1110A
Vinsol Emulsion
Defloc 50
Abitol
Paracol emulsion
Piccolyte Dipentene
Reten 421
Neuphor 100
Kymene 557
Landlock XA2440
Latex M145 or M166
Elvanol
Vinylac
ARQUAD 2HT
Chemical Identification
Stabilized abietyl
amine
High molecular-weight
amine ethylene oxide
adducts
Water emulsion of
aliphatic resin
Cationic polymer
Hydroabietyl alcohol
Wax-rosin emulsion
Polyterpene adhesive
resin
Anionic acrylic polymer
Anionic emulsion
Cationic polyamide-
epichlor-hydrin resin
Adhesive binder
Latex binder
Polyvinyl alcohol
Polyvinyl acetate,
tackified dispersion
Quaternary ammonium
compound
Supplier
Hercules, Inc.
Hattiesburg, Miss.
Hercules, Inc.
Hattiesburg, Miss.
Hercules, Inc.,
Kalamazoo, Mich.
Hercules, Inc.,
Milwaukee, Wise.
Hercules, Inc.,
Burlington, N.J.
Hercules, Inc.
Hercules, Inc.
Hercules, Inc.,
Hopewell, Va.
Hercules, Inc.,
Milwaukee, Wise.
Hercules, Inc.
3 M Co., St. Paul, Minn.
Dowell Div., Dow Chem-
ical Co., Tulsa, Okla.
E.I. Du Pont de
Nemours & Co.
Borden Chemical Co.
Armak Chemicals Co.
Cost,
Dollars/lb
0.625
0.780
0.126
0.170
0.8375
1.40
0.2335
2.00/gal
0.51
-195-
-------
Table 5 (continued)
Product
ARQUAD 2S
Ethomeen T/12
Curde Amine
Krilium CRD-186
Sodium alginate
Polyacrylic acid
Superfloc 16
Coherex
Chemical Identification
Quaternary ammonium
compound
Tertiary aliphatic
amine
Amine compound,
unpurified
Vinyl acetate/maleic
acid
Sodium alginate
Polyacrylic acid amine
Flocculant
Resinous binder
Supplier
Armak Chemicals Co.
Armak Chemicals Co.
Armak Chemicals Co.
Monsanto Chemical Co.,
St. Louis, Mo.
Rolakem Co.,
Teaneck, N.J.
Rohm & Haas Co.,
Philadelphia, Pa.
American Cyanamid Co.,
St. Louis, Mo.
Witco Chemical Co.,
Cost,
Dollars/lb
1.09
0.725
0.340
2.00
Rezosol 5411B
Dextran
Hammond , Inc .
Cationic resin emulsion E. F. Houghton Co.,
Philadelphia, Pa.
Dextran Howard Hall Co.,
Cos Cob, Conn.
0.36/gal
0.245/gal
4.00
-196-
-------
a soil layer must be applied. It is generally found to be
advisable to use soil types which fit in with the local area
and also to use plant seeds of indigenous plants.
Costs of Various Control Options
To estimate the cost of controlling emissions from
transfer operations at the waste and from the waste dump
itself, a hypothetical plant was considered. The plant dis-
posed of 13.2 metric tons of reject product and other waste
aggregates per day; in addition, its plant disposed of
0.9 metric tons of baghouse waste fines per day.
Fugitive emissions were assumed to result from four
basic operations:
• Daily dumping of fines onto "active" pile
• Crushing of reject pipe by bulldozer, once a month,
on active pile
• "Weathering" of active piles
• "Weathering" of inactive piles
The emission rates associated with each source were
based on reported fugitive dust emission rates and from
IITRI field experiences.
Control techniques for reducing emissions from the
waste disposal activities vary considerably in the annual
costs and the emission reduction achieved. Thirteen basic
control options were evaluated. They were considered
separately and in combinations to ascertain the lowest cost
methods for achieving emission reductions. Table 6 summarizes
the results of this cost analysis.
The cost of controlling emissions from fines dumping was
estimated for a water spray, water plus surfactant, pelleti-
aing of fines, water slurry, water slurry with chemical
binder, and bagging of fines. The bagging of the fines was
the most efficient control scheme for the fines dumping oper-
ation (100% efficient) found in this analysis while a water
-197-
-------
Table 6
SUMMARY OF CONTROL OPTIONS
Percent Reduction in Emissions From:
Total Annual Fines Aggregate Active Inactive Total
Control Method Cost, $ Dumpings Crushing Pile Pile Emissions
1. Water Spray at
Fines Dumping 2,800 10 — — — 2
2. Water and Surfactant
at Fines Dumping 3,400 20 — — — 4
3. Agglomeration of
Fines with Water 10,000 90 — 5 — 23
4. Agglomeration of
Fines with Binder 13,000 90 — 25 — 35
5. Water Slurrying of
Fines 4,100 85 — — — 19
6a.Chemical Binder
with Water + 0.25%
Binder 5,800 85 — 45 — 46
6b.Chemical Binder
with Water + 0.20%
Binder 5,400 85 — 27 — 35
6c.Chemical Binder
with Water + 0.10%
Binder 5,000 85 — 14 — 27
7. Bagging of Baghouse
Fines 10,500 100 — 45 22 52
8. Chemical-Vegetative
Control of Inactive
Pile 3,380 — — — 90 11
9. Water Spray on
Active Pile 3,570 . — — 50 — 30
lOa.Chemical Stabilize
Active Pile Once/
Week 8,970 -- — 90 -- 54
-198-
-------
Table 6 (continued)
Percent Reduction in Emissions From:
Total Annual Fines Aggregate Active Inactive Total
Control Method Cost, $ Dumpings Crushing Pile Pile Emissions
10b.Chemical Stabilize
Active Pile Once/
Month 3,970 - - 80 - 48
11. Landfilling Active
Pile Once/Month 8,700 -- - 73 20 46
-199-
-------
spray at the dump site was the least efficient (10% efficient).
The control costs for the fines dumping operation varied from
$2,800 per year (water spray) to $13,000 per year (pelletizing).
The fugitive emissions from an active pile could be re-
duced with methods similar to those used to obtain fines con-
trol. Water spray, chemical stabilization, foaming agent, and
landfill techniques were all considered as possible mitigation
methods. The cost of these control methods varied from $3,750
to $8,970 per year. The emission reduction from the active
pile ranged from 50% (water spray) to 90% (chemical stabili-
zation of pile once per week).
A permanent cover on the inactive pile was required to
prevent the source emissions from increasing each year as the
size of the pile increased. A combination chemical stabilization-
vegetive cover was assumed to be the method most likely to
yield a permanent cover. The cost for developing the permanent
cover on the inactive asbestos pile was calculated to be
$3,400 per year.
Acknowledgements
The work described in this paper represents a portion of
a study supported by the Environmental Protection Agency under
Contract Number 68-02-1872. Statements made are those of the
authors and may not represent the views of the EPA.
The guidance and interest of Mr. David Oestreich of the
EPA was very much appreciated. Kurt Gutfreund of IITRI was
helpful in guiding its selection of suitable polymers and
Dr. William Berg of Colorado State University provided
invaluable consultancy for revegetating waste dumps.
It should be noted that the views and comments expressed
in this paper are those of the authors and does not necessarily
reflect those of the EPA.
-200-
-------
REFERENCES
1. Murray, H. M. (1907) Statement before the committee in
the minutes of evidence, pp. 127-128. In Report of the
Departmental Committee on Compensation of Industrial
Disease. London:H. M. Stationery Office.
2. Merewether, E. R. A. (1930) The occurrence of pulmonary
fibrosis and other pulmonary affections in asbestos
workers. J. Ind. Hyg. . 12_, 198-22, 239-257.
3. Dreessen, W. C. , Dallavalle, J. M. , Edward, T. I.,
Miller, J. W. , and Sayers, R. R. (1938) A study of asbestos
in the asbestos textile industry. Public Health Bulletin 241,
Washington, D.C., U.S. Government Printing Office, 126 pp.
4. Wagner, J. C. , Sleggs, C. A., and Mar chard, P. (1960)
Diffuse pleural mesothelioma and asbestos exposure in the
north western Cape Province. Brit. J. Ind. Med. , 17, 260-271.
5. Willis, R. A. The pathology of tumors. 3rd Edition,
Washington, D.C., Pub.Butterworth.
6. Selikoff, I. J. , Nicolson, W. J. , and Langer, D. M.
(Oct. 5, 1970) Asbestos air pollution in urban areas.
Presented at the American Medical Association Air Pollution
Medical Research Conference. New Orleans.
7. Holt, P. F. , Yound, D. K. (1973) Asbestos fibers in the
air of towns. Atmos. Envir. , 7_, 481-483.
8. Urn, C. H. (1971) Study of the secular trend in asbestos
bodies in lungs in London 1936-1966. Brit. Med. J.,
2, 248-251.
9. Newhouse, M. L. and Thompson, H. (1965) Mesothelioma and
peritoneum following exposure to asbestos in the London
area. Brit. J. Ind. Med. , 22_, 261-269.
10. Lieben, J., Pistawka, H. (1967) Mesothelioma and asbestos
exposure. Arch. Environ. Health, 14, 559-563.
11. Borow, M. , Conston, A., Livornese, L. L. , and Schalet, N.
(1967) Mesothelioma and its association with asbestosis.
J.A.M.A., 201, 587-591.
12. Bohlig, H. , et al. (1970) Epidemiology of Malignant
Mesothelioma in Hamburg, Environ. Res. , 3_, 365-372.
-201-
-------
AN ASSESSMENT OF FUGITIVE EMISSIONS IN
THE PRIMARY^ALUMINUM INDUSTRY
BY
.William D. Balgord, Ph.D.
The Aluminum Association
PRIMARY ALUMINUM INDUSTRY
The primary industry supplies all new, or virgin, aluminum,
since aluminum is not produced as a by-product of other industrial
activity.
It extracts chemically bound aluminum from its commercial
*
ore, bauxite, and converts it to metal in two stages.
»
In the first stage, bauxite .is treated by the Bayer Process
to obtain aluminum oxide in a high"state of purity.
In the second step, metallurgical grade alumina as converted
electrolytically to the metal.
The aluminum industry, in common with other major materials
producing industries, must process' necessarily large amounts of
raw materials by economic methods. In 1974 approximately
5 million tons of aluminum was produced in the United States.
Since one pound of aluminum requires an input of 5.6 pounds of
raw materials, sta'rting with the refining of bauxite, the overall
magnitude of the materials handling is obvious.
At each successive stage in the production of aluminum, the
value of the product increases substantially. And with the
increase in value, the incentive to conserve materials. But the
incremental costs to control further losses beyond a certain level
of control tend to rise rapidly and finally reach a point of
extreme cost ineffectiveness.
-203-
-------
FLOW OF RAW MATERIALS AND PROCESSES
Domestic production of bauxite supplies less than 8% of
current U.S. demand for metallurgical grade alumina.
*
At the present time, domestic bauxite is actively mined
only in Arkansas. Arkansas bauxite is produced mostly from
•
underground mines. It is a fairly hard material and requires
special processing because of higher than normal amounts of
alumino-silicate materials.
Most other bauxite is a soft, clay-like deposit—not
readily distinguished from the other soil i*n areas where it is
found. It may be hauled by truck or rail from the mine site
to a nearby alumina plant or to shiploading site for export.
Foreign sources of bauxite now account for somewhat more
than 92% of U.S. alumina consumption—imported partly as bauxite,
partly as refined alumina.
At the Bayer plant bauxite is unloaded from ships by
bucket or clam-type crane, or in certain instances, by conveyor
mechanisms, and transferred to storage buildings.
As needed, bauxite is transferred from storage to a crush-
ing and milling operation. This step may be accompanied by
blending of two or more bauxites from different sources to
achieve desired properties.
The foregoing operations may be accompanied by the generation
of dust, depending on the moisture content of the bauxite and
weather conditions and on provisions for capture and control
of dusts that may be operated at a particular plant. It should
-204-
-------
be borne in mind in this regard that, with the one exception of
an alumina-plant operated in the Virgin Islands, the construction
of all domestic alumina plants were begun before the late 1950's.
The remaining steps in the Bayer Process, until final
calcining, are totally enclosed.
*
Bauxite is mixed with hot caustic and heated under pressure
to dissolve hydrated aluminum oxide to produce a solution of
sodium aluminate. This step is known as "digestion." Impurities
—mostly iron oxide and free silica or quartz—remain essentially
undissolved.
Insoluble impurities are removed by filtration.
The clear sodium aluminate solution is cooled, accompanied
by precipitation of alumina-trihydrate.
The trihydrate filter-cake is calcined at 1200°C removing
practically all chemically bound water and producing a metallurgical
grade alumina of about 99.4% purity.
Other raw materials used in the Bayer Process are lime and
caustic. The caustic is recycled to the process and is handled
as a liquid throughout.
Today the production of primary aluminum relies exclusively
on the Hall Process.
In the Hall Process, aluminum is produced by the passage of
direct current through a cell containing alumina dissolved in
molten electrolyte. The electrolyte is primarily cryolite.
The molten cryolite is contained in a carbon-lined refractory
box, called a pot. As electrolysis proceeds, molten aluminum
-205-
-------
collects in the bottom of the cell. Electrical current passes
through the electrolyte between a large block of carbon, suspended
over and partially immersed in the electrolyte—this is the anode—
and the cathode which is the pool of molten aluminum resting
on the carbon lining. During the course of electrolysis, the
«
anode carbon is gradually consumed by reaction with oxygen
generated by the cell.
Thus, there are two major materials input to the manufacture
of primary aluminum: alumina and anode carbon, prepared from
petroleum coke.
Alumina and petroleum coke typically arrive at most aluminum
smelters by rail, barge or ship.
Alumina may be unloaded from hopper cars pneumatically and
transferred to storage bins referred to as "ore" silos.
i
The petroleum coke is moved to the carbon plant where it is
stored until needed to manufacture pre-cast anodes or anode paste
for Soderberg-type plants.
Distribution of alumina to the individual cells requires
handling that tends to create some dust. Virtually all of this
dust is settleable.
Likewise, also the preparation of anode carbon from coke
involves operations which tend to create dust.
Cellroom and anode plant emissions are covered by current
or proposed EPA point source and fugitive emission regulations
and will not be further discussed.
-206-
-------
NATURE AND SIGNIFICANCE OF
At the source of bauxite
Operations:
Type of Materials:
Significance:
At the Bayer Plant
Operations:
Materials:
Significance:
Earth Moving
Loading
Transportation
The bauxite contains principally hydrated
oxides of aluminum with lesser, variable
amounts of iron oxides, free silica, hydrated
alumino-silicates, titanium dioxide, and
other minor impurities.
Generation of dust depends very much on
moisture content of ore—this, of course,
depends on weather conditions. Generally
speaking, bauxite dusts contain particles
larger than 20 microns in diameter and are
considered settleable—other than a poten-
tial nuisance in the immediate vicinity of
bauxite mining operations, these dusts
are considered to have a minor to insigni-
ficant impact beyond the locus of operations
—the dusts are virtually indistinguishable
in composition from endemic soils.
Unloading Bauxite
Transfer and storage of bauxite
Crushing and milling
Loading of calcined alumina
The composition of bauxite was just
described. A bauxite may be very similar
chemically and mineralogically, but
usually not identical to native soils at
the Bayer Plant. They may contrast
somewhat in color.
They will support vegetation.
Metallurgical alumina is high purity
aluminum oxide, snow-white in color.
Very little aerborne material actually
leaves the plant property; bauxite
itself is a soil which will support
vegetation. Bauxite can be thought of
as the last stage of evolution in the
-207-
-------
formation of laterite wherein most of the
former constituents have been leached
away over geologic time, leaving aluminum
oxide.
Alumina is exceptionally inert, chemically
and biologically, and is considered to have
very little if any adverse effect on the
environment.
3. At the Primary Smelter
Operations: Unloading and transfer of alumina
Distribution of alumina to the cells
Unloading and handling of petroleum coke
Materials: High purity alumina, petroleum coke is a
pure form of commercial carbon; mechanical
handling of P. coke produces coarse dust.
Significance: From the calcination step in the production
of alumina through the smelting of primary
aluminum, about 99% of the alumina is
converted to metal. This is a remarkable
efficiency considering the many opportunities
for losses: transportation and transfer,
housekeeping, and oxidation losses due to
formation of aluminum oxide skim and dross.
SUMMARY
Although the scope of presentation bars any detailed assess-
ment of fugitive emissions at particular plants, it is possible
to examine several factors that have a bearing on the matter:
- Siting In most instances, alumina and primary
aluminum facilities are situated well away
from population centers
Prior Action The aluminum industry has consistently
instituted measures to reduce fugitive
- - losses.
- Character Fugitive dusts, represented by large
particle size, tend not to leave the plant
site.
- Chemistry Calcined alumina is inert; the chemical
composition and mineralogy of bauxite so
strongly resembles that of native soils
at many plants that it is very difficult
to distinguish from soil derived dusts
from other sources such as local agriculture
-208-
-------
The aluminum industry has spent considerable money and
effort to contain fluoride emissions—these are the subject
of EPA New Source Performance Standards, published recently
EPA Guidelines to the States for Existing Aluminum Plants
are pending.
Finally, because of potential litigation on the new source
standards for aluminum, I have avoided reference to cell-
room emissions or of emissions from any other source that
might contain fluoride.
-209-
-------
MEASUREMENT OF IRON FOUNDRY FUGITIVE EMISSIONS
by
William D. Scott
Charles E-. Bates, Ph.D.
Southern Research Institute
Presented at the
SYMPOSIUM ON FUGITIVE EMISSIONS:
Measurement and Control
May 18, 1976 at Hartford, Conn.
-211-
-------
MEASUREMENT OF IRON FOUNDRY FUGITIVE EMISSIONS
INTRODUCTION
An iron foundry has traditionally been thought of as an unpleasant
place to work; foundry workers may be exposed to dust, free silica,
heat, chemical emissions, and noise during molding and casting.
Nonetheless, the foundry industry is one of the largest and
most basic industries in the United States; it supplies components
used in the manufacture of a great variety of products by other in-
dustries. The industry is the sixth largest of all U. S. manufactur-
ing industries, according to the 1970 U. S. Census. In 1974 the annual
U. S. casting production was about 22,000,000 tons with a direct product
value of approximately $13 billion.
This paper reports some cf the experimental techniques and
results employed to quantify point source emissions in the foundry .
The work included characterization of general particulate, free silica,
and chemical emissions from typical green sand foundry molds.
The detailed results of the chemical emissions portion of the
program have been reported elsewhere.1'2* Various organic materials
are added to the sand mixture so that their thermal decomposition
produces a reducing atmosphere in the mold, which is necessary for
good casting quality. The gases or products of decomposition are
then released into the foundry atmosphere. Adequate ventilation is
necessary to prevent build-up of noxious gases such as carbon monoxide,
hydrogen cyanide, ammonia, methane, and other more complex species.
This ventilation is often in the form of forced air drafts that
are exhausted directly from the plant without treatment. Design and
* Superscript figures refer to items in "References."
-212-
-------
operation of these ventilation systems determine the actual concen-
trations of gases present in the working environment and also those
released outside the plant. A fundamental knowledge of the point
source emissions in the foundry allows estimations of the total
amount of objectionable materials produced.
In addition to gaseous emissions, particulate emissions are
also generated in the green sand foundry process. It is estimated
that between 9 and 15 tons of sand must be processed for every ton
of castings made.
An estimate of the particulate emissions generated during
casting production in a typical foundry using sand molds is pre-
sented in Table I.3 The major plant areas that contribute are
(a) molding, pouring and shakeout, (b) cleaning and finishing, and
(c) sand conditioning areas. ** The total estimate of non-melting
operations indicates that 115 Ib, of emissions is generated for
each ton of metal poured. Normal collection equipment reduces this
to 60 Ib/ton in the plant atmosphere, of which 51 Ib. settles out.
About six pounds of this total is released outside the plant.
In addition to the dust from the sand handling, the emissions
from the shakeout and pouring operation include organic compounds
of potentially harmful character. These most often are expected
to condense on the particulate matter and thus are available for
respiration. The types of compounds that are included in this
portion of the emissions have also been reported elsewhere.5
-213-
-------
Table I
Particulate Emissions Factors
from Non-Melting Operations
(5)
(6)
i
Isi
Department
Scrap Yard
Molding, Pouring
and Sliakeout
Cleaning and
Finishing
Sand Conditioning
Coio Derailment
Pailum Shop
Tolal
Operation
Raw Material Handling
Charge Make-up
Charge Pro-heating
Molding
Magnesium Treatment
Pouring
Cooling
Shakeout
Shot Blast
Grinding
Annealing
Palming
Dry Sand Handling
Prepared Sand Handling
Screening
Mulling
Drying and
Sand Reclamation
Oil Sand Core Baking
Shull. Hot Box and
Cold Set Sand Cores
Wood Pallernmaking
Molal Patternmakino
Emissions
Generated
Lb./Ton Melt
.20
.IS
.20
.50
S.OO
5.10
10.30
32.20
15.50
1.60
.10
.02
10.30
.50
10.00
20.60
1.50
.10
1.02
.01
.0£
114.92
Normal
Collection
Percent
0
0
10%
0
10
10
10
60
99
95
0
95
10
0
20
60
tiO
5
0
60
60
Emissions
to Foundry
Environment
Lb./Ton Melt
.20
.15
.18
.50
4.50
4.59
9.27
12.80
.16
.06
.10
—
9.27
.50
6.00
6.24
.60
.05
1.02
—
~
60.21
Sellllnq
Factor
Percent
60%
80
40
90
75
60
90
90
60
80
30
40
90
90
90
90
80
0
0
50
50
Emission*
Released to
Atmosphere
Lb./Ton Melt
.04
.03
.11
.05
1.12
1.84
.93
1.28
.03
.02
.07
—
.93
.05
.80
£2
.12
.05
1.02
—
•H^^H*
fl.31_
Emissions to
Atmosphere
Nationwide'
Lb./Ton Moll
.04
,03
.01
.04
.11
1.45
.75
1.02
.02
.01
.03
— —
.75
.04
.64
.65
.01
.03
.20
—
J^l»
583
Note: 'incidence (actor hai wtn applied lu these *ml»slont rules to iclltct «clu«l industry application ol tacit operation.
-------
SAMPLING PROCEDURES
The sampling procedure used to determine the point source emis-
sions from typical foundry molds employed an exhaust hood and canopy
that covered the source. This arrangement has been referred to as a
quasi-stack method and it is shown schematically in Figure 1. An
orifice plate and DC motor-driven exhaust fan in the stack were used
to establish a known flow rate through the hood. The flow in the
stack was made turbulent to insure a homogeneous sample.
The draft air volume was controlled with a rheostat on the fan
motor and was set to maintain a constant pressure drop of 1/2 inch of
water, as monitored with a manometer. This pressure drop corresponded
to a volume flow of 35 cfm in the stack, or an air velocity of 50
ft/min over the mold surface. This velocity was made high enough to
prevent significant losses of effluent by diffusion under the hood.
Gaseous emissions were periodically sampled by drawing the ef-
fluent into 250-ml evacuated bulbs. This technique is shown in Figure
2. These grab samples were taken at intervals starting immediately
after the casting had been poured, while the castings cooled in the
sand molds, and as the castings were separated from the sand in a
simulated shakeout. These samples were analyzed by gas chromatography.
This procedure allowed the preparation of a time-concentration pro-
file.
A constant stream of gas from the stack was passed through dim-
pled bubblers containing specific absorbing reagents for the determi-
nation of average concentrations of gases such as ammonia and hydrogen
cyanide that were present in low concentrations. These bubblers are
shown in place on the stack in Figure 3. Figure 4 shows the metal
-215-
-------
TO
VACUUM
PUMP
STACK
VARIABLE SPEED MOTOR
AND EXHAUST FAN
FILTER
GAS SAMPLING
BULB
FLOW METER
ORIFICE PLATE
FLASKLESS SAND MOLD
JACKET
FLOOR
Figure 1. Schematic of portable gas sampling hood.
-216-
-------
Figure 2.
Taking a grab sample for gas analysis
above an open mold.
Figure 3.
Top of sampling hood, showing bub-
blers for trace gas detection.
-217-
-------
Figure 4.
Pouring an open mold.
~
Figure 5.
Placing the collection hood
open mold. Note mold gases
at the parting line.
-218-
over an
burning
-------
being poured into a mold. Figure 5 shows the hood being positioned
over the mold after the pour. Positioning the hood required approxi-
mately 15 seconds from the start of the pour.
The hood was also used to sample the air above the muller and
shakeout bin in a pilot plant. In these tests, the hood was lifted
into place over the operation of interest with a forklift, and then
a polyethylene canopy was suspended to cover the entire operation.
Cascade impactors were positioned in the stack to obtain parti-
cle mass-size distributions over the range of particle diameters of
0.5 to 14 ym. Brink and Andersen designs of cascade impactors were
used, depending on the dust loading of the atmosphere. The Brink
impactor, which has a low flow rate, was used at high loadings. At
I
low loadings, the Andersen impactor, which has a higher flow rate,
was used to keep the sampling time reasonable. In either impactor,
the particulate material entrained in the air is separated according
to particle size by suitable orifice designs in the impactor, and
deposited upon collector plates.
The Brink impactor uses aluminum foil plates that may later be
dissolved in acid as the first step in the free silica determination
by the Talvitie colorimetric method.6 By summing up the weights of
the particles on all of the collector plates, and dividing the re-
sult by the volume of air drawn!through the impactor, an average
dust loading was determined. The Andersen impactor will also allow
a dust loading to be calculated as well as size distribution, but
the particulate matter may not easily be chemically analyzed.
Particles in the 0.3 to 1.0 pm size range were counted with an
optical particle size analyzer. In operation, a sample of effluent
from the hood was continuously removed, diluted with clean air, and
-219-
-------
passed through an orifice and then through a collimated light beam.
The amount of scattering caused by individual particles was measured
with a detector tube. Since small particles scatter more light than
large particles, the signal from the detector could be electronical-
ly analyzed to both size and count the number of particles in the gas
stream.7 This allowed data in the form of concentration-versus-time
to be obtained.
RESULTS
Gaseous Emissions from Green Sand Foundry Molds
Gaseous emissions from green sand molds were determined as de-
scribed above by collecting emissions as a function of time after
pouring and during shakeout. A constant rate of air flow of 35 cfm
through the exhaust hood was maintained for these experiments. This
air flow dilutes the effluent and permits partial oxidation of the
combustible gases present in the effluent. This simulates normal
foundry practice where air flow from mechanical ventilation is used
and the burning of mold gases is observed.
Concentration-time plots of the carbon monoxide from green sand
molds containing about 5% seacoal are shown in Figure 6. The cast-
ings were 4" cubes with appropriate ingates and pouring basins, which
gave a total weight of approximately 30 Ib. The castings were poured
at a sandrmetal ratio of approximately 3:1. The first peak in the
concentration-time curve represents the carbon monoxide concentra-
tion in the effluent after pouring. The second peak represents the
CO concentration when the mold was broken open.
The carbon monoxide concentration increased from a low value
just after pouring to approximately 1900 ppm about five minutes
-220-
-------
I
to
M
I
2000
18QQ
1600
1400
1200
1000
800
600
400
200
8 12
24
28 32 36
Time in Minutes
40 44 48 52 56 60
Figure 6. Variation of carbon monoxide concentration with time for green
sand mold under the hood.
-------
after pouring, and then began to gradually decrease. The casting
was held in the mold for approximately 25 minutes and then was
manually removed from the mold under the hooded sampling system.
During the simulated shakeout operation, the CO content of the ef-
fluent increased as the mold was broken open and the hot sand ex-
posed to the atmosphere. The CO concentration reached a level of
about 1350 ppm during this operation and then began to decrease as
the sand cooled.
Similar data is shown in Figure 7 for the concentrations of
total hydrocarbons evolved after pouring and during shakeout of
green sand molds. The hydrocarbon concentration increased after
the mold was poured, to a value of about 1200 ppm at six minutes,
and then began to decrease. The casting was broken out of the mold
about 30 minutes after pouring and maximum hydrocarbon concentra-
tions of about 1500 ppm were observed.
The average concentrations of the major gaseous constituents of
the effluents from the green sand are shown in Table II. Two dif-
ferent castings were made: the 4-inch cube discussed above and a
set of bars on an ingate that weighed about 15 Ib. (with correspond-
ing sand-to-metal ratio of 7:1). The values listed in Table II rep-
resent the average maximum concentrations observed during the pouring
and breaking out of the castings. The size of the casting does not
appear to be of major importance when the mold is well ventilated,
and the values for the concentrations are not significantly different.
Hydrogen was not detected in these samples; it can be expected
to burn in the mold. This burning, shown in Figure 5, may also ac-
count for some of the variations in the values obtained for carbon
monoxide and carbon dioxide.
-222-
-------
NJ
N>
E
a
a.
2000
1800
1600
1400
1200
1000
800
600
400
200
Time in Minutes
Figure 7. Variation of total hydrocarbons concentrated with time
for green sand under mold.
-------
Table II
Green Sand Emissions for Uncored Castings
Bar Mold
Sand:Metal Ratio =7:1
Cube Mold
Sand:Metal Ratio = 3:1
Element
Carbon monoxide
Carbon dioxide
Total hydrocarbons
Methane
Cyanide
Ammonia
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
Pour
1350
4920
1780
630
0.6
1.4
Shakeout
230
2360
640
80
0.4
1.1
Pour
1510
-
1400
520
1.3
1.1
Shakeout
650
-
470
250
3.3
3.4
-------
Particulate Emissions from Green Sand Molds
Considerable amounts of participate material are evolved from
the green sand molds after pouring, and during the breaking open of
the molds to remove the castings. The particulate matter contains
carbonaceous material from the burning organic material in the sand,
silica fines and clay from the molding aggregate, other fines pres-
ent in the mold, and metallic fumes.
The particle-size distributions of particulate matter collected
in the cascade impactors during pouring and breakout of the green
sand molds are presented in Table III. These values are for the 4"
cube casting weighing about 30 pounds. During pouring, 95% of the
particles evolved were less than 5 microns in diameter. During
shakeout, 50% of the particles were less than 5 microns in diameter.
Particles in this size range are considered more hazardous to human
health than larger airborne particle^ because they penetrate into
the lungs and are deposited there.
The dust loading of the effluent averaged 0.0625 grains/scf after
pouring and during solidification, and 0.0968 grains/scf during the
breakout of the 30 Ib. casting. These calculations were based on
the total weight of material collected during the sampling time of
the impactors, and so represent average concentrations of dust in
the air.
However, a time profile of dust concentration, shown in Figure 8,
shows that loading peaks occur soon after pouring and again just after
the breaking open of the mold to release the casting. This type of
data was generated using an optical particle counter sampling over a
ten-second interval and scanning several size ranges. The data shown
in Figure 8 is for particles in the size range of 0.35 to 1.00 microns
-225-
-------
Table III
Particle Size Distributions of Green Sand
Emissions for 4" Cube Pattern
Pouring
Shakeout
Size (microns) Mass (grams) % of Total Mass (grams) % of Total
Less than 0.54 3.98
0.54 - 0.83 8.35
0.84 - 1.34 23.01
1.35 - 2.67 16.69
2.68 - 4.14 1.86
4.15 - 6.08 .97
6.09 - 8.95 .53
8.96 -14.36 .40
More than 14.36 .68
7.0
14.8
40.7
29.5
3.3
1.7
0.9
0.7
1.2
5.14
2.28
1.36
0.36
0.56
0.24
10.88
0.34
0.28
24.0
10.6
6.3
1.7
2.6
1.1
50.7
1.6
1.3
-226-
-------
o
M
O
-H
e
o H
c
i
c
O m
•H m
-P •
<0
M 0)
4-> N
C-H
-------
A similar peaking is observed for larger particles, but the concen-
trations are much lower.
Particle counts increase very rapidly after pouring to a count
of about 30,000 per cubic centimeter and then exponentially decrease
with time. After allowing the casting to cool for 30 minutes in the
mold, the mold was broken open in a simulated shakeout operation and
the number of particles in the effluent increased dramatically to
about 300,000 per cubic centimeter, an order of magnitude greater
than the particulate emissions observed during pouring. This agrees
with the impactor data that showed shakeout to have a higher dust
loading than pouring.
The dust samples of particles less than 5 microns in diameter
collected in the impactor were analyzed for free silica by the
Talvite colorimetric method recommended by NIOSH.6 The free silica
content of the dust ranged from 0.67% to 7.3%, with the mean value
slightly over 1%. Other constituents of the dust were not identified.
Pilot Plant Studies
The sand casting laboratories were used for evaluating emissions
from a small scale foundry operation. The molding sand used is high
in natural clay content and is referred to as "Yellow Velvet." This
sand has no seacoal in it, but 1% cereal binder is added. The water
content is nominally 5%. Emissions were measured during mulling, at
an automatic molding machine, and during pouring and shakeout.
In these trials/ a plate casting with a sand-to-metal ratio of
4:1 and an open riser was cast. No mold ignition was observed, either
during pouring or while on the runout line cooling, although steam was
visibly given off. Table IV lists the results for the pilot plant
-228-
-------
Table IV
Pilot Plant Emission Data
Muller:
Dust loading .00396 gr/scf
CO <50 ppm
CO2 850 ppm
Total Hydrocarbons 200 ppm
CH4 7 ppm
Hunter Molding Machine:
Dust loading .00373 gr/scf
CO <50 ppm
CO2 775 PPm
Total Hydrocarbons 180 ppm
CHi, 2 ppm
Pouring:
Dust loading .00291 gr/scf
CO I500 PPm
C02 250° PPm
Total Hydrocarbons 250 ppm
15 PPm
Shakeout:
Dust loading .01654 gr/scf
CO 2
co 670 ppm
Total Hydrocarbons 215 ppm
5
-229-
-------
study. The pouring emissions, collected under the hood, show a
grain loading in the effluent of .00291 gr/scf. Carbon monoxide
reached a maximum value of 1500 ppm, and the carbon dioxide level
was a maximum of 2500 ppm. The total hydrocarbon content, calcu-
lated relative to methane, was 250 ppm.
The shakeout was accomplished by dumping the molds into a basket'
that moved up and down vigorously, so that the sand fell away from
the castings and out the bottom. The atmosphere in this operation
was extremely humid. The dust loading measured in the impactor was
.01654 gr/scf. However, the gaseous effluent using this sand was
much lower, with a carbon monoxide level less than 50 ppm, and car-
bon dioxide registering at only 670 ppm. Total hydrocarbons were
215 ppm.
Table IV also contains data on several other sand mold prepara-
tion operations found in the foundry. In mulling, batches of 150
Ib. were prepared by adding the sand, then the clay, and finally the
water, and mulling for three minutes. Total mulling cycle time was
approximately six minutes, and sampling was continued through four
complete cycles to insure that impactor catches were large enough
for accurate weighing. Average sand temperature was 80°F.
The grain loading was observed to be .00396 gr/scf. The carbon
monoxide level was below 50 ppm, and the carbon dioxide was near 850
ppm. Total hydrocarbons were present at 200 ppm. The operation in
the pilot plant was considerably cleaner than with the green sand
containing seacoal that was sampled in the laboratory muller. Table
V lists the effluent analysis from the laboratory tests for the
various sand preparation steps. The moisture content of the sand
is the largest variable. A hot, dry sand as it is returned from the
-230-
-------
Table V
Laboratory Sand Preparation Effluent
Muller
Dust loading (dry) .192 gr/SCF
(wet) .007 gr/SCF
Average CO 18 ppm
Average CO2 1530 ppm
Average Total Hydro-
carbons 115 ppm
Jolt-squeeze molding
Average CO <20 ppm
Average C02 800 ppm
Average Total Hydro-
carbons <10 ppm
Shakeout
Dust loading .149 gr/SCF
Average CO 20 ppm
Average CO2 510 ppm
Average Total Hydro-
carbons 80 ppm
-231-
-------
shakeout area produces a much higher grain loading than the wet
ready-to-mold sand.
Discussion
This research has shown that a number of undesirable substances
can be emitted into the foundry atmosphere from the green sand mold-
ing operation as it is commonly employed. Data from this research
should be useful in developing suitable designs of ventilation sys-
tems, and it certainly shows the necessity of proper ventilation in
the pouring area and in the run-out and shakeout areas of the foundry.
A typical foundry may pour over a hundred molds an hour. Each
mold can produce an average of about 750 ppm of CO, or 150 ft.3, over
a period of an hour, as estimated by integration of the area under
the curve in Figure 6. The other gaseous constituents will be a frac-
tion of this amount, but the combined amounts are large enough to be
important.
In a foundry where the ventilation system is good, these species
may well be diluted to the point where their discharge is harmless.
The peak concentrations will tend to be levelled out as a new mold
is poured every 30 to 45 seconds. In those foundries that have not
considered the problem, the amount and nature of the environmental
pollution may be significant.
Likewise the particulate matter introduced into the foundry at-
mosphere by the sand casting process is approximately 5.5 grams per
mold for one of the castings selected in this investigation. The
particle size distribution in Table III indicates that much of this
material, especially that from the shakeout, will settle out inside
the plant, or in the vicinity if it is entrained in the ventilation
-232-
-------
system and exhausted from the plant. However, there are appreciable
amounts of fine particles in these emissions that may present environ-
mental problems. The low content of free silica in this material in-
dicates that it may not present a serious health problem. These parti-
cles are small enough to remain airborne for several days and they are
in the respirable size range. If some of the organic compounds emit-
ted in the molding operation (such as the polynuclear aromatic hydro-
carbons) are condensed on them, as is quite possible, they may present
a hazard to human health. Much work needs to be done in the area
of qualification and quantification of these compounds in the foun-
dry emissions before the scope of the problem is properly understood.
There are several different ways to produce molds, and perhaps
one of these will decrease the severity of the problem. There are sea-
coal substitutes being marketed that, it is claimed, will decrease
the amount of carbon monoxide emitted. However, these are polysty-
renes and it is possible that some of their thermal decomposition
products could be equally undesirable. Unfortunately, the reducing
atmosphere at the mold-metal interface which is needed to make a
good casting almost necessarily results in the production of noxious
organic emissions.8'9
In addition to the traditional clay bonded molding aggregates,
there have been a number of chemical binders introduced in the last
fifteen years. It is possible that these no-bake binders may im-
prove the emissions. There are several of these binders that have
little or no organic constituents, and perhaps one of these can be
developed to where it can produce acceptable castings with relatively
few harmful emissions.
-233-
-------
Recent work with permanent mold casting processes for iron cast-
ings has shown that an increasing number of castings may be made by
this process without loss of desired physical properties and with
significantly less environmental contamination.2'10 The weight of
particulate introduced into the atmosphere on a per casting basis
was reduced from 5.5 grams to 0.15 grams for one particular configu-
ration. Likewise, there were significant reductions in both peak
and average gas concentrations. The average carbon monoxide con-
centration was less than 35 ppm. Unfortunately, this process is not
suitable for all castings, but it could be used more extensively
than at the present.
In conclusion, the foundry industry as a whole may have another
problem in that their fugitive emissions appear to present an unde-
sirable environmental contamination. Control will be difficult and
expensive, due to the large amounts of air involved. The cost of
control will have to be borne on top of the costs of control of stack
emissions and requirements for meeting Department of Labor (OSHA)
compliance. Many in the industry feel that they may not be able to
afford another costly compliance program. There is evidence that a
number of foundries have already been forced to close due to the cost
of EPA and OSHA compliance requirements. In addition, the money spent
by the foundries for compliance has been diverted from needed capital
expenditures that would normally upgrade the technology in the foundry.
A great deal of work needs to be done to identify and quantify the
fugitive emissions from a foundry and to determine the importance
of their contribution to the whole environment as weighed against
-234-
-------
the importance of the industry to our society. If control should still
be necessary, it then becomes paramount that the methods be fully de-
veloped before being implemented to insure the best capital utilization,
Acknowledgements
The experimental work reported in this paper was supported under
Grant 1 RO1 OH 00456-01 from the National Institute for Occupational
Safety and Health, Public Health Service, Department of Health, Educa-
tion and Welfare. The authors also gratefully acknowledge the
assistance of Mrs. Ruby James and Mr. Joe McCain and their staffs.
cbf
-235-
-------
REFERENCES
1. Scott, W. D., James, R. H., and Bates, C. E., "Foundry Air Con-
taminants from Green Sand Molds," Journal of American Industrial
Hygiene Association, April, 1976.
2. Bates, C. E. and Scott, W. D., "Better Foundry Hygiene Through
Permanent Mold Casting," Final Report to NIOSH on Contract
1 R01 00456-01, January 30, 1976.
3. Gutow, B., "An Inventory of Iron Foundry Emissions," Modern
Castings, January, 1972, p. 46-48.
4. Bates, C. E. and Scheel, L. D., "Processing Emissions and Occu-
pational Health in the Ferrous Foundry Industry," Journal of
Am. Indus. Hygiene Assn., August,,1974, p. 452-462.
5. Gwin, C., Scott, W. D., and James, R. H., "A Preliminary Inves-
tigation of the Organic Chemical Emissions from Green Sand
Pyrolysis," to be presented at the 1976 Annual Conference of
the Am. Indus. Hygiene Assn., Atlanta, Goergia, in May, 1976.
6. "Criteria for a Recommended Standard for Occupational Exposure
to Crystalline Silica, " HEW Publication No. (NIOSH) 75-120,
1974.
7. Smith, W. B. and McCain, J. D., "Particle Size Measurement in
Industrial Flue Gases," Air Pollution Control, Part II, John
Wiley & Sons, 1976.
8. Scott, W. D. and Bates, C. E., "Decomposition of Resin Binders
and the Relationship Between Gases Formed and the Casting Sur-
face Quality," AFS Transactions, 1975, p. 519-524.
-236-
-------
9. Draper, A. B. and Gaindhar, J. L., "The Role of Mold Atmospheres
in the Penetration of Steel in Sand Molds," AFS Transactions,
1975.
10. Jones, C. A. and Bates, C. E., "Permanent Mold Casting of Gray,
Ductile and Malleable Iron," AFS Transactions, 1972, p. 547-559.
-237-
-------
Tuesday Afternoon, May 18, 1976
SESSION IV: CONTROL TECHNOLOGY
CONTROL OP FUGITIVE EMISSIONS IN PETROLEUM REFINING
John H. Weiland, Texaco Inc.
(Representing the American Petroleum Institute)
-239-
-------
Mr. Weiland holds the title of Coordinator in the
*
Environmental Protection Department of Texaco Inc. His pri-
mary responsibility is coordinating Texaco"s activities in
the areas of stationary source emissions and environmental
noise. Mr. Weiland has been actively involved in the American
Petroleum Institute's environmental affairs activities for the
past 10 years, having served as Chairman and member of a number
of task forces investigating various industry problems asso-
ciated with stationary source emissions. He is currently
Chairman of the Stationary Source Emissions Committee of the
API Division of Environmental Affairs.
•X-*********-****-**
For the purpose of this discussion fugitive emis-
sions are defined as any emissions which are not released
through a stack or duct. In petroleum refining,under this
definition,volatile hydrocarbons are the primary futitive
emissions of concern. It is proposed to review some of the
emission sources, discuss briefly the emission factors that
are commonly used to attempt to get some fix on these emis-
sions, and then discuss some of the control methods that may
be used.
There are a great many possible sources of fugitive
hydrocarbon emissions in. a refinery. A modern refinery is a
very complex installation consisting of numerous different
processes, all with associated pumps, compressors, pipe line
flanges, valves, including relief valves, tanks, and so on.
There may be thousands of flanges and valves in a large complex
refinery. Pumps and compressors can number in the hundreds.
-240-
-------
Other possible sources of fugitive emissions, much fewer in
number, include vacuum jets, cooling towers, blowdown systems,
hydrocarbon loading operations, sampling of the hydrocarbon
streams, process drains, and oil-water separators.
The emission factors in EPA publication AP-42*
basically go back to studies made in the Los Angeles area in
the last of the fifties. These studies were conducted by the
U.S. Public Health Service in cooperation with the Western Oil
& Gas Association and the Los Angeles Air Pollution Control
District. This was a very excellent study in that an effort
was made to thoroughly inventory the equipment and facilities
which actually existed in these plants and the services in
which the equipment was operating. Then "representative"
individual items were selected for emission testing. Various
procedures were used, including enclosing the equipment in
plastic "tents" where feasible to capture and measure the
emissions.
There have been a number of comments made during the
course of this conference which have emphasized that an emis-
sion factor for a given piece of equipment or process is not
necessarily applicable across the board to all superfically
similar equipment in a whole industry. Thus, it is important
to keep in mind that, in an absolute sense, the factors that
were derived for petroleum refineries were specific to the
^"Compilation of Air Pollutant Emission Factors," U.S. Envi-
ronmental. Protection Agency, March, 197';;.
-241-
-------
pieces of equipment which were measured and the extrapolation
of these factors to any refinery, in the gross sense, or to
any other petroleum operation is not technically completely
accurate. The results, however, have been useful to give a
general idea of where major problem areas may be and to help
evaluate the possible order of magnitude of refinery emissions.
At the present time the emission factors developed
in the late 1950fs are no longer believed generally applicable
in many cases. The technology's improved a very great deal:
the operating and maintenance procedures have improved over
the last 15 years and some of these factors are undoubtedly
high by an order of magnitude or more when applied to modern
technology and operations. Consideration has been and is
being given by industry and various regulatory agencies to
the updating of these emission factors. This may indeed be-
come necessary in order to obtain a current assessment of the
present order of magnitude of fugitive hydrocarbon emissions
from modern refineries. However, it is not an easy or inex-
pensive thing to do.
Now, let's consider possible control methods. Good
housekeeping and good maintenance is the key to eliminating
or minimizing hydrocarbon losses from many of the sources men-
tioned. It is necessary that operating personnel realize that
hydrocarbons must be contained within the appropriate lines
and vessels and that any leaks or malfunctions should be cor-
rected promptly. Specific comments on the various sources
follow:
-242-
-------
• Flanges and valves - Leaks can be eliminated or held
«
to an absolute minimum with adequate inspection procedures and
prompt maintenance to correct any leaks noted.
• Relief valves - Venting of relief valves to the
refinery flare system to the maximum extent feasible will
reduce hydrocarbon emissions from this source to a minimum.
• Pumps and compressors - There has been a great deal
of both technical and housekeeping improvement in this area.
The industry is going more and more to the use of mechanical
seals on new equipment as opposed to packing glands. Properly
designed and maintained mechanical seals reduce emissions to
insignificance in pumps and compressors. In cases where there
are older pumps with packing glands or where packing glands
are needed because of the type of service a technique that can
be used is to collect any drips and route them into an under-
ground slop oil tank for later pumpout to a slop oil system,
• Tankage - The control of volatile hydrocarbons emis-
sions from tankage is accomplished either by the use of float-
ing roofs or vapor recovery systems. Both of these are very
effective in controlling these emissions, and over the years
the industry has increasingly utilized such systems.
• Vacuum jets - In the past many vacuum towers on
crude units used a barometric condenser system to condense
the steam from the vacuum jets. This requires the injection
of large volumes of water into the steam jet. The water and
condensed steam was discharged into a sump at the base of the
jets. There is some carryover of hydrocarbon
-243-
-------
fractions into the barometric sump resulting in some hydrocar-
bon evaporation into the atmosphere. This can be corrected by
replacing barometric condensers using direct water injection
with surface condensers.
• Cooling Towers - The control of cooling towers is
obviously a question of maintenance. There will be no hydro-
carbons in cooling towers unless there is a leak in condensers
or coolers somewhere in the system and unless the pressure
differential is such that the hydrocarbon pressure in the
equipment is in excess of the water pressure.
• Slowdown- systems - In preparing a unit for test and
inspection,any vessels, exchangers, lines, etc. that are going
to be opened must, as a matter of safety, be thoroughly purged
of hydrocarbons. To minimize hydrocarbon emissions the unit
can be deprassured to a flare and the liquid hydrocarbons
that remain pumped to a slop tank. The jremianing hydrocarbons -
and there'll still be little residues here and there -- can be
drained to an underground slop tank.
»
• Process drains, oil/water sewers, and oil/water sep-
arators - The best solution to these, of course, is to keep
the amount of hydrocarbons entering them to a minimum. Over
and beyond that, use of covered-drain systems and covered oil/
water separators will prevent or minimize hydrocarbon emissions
to the atmosphere. In addition, segregation of steam conden-
sate and any other high temperature water to prevent its dis-
charge into oil/water sewers or process drains will reduce
volatilzation of any hydrocarbons in the system.
-244-
-------
• Hydrocarbon loading operations - Hydrocarbon emis-
sions during loading of tank cars and tank trucks can be con-
trolled by use of a vapor-recovery system which will recover
the hydrocarbons during the loading operation.
• Fugitive dust emissions - The major possible source
of such emissions in a refinery is from coke piles associated
with coking units. There are at least three ways of handling
any problem of dust emissions from a coke pile. One involves
the use of a large building in which to store the coke under
cover; If outside storage is used, the pile can be wet down
with water. This can create a problem in that any water run-
off must be settled to remove coke dust. Another way that
has been used where the pile is going to be inactive for a
time is to spray the pile with some type of a polymer solu-
tion which will seal the surface.
In summary^ a well controlled and maintained refin-
ery should have minimal problems with the types of emissions
discussed. One of the ways to measure the progress that a
given refinery has made over the years is to compare the odor
eminating from a well-controlled refinery today with how the
refinery smelled say 35 years ago. Fart of the improvement,
of course,, is associated with control of stack emissions.
But a great deal of it is due to the better housekeeping and
to the use of improved equipment 'and procedures. This improve-
ment is certainly indicative that the refinery Is indeed doing
a good job of controlling fugitive emissions.
-245-
-------
THE COST EFFECTIVENESS OF
COKE OVEN CONTROL TECHNOLOGY
Robert E. Kenson
Norman E. Bowne
William A. Cot£
TRC - THE RESEARCH CORPORATION
of New England
125 Silas Deane Highway
Wethersfield, CT 06109
-247-
-------
THE COST EFFECTIVENESS OF COKE OVEN CONTROL TECHNOLOGY -
STUDY OF AMBIENT AIR IMPACT
1.0 Introduction
The state-of-the art of coke oven emissions has progressed dramat-
ically in the last decade. Substantial reductions in charging, pushing,
and quenching emissions are possible; however, there are large costs as-
sociated with this new control technology. Four questions need answering:
1. What effect do controls have on ambient air quality?
(Primary standards and/or allowed significant deteriora-
tion increment)
2. What is feasible for retrofit on present coke ovens?
3. What is feasible for installation with new coke ovens?
4. What controls are most cost effective?
TRC air quality studies for industrial clients have shown that
low level or fugitive emissions, such as those usually found in coke oven
operations, are often the critical factor in determining local ambient air
quality rather than the stack emissions.
The steel industry is faced with the following dilemma:
o Ambient air quality standards are not being met in many
urban areas where integrated steel mills are located.
o Substantial sums of money have already been expended by
steel companies to control stack emissions.
o The highly visible fugitive emissions from operations
such as coke oven batteries have been identified as the
next targets for control equipment installation.
o The cost of fugitive emission controls is high and not
as well proven as those for stack emissions.
o If, after installation of these controls, ambient air
quality does not substantially improve, more pressure
will be on for tighter controls on all steel mill air
emission sources.
It is the purpose of this paper to show how to assess the cost effec-
tiveness of coke oven emission controls. The approach uses measured coke
oven emission rates or best estimates of them as inputs to a computer model
which predicts their impact on amtjient air quality. Knowledge of the costs
of each control concept and the degree of control of coke oven emissions
for each concept permits the determination of the incremental cost of in-
cremental improvements in ambient air quality. When these are plotted,
cost effectiveness curves are developed which answer two serious questions.
These questions are:
-248-
-------
o Will coke oven emission controls solve the ambient air quality
problem? n *
o If so, what is the minimum cost solution to the problem?
If emission controls are installed, they must solve the right problems.
2.0 Coke Emission Control Technology
Although the purpose of this paper was not to establish the state-of-
the-art in coke oven control technology, an evaluation of the feasibility
(commercial and technical) of various proposed control methods was re-
quired. The most proven technologies were identified for control of:
o Coke oven charging emissions
o Coke pushing emissions
o Coke quenching emissions
OTHER SOURCES (door leaks, underfiring, coal and coke handling)'were not
evaluated here.
2.1 Coke Oven Charging
i
The charging of coal into coke ovens results in a fugitive emission
release consisting of coal dust, tars and gases from the charging hole.
Control technologies considered commercially feasible for prevention of
substantial charging emissions include:
o Staged charging with oven evacuation
o Larry cars equipped x^ith gas collecting systems and wet
scrubbers
o Pipeline charging
The former two have been considered as retrofits for present coke ovens as
well as feasible for new coke oven battery construction. In some cases,
pipeline charging, which is a technology considered suitable for new con-
struction, has been installed in rebuilt batteries to meet the need for
strict control of particulate emissions.
2.2 Coking
Leakage of emissions (gases, fumes) from the coke oven doors and other
openings in the ovens are minor but hard to control sources of emissions.
Although improved door sealing is a potential control, it is hard to esti-
mate the degree of control achieved by this technique. The only controls
considered here were sheds with scrubbers whose primary purpose was to con-
trol pushing emissions.
2.3 Coke Pushing
The pushing of the incandescent coke from the oven into the quench
-249-
-------
car results in emission of hot coke particles and tars as well as gases
from the coke as it leaves the oven and dumps into the quench car. Al-
though there have been commercial control equipment installations, the
technology is undergoing change and new concepts are now in design stage.
Commercially feasible controls include:
o Coke side sheds ducted to wet scrubbers or electrostatic
precipitators
o Coke guide and hooded quench car
Both have been considered for retrofit to present ovens and for new con-
struction. In new construction, the hooded quench car can contain a mo-
bile quench station which eliminates quench towers.
2.4 Coke Quenching
Although changes in this technology may be more related to water
reuse and pollution, they do result in lower air pollution emissions.
Commercially feasible control technologies include:
o Dry quenching
o Coke guide and hooded quench car (with mobile quench station)
t
Both technologies have been primarily considered for new installations
rather than retrofit.
2.5 Effectiveness of Controls
"Because of the problems associated with fugitive emission measurement
and the small number of present control equipment installations, the
effective percent emission control of many of the technologies described
can only be estimated. Conservative estimates were used for calculations
in this paper, and deviations from these in actual practice can be shown
not to be critical for analysis of the data. Comprehensive and accurate
measurement programs will be required to give us better emission factors
for coke oven operations. Table II-I sums up the emission estimates used
in the prediction of the effect of coke oven emissions on ambient air
quality.
3.0 Air Quality Prediction
3..1 Methodology for Prediction of Impact
The basic procedures for the accurate prediction of the impact on
ambient air of any emissions, whether fugitive or stack, include:
o Identification of all significant emission points
o Estimation of emission release heights, temperatures and
exit velocities
o Estimation of emission rates, and particle size/density for
particulates (latter for deposition models)
-250-
-------
o Evaluation of availability and validity of local ambient
air quality and meteorology data
o Application of a reliable, tested short term (1 - 24 hour)
or long term (annual average) diffusion model to worst case
air quality time period or periods
o Comparison of measured and predicted air quality to establish
the accuracy of the calculations
o Use of auxiliary techniques (tracer studies, morphological/
chemical analysis of particulates, etc.) to enhance accuracy
of predictions if required.
TRC has applied these techniques to the study for the steel industry
of coke emission impacts for several projected and existing facilities.
The predicted and observed air quality data (in the case of existing facil-
ities) has shown the validity and the basic accuracy of the methodology
mentioned above.
3.2 Case Studies
As a case study, TRC chose a hypothetical coke oven facility which
could represent a proposed or present facility. The facility specifica-
tions are given in Table III-l. This represents a medium size "grassroots"
mill or an expansion of a present integrated steel mill,
The emission factors of Table II-l were used to represent those of
the actual facilities with and without various add-on controls. A TRC
developed short-term air quality diffusion model was used in conjunction
with "worst case" 24-hour meteorology from typical steel mill locations in
East Coast and Pittsburgh areas to show the impact of coke oven emissions
on ambient air quality at two different site types (coastal and river
valley). Particle size was. assumed to be 1.0 y and particle density was
assumed to be 1.0 gm/cc. Emissions from coal handling or oven underfiring
were not modeled.
3.3 Results of Modelling
Figure 3-1 is an isopleth map of the "worst case" 24-hour particulate
concentrations for the East Coast location with no coke oven emission con-
trols. As a comparison, Table III-2 shows the "fenceline" air quality re-
sults and results at 1/4, 1/2 and 1 mile from the "fenceline" for both
controlled and uncontrolled coke oven emissions.
The results show that uncontrolled coke oven emissions have a signifi-
cant impact on local ambient air quality. In order to ensure that off-
plant ambient air samples met the 24-hour primary particulate standards
(considering here also the other potential fugitive sources in-plant and
background particulate levels), an uncontrolled coke oven would have to be
located about 1 1/2 miles from any steel plant boundary. In fact, it
would be imperative that significant modelling predictions be run before
the location for such a facility is chosen. The impact decreases drastic-
ally with distance from the plant "fenceline" and is also directionally
dependent.
-------
A similar analysis can be given to the predicted particulate concen-
trations for the Pittsburgh area location. Figure 3-2 is an isopleth map
for particulate concentrations for this case, and Table II-3 summarizes
particulate concentrations as a function of distance from plant "fence-
line". The comparison of controlled vs uncontrolled coke oven emissions is
similar to that for the East Coast location.
One point worth noting is the impact of quench tower emissions on am-
bient air quality. Because they act as a high velocity elevated plume from
a stack (except for "rain out" close to the tower) rather than as a fugi-
tive emission, their impact would be at a greater distance from the plant
than low level fugitive emission from changing or pushing. In fact, their
local impact might be increased by their control by a mobil quench facility
whose off-gasses are wet scrubbed even though overall impact is decreased
off plant property. The low buoyancy, low velocity plume if released from
a stub stack will be transported in a similar manner to the true fugitive
emissions.
The two case studies shown are just examples of potential impact of
coke over emissions on ambient air quality. Each specific site may lead
to quantitatively different results. However, the two cases cited show,
for locations of many present and proposed steel mill facilities, the en-
vironmental problems of coke oven site selection.
4.0 . Cost Effectiveness of Controls
In order to determine the cost effectiveness of the potential coke
oven emission control technologies, it was necessary to compare the costs
of each of these. Table IV-1 shows updated cost data developed from cur-
rent literature sources. It shows that there are a wide range of costs
associated with the emission controls considered for present and proposed
coke ovens. Their cost effectiveness is determined by plotting the pre-
dicted ambient air quality at a specific location against the cost of the
control technology required to obtain that air quality.
Figures 4-1 and 4-2 show such plots for two ambient locations beyond
plant "fenceline" (East Coast and Pittsburgh locations are plotted on the
same curves). As better and better air quality is required, the costs
rise significantly and not in any direct proportion to the air quality im-
provement. These cases have in addition neglected the fugitive emissions
from such integrated steel mill sources as blast furnaces, EOF shops, elec-
tric furnace shops, sinter plants and ore/limestone/coal/coke storage.
The cost curves are similar for control of these. It is therefore impor-
tant to ask the question: Is there a significant overall benefit to the
population in reducing coke oven (and other fugitive source) emissions
below a certain level where controls are no longer cost-effective in order
to meet air quality standards? An additional question is: Is there
another control strategy which would achieve the same result?
5.0 Conclusions
The results of this,case study have the following implications for
present coke oven operations:
o Stringent controls may be required on coke ovens to help
-252-
-------
the attainment of ambient air quality standards in the
local area.
o It is important to identify the sources which signifi-
cantly impact the ambient air so that the controls can
be applied to the proper sources.
o The controls may be costly and require rebuilding of the
ovens.
o It is possible, even after the implementation of con-
trols, especially where property lines are close to the
coke ovens, to have ambient air quality standard vio-
lations partially caused by coke oven emissions.
o It is important for steel mill management to examine
the cost-effectiveness (air quality benefit vs incre-
mental control costs) of several control system alter-
natives for coke oven emissions before implementing any
of'them. *
o Attention should also be paid to control of other in-
plant and non-plant fugitive sources.
• For proposed coke oven operations, the aforementioned hold and, as
well, there are the following implications:
o The environmental as well as construction/process
engineering aspects of coke oven battery placement
must be taken into account in design and placement
of a new facility.
o Even with stringent coke oven controls, problems may
occur in meeting the air quality deterioration criteria
for new facilities even in areas where such industrial
expansion would be allowed.
The technology of coke oven emission control is advancing, however it
is important to assess the impact on ambient air quality of the expendi-
ture of such large sums of capital. TRC has identified, and is now pur-
suing, the following areas of future needs to improve our ability to pre-
dict the impact of fugitive emissions (such as coke ovens) on ambient air ,
quality:
o Methods for measurement of mass rate, particle size/
density of fugitive emissions.
o Methods for the measurement of time dependent and ill-
defined fugitive emission sources.
o Methods for the calibration of models using tracers to
develop site-specific detailed models for industrial
sites.
o Methods for better defining non-industrial emission
-253-
-------
sources or ill-defined fugitive sources to allow
better predictions of ambient air quality.
hese areas of development will allow us to give a more thorough analysis
f what affects local ambient air quality and how to improve it in the
ost cost-effective manner.
-254-
-------
REFERENCES
1. R. E. Kenson, P. W. Kalika, and J. E. Yocom, "Fugitive Emissions
from Coal", NCA/BCR Coal Conference and Expo II, Louisville, KY,
October, 1975.
2. P. W. Kalika, P. T. Bartlett, R. E. Kenson and J. E. Yocom,
"Measurement of Fugitive Emissions", 68th Annual APCA Meeting,
Boston, Mass., June 1975.
3. L. F. Kutcher and B. Linsky, "Economics of Coke Oven Charging
Controls", Air Pollution Control Association, 24:765 (1974).
4. Air Pollution Emission Factors, EPA Publication AP-42, April 1973.
-255-
-------
Table 11-1
Emission Factors* Used for the Prediction of the
Impact of Coke Oven Emissions on Ambient Air Quality
Operation
Charging
Charging
Charging
Charging
Coking
Coking
Coking
Pushing
Pushing
Pushing
Quench
Quench
Quench
Controls
technology
None
Staged charging
Scrubber on Larry car
Pipeline charging
None
Coke side shed/scrubber
Push/coke side shed/scrubber
None
Coke side shed/scrubber
Coke guide/hood
None
Coke guide/hood
Dry quench
Estimated
per cent
control
0
90
90
98
0
42
83
0
83
95
0
98
98
Particulate
emissions
(Ibs./ton coal)
1.5
0.15
0.15
0.03
0.1
0.06
0.02
0.6
0.1
0.03
0.9
0.02
0.02
* Uncontrolled taken from
EPA Publication AP-42
-256-
-------
Table III - 1
Hypothetical Coke Oven Facility For Case Study
Steel Production
Coke Production
Coke Batteries
Ovens/Battery
Oven Capacity
Coking Cycle
Oven Reliability
2.5 mm tons/year (6850 tons/day)
4100 tons/day
2
100
25 tons
16 hours
90%
Oven Site Size: 1000' x 3000' (including coal storage)
Steel Plant Size: 6000' x 8000' (centered on coke ovens)
-257-
-------
15,000
10,000
Distance
in feet
5,000
N
t
Figure 3 - 1
Particu late Concentration Isopleths
(in Micrograms/Cubic Meter) for
Coke Oven Impact Studies
24 Hour Worst Case - East Coast Location
Coke
oven
site
5,000
10,000
15,000
Distance in feet
-258-
-------
Table
Predicted Ambient Air Quality (Micrograms/Cubic Meter)
for Particulates - 24 Hour Worst Case for East Coast Coke Oven Location
Uncontrolled
Staged charging
Staged charging
plus
coke side shed
with scrubber
Staged charging
plus
coke guide/hood
Pipeline charging
plus
coke side shed
with scrubber
Pipeline charging
plus
coke guide/hood
Pipeline charging
plus
push/coke side
sheds with scrubber
plus
coke guide/hood
plus
dry quench
Fenceline
sampler
• • "• 1 !!••
1290
500
182
165
110
94
53
1/4 mile
from
Fenceline
1275
490
180
160
109
93
52
1/2 mile
from
Fenceline
295
115
40
37
25
22
13
1 mile
from
Fenceline
30
12
4
4
3
2
1
-259-
-------
Table III - 3
Predicted Ambient Air Quality (Micrograms/Cubic Meter)
for Participates - 24 Hour Worst Case for Pittsburgh Coke Oven Location
Uncontrolled
Staged charging
Staged charging
plus
shed with
scrubber/precipitator
Pipeline charging
plus
coke side shed
with scrubber
Staged charging
plus
coke guide/hood
Pipeline charging
plus
push/coke side
sheds with scrubber
plus
coke guide hood
plus
dry quench
Fenceline
sampler
750
290
95
65
55
30
1/4 mile ,
from
Fenceline
480
195
60
40
35
20
1/2 mile
from
Fenceline
315
125
40
27
23
13
1 mile
from
Fenceline
67
28
9
6
5
3
-260-
-------
15,000
10,000
Distance
in feet
5,000
Figure 3 - 2
Particulate Concentration Isopleths
(in Micrograms/Cubic Meter) for
Coke Oven Impact Studies
24 Hour Worst Case - Pittsburgh Location
1
5,000
Distance in feet
10,000
15,000
-261-
-------
Table IV - 1
Estimated Costs of Emission Control Systems for
Coke Oven Cases Studied*
Controls
installed
1. AISI staged
charging
2. Coke side shed
a) with scrubber
b) with preciptator
3. Charging air
with scrubber
4. Enclosed coke
pushing and
quench car
with scrubber
5. Pipeline charging
(Retrofit)
6. Dry quenching
Capital costs
$
2 batteries
1,500,000
3,000,000
6,000,000
1,500,000
8,000,000
20,000,000
10,000,000
4 batteries
3,000,000
6,000,000
12,000,000
3,000,000
16,000,000
40,000,000
20,000,000
Operating, maintenance
and repair costs
$/year
2 batteries
200,000
450,000
200,000
225,000
800,000
2,000,000
650,000
4 batteries
400,000
900,000
400,000
450,000
1,600,000
4,000,000
1,300,000
*Using updated literature data and equipment
vendor quotations
-262-
-------
1000
Participate
concentrations C
at "Fenceline"
500
0
(
>- • East Coast Location
I O Pittsburgh Location
-*•
X
°^°- S ft
) 5 10 15 20 25 30 35 40 4
Capital cost for coke oven controls ($mm)
Cost Effectiveness of Coke Oven
Emission Controls
Figure 4 - 1
-263-
-------
300
O
200
Participate
concentrations
1/2 mile from
"Fenceline"
100
O
• East Coast Location
O Pittsburgh Location
10
20 30
Capital cost for coke oven controls ($mm)
Cost Effectiveness of Coke Oven
Emission Controls
Figure 4 - 2
40
-264-
-------
Session V:
FUTURE NEEDS FOR MEASUREMENTS AND CONTROL TECHNOLOGY
Robert M. Statnick, Ph.D.
Session Chairman
-265-
-------
FUTURE NEEDS FOR MEASUREMENT AND CONTROL OF FUGITIVE DUST
F. A. Renninger, P.E.
Vice President-Operations
National Crushed Stone Association
* * *
This is not a technical paper in any sense of the word.
It is a review of how the crushed stone industry looks at a problem,
which is basically a dust problem...fugitive dust problems.
Dr. Thomas Blackwood has laid some very excellent groundwork for
this review with his studies of the impact of open sources on over-
all air quality.
To begin-this presentation, here is a bit of background
on the crushed stone industry. The crushed stone industry is the
largest non-fuel mining industry in the country. Approximately
one billion tons of rock are processed per year. These billion
tons are produced from approximately 4500 to 4700 individual quarry
sites around the country. These quarries vary from extremely small,
independently-owned operations to extremely large, corporate-held
locations. For instance, the 1900 smallest quarries account for
less than 2 percent of the total production; the 180 largest quarries
account for over 50 percent of the production. There is a tremendous
size variation and there is also a tremendous difference in the
ability of individual companies to deal with the many technical pro-
blems facing the industry today.
BACKGROUND
Concern with the ambient dust situation in quarries started
in the early 1960's. In 1963 it was decided that the time had come
to try to determine what was meant by the numbers everybody was talking
about with respect to a rock quarry.
Considerable time was spent in the field and some very inter-
-267-
-------
esting things were discovered. The only types of applicable
measuring procedures available at that time were, of course, the
ambient air measuring techniques. But, it was found that there is
a very definite area of influence around a rock quarry. It doesn't
really make much difference whether one is upwind or downwind of
the site. The fact that there is a big hole in the ground tends
to set up a micrometerology in the area and an upwind/downwind
relationship does not truly develop. A scattering phenomenon in
all directions is more common. If one were to take a compass and
draw a circle approximately 3500 feet in radius from some central
point in the operation, that circle would encompass, for all practical
purposes, the impact area of that rock quarry. In other words, at
about this distance both the suspended air levels and the settled
dust levels essentially return to the normal background in the area.
High volume samples within this area literally have dust particles
falling off the filter paper. But, as Dr. Blackwood previously
stated, one is not really measuring suspended dust, but rather both
suspended dust and a substantial amount of dust that is on its way
out of the atmosphere...very heavy, large pieces of rock.
Since the above original work, this same pattern has
developed at a number of locations. There seems to be almost a
halving effect for every thousand feet one moves away from some
central point within the operation. Within this area, however,
there is without question, some impact not a toxic situation, but
rather a nuisance situation.
For years our industry has talked about the captive or
capturable dust, which is generated as a result of the -processing
-268-
-------
operation...crushing, screening and conveying. This is opposed to
the fugitive dust, which is generated from things like the unpaved
roads, the open quarry floor, the area that is stripped bare of
vegetation, etc. The plant-generated dust is fairly easily control-
lable and is not truly a problem. Collecting and suppression equip-
ment can be installed. However, there is some question as to whether
or not this truly solves the problem. There was a study conducted a
number of years ago by the Manatee County Health Department in the
State of Florida around a dolomitic lime plant -- quarry and plant
operation. The significant finding of that study was that regardless
of whether the control equipment was operating or not, the impact
on the downwind ambient air concentrations was negligible. In
other words, no real difference could be detected between the ambient
levels in that plant area when that plant's control equipment was
operating.
An assumption is that 75 to 80% of the crushed stone industry's
problem is probably due to what was defined above as fugitive dusts.
This has never truly been documented because normal test methods
have not permitted isolation of the various fugitive sources.
It has been inferred from looking at an operation that open road
areas and the open quarry floor were probably the primary problems.
Some of the data that was presented at this conference by
Dr. Blackwood tends to indicate that this assumption is correct. As
an example, he demonstrated that for a typical limestone quarrying
operation, there was a total emission in the neighborhood of .007
pounds per ton, and that roughly two-thirds of that was attributable
to the fugitive sources, primarily the haul roads and quarry floor areas
-269-
-------
CURRENT STATUS
The crushed stone industry has two control options. The
first option is the dry collection, the second is wet suppression.
Each option has certain inherent problems.
The major problem with the dry collection option, of course,
is the creation of a substantial solid waste disposal problem. There
are quarries that have collection equipment in operation and are
collecting from 50 or more tons of dry dust a day. They have a
gigantic solid-waste disposal problem. Much of this dry dust is
not saleable; therefore, it must be disposed of, dumped, or other-
wise handled. Another problems created by the installation of
dry collection equipment is the creation of a stack or stacks
which in turn subject the operation to stack emission codes. Since
there is evidence that there is very little measurable effect on
ambient air levels whether the control equipment is operating or not
not, one must ask whether a problem has really been solved. Just
recently the stone industry has been involved with the Environmental
Protection Agency as they have attempted to develop new source
performance standards. Two plants that were on the drawing board
were studied. They were roughly 300-350 ton-an-hour operations.
Note what would happen, not in terms of cost, but rather in terms
of power consumption, if adequate dry collection equipment to meet a
0.02 to 0.03 grain loading standard were installed. On these two
plants, the power required to operate the control equipment approached
20 to 26% of the power to operate the plant. That is a substantial
amount of additional power. Forget about the cost of the control
equipment; forget about the cost of that power. Just consider the
-270-
-------
total number of kilowatt hours in that 20 to 26% excess power
required to operate this collection equipment. Under those
circumstances, has the problem been solved or has a new problem
been created? In other words, somewhere, someplace, somebody
has to burn some coal or other fuel to generate that power being
used to operate the control equipment to collect the dust which
doesn't have an impact beyond a half mile. This is a very real
question and one which warrants serious consideration right now!
The stone industry currently has a program underway to try
to determine those areas in which the power currently used in the
production process can be conserved. There are projections with
respect to stone production which tend to indicate that by the
year 2000, the industry's ability to supply the demand will be
severely limited if the same amount of power per ton for production
is used. By the year 2000 the industry's ability to meet demands
will be jeopardized if power conservation is not achieved. The dry
collection option, in effect, nullifies power conservation techni-
ques. Roughly increased power must be saved before reaching the
break-even point, but even at the break-even point, the production
demands will not be met by the year 2000.
The second control option is the wet suppression system.
This is a highly-effective system. Go out to a typical plant oper-
ation and look at it with those suppression systems operating. Have
the operator turn them off and look at it again, 15 minutes later.
One can definitely see what those suppression systems are doing.
The biggest single problem is that it" cannot be measured quantitatively,
Are suppression systems 501 effective--901 effective or are they
-971-
-------
95% effective? As yet, there is no answer. They are economical
to use from the standpoint of capital investment, operating
cost and negligible power consumption.
Some 60 operations with only suppression equipment
operating, were observed, with ambient air samples taken outside
the plant boundaries. (Remember, there was the full effect of
all the fugitive sources, and the only control in the plant itself
was the wet suppression equipment.) The same dispersion model
used to measure the impact of stack emissions from crushed stone
operations was taken and worked backwards. The ambient air levels
converted to an equivalent stack emission of approximately 0.02
grains per standard cubic foot. This was a measure of the full
impact of all fugitive sources, as well as a plant controlled by
simply wet suppression equipment.
One other area of importance is types of regulations...not
specific regulations, but the types of regulations which have been
applied over the past 10 or 12 years. There are the stack emission
codes which can be applied if the industry either elects or is
forced to create stacks. Those are very readily understandable.
One can measure whether or not one complies with a stack emission
code. The fugitive sources have been handled by rather subjective
regulations. A typical one might say that an operator or an individual
shall take reasonable precautions to control fugitive emissions and
such precautions might include a whole list of things. The problem
with this is that nobody--not the control agency enforcing the regu-
lation nor the operator upon whom the regulation is being enforced--
can determine whether or not compliance is achieved. These types
of regulations have presented considerable problems.
-272-
-------
Dr. Blackwood mentioned the respirable fraction, which
he defined as that below 10 microns, and then mentioned the total
suspended portion as being less than 50 microns making no refer-
ence at all to those dusts above 50 microns. The question was raised,
from the audience, that those dusts are, in fact, nuisance dusts and
ought to be controlled. That is true. However, whether or not
they ought to be controlled under air pollution regulations is
another question, particularly as they relate to the stone industry.
It is almost virtually impossible, for one reason or another, in
the country today, to open a new crushed stone or sand and gravel
operation, the reasons being zoning restrictions, land use planning
restrictions, or general public apathy. All plants operate by virtue
of operating permits, which are reviewable. These larger dusts that
settle in the immediate area of a plant are viewed primarily as a
public relations problem. It just simply makes good sense to do
something about them if one expects to continue to operate. The
industry is making an effort to control fugitive emissions by
oiling, watering, or even paving haul roads, and quarry floor areas,
by shielding stockpile areas and by planting substantial vegetation.
Incidentally, a tree buffer is an extremely effective mechanism for
controlling large fugitive dust particles. During one of the first
studies it was observed in the fall of the year, the settled dust
concentrations tended to increase, even though the plant operations
were decreasing. In the spring season, they tended to decrease,
even though the activity at the plant was increasing. This phenomina
was observed for a period of about two years and correlated very well
with the foliation and defoliation of the trees.
-273-
-------
SUMMARY
The above are some of the problems and concerns the
crushed stone industry is facing. The biggest concern has to do
with the tendency to more or less force an industry, by virtue
of regulation, to move towards collection equipment. Our industry
is not convinced that, all things considered, that this is the
'>
best approach for the stone quarry. The entire plant area and
the quarry operation can be treated and controlled as a fugitive
source. Its area of impact is extremely limited. The number of
individuals involved or affected is another area that more
emphasis ought to be placed upon. Dr. Blackwood touched on it
in his remarks. For the most part, a great majority of quarry
operations tend to be located in rural low-density population
areas. The number of individuals affected and the effect upon
those individuals should dictate more strongly the nature and
type of controls to be demanded by various control agencies. Test
methods to truly measure the impact of open sources are sorely needed.
Most of the work up to now has been done by inference. Available
tests are run and based upon the results of those tests the impact
and/or true levels have been inferred. A method of assessment must
be developed so that a move away from subjective toward more defini-
tive control regulations can be achieved. Industry, the public,
and control officials should know the true impact of a given fugitive
dust source.
-274-
-------
DETERMINING EMISSIONS MEASUREMENTS NEEDS FOR AN EMERGING
INDUSTRY-ADVANCED,FOSSIL FUELS UTILIZATION*
M. R. Guerin and J. L. Epler
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37830
June 17, 1976
Presented at the First Conference on
"Determining Fugative Emissions Measurements Needs"
Hartford, Connecticut
May 17-19, 1976
* Research sponsored jointly by the Environmental Protection Agency and the
Energy Research and Development Administration under contract with the
Union Carbide Corporation.
By acceptance of thi* article, the
publisher or recipient acxnowledges
the U.S. Governmeni'i right to
retain a nonexclusive, royalty-free
license in and to any copyright
covering the article.
-275-
-------
DETERMINING EMISSIONS MEASUREMENTS NEEDS FOR AN EMERGING
INDUSTRY-ADVANCED FOSSIL FUELS UTILIZATION
M. R. Guerin1 and J. L. Epler2
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37830
Introductory Remarks
Converting coal or the organic matter in oil shale to petroleum-like
liquids or to fuel gases is predicted (1) to be a major component of the
nations energy base. Liquid products may also constitute a significant
source of raw materials for the petrochemical industry. Successful pro-
cesses are likely to constitute a new major industry in this country.
The many processes (2-5) now under study have the common objective of
extracting hydrocarbons with an atomic hydrogen to carbon ratio approxi-
mating that of natural crude oils or natural gas from the solids. Pyrolysis
and hydrogenation at elevated pressure and temperature are therefore required
to chemically convert the solids to usable liquids or gases. The COED Pro-
cess, designed to produce a high quality char, is illustrated in Figure 1 as
an example. The complexity of the starting material and of the chemical re-
actions occurring during conversion suggest the possibility of both point
and fugitive emissions. Evidence of potential occupational health problems
is available in the literature (6-10) for coal hydrogenation.
The existence or extent of environmental and health problems cannot be
determined at present because commercial scale plants do not exist. Few pro-
cesses have yet reached the demonstration plant scale. Methods are required,
however, to quickly and reliably assess the extent of the problem as large
1 Analytical Chemistry Division
2 Biology Division
-276-
-------
scale operations are initiated. Data generated in the course of methods
development can also be useful in designing control technologies. The find-
ing (11) that conditions can be modified in the Synthane coal gasification
.process to reduce emissions without sacrificing conversion efficiency pro-
vides additional impetus for early emissions characterization.
This paper presents one approach to establishing reliable methods and
generating data of value in prioritizing environmental and health studies.
Measurements needs are identified through an experimental assessment of exist-
ing methods.
Chemical Screening of Complex Mixtures
Solid, aqueous, gaseous, and liquid emissions (4) are theoretically possi-
ble from each process. Liquid products themselves must be considered until it
is demonstrated that they can be transported and handled as are natural crudes.
Each of these materials is expected to consist of a highly complex mixture of
chemicals varying in chemical type and concentration prior to applying emission
control methods.
At least three methods are commonly used to determine the chemical nature
of complex mixtures. Individual constituents, e.g., benzo(a)pyrene, indicative
of chemical classes, e.g., polynuclear aromatic hydrocarbons, can be determined.
The material can be subjected to class fractionation to obtain a weight percent
distribution of constituent types. Multicomponent chromatographic profiling or
m
direct spectral analysis of the untreated material can be carried out to esti-
mate chemical nature and complexity.
Each of these approaches has been used in the work reported here. Multicom-
ponent chromatographic profiling, particularly when combined with class fractiona-
tion or the isolation of specific subfractions has been found most useful. In-
dividual constituents must also be determined because biological effects are
stereospecific.
-277-
-------
Methodology
Samples which have been surveyed in varying detail to date (12-16) and
their sources are summarized as follows:
(a) Coal liquefaction product from the Synthoil Process and the
aqueous condensate from the Synthane Gasification Process
courtesy of the Pittsburg Energy Research Center.
(b) Samples from the points in the COED Pyrolysis Process num-
bered in Figure 1 courtesy of FMC.
(c) Shale oil from an above ground simulated insitu retorting
courtesy of the Laramie Energy Research Center. Product water
was obtained in-house by batch centrifugation of the oil-water
emulsion.
(d) A Louisiana-Mississippi Sweet crude oil courtesy of Dr. J. A.
Carter of the Analytical Chemistry Division, Oak Ridge National
Laboratory.
Gas samples and product headspace volatiles are profiled using on-column
cryothermal trapping and subsequent temperature programmed capillary column
gas liquid chromatography (12). Polynuclear aromatic hydrocarbons are isolated
as a group by sequential liquid-liquid extraction followed by florisil and
alumina column chromatography (15). Carbon-14 isotope dilution is used to lo-
cate eluting constituents and compute recoveries. The isolate is profiled gas
chromatographically using a'22 ft. long by 1/8 inch O.D. glass column of 3% Dex-
sil 400 on 80/100 mesh HP Chromosorb G programmed from 110°C to 320°C at a rate
of l°C/min. Benzo(a)pyrene is determined using the same extraction procedure
but is isolated from benzo(e)pyrene and other isomers by acetylated cellulose
paper or column chromatography and determined spectrophotofluorimetrically.
Alkyl chrysenes are determined using the method of Hecht (17). Aqueous samples
-278-
-------
are gas chromatographed without prior treatment using Tenax as the column
packing (13). The class fractionation method (16) used most extensively is
that developed (18) to elucidate the carcinogenic properties of condensed
tobacco smokes. Most of the constituents listed have been identified only
by isolation and co-chromatography. Identifications must therefore be con-
sidered preliminary.
Preliminary Results
Figure 2 illustrates the results of direct gas chromatographic profiling
of a stack gas sample and of the material volatilized from Synthoil at 50°C.
The stack sample is illustrative of a point emission and the volatiles are
illustrative of a possible fugitive emission. The profiles visualize organic
constituents containing from one to approximately seven carbon atoms. Very
low molecular weight compounds, hydrocarbons containing from one to four car-
bon atoms, are not resolved under the conditions used. Both samples are seen
to consist of a large number of isomeric organic compounds. Use of the flame
photometric detector allows selectively visualizing sulfur containing
constituents including hydrogen sulfide, carbonylsulfide, methyldisulfide, and
thiophene in gas samples (12).
Aqueous samples from the COED Process, Synthane Process, and simulated in-
situ oil shale retorting also contain a large number of constituents (Figure 3)
at high concentrations (Table 1). Waters derived from coal processing are gen-
*
erally found to contain phenol, the cresols, and other isomeric phenols as the
primary contaminants. Oil shale derived water contains a homologous series of
saturated carboxylic acids. Flame-photometric' detection indicates (12) the
presence of at least ten sulfur containing constituents in a product separator
liquor from the COED process. All of these constituents would be greatly re-
duced in effluent waters by standard control methods.
-279-
-------
Direct chromatographic profiling easily and rapidly provides a great
deal of information but is limited in two important ways, (a) chromatographic
conditions optimized to visualize non-polar compounds often preclude visualiz-
ing polar compounds and vice versa, and (b) compounds present at low concen-
trations are obscured by the major constituents. Procedures which combine
isolating the compound class of interest and chromatographic conditions opti-
mized for profiling that class represent the next level of complexity.
Figure 4 illustrates the chromatographic profiles of the polynuclear aro-
matic hydrocarbon (PAH) isolates from condensed cigarette smoke and a coal
liquefaction process. The comparison is more than academic because this frac-
tion is thought to be the primary contributor to mouse skin carcinogenicity of
condensed smoke and skin contact is a primary health concern (9) in handling
coal liquids. For purposes of quantitative comparison, the coal derived pro-
file was obtained at a detector sensitivity one eighth that used for the smoke
condensate. The PAH isolate from the coal product contains a wider variety of
constituents present at concentrations at least an order of magnitude greater
than in smokes.
Multicomponent profiling of isolates is readily carried out for qualita-
tive or semi-quantitative analytical purposes. Table 2 summarizes estimates
of PAH's in products and related aqueous samples obtained using the profiles.
Multicomponent quantitative determination is also possible if recoveries and
identities are known. Recoveries of 80% or more and relative standard devia-
tions of 10% for triplicate determinations are common with experience.
The complexity of the isolates and the importance of specific isomers
limits the utility of even quantitative multicomponent analyses. In the ex-
amples shown, benzo(a)pyrene cannot be distinguished from benzo(e)pyrene.
Methods optimized for the determination of selected constituents must supple-
-280-
-------
ment profiling. We find (12) approximately 40 ppm benzo(a)pyrene in one syn-
thetic crude oil by extractive and chromatographic isolation, isotope dilution,
and fluorescence measurement. Interest in methyl chrysene concentrations re-
quires a comparable isolation procedure plus the use of Diels-Adler adducts
to remove interfering benzanthracenes. Results of this study indicate the
presence of chrysene (98 ppm), 2 methyl chrysene (102 ppm), 3-methyl chrysene
(106 ppm), and 6-methyl chrysene (64 ppm). The highly carcinogenic (17) 5-
methyl chrysene has not yet been adequately resolved from the 4-methyl isomer
to establish its presence but the combined concentration of both constituents
is less than 20 ppm.
Figure 5 illustrates a traditional approach to the class fractionation of
complex organic mixtures. The percentages included in the figure correspond
to the weight percentage of those fractions obtained when syncrude from the
COED Process is subjected to the procedure. The primary results of this ap-
proach are weight percent distributions of constituent types and partially de-
fined fractions for further study. The procedure has !been applied to products
and aqueous samples from both shale and coal processing. Table 3 illustrates
results obtained for the COED and Synthoil products. Reproducibility is typi-
cal of qualitative procedures requiring much manual manipulation. Studies
are presently underway of column chromatographic fractionation methods which
promise to be more reproducible and less likely to produce chemical artifacts.
Sufficient sensitivity-and resolution is likely to allow the detection of
chemicals of every type. High molecular weight paraffins can be detected in
shale derived waters, as is illustrated in Table 4, for example. Information
is required to prioritize chemical types and sample types for measurements
research. Chemical fractionation followed by biological screening is in use
here (16) to prioritize chemical needs while developing the bioassay systems.
-281-
-------
Mil tag en 1 city Screening of Complex Mixtures
The potential predictive value of assays for genetic damage and the
correlation with carcinogenic damage has been emphasized in a number of re-
cent reports (19, 20, 21). The wide applicability of the bacterial test
system developed by Ames has been illustrated for a large number of "pure"
compounds but also may be used as a prescreen for ascertaining the genetic
and potential carcinogenic hazard of complex environmental effluents or pro-
ducts e.g., tobacco smoke condensates (22), soot from city air (23), hair dyes
(24) and in our preliminary work, synthetic crude oil (16). Furthermore, the
overall economy and rapidity of the Ames test, along with the high resolution,
also allows the detection of potential mutagenic/carcinogenic activity in
human body fluids, e.g., monitoring urines from occupationally exposed workers
(20).
The Ames test system is a highly sensitive yet simple bacterial assay
for chemical mutagens. Compounds (or mixtures) are tested with a group of
well-characterized mutants of Salmonella typhimurium requiring histidine for
growth. Simply stated, the assay detects genetic damage induced by chemicals
by the reversion of these specific strains to the wild-type or prototrophic
state (the ability to grow without histidine supplement). Additionally, the
particular type of genetic damage - base alterations, frame shifts (addition
or deletions) - can be detected with the appropriate strain.
Furthermore, since we now realize that many carcinogens and mutagens
require metabolic activation in order to reach their ultimate form, the test
can be modified to include an in vitn? activation of the chemical under test.
Here, homogenates from rat or human liver or other tissues can be applied
and the mutagenic activity of the activated form can be detected. The inclu-
sion of this metabolic activation has led to the detection of a wide array of
carcinogens as mutagens (21).
-282-
-------
The Salmonella strains and procedures used have been described in detail
by Ames (23). For the study of the feasibility of application of mutagenicity
testing to environmental effluents and crude products from the synthetic fuels
technology, we attempted to perform wide range screening with two of the highly
sensitive strains known to respond to a wide variety of known mutagens/carcino-
gens. The working hypothesis was that sensitive detection of potential muta-
gens in fractionated complex mixtures could be used to isolate and identify
the biohazard. In addition, the information could be helpful in establishing
priorities for further testing, either with other genetic assays or carcino-
genic assays. Finally, the procedures might show utility in monitoring plant
processes, effluents, or personnel early in the formation of the engineering
and environmental technology that will eventually evolve in the synthetic fuels
industry. The approach and preliminary results cited here show that the coupled
analytical-biological scheme is a feasible research mechanism and is applicable
to the ascertainment of potential human health hazard of a wide variety of en-
vironmental exposures, either occupationally or to the population in general.
Methodology
The Salmonella strains used in the various assays are listed below. All
strains were obtained through the courtesy of Dr. Bruce Ames, Berkeley, Cali-
fornia.
>•
Salmonella typhimurium Strains
TA 1535 hisG46, uvrB, rfa (missense)
TA 100 hisG46, uvrB, rfa (missense plus R factor)
TA 1537 hisC3076, urvB, rfa (frameshift)
TA 1538 hisD3052, uvrB, rfa (fr-ameshift)
TA 98 hisD3052, uvrB, rfa (frameshift plus R factor)
-283-
-------
In the screening of fractionated materials the two strains TA98 and
TA100 were generally employed. Standard experimental procedures have been
given by Ames, McCann, and Yamasaki (23). Briefly, the strain to be treated
with the potential mutagen(s) is added to soft agar containing a low level
of histidine and biotin along with varying amounts of the test substance.
The suspension containing approximately 2 x 108 bacteria is overlaid on mini-
mal agar plates. The bacteria undergo several divisions with the reduced level
of histidine, thus forming a light lawn of background growth on the plate and
allowing the mutagen to act. Revertants to the wild-type state appear as ob-
vious large colonies on the plate. The assay can be quantitated with respect
to dose (added amount) of mutagen and modified to include "on-the-plate" treat-
ment with the liver homogenate required to metabolically activate many compounds.
Fractions and/or control compounds to be tested were suspended in dimethyl-
sulfoxide (DMSO, supplied sterile, spectrophotometric grade from Schwarz-Mann)
to concentrations in the range of 10-20 mg solids. The potential mutagen was
in some cases assayed for general toxicity (bacterial survival) with strain
TA1537. Generally, the fraction was tested with the plate assay over at least
a 1000-fold concentration range with the two tester strains TA98 and TA100.
Revertant colonies were counted after 48 hours incubation. Data were recorded
and plotted versus added concentration and the approximate slope of the induc-
tion curve was determined. It is assumed that the slope of the linear dose-
response range reflects the-mutagenic activity. Positive or questionable re-
sults were clarified using a narrower range of concentrations. All studies
were carried out with parallel series of plates plus and minus the liver enzyme
preparation for metabolic activation. The background lawn of bacterial growth
was routinely examined so that any effects attributed to massive cell death and
subsequent growth of the few surviving bacteria (availability of more histidine)
-284-
-------
could be differentiated from mutation induction. Routine controls demonstra-
ing the sterility of samples, enzyme or S-9 preparations, and reagents were
performed. Positive controls with known mutagens were carried out in order
to recheck strain response and enzyme preparations.
Preliminary Results
To demonstrate the feasibility of the coupled analytical-biological ap-
proach, we derived primary fractions from a number of crude products and ef-
fluents from various fossil fuel technologies. For example, Table 5 lists
the mutagenicity testing results from three crude oils: Syncrude from the
COED process, shale oil, and a natural crude. The important neutral fraction
was subfractionated and the individual values are listed. Since in most cases
the original crude product was too toxic to test, the total represents the
summation of the assays of all fractions. We have assumed that the most accu-
rate measure of the total potential of the neutral fraction is the sum of the
tested subfractions.
An overview of the preliminary results points to a number of consistencies:
(a) all crudes showed some mutagenic potential, (b) the neutral and basic frac-
tions showed activities regardless of the source of the sample, and (c) the
relative total mutagenic potentials vary from 516 revertants/mg for Syncrude,
178 revertants/mg for shale oil, to 76 revertants/mg for the natural crude oil.
Whether these results reflect a comparative biohazard is not the point in ques-
tion here. The results simply show that biological testing - genetic reversion
assays in this case - can be carried out with the newly developed tester sys-
tems but only when coupled with the appropriate analytical separation schemes.
In addition to the obvious bias that could accompany the choice of samples
and their solubility, or the time and method of storage, a number of biological
-285-
-------
discrepancies can also enter into the determinations. For example, concomitant
bacterial toxicity can nullify any genetic damage assay that might be carried
out; the choice of inducer for the liver enzymes involved can be wrong for
selected compounds; the choice of strain could be inappropriate to, again,
selected compounds; and additionally, the applicability of the Salmonelle test
to other genetic assays and the validation of the apparent correlation between
mutagenicity and carcinogenicity remains a point of significant fundamental
research. Furthermore, the short term assays chronically show negative results
with, e.g., heavy metals. Similarly, compounds involved in or requiring co-
carcinogenic phenomena would presumably go undetected. However, in the context
of a prescreen to aid the investigators in ordering their priorities, the short-
term tests appear to be a valid testing approach to address the dilemma of the
magnitude of the number of hazardous compounds and complex mixtures that man
encounters in his environment.
Perhaps more appropriate to environmental effects, we have extended the
fractionation procedures and mutagenicity assay to the organics recovered from
a number of aqueous samples from various technologies. In parallel with the
assays listed above, we have investigated product water from the shale oil
process, the separator liquor from the COED process, and condensate from the
Synthane process. Again, the mutagenic potential of the various fractions can
be ascertained. The studies are not being extended to the isolation and iden-
tification of the active components. In correlation with the biological effort,
the chemical analyses of the materials and fractions are being extended. A con-
centrated effort designed to genetically assay the known or predicted constit-
uents of effluents from various synthetic fuel processes is also underway,
utilizing a number of biological systems in a "tier approach" to mutagenicity
testing.
-286-
-------
Comments
The single most important measurement "need" encountered in our chemi-
cal studies to date is the need to prioritize measurements research activi-
ties. The number of combinations of processes, sample types, and chemicals
available for study is staggering. Bioassay data when health effects are
of interest and environmental data when environmental impact is of interest
provide a rational basis for prioritizing research. If carcinogenesis is of
interest, results of the Ames test are accepted as indicators of carcinogene-
sis, and the additivity of mutagenic activities of chemical fractions are
accepted, bioassay results reported here suggest that the neutral and ether
soluble basic constituents of coal derived samples should receive priority
attention. The assumptions leading to this conclusion and the bioassay
methodologies require considerable basic research for validation but any
guidance at this point is invaluable.
A related problem (raised by Dr. James Dorsey, EPA/RTP following this
presentation) is that of "decision points" - quantifying bioassays responses,
at least in a relative sense, so that subsequent action is indicated. Such
action can range from prioritizing samples for study to dictating control
technology requirements. "Natural" materials of epidemiologically known health
effect, cigarette smoke for inhalation' exposure or soot for skin contact as
examples, might be used to establish baseline biological responses.
The large number of questions associated with the relationship between
bioassay systems, human health effects, and exposure to complex mixtures of
materials, preclude positive judgments of environmental and health impacts of
advanced fossil fuels processes. Judgments are made essentially impossible
by the absence of commercial scale conversion plants and uncertainties about
the relationship between aged untreated samples from pilot plant experiments
-287-
-------
and materials issued from commercial scale operations which will incorporate
environmental control methods. Studies carried out now must emphasize methods
development in preparation for process evaluation.
Measurements research needs identified from work referenced and reported
here include:
(a) Development of chemical class fractionation procedures for bio-
testing support which are free of artifact formation and routinely
applicable at preparatory (100+ gm) levels.
(b) Development of multicomponent quantitative analytical methods for
carcinogenic, co-carcinogenic, and tumor promoting fractions.
(c) Identification and quantisation of stereoisomers in the polynuclear
aromatic hydrocarbon subfractions of liquid products and aqueous
effluents.
(d) Identification of individual constituents or physico-chemical pa-
rameters (e.g., fluorescence) which can be monitored as indicators
of the polynuclear aromatic hydrocarbon, weak acid, and basic
fractions of liquid products and effluents. Development of moni-
toring instrumentation follows.
(e) Identification of constituents in aqueous leachates from spent
shale or solid residues from coal processing.
(f) Development of methods for the rapid, highly selective determination
of stereoisomers in environmental samples.
(g) Identification of individual constituents in airborne materials avail
able for occupational inhalation exposure.
-288-
-------
REFERENCES
1. The Nations Energy Future, A report to Richard M. Nixon, President of
the United States, Wash-1281, Dixy Lee Ray, December 1973.
2. Evaluation of Coal Conversion Process to Provide Clean Fuels, Electric
Power Research Institute (EPRI) Report 206-0-0. Parts I, II, III, February
1974. Prepared by Donald L. Katz, et.al., University of Michigan, College
of Engineering.
3. Synthetic Fuels Data Handbook, Cameron Engineers Inc., Denver Colorado,
Thomas A. Hendrickson, 1975.
4. Evaluation of Pollution Control in Fossil Fuel Conversion Processes.
Liquefaction: Section 2. SRC Process, National Technical Information
Service Report PB-241 792, U.S. Dept. of Commerce, C. E. Jahnig, March
1975.
5. Klass, Donald L., Synthetic Crude Oil From Shale and Coal, Chem tech,
499-510, August 1975.
6. Sexton, Richard J. The Hazards to Health in the Hydrogenation of Coal.
I. An Introductory Statement on General Information, Process Description,
and a Definition of the Problem, Arch Enviro Hlth. 1_, 181-186 (1960).
7. Weil, Carrol S., and Condra, Niomi L. The Hazards to Health in the Hydro-
genation of Coal. II. Carcinogenic Effect of Materials on the Skin of
Mice, Arch Enviro Hlth, 1, 187-193 (1960).
8. Ketcham, N. H. and Norton, R. W. The Hazards to Health in the Hydrogena-
tion of Coal. III. The Industrial Hygiene Studies, Arch Enviro Hlth, 1_,
194-207 (1960).
9. Sexton, Richard J. The Hazards to Health in the Hydrogenation of Coal.
IV. The Control Program and the Clinical Effects, Arch Enviro Hlth, 1_,
208-231 (1960).
10. Eckardt, Robert E. The Hydrogenation of Coal, Arch Enviro Hlth. 1, 232-
233 (1960).
11. Massey, Michael J., Nakles, David V., Forney, Albert J., and Haynes,
William P. Role of Gasifier Process Variables in Effluent and Product
" Gas Production in the Synthane Process, presented at the Environmental
Aspects of Fuel Conversion Technology II, Hollywood, Florida, December
15-18, (1975).
12. Preliminary Results: Chemical and Biological Examination of Coal-Derived
Materials, ORNL/NSF/EATC-18, Oak Ridge National Laboratory, Oak Ridge,
Tennessee, March 1976. W. D. Shults, ed.
-289-
-------
13. Ho, C.-h., Clark, B. R., and Guerin, M. R. Direct Analysis of Organic
Compounds in Aqueous By-Products from Fossil Fuel Conversion Processes:
Oil Shale Retorting, Synthane Coal Gasification and COED Coal Liquefac-
tion, Environmental Letters (in press).
14. Guerin, M. R., Griest, W. H., Ho, C.-h. Ho, Shults, W. D. Chemical Charac-
terization of Coal Convers-ion Pilot Plant Materials, presented at the
Third ERDA Environmental Protection Conference, Chicago, Illinois, Septem-
ber, 1975.
15. Kubota, H., Griest, W. H., and Guerin, M. R. Determination of Carcinogens
in Tobacco Smoke and Coal Derived Samples-Trace Polynuclear Aromatic Hydro-
carbons, pp 281-290, proceedings of the Ninth Annual Conference on Trace
Substances in Environmental Health, University of Missouri Press, Delbert
D. Hemphill, November, 1975.
16. Rubin, I. B., Guerin, M. R., Hardigree, A. D., and Epler, J. L. Fractiona-
tion of Synthetic Crude Oils from Coal for Biological Testing, Environmental
Research (in press).
17. Hecht, S. S., Bondinell, W. E., and Hoffmann, D. Chrysene and Methylchry-
senes - Presence in Tobacco Smoke and Carcinogenicity, J. Nat! Cancer Inst.
53, 1121 (1974).
18. Swain, A. P., Cooper, J. E., and Stedman, R. L. Large Scale Fractionation
of Cigarette Smoke Condensate for Chemical and Biologic Investigations,
Cancer Res. 29, 579-583 (1969).
19. Comittee 17, Environmental Mutagenic Hazards, Science, 187 (1975), 503-514.
20. Commoner, B., A. J. Vithayathil and J. I. Henry. Detection of Metabolic
Carcinogen Intermediates in Urine of Carcinogen-fed Rats by Means of
Bacterial Mutagenesis, Nature, 249, 850-852, (1974).
21. McCann, J., E. Choi, E. Yamasaki, and B. N. Ames. Detection of Carcinogens
as Mutagens in the Salmonella/microsome Test: Assay of 300 Chemicals, Part
I, Proc. Natl. Acad. Sci. USA, 72., 5135-5139 (1975).
22. Kier, L. D., E. Yamasaki and B. N. Ames. Detection of Mutagenic Activity
in Cigarette Smoke Condensates, Proc. Natl. Acad. Sci., USA, 71, 4159-4163
(1974). '
23. Ames, B. N., J. McCann, and E. Yamasaki. Methods for Detecting Carcinogens
and Mutagens with the Salmonel1 a/Mammalian-microsome Mutagenicity Test,
Mutation Res., 31_, 347-364 (1975).
24. Ames, B. N., H. 0. Kammen and E. Yamasaki. Hair Dyes are Mutagenic: Iden-
tification of a Variety of Mutagenic Ingredients, Proc. Natl. Acad. Sci.
USA, 72, 2423-2427 (1975). '
-290-
-------
Figure 1. 75-4852 Schematic Representation of the COED Coal Conversion
Process with Sampling Points Indicated
Figure 2. 75-12340 Capillary Column Gas Chromatographic Profiles of
Organic Constituents in Untreated Stack Gas and Headspace
Figure 3. 76-2806A Direct Injection Gas Chromatographic Profiles of Coal
and Shale Derived Waters
Figure 4. 75-7368R Gas Chromatographic Profiles of Polynuclear Aromatic
Hydrocarbon Isolates
Figure 5. 75-9451 Extractive Fractionation of COED Syncrude
-291-
-------
ORNL-DWG 75-4852
11
10
i
NJ
VO
NJ
I
LJ
STACK
COAL
GAS _ SULFUF
i — WATER — i
SCRUBBER
i \
ST/
NC
-
h-J_0
LIQUOR
\GE CH/
D.I
t
FLUIDIZING
GAS
PROt
RECO
RE M OVA
3UCT U'L riLTPATK
VERY / TILTRATK
©—•^4 '
LIQUOR
K T^K v^'/ 4
FINES FILTER
CAKE
VD PYROLYSIS
*"» ^ ^Trtrr^ ___ ^ nr?nni ITT
2-4 / CHAR
J A
STEAM
AND
OXYGEN
[ 1
- c,&^
1 (JAb
i
-\n , , .k, SSB* HYDRnTRFATING
A /^"N NH-t
-A C4J NH3
?
f5")
SYNCRUDE
COED Process.
-------
ORNL-DWG 75-12340
(a) STACK GAS FROM
COED PROCESS
ro
vo
Co
HEADSPACE VOLATILES (50°C)
FROM SYNTHOIL PRODUCT
180
150
120
90
TIME (min)
-------
ORNL-DWG. 76-2806A
COED LIQUEFACTION PROCESS
FIRST STAGE SCRUBBER WATER
20
SYNTHANE GASIFICATION PROCESS
UNTREATED CONDENSATE WATER
PRODUCT WATER FROM SIMULATED
IN SITU OIL SHALE RETORTING
100
60 40
RETENTION TIME . minutes
-294-
-------
ORNL-DWG 75-7368R
I
K>
^£>
Ul
(a) CONDENSED CIGARETTE SMOKE
4 3
(t>) COAL LIQUEFACTION PRODUCT
(C) PAH STANDARD
TIME (hr)
-------
ORNL-DWG 75-9451
Orgaal«
laetc
and
Neutral
1 N HC1
pB 11
Ether
Aq
pH 9
Org
Partition:
Ether (or MaCl.)
1 I MOB
Aqueoua
Ppc
0.91
Org
Neutral
Bexane
74.22
Strong
Acldl
HO
2 1.6Z
Plorlsll Colun
Hexan*/3enz ene
8/1
4.9Z
pH 6.1
Ether
Pot
0.2Z
pH 1.0
Ether
Org
Ppt
0.2Z
Strong
Acidl
HZ
Benzene/Ether
4.7Z
Hethaool
2.4Z
Org
Weak
Ac Ida
Et20
1.9Z
-2y&-
-------
Table 1
Organic Compounds Determined in By-Produce Waters
from Fossil Fuel Conversion Processes
Co-Chromatographic
Peak1 Identification
1 Acetic Acid
2 Propanoic Acid
3 n-Butanoic Acid
4 Acetamide
5 n-Pentanoic Acid
6 Propionamide
7 n-Hexanoic Acid
8 Butyramide
9 Phenol
10 n-Heptanoic Acid
11 o-Cresol
12 m & p-Cresols
13 n-Octanoic Acid
14 2,6-Dimethylphenol
15 o-Ethylphenol
16 2,5-Dimethylphenol
17 3,5-Dimethylphenol
18 2,3-Dimethylphenol
19 n-Nonanoic Acid
20 3,4-Dimethylphenol
21 n-Decanoic Acid
22 a-Napthol
23 3-Napthol
Concentration (yg/ml)
Oil Shale Synthane Coal COED Coal
Retorting Gasification Liquefaction
600
210
130
230
200
50
250
10
10
260
30
20
250
—
—
_.,
--
—
100
--
50
—
«• *—
620
60
20
—
10
—
20
—
2100
—
670
i
1800
--
40
30
250
230
30
—
100
—
10
30
600
90
40
--
30
--
30
—
2100
—
650
1800
--
30
30
220
240
30
—
900
—
--
__
1 Peak number in Figure 3
-297-
-------
Table 2
Estimation of Polynuclear Aromatic Hydrocarbons
in Coal and Shale Derived Samples
oo
Peak Co-Chromatographic
Number1 Identification
1 cis-and trans-Decahydronapthalene
2 1,4-Dihydronapthalene
3 Napthalene
4 2-Methylnapthalene
5 1-Methyl napthale,ne
6 Azalene
7 Biphenyl
8 2,6-Dimethylnapthalene
9 1,3 + 1,6-Dimethylnapthalene
10 Butylated hydroxytoluene
1,5 + 2,3-Dimethylnapthalene
1,2-Dimethylnapthalene
11 Acenapthalene
12 Acenapthene
13 Fluorene
14 9,10-Dihydroanthracene
15 9-Methylfluorene
16 9,10-Dihydrophenanthrene
17 Octanthrene
18 1-Methylfluorene
19 Phenanthrene + 1,3,6-Trimethylnapthalene1
20 Anthracene J
21 1-Phenylnapthalene
22 2-Methylanthracene
23 1-Methylphenanthrene
24 2-Phenylnapthalene
25 9-Methylanthracene
26 Fluoranthene
Products
micrograms per gram (ppm)
Crude Shale Oil Synthoil
440
1400
430
84
250
170
180
IR2
IR
345
1700
90
1900
1400
ND
ND
IR
380
micrograms
Shale Retort
Waters
per liter (ppb)
Synthane Condensatt
200
650
290
23
250
670
220
37
IR
220
ND3
ND
790
4600
2900
190
3400
5100
2000
1100
280
940
26
410
140
290
ND
330
3
68
980
620
IR
IR
ND
400
86
330
110
33
46
250
100
38
45
180
27
IR
160
260
210
130
35
18
160
1300
32
2
48
66
24
10
IR
21
2
ND
ND
ND
ND
ND
ND
ND
lumber of chromatographic peak in Figure 4
2IR = incomplete resolution
3ND = Not detected
-------
Table 2 (Cont'd)
i
N5
VO
Peak Co-Chromatographic
Number Identification
27 Pyrene
28 1,2-Benzofluorene
29 2,3-Benzofluorene
30 4-Methylpyrene
31 1-Methylpyrene
32 5,12-Dihydrotetracene
33 1,2-Benzanthracene
34 Chrysene + Triphenylene
35 2,3-Benzanthracene
36 7,12-Dimethylbenz(a)anthracene
37 1 ,3,5-Triphenylbenzene ,
38 1,2-+3,4-Benzopyrene '
39 Perylene
40 3-Mcthylcholanthrene -i- Unknown
41 1,2,5,6-+!,2,3,4-Dibenzanthracene
42 o-Phenylene pyrene
43 Picene
44 1,12-Benzoperylene
45 Anthanthracene
46 3,4,9,10-Dibenzopyrene + Coronene
Products
micrograms per
Crude Shale Oi
IR
22
13
36
IR
IR
20
31
ND
ND
ND
ND
170
53
140
70
IR
IR
180
22
20
ND
ND
ND
gram (ppm)
1 Synthoil
4300
IR
IR
620
270
130
IR
ND
200
380
1500
80
micrograms
Shale Retort
Waters
per liter (ppb)
Synthane Condenjat
120
10
5
61
37
5
5
ND
6
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
-------
Table 3 ORNL DWG. 76-5075
Fractionation of Coal Liquefaction Products
Fraction
Quantity
COED
Found, 9
x RSD, %
1. NaOHInsol.
2. WA,
3. WAE
4. SA,
5. SAE
6. SAW
7. BIa
«• Bib
9. BE
10. Bw
Neutrals
11. Hexane
12. Hexane/Benzene
13. Benzene/Ether
14. Methanol
Total Recovery
n
Sample wt., g
0.9
0.2
1.9
0.2
1.1
1.6
0.2
0.2
2.2
7.3
74.2
4.9
4.7
2.4
101.7
4
4.4-11.9
34
86
18
41
50
82
29
38
16
89
2
47
17
17
9
'o
Synthoil
14.1
2.0
6.5
0.1
1.9
1.7
3.6
0.3
1.6
0.6
27.6
7.4
21.7
11.1
100.2
1
15.2
-300-
-------
_ L, „ ORNL DWG. 76-7874
Table 4
Determination of n-Alkanes
in Shale Oil and Shale Oil By-Product Water
Concentration
Peak
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Compound shale Oil
(mg/g)
CnH24 7.55
^•12^26 6.65
C13H28 6.20
C14H30 4.80
C1SH32 3.30
C16H34 4.50
£17^36 8.60
C18H38 8.80
GI 9H4Q 5.60
^20^42 4.80
C21H44 5.70
C22H46 4.50
Co c nn
2 3 **4 8 J.UV/
C24H5o 3.60
C25H52 3.90
C26HS4 3.45
CH ^ "3^
2 7*^5 6 J.JJ
C28H58 2.90
C29H6o 3.20
C30H62 1-00
C31H64 0.20
C32H66 0.10
C33H68 0.65
C34H70 0.40
C3SH72 0.15
Shale Oil
By-Product Water
67
66
65
62
45
53
105
106
70
62
73
59
67
48
53
50
51
48
62
30
27
9
' 12
5
2
-301-
-------
Table 5
Mutagenicity Assays of Fractionated Crude 011s *
to
I
Fraction
NaOH Insoluble
Keak Acids, Insoluble
Weak Acids, Ether Soluble
Strong Acids, Insoluble
Strong Acids, Ether Soluble
Strong Acids, Water Soluble
Bases, Insoluble (a)
Bases, Insoluble (b)
Bases, Ether Soluble
Bases, Water Soluble
Neutral
2
Fractionated Neutrals
Hexane A
B
C
Hexane/Benzene A
B
C
Benzene/Ether A
B
C
Methanol A
B
C
MeC12
Subtotal (Neutrals)
Total
Syncrude (COED)
% of
Total
1.0
0.1
1.8
0.1
0.9
0.4
0.2
0.2
2.6
0.4
82.3
(it-Neutrals)
87.1
2.6
1.1
1.6
0.7
0.6
4.1
0.4
0.2
1.1
0.4
0.1
0.2
100.2
91.8
Fraction
rev/mg3
0
0
0
0
0
0
8300
0
1500
0
559
455
3100
760
2120
2400
0
0
200
160
1520
400
300
200
—
—
Total
rev/mg
—
__
-.
—
__
17
—
39
—
460
396
81
8
34
17
—
—
1
<1
18
2
<1
<1
559
516
% of
Total
1.0
0.1
1.2
0.1
0.3
0.6
0.2
0.3
7.1
0.3
86.7
(%-Neutrals)
58.7
2.1
1.3
4.4
1.9
1.4
12.4
2.2
1.3
15.1
0.5
0.9
102.2
97.9
Shale 011
Fraction
rev/mg
256
185
52
0
159
160
1377
800
952
223
112
40
625
750
238
340
320
65
142
253
179
684
263
—
—
Total
rev/mg
3
<1
1
<1
1
3
2
68
1
97
23
13
10
10
6
4
8
3
3
27
3
2
112
178
% of
Total
2.9
0.2
0.8
0.2
0.5
0.1
0.4
0.1
0.2
0.1
80.7
U-Neutrals)
82.0
2.0
0.6
3.6
0.7
0.5
3.5
0.3
0.2
2.4
0.4
0.2
96.4
86.2
Crude 011
Fraction
rev/mg
0
0
0
0
115
236
0
0
175
0
90
92
50
168
150
254
70
32
53
Not Tested
32
82
74
--
--
Total
rev/mg
. —
-
„ _
<1
<1
--
<1
73
75
1
1
5
2
<1
1
<1
—
1
<1
<1
90
76
-------
Table 5 (Cont'd)
Initial sample weights: Syncrude, 11.8862g; Shale Oil, 24.0424g; and Crude Oil, 22.3099g.
2
Neutral Fraction chromatographed: Syncrude, 9.7823g; Shale Oil, 10.1314g; and Crude Oil, 10.0130g.
7 8
rev/mg * revertants/milligram: number of histidine from Salmonella strain TA98 using plate assay employing 2X10
bacteria-per plate; values derived from assumed slope of induction curve extrapolated to milligram value. All assays carried
out in the presence of crude liver S-9 from rats induced with Aroclor 1254 (gift of Monsanto).
* Louisiana-Mississippi Sweet Crude. Not necessarily representative of all crude.
-------
ACKNOWLEDGEMENTS
The authors wish to acknowledge Dr. Jack Sharkey and his colleagues at
the Pittsburg Energy Research Center and Drs. Larry Jackson, Dick Poulson,
and their colleagues at the Laramie Energy Research Center for assistance
in acquiring samples and for technical suggestions! We also wish to acknowl-
edge A. A. Hardigree, C.-h. Ho, B. R. Clark, W. H. Griest, and I. B. Rubin
of this Laboratory for their experimental contributions to the work reported
here.
-304-
-------
NONPOINT SOURCE WATER EMISSIONS:
ENERGY AND INDUSTRY PROCESSES
Robert M. Statnick, Ph.D.
Environmental Protection Agency
Gordon T. Brookman
TRC - The Research Corporation
of Hex* England
To Be Presented May 19, 1976
at the
EPA Symposium
on
Fugitive Emission: Measurement and Control
Hartford, Connecticut
-305-
-------
NONPOINT SOURCE WATER EMISSIONS; ENERGY AND INDUSTRY PROCESSES
INTRODUCTION
Since the enactment of P.L. 92-500 (Federal Water Pollution Control Act
Amendments of 1972), the U. S. Environmental Protection Agency (EPA) has largely
directed its water pollution control program at point sources such as process
wastewaters discharged through pipes to receiving bodies of water. Most indus-
tries and many municipalities will meet the standards of the interim goal of 1977.
However, there are many areas in the United States where water quality has not
significantly improved even though point sources have been controlled. In such
areas nonpoint source water pollution has a major influence on water quality.
You may ask: what is nonpoint source water pollution? While point sources
are defined in P.L. 92-500, nonpoint sources are not defined. However, by in-
ference we can define nonpoint sources as the accumulated pollutants in a receiving
body of water from runoff due to snow melt and rain, seepage, and percolation con-
tributing to the degradation of the quality of surface and groundwaters. Some of
the characteristics of nonpoint sources are:
diffuse in nature
intermittent
site specific
not easily monitored at their exact source
related to uncontrollable climate events
not usually repetitive in nature from event
to event.
Figure l^1' shows a representation of the runoff cycle.
Since nonpoint source is defined by inference, everyone does not agree on a
common definition. Some experts define nonpoint source pollution as water acti-
vities not requiring an NPDES (National Pollution-Discharge Elimination System)
permit. This definition excludes one important class of intermittent, diffuse,
site specific sources. That is; raw material pile runoff, process spills re-
sulting in runoff, overflow from impoundments, and any other source caused by
rainfall such as solids accumulation, which in several cases have been regulated
by the NPDES permit system. This definition also includes a source which may be
intermittent, but is not caused by climate conditions, and can be monitored at
its source. That is; point sources (which may number in the thousands) which
have too low a flow to require an NPDES permit.
Based on these definitions, it is easy to see why the emphasis has been on
point source controls. Nonpoint sources by their nature are very difficult and
costly to
predict
monitor
control
-306-
-------
i
U)
o
Interception
Ijiiri;!!^!:. 'Vjtt^^
Surface detention = sheet of water
Infiltration
storage
'."•;.'•'•."•'•.'• Soil moisture ••.'•'•/•'•/•'• V- •/•'•
Perched
water table
Interflow
"-~-^r-'-£"-Z-r->>"r>--.---I Water table---------r- - T--_-_-_-
Impervious lens Groundwater^
flow
Stream
channel
Fig. 1. Simple representation of the runoff cycle.
-------
Now that we have defined nonpoint source water pollution, what are the major
sources? Figure 2 is a sketch of potential nonpoint sources which include run-
off from urban areas, agricultural and silvicultural activities, construction and
recreation areas, wetlands and industrial and mining sites, and salt water intrusion,
hydrographic modifications and unregistered point sources. Runoff from urban areas
and agricultural activities are known to be of prime concern since much quantifi-
cation has already been performed. This paper will present a brief overview of a
portion of a nonpoint source program performed by TRC for the Industrial Environ-
mental Research Laboratories of EPA. It will address industrial (including mining
and energy) processes and look at
potential nonpoint sources
measurement methodology
prediction methodology
Industrial Nonpoint Sources
In looking at potential industrial nonpoint sources, we have selected the
following industries:
energy generation
timber products
pulp and paper
iron and steel
inorganic chemicals
petroleum
fertilizer
mining (coal, mineral, ore)
cement
, feedlots
phosphate manufacturing,
nonferrous metals
Very little work (except acid mine drainage) has been performed in trying to
isolate and quantify nonpoint source pollution from industries. Therefore, for each
industry we have developed a matrix relating potential sources to categories of
pollutants. These pollutants include:
sediment (suspended and dissolved solids)
organic materials (including oil and grease)
metals
nutrients (nitrogen and phosphorous compounds)
heat
sulfates
acids
pesticides
Energy Generation
Nonpoint sources depend on the type of power generation, whether fossil
fueled (coal, oil, gas) or nuclear powered.
-308-
-------
fallout
CO
o
~——
Residential
Commercial
Sanitary landfills
Septic tanks
t'^o^K^VK \v4fK \M \\v4 A /\ <%A 1
V\N^^>^^\V^^^\\N\\\H^ ^4';tf
Precipitation <$fcVt/vtv
Sr%. _ xv^v\lMi%^'M'
<"•* v • \> i\- i -v -^ v ,,r «N?\K»/'^{
Precipitation vl&^t
^_ t^^^lS
^a=>~.. Vv^-rt ' ^c-.v^V-tV* *J'\^
'^^^ »*M^fe-S
L\\ \*."^^'^~J } s v j**^ i * § •• •
;t{\v- ?n A^*J Woodlands
> |-|i^Rfl (silviculture)
Construction "^l^
tiftaHlJ
Groves and orchards
Agricultural
Hydrographic ^.^
modifications "^i
v •_/_., ^J. '-&**\
rv ..; —^_--- __
»•'(
Salt water
intrusion
Fig. 2. Monpoint sources.
-------
Coal fired plants have runoff from coal storage piles and both coal and
oil fired plants may have runoff from fly ash storage piles. All fossil fuel
plants have the potential of deposited drift from cooling towers and deposited
blowdown from steam vents and fallout from fugitive air emissions. Nuclear
plants may have deposited drift from cooling towers and deposited blowdown
from steam vents. The nonpoint sources from energy generating plants will
likely contain sediments (suspended and dissolved solids), pesticides, sul-
fates, organic materials and acids.
Mining (Coal, Ore & Mineral) - Nonpoint sources include mine drainage,
spoil material drainage, runoff from storage piles, runoff from tailing piles,
and fallout from fugitive air emissions. Nonpoint source pollutants include
metals, organic materials, sediment (suspended and dissolved solids), sulfates
and acids.
Iron & Steel - Nonpoint sources include runoff from coal, limestone and
iron ore storage piles, process water runoff from ingot and pig casting and
process water spills and fugitive and point source air emission fallout.
Pollutants include metals, nutrients, acids, organic materials, sediment (sus-
pended and dissolved solids), and sulfates.
• Petroleum - Fugitive air emission and air point source emission fallout,
leaching from waste ultimate disposal sites, process spills and leaks are the
most probable nonpoint sources. These sources contribute organic materials
(including oil and grease), pesticides, sediment (suspended and dissolved
solids), sulfates, nutrients, and heat.
Timber Products - Erosion from roadways, and timber harvesting fugitive
air emission fallout from cutting and transportation and fertilization are
potential nonpoint sources. The pollutants emitted include organic materials,
acids, pesticides, sediment (suspended and dissolved solids), and heat.
Pulp & Paper - Nonpoint sources are runoff from coal storage piles, log
piles, chip piles and bark disposal piles. Sources also include fugitive and
point source air emission fallout and erosion from roadways. Organic materials.
acids, sediment (suspended and dissolved solids), and sulfates are the most
probable contaminants.
Inorganic Chemicals - Runoff from raw material and intermediate storage
piles, leaching and runoff from ultimate disposal areas of concentrated wastes
and accumulated materials from spills, and fugitive and points source air emis-
sion fallout are potential nonpoint sources. Pollutants include sediment
(dissolved and suspended solids), organic materials, metals, sulfates, acids
and heat.
Fertilizer - Nonpoint sources include runoff from storage piles, accumu-
lated materials from spills and leaks and fallout from fugitive and point
source air emissions. Organic materials, nutrients and sediment (suspended
and dissolved solids) are the primary pollutants.
Cement - Runoff from storage piles and fugitive and point source fallout
are the most probable nonpoint sources. Sediment (suspended and dissolved
solids) is the primary contaminant.
-310-
-------
Feedlots - Runoff from accumulated materials (waste and feed) Is the
major nonpoint source. Pollutants include organic materials, nutrients, and
sediment (suspended and dissolved solids) .
Phosphate Manufacturing - Nonpoint sources are runoff from material piles
and fallout from fugitive and point source air emissions. Sediment (suspended
and dissolved solids), nutrients and organic materials are principal pollutants.
Nonferrous Metals - Runoff from storage piles and accumulated materials
from process water spills, and fallout from fugitive and point source air emis-
sions are most probable nonpoint sources. Contaminants include metals, nutrients
and sediment (suspended and dissolved solids) .
While regulations now exist for controlling some of these sources, as
stated earlier, little has been done in trying to isolate and quantify industrial
nonpoint sources. Before regulations imposing controls are to be enforced, there
is a need for defining the potential problem. Therefore, there is a need for
collecting field data. We would like to discuss the major elements of a field
program for sampling industrial nonpoint sources.
Measurement Methodology
As stated earlier in the definition of nonpoint sources, they are not easily
monitored at their source. Most of the nonpoint work performed (urban and agri-
cultural) , has focused on monitoring changes in the receiving water body. Most
of these studies have tested large drainage basins and usually no effort has been
made to isolate individual sources. The receiving waters were monitored upstream
and downstream of the test areas in both dry and wet conditions. The difference
in parameters minus any point source contribution was the nonpoint source loading.
Since industries are usually in urban areas or within industrial parks, it
is much more difficult to isolate a particular -industry. Sampling the receiving
water body input and output and assuming the difference, minus point source, is
due only to the particular industry's nonpoint sources could result- in gross
errors. Therefore, some quantification and qualification of the industrial non-
point sources must be performed in parallel with receiving body monitoring.
~
„
of the common parameters monitored. In addition, it " rhemical The
choose a particular organic, such as phenol as an indicato r ^e^' ™
choice of a sampling technique either manua -^^J^ilfinTeJeloping
sample, either composite or discre te «« ££ " but can be an advantage
a
-311-
-------
TABLE 1
PARAMETERS COMMONLY MONITORED IN NONPOINT SOURCE PROGRAMS
Suspended Solids
Dissolved Solids
Turbidity
Dissolved Oxygen (DO)
pH
Total Organic Carbon (TOC)
Biochemical Oxygen Demand (BOD)
Chemical Oxygen Demand (COD)
Oil & Grease
Metals
Phosphate (Ortho, para)
Total Kjeldahl Nitrogen (TKN)
Ammonia Nitrogen
Sulfates
Pesticides
-312-
-------
advantage.
A discrete sample is collected over a certain time interval and is kept
separate from other samples. This type of sample allows comparison of runoff
water quality over a period of time. The problem with discrete samples is the
large number of samples to be collected and analyzed. A composite sample con-
sists of a series of smaller samples which are consolidated to form a larger
sample. A composite sample reduces the number of samples to be analyzed, but
it does not allow comparison of runoff from specific time intervals. A test
program should incorporate both types of samples, discrete samples being used
on the parameters of major interest and composite samples on the parameters
of minor interest, depending on the objectives of the test program. If inter-
est is in a single storm event then discrete sampling would be required fo?
adequate definition. If interest is in long-term or average storm conditions
then composite sampling is adequate.
Samples should be taken in both dry and wet weather conditions. The
sampling time is critical during storm events. Because of the "first flush"
effect of storms, sampling must begin at the first instance of rainfall and
discrete samples should be taken at intervals of 5-10 minutes. Composite
samples should be taken at intervals of 15-30 minutes. An automatic sampling
system which is electronically signalled to start by increases in water level
in a rain gauge is a must if the "first flush" is to be sampled. Figure 3^2) is
a flow chart of such a system.
Figure 4 shows a sketch of a hypothetical test area located along a river.
Once a test area has been selected, upstream and downstream river stations are
located during dry weather conditions for sampling and flow measurements.
These stations are sampled at various depths and distances across the river
during both dry and wet weather conditions. If desired, intermediate river
stations can be set up to isolate individual nonpoint sources. In addition,
if tributaries flow into the river within the test area, they too must be sam-
pled. River sampling techniques such as those used by Whipple , DiGiano ,
Randall5 and Colston have been used for several years.
During storm events, meteorological data including rainfall intensity and
duration must be collected.
Topographic plots are developed to divide the test area into drainage
basins. (See Figure 4.) Runoff sampling stations are located based on these
plots. The major problem with this type of test program is quantification and
qualification of runoff. Since we know runoff is diffuse, it is impossible to
collect all runoff; therefore total quantification will have to be estimated
from representative samples. Systems for collecting these samples are only in
an early state of development. In any event a test program of this proportion
is costly and since nonpoint sources are site specific and non-repetitive, it
is conceivable that it would have to be carried out on several sites and several
storm events for a particular industry. Since this is impractical from cost
and time standpoints, the use of mathematical models for prediction of nonpoint
source pollution as a replacement of a majority of the sampling could be more
efficient.
-313-
-------
Rain
i
UJ
Rain
gauge
Rain
^
r
-
High
intensity
by-pass
Flow
gauge
—
•>-
Rainfall
recorder
Flow
recorder
„>.
_J
Delay
circuit
i
i
Y
Sampler
activator
r*
__,
Refrigerated
sample
containers
>-
*-
-<
Elapsed
time
indicator
Sequential
sampler
*—
O
i
T
Indexer
distributor
Pump
ut I A In
— -j
Stream
intake
Stream flow
Water from stream or rainfall
Electrical signal
Transport to lab
Fig. 3. Flow chart - automatic receiving water
sampling during storms.
-------
Upstream
river -J£
sampling
River flow
i»j._s~— Intermediate sampling
>f* ^-—
Downstream
river
sampling
$H Runoff sampling locations selected based on
topography of drainage basin.
Fig. 4. Field program
-------
Prediction Methodology
Mathematical models properly applied provide a cost effective means of
quantifying impacts on water quality resulting from nonpoint source pollution
and of evaluating alternatives for the control of nonpoint sources of pollu-
tion. In recent years many mathematical models have been developed to simulate
the quantity and quality of stormwater runoff and the impact of such runoff on
the quality of natural water bodies. Each model, however, was developed to
satisfy a different need ranging from the design of municipal storm sewer sys-
tems to the assessment of land use as it influences flooding and water quality.
There does not yet exist a model developed specifically for industrial runoff
(except mining) although some models can be adapted. There are many criteria
that can be used when selecting a model. In general, the simplest model which
satisfies the project needs should be selected for use since such a model is
normally the most economical choice. Figure 5^7' serves to illustrate..one aspect
which contributes to model complexity - the choice of parameters to be modeled.
For instance, in a relative sense, it is more difficult to model toxicity
relationships than to model dissolved oxygen levels.
Once a model has been selected it must be adapted to the specific site or
area being studied. A model is so adapted through the processes of calibration
and verification. Calibration is achieved by adjusting the model to reflect
site.specific field data. After the model has been calibrated, it should be
tested against a second set of field data. If the second set of field data and
the modeled results compare favorably, the model is considered to be verified
and ready for application.
For a model to be adaptable to industrial applications it must be capable
of predicting the quantity and quality of stormwater runoff, the transport of
such runoff to a receiving body of water, and the impact of such runoff on the
quantity and quality of the receiving water. Pollutants of primary importance
for model simulation include sediment (suspended and dissolved solids), nutrients
(compounds of nitrogen and phosphorus), pesticides, acidty (pH and sulfuric acid),
organic material (biochemical oxygen demand, chemical oxygen demand, dissolved
oxygen), and heat (temperature). In addition, since storm events are dynamic,
a model must also be capable of simulating functions in a dynamic, i.e., time
dependent fashion.
To predict the quantity and quality of stormwater runoff a model must be
capable of simulating the effects of such items as the intensity and the dura-
tion of the storm event, infiltration and drainage characteristics, the accumu-
lation of pollutants between storms, and the washoff of such pollutants during
storms. For continuous simulation of multiple storms, a model must be capable
of simulating dry weather flows as well cis storm flows.
To predict the transport of stormwater runoff for industrial land use, a
model must be capable of simulating overland flow and routing in man-made sys-
tems (channels, sewers, etc*).
To describe the impact of the stormwater runoff on a receiving body of
water, a model must be capable of simulating the quantity and quality responses
of the receiving water to the runoff impact. Again, for continuous simulation
of multiple storms, a model must be capable of simulating dry weather flows as
well as storm flows. For increased flexibility a model should be capable of
-316-
-------
Erosion
Entrophication
Toxicity relationships
Algal growth » Metal transport
Nutrient and pesticide transport
Indicator bacteria e Sediment transport
DO • Temperature • Dissolved solids
Fig. 5. Relative difficulty of applied modeling.
-317-
-------
simulating various types of receiving waters including rivers, lakes and
estuaries.
We have reviewed the prominent runoff mathematical models and those which
can be adapted to industrial sites are listed in Table 2. The models were
evaluated for suitability, adaptability, complexity, cost and availability for
predicting the impact of industrial nonpoint sources on receiving body water
quality using the following criteria:
Wastewater (Runoff) - quantity, quality, dry weather flows,
storm runoff;
Receiving Water - quantity, quality, river, lake, estuary;
Quality Parameters - temperature, suspended solids, total
dissolved solids, biochemical oxygen demand (BODs),
chemical oxygen demand (COD), dissolved oxygen, nitrogen,
phosphorus, pH, oil and grease, pesticides;
Simulation of Single Storm
Simulation of Multiple Storms
Computer Program Availability - Public or proprietary;
Complexity - high, moderate, low;
Costs - high, moderate, low.
The last four models listed in Table 2 were eliminated because they can
not presently model the receiving water although STORM is currently being
adapted to include the receiving water. The WRE, HSP, Dorsch Consult, SWMM -
Release II and SSWMM - Receiv II models seem best suited for adapting to indus-
trial applications. However, WRE, HSP and Dorsch Consult models are proprietary.
Table 3 shows an evaluation of SWMM - Release II and SSWMM - Receiv II models.
Conclusions and Recommendations
As a result of this overview evaluation of nonpoint source water pollu-
tion from industrial activities, we have found the following:
1. Little or no quantification data for any of the 12
industries studied except acid mine drainage.
2. Nonpoint source measurement has been performed for
urban and agricultural activities with the monitoring
effort centered on the receiving water body.
3. Very little quantification and qualification of runoff
has been performed.
4. Several mathematical models are available for predict-
ing impact of runoff on receiving water bodies but none
have been applied to industrial activities except mining.
However, there are models which can be adapted to indus-
trial activities.
-318-
-------
TABLE 2
MATHEMATICAL MODELS ADAPTABLE TO INDUSTRIAL SITES
EPA Stormwater Management Model - Release II (SWMM)
Water Resources Engineers Stormwater Management Model (WRE)
Short Stormwater Management Model - Receiv II (Short SWMM)
Hydrocomp Simulation Program (HSP)
Dorsch Consult Hydrograph Volume Method
Corps of Engineers Storage, Treatment, Overflow and
Runoff Model (STORM)
Battelle Wastewater Management Model
Metcalf & Eddy Simplified Stormwater Management Model
Pyritic Systems: A Mathematical Model
-319-
-------
TABLE 3
MODEL COMPARISON
Wastewater
Quantity
Quality
Dry Weather Flow
Storm Runoff
Receiving Water
Quantity
Quality
River
Lake
Estuary
Quality Parameters
Temperature
Suspended Solids
Total Dissolved Solids
BODs
COD
Dissolved Oxygen
Nitrogen
Phosphorus
PH
Oil & Grease
Pesticides
Simulation of Single Storm
Simulation of Multiple Storms
Complexity
Cost
EPA SWMM
Release II
X
X
X
X
X
X
X
X
X
X
X
w
X
w
w
w
X
X
H
H
Short SWMM
Receiv II
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
M
M
KEY: X = Yes
W = Wastwater Only
H = High
M = Moderate
-320-
-------
Since many industries are faced with meeting regulations regarding non-
point source water pollution control, there is a need for determining the
extent of the problem so that cost-effective control measures can be taken.
The program performed by TRC for the Industrial Environmental Research
Laboratory of EPA will be extended to include:
1. The design and performance of a sampling program for
runoff and receiving waters for a particular industry.
2. The adaptation of one of the SWMM models to be used in
conjunction with the industry test program to model
the impact of runoff from an industrial site on a
receiving water body.
-321-
-------
REFERENCES
1. Donald M. Gray, editor in-chief, Handbook on the Principle of
Hydrology (Port Washington, NY: Water Information Center and
National Research Council of Canada, 1970)
2. M. P. Wanielista, Y. A. Yousef, and W. M. McLellan, "Transient
Water Quality Responses from Nonpoint Sources," paper presented
at the 48th Annual Pollution Control Federation Conference,
Miami Beach, Fla, October, 1975.
3. Interview with William Whipple, Jr., Director of the Water Resources
Research Institute of Rutgers University, New Brunswick, NJ 3/16/76.
4. Interview with Francis DiGiano, Associate Professor of Civil Engi-
neering, University of Massachusetts, Amherst, Massachusetts,
3/9/76.
5. Interview with Clifford Randall, Professor of Sanitary Engineering,
Virginia Polytechnic Institute and State University, Virginia, 3/8/76.
6. Colston - Characterization and Treatment of Urban Land Runoff, Newton
V. Colston, Jr., NERC, EPA Report # EPA-670/2-74-096. December, 1974.
7. W. G. Hines, et. al., Formulation and Use of Practical Models for
River Quality Assessment (U. S. Geological Survey Circular 715-
B[1975]), pg. B2.
-322-
-------
TECHNICAL REPORT DATA I
(f 'lease read Instructions on the reverse before completing) \
1.REPORTNO. 2
EPA-600/2-76-246
4. TITLE AND SUBTITLE
SYMPOSIUM ON FUGITIVE EMISSIONS MEASURE-
MENT AND CONTROL (May 1976, Hartford, CT)
7. AUTHOR(S)
E.M. Helming, Compiler
9. PERFORMING OR8ANIZATION NAME AND ADDRESS
TRC, The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
3. RECIPIENT'S ACCESSION NO. 1
S. REPORT DATE 1
September 1976
6. PERFORMING ORGANIZATION CODE 1
8. PERFORMING ORGANIZATION REPORT NO. 1
10. PROGRAM ELEMENT NO. I
1AB015; ROAP 21AUY-095
11. CONTRACT/GRANT NOV 1
68-02-2110
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings; 6/75-6/76
14. SPONSORING AGENCY CODE
EPA-ORD
is. SUPPLEMENTARY NOTES IERL_RTP project officer for this report is R. M. Statnick, Mail 1
Drop 62, 919/549-8411 Ext 2557.
16. ABSTRACT
The proceedings are a compilation of technical papers presented at a symposium on
the measurement and control of fugitive emissions (or non-point sources). They
discuss techniques which have been used to measure fugitive emissions, as well as
systems which have been used to control the emissions.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
Pollution
Measurement
Emission
Effluents
Sources
18. DISTRIBUTION STATEMENT
Unlimited
EPA Form 2220-1 (9-73)
Pollution Control
Stationary Sources
Fugitive Emissions
Non-Point Sources
19. SECURITY CLASS (ThisReport)
Unclassified
13B
14B
21. NO. OF PAGES
327
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
-323-
------- |