&EFK
United States
Environmental Protection
Agency
Industrial Environmental Research EPA-600/7-79-182
Laboratory August 1979
Research Triangle Park NC 27711
Third Symposium on
Fugitive Emissions
Measurement and Control
(October 1978,
San Francisco, CA)
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-79-182
August 1979
Third Symposium on Fugitive
Emissions Measurement and Control
(October 1978, San Francisco, CA)
J. King, Compiler
TRC - The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
Contract No. 68-02-2615
Task No. 201
Program Element No. INE623
EPA Project Officer: D. Bruce Harris
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
The technical papers included in this volume were prepared for presentation
at the "Third Symposium on Fugitive Emissions: Measurement and Control,"
held in San Francisco, California on October 23-25, 1978.
The objective of the Symposium was to provide a forum for the exchange of
information among concerned representatives of industrial, research and
governmental organizations relative to recent developments in industrial
and energy-related fugitive emissions measurement and control. The
Symposium was sponsored by the Environmental Protection Agency's Industrial
Environmental Research Laboratory at Research Triangle Park, North
Carolina as part of its continuing effort to develop methods for the
measurement and control of airborne and waterborne fugitive emissions.
D. Bruce Harris of the Industrial Environmental Research Laboratory was
the Project Officer and General Chairman of the Symposium.
Joanne King of TRC - The Research Corporation of New England was the
Symposium Coordinator and Compiler of these Proceedings.
-------
TABLE OF CONTENTS
(Speakers' names appeal' in italics.)
October 23. 1978 Page
SESSION I
Session Chairman: D. Bruoe Harris, EPA/IERL-RTP
OVERVIEW OF THE FUGITIVE EMISSION PROBLEM - 1979 SIP REVISIONS 3
David R. Dunbar, Chief, Policy Development, EPA/OAQPS
REGULATION OF FUGITIVE EMISSIONS UNDER THE CLEAN AIR ACT'S
PREVENTION OF SIGNIFICANT DETERIORATION (PSD) AND NON-
ATTAINMENT REQUIREMENTS 13
Bradley I. Raffle, Environmental Counsel, TRC
THE IMPACT OF FUGITIVE EMISSIONS ON TSP CONCENTRATIONS IN
URBAN AREAS 37
Ronald G. Draftz, IIT Research Institute
SESSION II
Session Chairman: Dennis C. Drehmel, EPA/IERL-RTP
EMISSIONS AND EFFLUENTS FROM COAL STORAGE PILES . -. (39J
T. R. Blaokwood and R. A. Wachter, Monsanto Research Corporation -^
URANIUM MILL TAILINGS AREA FUGITIVE EMISSIONS 51
Peter Piersol and P. G. Complin, James F. MacLaren Limited
IDENTIFYING SOURCES AND QUANTITIES OF FUGITIVE EMISSIONS IN
BALTIMORE 73
Robert C, Koah and John T. Schakenbach, GEOMET, Inc., and
Kathryn G. Severin, IIT Research Institute
FUGITIVE EMISSIONS FROM THE BY-PRODUCT COKE OVEN PUSHING OPERATION ... 91
Robert B. Jaoko, Purdue University
ESTIMATING DUST PRODUCTION FROM SURFACE MINING 103
Karl F. teller, Bureau of Land Management, Douglas G. Fox,
USDA Forest Service, and William E. Marlatt, Colorado State
University
-------
TABLE OF CONTENTS (Continued)
October 24. 1978 Page
SESSION III
Session Chairman: Lyle D. Randen, AMAX Coal Company
COMPARISON OF PREDICTED AND OBSERVED EFFECTS OF FUGITIVE DUST
FROM COAL OPERATIONS 121
Alexis W. Lemmon, 3v.3 Ronald Clark, and Duane A. Tolle,
BATTELLE Columbus Laboratories
INDUSTRIAL NON-POINT SOURCES: ASSESSMENT OF SURFACE RUNOFF FROM
THE IRON AND STEEL INDUSTRY 137
Gordon T. Brooknan3 John A. Ripp, and Bradford C. Middlesworth,
TRC
DIFFERENTIAL TRACING OF OILY WASTE AND THE ASSOCIATED WATER MASS
BY TAGGING WITH RARE EARTHS 177
D. L. MoCown, Argonne National Laboratory
ESTIMATION OF FUGITIVE HYDROCARBON EMISSIONS FROM AN OIL REFINERY
BY INVERSE MODELING 201
Andrew A. Huang and Sara J. Head, AeroVironment Inc.
MEASUREMENT OF FUGITIVE EMISSIONS FROM PETROCHEMICAL PLANTS 229
D. R. Tierney, I. S. Khan, and T. W. Hughes, Monsanto Research
Corporation
SESSION IV
Session Chairman: James A. Dorsey, EPA/IERL-RTP
DEVELOPMENT OF COATINGS TO REDUCE FUGITIVE EMISSIONS FROM COAL ^^
STOCKPILES 247
/?. S. Valentine, R. V. Kromrey, R. Naismith, and R. S. Scheffee,
Atlantic Research Corporation
NEW CONCEPTS FOR CONTROL OF FUGITIVE DUST 271 '
Dennis C. Drehmel, EPA/IERL-RTP, and Thomas Blackwood, Monsanto
Research Corporation, and S. Calvert, R. G. Patterson, and S. C.
Yung, Air Pollution Technology, Inc.
BEST AVAILABLE CONTROL TECHNOLOGY (BACT) FOR FUGITIVE EMISSIONS
CONTROL IN THE STEEL INDUSTRY 281
Arthur G. Nioola, Pennsylvania Engineering Corporation
CONTROL OF FUGITIVE EMISSIONS AT REVERBERATORY FURNACES AND
CONVERTERS 313
L. V. yerino, R. T. Price, and T. K. Corwln, PEDCo Environmental,
Inc., and Alfred B. Craig, Jr., EPA/IERL-CI
iv
-------
TABLE OF CONTENTS (Continued)
Page
CORE ROOM EMISSIONS IN FOUNDRIES 363
William D. Soott* Southern Research Institute and Robert C.
Cummisford, Krause Milling Company
October 25, 1978
SESSION V
Session Chairman: Henry J. Kolnaberg3 TRC
FUGITIVE EMISSIONS PROBLEMS AND CONTROL AT A SURFACE COAL MINE 383
Lyle D. Randen, AMAX Coal Company
AIRCRAFT TURBINE ENGINE PARTICULATE EMISSION CHARACTERIZATION., ..... 391
John D. Stookham and Erdmann H. Luebcke, IIT Research Institute,
and Larry Taubenkibel, DOT-Federal Aviation Administration
MEASUREMENT OF FUGITIVE DUST EMISSIONS FROM HAUL ROADS 415v
Chatten Cowherd* Jv.f Midwest Research Institute
DEVELOPMENT OF MEASUREMENT METHODOLOGY FOR EVALUATING FUGITIVE
PARTICULATE EMISSIONS 431
Edward E. uthe, Charles E. Lapple, Clyde L. Witham, and Robert
L. Mancuso, SRI International
DEVELOPMENT OF A FUGITIVE ASSESSMENT SAMPLING TRAIN FOR PARTICULATE
AND ORGANIC EMISSIONS 443
Roland L. Severance and Henry J. Kolnsberg, TRC
-------
OCTOBER 23. 1978
Monday Morning - SESSION I
Session Chairman: D. Bruce Harris
Sanitary Engineer
EPA/IERL-RTP
Monday Afternoon - SESSION II
Session Chairman: Dennis C. Drehmel
Research Chemical Engineer
EPA/IERL-RTP
-1-
-------
OVERVIEW OF THE FUGITIVE EMISSION PROBLEM -
1979 SIP REVISIONS
By
DAVID R. DUNBAR
OFFICE OF AIR QUALITY PLANNING AND STANDARDS
U. S. ENVIRONMENTAL PROTECTION AGENCY
Presented at
Third Symposium on Fugitive Emissions
Measurement and Control
October 23 - 25, 1978
San Francisco, California
-3-
-------
ABSTRACT
Fugitive emissions and fugitive dust and their associated impacts
will be a major consideration in the development of the 1979 State Im-
plementation Plans for those areas which are currently not attaining the
National Ambient Air Quality Standards. Fugitive emissions and fugitive
dust will also play an important role in obtaining new source permits both
under PSD and new source review in nonattainment areas. Specific guidance
will be presented on how fugitive emission/dust control will be factored
into the State plans for existing, as well as new, sources considering the
possible revision to the particulate matter standard.
-4-
-------
This paper summarizes some of the more important requirements of the
new Clean Air Act of 1977, especially those which relate to the fugitive
particulate matter problem—provisions that are already affecting, or
likely to affect, everyone in the near future whether they represent a
control agency, industry, or the general public.
The past seven years have served as a major testing ground for the
fundamental goals of the Act. Seven years after passage of the Act, vio-
lations of the National Ambient Air Quality Standards for one or more pol-
lutants still exist in more than half of the 247 air quality control regions.
As a result of the recent efforts required by Section 107 of the Act, to
designate areas as meeting, not meeting, or unclassified with respect to
the NAAQS, it was found that of the 3,215 counties in the United States 408
are listed as nonattainment for TSP. (Figure 1)
These standards were to have been attained by 1975; however, substan-
tial progress has been made. For example, the estimated number of people
exposed to total suspended particulates in excess of the annual primary
standard decreased from 74 million in 1970 to less than 50 million in 1975;
in other words, the exposed population has been reduced by more than one-
third since the passage of the Clean Air Act of 1970.
Despite the progress that has been made in reducing nationwide emissions
of air pollutants, failure to achieve the National Ambient Air Quality
Standards in many areas of the country represents the seven-year reality of
the implementation of the 1970 Act. The Congress began to address the
reality of this situation more than three years ago and initiated the
drafting of the first comprehensive amendments to the Act since 1970.
During the course of these Intense deliberations in Congress, the Agency
merely tried to maintain an even keel while we were forced to balance the
realities of the non-attainment situation against the demands of the
nation's emerging energy problems.
Finally, with the recent enactment of the Clean Air Act Amendments of
1977, the Congress has reaffirmed its commitment to the goals originally
established by the 1970 legislation. The new Amendments provide a strong
mandate and make clear that Congress does not intend that the nation's
energy problem be allowed to compromise environmental quality. The Act
strongly reaffirms our goals of attainment and the prevention of signifi-
cant air quality deterioration.
-5-
-------
Figure 1.
PRIflflRY NflROS
NOT RTTHINEO
IN tITIM COUN1T
IN PMT OF COUM1T
SECONORRY NflROS
NOT flTTRINEO
IN CUT me COUNTY
IK mi Of COUNtT
PROMULGRTEO LIST OF TSP NON-RTTRINING COUNTIES - 1978
-------
To a large extent, the hardest phase of our efforts to meet ambient
air quality standards is yet to come. The previously mentioned progress
has been achieved through the application of controls to sources generally
amenable to control requirements. During this next round of control
efforts, agencies will be forced to concentrate on the less conventional
sources (fugitive dust, fugitive emissions) with which we have far less
control experience. In other words, we have done the so-called easy things,
and what is left are the more difficult, less tried methods of air pollu-
tion control. This is particularly true with respect to the attainment of
the TSP standard, as it must be attained by 1982. Conventional stationary
sources have generally been controlled; however, in many cases, fugitive
emissions constitute a large percentage of total emissions. A recent EPA
study on National Assessment of TSP1 indicates fugitive emissions add 15 to
25 ug/m3 to citywide average TSP levels. However, significantly greater
impacts in vicinity of industrial sources were noted. Additionally, as
part of this study, the contractor visited a number of industrial sites
within 14 cities. Of all industrial sites, 32% were greatly impacted by
fugitive emissions. Of industrial sites with annual means twice national
standard, all were influenced by fugitive emissions. To some extent, the
new Act recognizes these difficulties and provides EPA with tools which
will enable the Agency to establish new regulatory requirements and new
approaches to difficult problems but also provides the flexibility to re-
solve isolated conflicts between environmental goals and energy, economic,
and employment concerns.
The heart of the new Amendments focuses on the problem of attainment
in areas where the National Ambient Air Quality Standards continue to be
violated. Each State that includes a nonattainment area must submit a plan
by January, 1979, for EPA approval by July, 1979, which provides for attain-
ment of the standards as expeditiously as practicable. Congress took a
firm position on the issue of attainment. The new State plan requirements
will necessitate in most cases the application of reasonable control meas-
ures and strict enforcement of these requirements for existing sources.
The plan will have to provide sufficient reductions in emissions to allow
room for growth, otherwise construction permits for new sources can be
granted only when more than offsetting emission reductions are secured on a
case-by-case basis prior to the facility start-up date.
Control programs must be developed for each area which has been de-
signated as nonattainment. These control programs must indicate how much
control will be needed.
In developing each control program, the control agencies will be con-
sidering various alternatives for attaining the standards. Once these
control alternatives have been considered, they should be screened to
determine what limitations may exist regarding each alternative. However,
it should be noted that in most cases, because of the magnitude of the
problem, it will take all that one can think of to attain the standards.
-7-
-------
Emission reductions for each alternative should be developed to deter-
mine the effectiveness of each program. For photochemical oxidants, EPA
has developed control techniques guideline (CTG) documents to provide some
information on the percent of reduction, etc., available for various hydro-
carbon control measures. However, no specific CTG's have been developed
for TSP. The Agency, however, has published various documents listed in
Table 1 which will be useful in determining the percent control for various
particulate matter control measures. However, the most important factor to
consider in developing a control strategy is that it be technically feasi-
ble.
The given strategy must provide for both attainment of the primary and
secondard standards for TSP. The primary standard must be attained as
expeditiously as practicable but no later than 1982 and the secondary
standard must also be met by 1982. However, if more than RACT is needed to
attain the secondary standard, a reasonable time is permitted.
Once the strategy has been selected and it provides for both attain-
ment and maintenance of the NAAQS, the strategy must be adopted in the form
of legally enforceable procedures and submitted as a revision to the State
Implementation Plan with the Governor's approval. This constitutes a com-
mitment on the part of the State to implement and enforce the plan.
Where adoption by 1979 is not possible (nontraditional controls for
particulate matter), a schedule for expeditious development, adoption, sub-
mittal, and implementation of these revisions would be acceptable. These
schedules must provide for implementation as expeditiously as practicable.
Prior to attainment, these measures must be implemented rapidly enough to
provide for emission reductions necessary to maintain reasonable further
progress as required by the action. Schedules would be a part of the
applicable SIP and would represent a commitment on the part of the State to
meet the key milestones set forth in the schedule.
Each plan must provide estimates of emission reductions for each
adopted or scheduled control measure or for related groups of measures
where estimates of individual measures are impractical. The Agency re-
cognizes that estimates may change as measures are more fully analyzed and
implemented. As estimates change, these should be revised and included in
subsequent plan submissions to ensure the plan remains adequate.
Thus, the 1979 SIP submission for TSP must contain a demonstration of
attainment of the primary standard as expeditiously as practicable but no
later than 1982, and emission limitations or regulations for the control of
traditional, as well as industrial, process fugitive particulate emissions.
The plan may also contain schedules for the development, adoption, and
submittal of controls of nontraditional sources. These may include demon-
stration studies, etc., prior to adoption for many nontraditional measures.
More information on demonstration studies can be found in "Guidelines for
Development of Control Strategy in Areas with Fugitive Dust Problems,"
OAQPS 1.2-071, pp 6-1 to 6-10.
-8-
-------
TABLE 1.
INFORMATION FOR DETERMINING
PERCENT CONTROL FOR TSP
TSP
— CONTROLLED AND UNCONTROLLED EMISSION RATES AND
APPLICABLE LIMITATIONS FOR EIGHTY PROCESSES, SEPTEMBER 1976.
- EXTERNAL COMBUSTION
- SOLID WASTE DISPOSAL
- FOOD AND AGRICULTURE
- METALLURGICAL
- MINERAL PRODUCTS
- WOOD PROCESSING
GUIDELINE FOR DEVELOPMENT OF CONTROL STRATEGIES IN AREAS
WITH FUGITIVE DUST PROBLEMS, OAQPS #1.2-071, OCTOBER 1977
- UNPAVED ROADS
- ENTRAINED STREET DUST
- CONSTRUCTION AND DEMOLITION
- AGRICULTURE
- TAILING PILES
- PARKING LOTS
TECHNICAL GUIDANCE FOR CONTROL OF INDUSTRIAL PROCESS
FUGITIVE PARTICULATE EMISSIONS, MARCH 1977, EPA 450/3-77-010
- COMMON DUST SOURCES
- IRON AND STEEL PRODUCTION
- PRIMARY NON-FERROUS SMELTING
- SECONDARY NON-FERROUS SMELTING
- FOUNDRIES
- MATERIALS EXTRACTION AND BENEFICIATION
- GRAIN ELEVATORS
- PORTLAND CEMENT
- LIME
- CONCRETE BATCHING
- ASPHALT CONCRETE PRODUCTION
- LUMBER AND FURNITURE
-9-
-------
Impact of Potential Revision to Partlculate Matter NAAQS
The Clean Air Act and EPA policy require a periodic reevaluation of
each NAAQS. Under requirements, a program has begun to revise the air
quality criteria for particulate matter. It is anticipated that a draft
revised document will be available for external review in late 1979. Any
revisions to the NAAQS for particulate matter that result from revision to
the criteria document would probably be proposed in mid-1980 and promul-
gated in late 1980.
Contrary to the position taken by some, the Agency contends that the
current TSP standards are reasonable in light of available evidence. The
standards have been under essentially constant review over the past several
years, and the Agency believes this position is widely supported by the
scientific community. This is not to suggest that the TSP standards could
not be improved upon. Many reviews of the standards have expressed concern
for the use of TSP as an index to health effects and indicate a need for
better characterization of particulate matter by size and chemical composi-
tion. The Agency shares the view that any revision of the particulate
matter criteria document should consider the possibility of defining a more
precise index of particulate matter pollution.
It is possible that the revised criteria document will result in some
form of an "inhalable" particulate matter primary standard based on par-
ticle sizes of less than 15 micrometers.
Of course, until the criteria document is revised, any estimate of the
effect is only speculation and therefore should not be allowed to disrupt
the current efforts to develop SIP revisions in areas designated nonattain-
ment for particulate matter. The SIP revisions must still be submitted by
January 1, 1979, must demonstrate attainment of the current particulate
matter standard by December 31, 1982, and must include emission regulations
for conventional sources and programs to subsequently develop controls for
unconventional sources as required.
If the reevaluation of the criteria document results in the kind of
change to the NAAQS discussed above, however, some controls that would be
necessary to attain the current NAAQS may not be necessary to attain a
revised primary NAAQS. Generally, the sources in this category are those
such as storage piles and materials handling operations that emit rela-
tively large particles.
The 1979 SIP revisions must contain or provide for the development of
all measures necessary to attain the current particulate matter standard.
However, it is legally permissible for States which adopt new regulations
covering sources of predominantly large particles to recognize the possi-
bility of a revision of the current standard in establishing compliance
schedules for such sources. As long as compliance is required not later
than 1982, such schedules may include dates which are late enough so that
the uncertainty over the particulate matter standard can be resolved prior
to significant expenditures for control.
-10-
-------
The possible revision of the particulate standard should not be a
factor which is considered in setting compliance schedule for sources other
than those described above. Nor can this possibility affect any existing
compliance schedules.
New Source Review and Fugitive Dust
•
The Agency has set forth a position regarding the impact of industrial
process fugitive emissions and fugitive dust emissions on new source review
in general in the August 16, 1978, Fugitive Dust Policy, and for PSD in
particular in the June 19, 1978, final Regulations for the Prevention of
Significant Deterioration.
Briefly, the Fugitive Dust Policy indicates that for new source review
in nonattainment urban areas, the current program which includes the emis-
sion offset concept remains unchanged. However, since fugitive dust is re-
cognized as a significant air pollution problem in urban areas, it is
appropriate to allow sources to minimize either existing fugitive dust
sources or particulate stack emissions in order to satisfy their emission
requirements.
There has been considerable concern about the location of major new
stationary sources in rural areas where fugitive dust has been determined
to be the major source. New sources that wish to construct in rural areas
with infrequent short-term violations of the TSP standard should be allowed
to construct without the need of an emission offset, as long as they comply
with the appropriate emission regulation (NSPS, state regulation or BACT
for PSD source) and when considering their emissions, plus "non-urban"
background and the emissions from other stationary sources in the vicinity
of the proposed location, they do not cause violations of the NAAQS or
appropriate PSD increments.
The June 19, 1978, regulations for PSD set forth a two-tier review
system. Under the second tier, a source must apply best available control
technology. In submitting a permit application with regard to BACT, the
source must propose BACT, set forth alternative systems, and then defend
the BACT selected. BACT can include several technology options. It must
cover all emission points, including stack emissions, fugitive process and
fugitive dust emissions. It must be at least as stringent as NSPS, NESHAP,
and can include design, equipment or operating standards.
Additionally, under the second-tier review, the source must conduct an
ambient impact review against both the NAAQS and increments. However, the
June 19, 1978, regulations for PSD contained several exemptions regarding
ambient impact review. One of those exemptions relates to fugitive dust.
Several comments were received regarding the proposed PSD regulations as
they related to fugitive dust emissions. As a result of the comments, the
Agency will exclude from any air quality impact assessment of a source or
modification any fugitive dust emissions. Additional support for this
exclusion can be found in the legislative listing. It points to the utili-
zation of "administrative good sense" regarding the treatment of fugitive
dust.
-11-
-------
Certain aspects of this exclusion for fugitive dust should be noted.
First the burden of showing to what extent emissions from the proposed
source or modification would be made up of fugitive dust rests with the
applicant. Second, the regulations do not exclude fugitive dust from the
determination of potential emissions. Any source or modification which,
taking into account emissions of fugitive dust, would have potential emis-
sions equal to or greater than 250 tons per year would be subject to the
applicable PSD requirements, especially in many instances the BACT require-
ment. Finally, EPA will treat emissions of fugitive dust as not consuming
increment for the purpose of evaluating other sources under PSD.
It should be emphasized that EPA intends to implement the above policy
of excluding the fugitive dust only on an interim basis. EPA will reassess
the implications of the policy and any possible technical improvements in
modeling fugitive dust, and will adjust the policy as appropriate.
The Agency realizes the difficulty of the task that lies ahead to
develop adequate SIPs by January, 1979, and to fully implement those SIPs
by 1982. The Agency, in developing the requirements for an acceptable SIP,
has tried to develop reasonable and achievable goals that still meet the
intent of Congress. We have tried to develop requirements that overcome
barriers and make it as easy as possible to develop SIPs that will meet the
intent of Congress. Between now and January many tough decisions are going
to have to be made.
The challenge we are faced with is to achieve the cleanest air possible
while maintaining the high standard of living the citizens of America have
attained. Fugitive emissions control must begin now—we are beginning to
more fully appreciate the problem, and fugitive emissions control must be
considered in the current round of SIP revisions. If not, we are only
dealing with part of the air pollution control program, and we will fall
short of our goal of attaining the NAAQS as expeditiously as practicable.
REFERENCES
1. National Assessment of the Urban Particulate Matter Problem,
EPA 450/3-76-025, July, 1976.
-12-
-------
REGULATION OF FUGITIVE EMISSIONS
UNDER THE CLEAN AIR ACT'S
PREVENTION OF SIGNIFICANT DETERIORATION
(PSD) AND NONATTAINMENT REQUIREMENTS
BY: BRADLEY I. RAFFLE, ESQ.
ENVIRONMENTAL COUNSEL
TRC - THE RESEARCH CORPORATION OF NEW ENGLAND
125 SILAS DEANE HIGHWAY
WETHERSFIELD, CONNECTICUT 06109
-13-
-------
ABSTRACT
The Clean Air Act Amendments of 1977 have forced EPA and the States to
come to grips with the fugitive emissions problem. Clearly, two of the most
significant forcing functions are the Act's new requirements for preventing
significant deterioration (PSD) and the new provisions relating to non-
attainment areas. This paper focuses upon the control strategies for
fugitive dust and fugitive hydrocarbon emissions which EPA is using to
implement the PSD and nonattainment requirements of the Act for total sus-
pended particulate and photochemical oxidants/ozone. These strategies focus
upon two basic areas:
1) Control of new fugitive emission sources under EPA's PSD
preconstruction review requirements and/or the provisions
of EPA's offset policy.
2) Control of new and existing fugitive emission sources
under revised State implementation plans called for by
the Act's PSD and nonattainment provisions.
A separate issue concerns the extent to which ambient particulate con-
centrations attributable to fugitive dust should be considered in the ambient
air quality analyses required under the PSD and nonattainment procedures for
proposed new and modified stationary sources. In areas subject to PSD require-
ments, the principal difficulty lies in distinguishing between natural and man-
induced fugitive dust. In nonattainment areas, distinguishing between rural
and urban fugitive dust problems is a major problem issue.
This paper explores the pertinent statutory and regulatory programs under
which EPA is attempting to deal with these problems.
-14-
-------
1.0 INTRODUCTION
Of the many regulatory consequences of the 1977 Clean Air Act Amendments,
one of the most significant Is the Increased attention they have generated on
the fugitive emissions problem. Historically, EPA's implementation of the
Clean Air Act has focused on emissions from new automobiles and industrial
installations. With respect to non-mobile sources of particulate matter,
regulations have been directed primarily at limiting the opacity and mass
loading of smokestack emissions from incinerators, fossil fuel-fired boilers
and major process industries. Non-stack particulate controls have traditional"
ly been applied only to open burning and mineral extraction operations.
Recent studies indicate that traditional stack-oriented control strate-
gies for particulate matter are insufficient by themselves to achieve the
ambitious air quality goals of the Clean Air Act. Control of industrial
fugitive emissions and nontraditional sources of fugitive dust will therefore
be an essential ingredient in future control strategies for particulate
matter. Control of new, and possibly existing, fugitive dust sources will
also be required in order to comply with EPA's new regulations on Prevention
of Significant Deterioration (PSD). This paper attempts to summarize the
federal regulations and policies applicable to fugitive dust sources under
the recently amended Clean Air Act. Pertinent provisions of EPA's regulations
and guidelines for nonattainment areas and PSD areas will be discussed within
the context of the EPA Fugitive Dust Policy.
-15-
-------
OVERVIEW OF PSD AND NONATTAINMENT - SPECIAL RULES FOR FUGITIVE DUST
2.1 National Ambient Standards
Any discussion of PSD and nonattainment should be preceded by an overview
of the fundamental statutory framework of the Clean Air Act for achieving and
maintaining the National Ambient Air Quality Standards (NAAQS). The primary
purpose of the Act is to achieve and maintain air quality levels which are
protective of public health and welfare. To this end, EPA is required to
establish National Ambient Air Quality Standards (NAAQS) for selected pollu-
tants emitted from diverse sources which could endanger public health or
welfare (criteria pollutants).1 (See Table I.) The ambient standards are of
two types. "Primary" standards are established at levels which will protect
public health while "secondary" standards are designed to protect public
welfare. As of October, 1978, EPA has promulgated NAAQS for six of the most
ubiquitous air pollutants including particulate matter (PM); sulfur oxides
(S0x); nitrogen oxides (N0x); carbon monoxide (CO); photochemical oxidants
(measured as ozone); and non-methane hydrocarbons (HC)*.2 An ambient standard
for lead was established in the Fall of 1978.3 It is the particulate matter
NAAQS which forms the statutory basis of fugitive dust control requirements
under the Clean Air Act.
Attainment and maintenance of the ambient standards for the criteria
pollutants is the primary responsibility of the States. Under Section 110 of
the Clean Air Act, each State is required to submit a State Implementation
Plan (SIP) for EPA's approval. These plans set forth the State's strategy for
attaining and maintaining the standards within the time frames established by
the Act. The required strategy must satisfy the eleven enumerated require-
ments contained in Section 110 of the Act, including programs for preventing
significant deterioration of clean areas and ensuring timely clean-up of
polluted areas. Under the 1970 Clean Air Act Amendments (the foundation of
the current statute) all States were to have attained the primary (health-
related) standards by May 31, 1975 with the exception of a relatively few
areas where an extension to mid-1977 was granted. Secondary standards were
to be attained within a "reasonable time," defined by most state plans to
coincide with the primary standard attainment date.4
2.2 Nonattainment
Unfortunately, the 1970 Act did not specify the consequences of a state's
failure to attain the primary standards by the statutory deadline. As the
deadline approached, however, at least 160 of the nation's 247 air quality
control regions (AQCRs) had monitored violations. Many urban areas failed to
attain the particulate standards.
*The hydrocarbon standard is designed only as a guide for assessing the
adequacy of State plans in attaining the photochemical oxidant standard.
-16-
-------
TABLE I
NATIONAL AMBIENT AIR QUALITY STANDARDS (NAAQS)
40 CFR 50
NATIONAL PRIMARY AMD SECONDARY AMBIENT AIR QUALITY STANDARDS
(Expressed as Mlcrograns per cubic meter [ug/ra3] at 25"C, 760 mm pressure)
-sj
I
POLLUTANT
Sulfur Oxides (SOx)
(measured as 802)
Participates
Carbon Monoxide (CO)
2
Photochemical Oxldants
3
Hydrocarbons (I1C)
14
Nitrogen Dioxide (N02)
PRIMARY STANDARD
Annual
Mean
80
75
—
—
—
100
Maximum Concentration
(Allowed Once Yearly)
365
(over 24 hours)
260
(over 24 hours)
10 milligrams/m3
(over 8 hours)
40 milligrams/m3
(over 1 hour)
157
(over 1 hour)
160
(over 3 hours —
6-9 a.m.)
.»
SECONDARY STANDARD
Annual
Mean
1
60
Maximum Concentration
(Allowed Or.ce Yearly)
1300
(over 3 hours)
150
(over 24 hoursj
Same an Primary Standard
Same as Primary Standard
Same as Primary Standard
Same a;
i Primary Standard
1. The 60 mg/n3 particulate standard is to be used only as a guide In assessing the adequacy of State
plans in achieving the 24-hour standard.
2. EPA has proposed to relax the primary oxidant standard from 157 up to 196 ng/rn3 (i.e., from 0.08
to 0.10 ppm).
3. The hydrocarbon standerd Is to be used as a guide in assessing the adequacy of oxidant control plans
4. Pursuant to a mandate in the Clean Air Act, EPA is currently investigating the need for a short-term
(i.e. 1-3 hour) NOx standard. EPA Is currently considering a one-hour standard ranging from 470 co
940 mp,/m3. The agency is expected to nuke a final decision on this matter in 1979.
-------
The obvious legal dilemma created by these widespread failures concerned
the legal "approvability" of new sources which would add to levels of pollu-
tion already in violation of the law. Section 110 of the Act, which pres-
cribed the July 1, 1975 attainment deadline, did not provide for the conse-
quences of failure. EPA's regulations implemented this inflexible mandate by
prohibiting the construction or modification of any facility which would
interfere with the attainment or maintenance of a national ambient standard.5
Thus, a strict reading of the law would have prohibited new sources from
locating in any area which had failed to attain the ambient standard for the
pollutant(s) it emitted. The practical effect of such an interpretation was
an end to growth in the developed areas of the United States. EPA realized
that the American public was unwilling to pay such a price, even for clean
air.
EPA responded to this problem with promulgation of its national Offset
Policy on December 21, 1976.*6 The essence of the Offset Policy is that major
new growth should be allowed in nonattainment areas only if air quality is
improved as a result of that growth. This improvement comes by way of contem-
poraneous emission reductions from existing sources so as to more than "off-
set" the emissions which will be added by the proposed new source.
The Offset Policy imposes four substantial conditions on major new or
modified sources, including major sources of fugitive dust, seeking permits to
expand in and around nonattainment areas. First, the source must reduce its
emissions to the lowest achievable emission rate (LAER). Second, it must be
able to certify that all sources which it owns or controls in the same state
are either im compliance or on a schedule of compliance with the State imple-
mentation plan. Third, the new source must obtain emission reductions from
existing area sources which more than "offset" the pollution to be added by
the new LAER-controlled source. These offsetting emission reductions must
exceed the proposed new emissions by enough to represent "reasonable further
progress" toward attainment of the ambient standards. The fourth and final
condition, closely related to the third, requires the owner to demonstrate
that the combination of LAER and offsets will lead to a net air quality
benefit in the affected area.
Congress met the nonattainment issue head-on in the 1977 Clean Air Act
Amendments by continuing the Offset Policy until mid-1979 and establishing a
State implementation plan revision process to deal with the problem after that
date. Specifically, Section 129 of the 1977 Amendments provided that EPA's
Offset Ruling, as it may be amended, was to remain in effect, except that the
baseline for determining emission offset credit was changed and the source
applicability was expanded. By January 1, 1979, the States must submit plan
revisions to EPA which conform to new Part D of the Act. These revised plans
are designed to replace EPA's Offset Policy and must be approved by no later
than June 30, 1979. Failure to adopt and receive EPA approval of the revised
*EPA is currently considering several significant changes to the policy.
These changes are not, however, expected to revise the basic format of the
current policy. Special rules will be proposed for fugitive dust sources,
however, as discussed in Section 2.4 of this paper.
-18-
-------
CHRONOLOGICAL DEVELOPMENT
OF EPA'S OFFSET POLICY
AND THE ACT'S NONATTAINMENT PROVISIONS
7/1/75: Original Deadline for Attaining the Primary NAAQS
Promulgation of EPA's Offset Policy
12/21/76
The 1977 Clean Air Act Amendments
8/7/77
EPA Administrator's Memorandum Detailing the Criteria
for Approving SIP Revisions for Nonattainment Areas
2/24/78
EPA Promulgates List Designating PSD and Nonattainment
Areas Pursuant to Section 107 of the Act
3/3/78
EPA Promulgates Major Revisions to Its Offset Ruling
12/78
Revised SIPs Due to EPA
1/1/79
Deadline for EPA Approval of the Revised SIP
(Failure to Obtain Approval by Deadline Triggers the
Growth Ban Provision of Section 110(a)(2)(I).)
6/30/79
New Deadline for Attaining the Primary NAAQS
12/31/82
Extended NAAQS Deadline for Areas Unable to
Attain the Oxidant and/or Carbon Monoxide Standards
12/31/87
-19-
-------
plan by this deadline triggers a statutory ban on major new sources (of the
nonattaining pollutant) in the nonattainment area.7 Until the revised SIPs
are approved and before July 1, 1979, the Offset Policy remains in fe-ree.
2.3 Prevention of Significant Deterioration
The Act's silence on the nonattainment problem was matched by an equal
legislative vacuum on the issue of whether an area with air quality already
better than the national standards could allow its air to be "degraded" to the
national standards. While the objective of the nonattainment provision is to
achieve and maintain the ambient standards, the purpose of PSD is to prevent
significant deterioration of air already cleaner than the ambient standards.
The PSD controversy began as a result of a seemingly innocuous phrase in
the original Act stating that one of the Act's four basic purposes was to
"protect and enhance" the quality of the nation's air. Relying primarily upon
this provision, environmental groups brought suit against EPA on May 24, 1972
to prevent the agency from approving State plans which failed to prevent
significant deterioration of clean air.8
This challenge was ultimately successful and on December 5, 1974, EPA
promulgated regulations to prevent emissions of sulfur dioxide and particulate
matter from significantly deteriorating air quality in areas where concentra-
tions of those pollutants were lower than the applicable national ambient
standards. (39 FR 42510, Codified at 40 CFR 52.21). EPA incorporated its PSD
regulations into the implementation plan of each State pursuant to Section
110(c) of the Act and established a procedure whereby EPA could delegate its
PSD responsibility to States.
The regulations prohibited construction of stationary sources in any of
nineteen specified categories unless EPA (or a delegate State) had issued a
permit evidencing that the source would apply "best available control techno-
logy" (BACT) for S02 and for particulate matter and that emissions of those
pollutants would not cause significant deterioration of clean air. For
determining what levels of deterioration were "significant," the regulations
set out an area classification system. Under it, clean air areas could be
classified as Class I, II, or III. In Class I areas, small increases of S02
and particulate matter would be significant; in Class II areas moderate
increases; and in Class III areas, increases up to an ambient standard. The
regulations initially classified all clean areas as Class II, but gave States,
Indian Governing Bodies, and Federal Land Managers the opportunity to reclas-
sify their lands under specified procedures.
The 1977 Amendments affirm the PSD concept.9 The new statutory scheme
follows the outline of the pre-existing regulations, but is generally more
comprehensive and restrictive. Some of the more significant changes intro-
duced by the 1977 Amendments include:
-20-
-------
o Formal Designation of PSD Areas - Under Section 107(d) of the new
Act, all areas with air quality better than the national standards
(or of indeterminate status) must be formally listed by EPA. EPA
issued this list on March 3, 1978 (see 43 FR 8962). Several revi-
sions have been made since March 3, 1978. (See 40 CFR 81.)
o More Restrictive "Increments" - The new law reduces the allowable
short-term increments for S02 and particulates in Class II and III
areas. Class I increments remain the same.1^
o New Mandatory Class I Areas - The new Act retains the 3-class
concept of the original PSD regulations and affirms the automatic
Class II designation for most PSD areas. However, large national
parks above a certain size which were in existence when the Amend-
ments were enacted (8/7/77) are now designated as mandatory Class I
areas. These areas may not be redesignated.1*
o More and Different Sources Subject to PSD Review - The new Act
increases the number of source categories subject to PSD precon-
struction review from 19 to 28. New and modified sources within one
of these 28 categories are subject to PSD if they have the poten-
tial* to emit or increase emissions by 100 tons per year of any
pollutant regulated under the Act. Furthermore, all new or modified
sources, regardless of their category, are covered if they have the
potential to emit 250 tons or more per year of any regulated pollu-
tant.12 Major sources of fugitive dust are covered by the 250-ton
criteria.
o More Restrictive Definition of BACT - Under the original PSD regula-
tions, Best Available Control Technology (BACT) could not be more
restrictive than the New Source Performance Standards (NSPS) ap-
plicable to the source category being proposed. Where NSPS had not
yet been established for the applicable category, BACT was a case by
case determination which weighed economic, technological, energy and
geographic factors. Under the new law, BACT is always a case by
case determination. Where New Source Performance Standards (or
hazardous emission standards) apply to the proposed source, they
represent a minimum, rather than a maximum, level of required
control. More importantly, BACT now applies to all pollutants
regulated under the Act, not just S02 and particulate matter as
under the prior regulations.^
o Substantially Increased Monitoring and Modeling Requirements -
Under the original PSD Regulations, applicants merely had to demon-
strate, through atmospheric diffusion modeling, that the proposed
emissions would not violate the allowable increments. The new la'w
*The Amendments define a major source in terms of its "potential to emit"
but do not define the term "potential." EPA's final PSD regulations define
"Potential to emit" as the capability at maximum capacity to emit a pollutant
in the absence of air pollution control equipment.
-21-
-------
Imposes far more sophisticated modeling requirements. In addition,
air quality and meteorologic monitoring requirements may now be
required both prior to and after construction.11*
o Other Criteria Pollutants to be Regulated in the Future -
EPA1s original PSD regulations applied only to S02 and particulates.
While the new law follows this policy on an interim basis, it
directs EPA to promulgate PSD regulations for the other criteria
pollutants (hydrocarbons, carbon monoxide, photochemical oxidants,
and nitrogen oxides) by no later than August 7, 1979. These regula-
tions will take effect one year after promulgation and must ulti-
mately be incorporated into all State Implementation plans. PSD
regulations for lead must be promulgated in I960.15
On November 3, 1977, EPA took four regulatory actions toward implementing
the Act's new PSD requirements.16 First, the agency promulgated amendments to
the pre-existing PSD regulations, conforming them to Section 168(b) of the new
Act. That section expressly made certain changes "immediately effective" as
of August 7, 1977 (i.e., new mandatory Class I areas; the new increments; and
new Class III reclassification procedures). The second action was the pro-
posal of regulations giving guidance for the preparation of SIP revisions
called for by the new PSD requirements of the Act. The third action was the
proposal of comprehensive changes to EPA's pre-existing PSD regulations,
conforming them to the new preconstruction requirements of Section 165. The
fourth action was a decision to implement the new preconstruction review and
BACT requirements of Section 165 as of March 1, 1978. Environmentalists had
argued for an August 7, 1977 effective date while Industry contended that the
States had sole authority to implement these new requirements through SIP
revisions.
EPA formally promulgated its new PSD regulations and SIP revision guide-
lines on June 19, 1978, three and one-half months after its original self-
imposed deadline of March 1, 1978.17 Due to the widespread public awareness
of the March 1st date, however, EPA made the new PSD regulations retroactive
to that date. Thus, sources which failed to obtain all necessary air pollu-
tion permits prior to March 1, 1978 will have to comply with the new PSD
regulations. Furthermore, even those sources which did receive the necessary
permits by March 1 will have to commence construction on or before March 19,
1979 in order to be exempt from the new requirements. This is the deadline
for the submission of revised SIPs to EPA.
As defined by the new PSD regulations "fugitive dust" consists of native
soil particles, uncontaminated by industrial pollutants, which become airborne
through the forces of wind or human activities.18 EPA has formally recognized
the greater health impact of fugitive dust in urban as opposed to rural
areas. In general, the particulate matter found in rural areas is composed of
non-respirable native soil particles. Such particles are usually not exposed
to potential contamination by Industrial pollutants and, therefore, present an
insignificant threat to public health. By contrast, the native soil in indus-
trialized urban areas is typically contaminated by a variety of potentially
harmful substances.
-22-
-------
CHRONOLOGICAL DEVELOPMENT OF
PREVENTION OF SIGNIFICANT DETERIORATION
Sierra Club vs. Ruckelshaus
6/11/73
EPA's Original PSD Regulations
12/5/74
The 1977 Clean Air Act Amendments
8/7/77
EPA Proposes Its Revised PSD Regulations
and SIP Revision Guidelines
11/3/77
Effective Date of EPA's Revised PSD Regulations
(Sources Receiving Permits After This Date Must Comply)
3/1/78
EPA Promulgates List Designating PSD and Nonattainment
Areas Pursuant to Section 107 of the Act
3/3/78
EPA Formally Promulgates Its Revised PSD Regulations
and SIP Revision Guidelines
(Note: The Revised PSD Regulations
Are Retroactively Effective as of March 1, 1978)
6/19/78
Major Sources with Allowable Emissions in Excess
of 50 Tons Per Year Will Have to Begin Submitting
One Year's Continuous Monitoring Data
7/7/78
Revised SIPs Due to EPA
(Note: All Sources Commencing Construction After This
Date Are Subject to the New PSD Requirements, Regardless
of When They Received Their Permit)
3/19/79
Deadline for EPA Approval or Disapproval of the Revised SIPs
7/19/79
-23-
-------
In light of such considerations and in recognition of a substantially
higher level of human exposure in urban areas, EPA has decided to focus its
fugitive dust control efforts in urban areas. In rural areas, controls are
recommended for large man-made sources of fugitive dust such as tailing piles
and surface mining operations which themselves are causing or contributing to
NAAQS violations.*19
For purposes of PSD, EPA will not require applicants to analyze the air
quality impacts of emissions which qualify as "fugitive dust." Nor will
fugitive dust emissions consume the allowable PSD increments for particulate
matter. In order to qualify for this exemption, however, applicants will
have to demonstrate that the emissions are indeed composed of uncontaminated
native soil. It should also be noted that for purposes of calculating poten-
tial emissions, fugitive dust emissions will be included. For example, while
a proposed surface mining expansion with a potential fugitive dust emission
rate in excess of 250 tons per year would be exempt from the PSD ambient air
quality analysis requirements, it would not be exempt from the other require-
ments of PSD such as BACT.20
2.4 Formal Attainment Status Designations
The State Implementation Plan revisions called for by the Act's new PSD
and nonattainment provisions are phrased in a way which requires the respec-
tive boundaries of a State's PSD and nonattainment areas be accurately de-
fined. Specifically, Section 107(d) directs each State to submit a list of the
NAAQS attainment status of all State areas to EPA. EPA promulgation of the
formal list, with any necessary modifications, was required within 60 days of
the submlttal of the State lists. This promulgation appears in the Federal
Register of March 3, 1978 (43 FR 8962, Codified at 40 CFR 81, Subpart C).
Section 107(d) specified that designations should be based upon air
quality levels on the date of enactment of the Amendments (August 7, 1977).
States were required by EPA guidance to consider the most recent four quarters
of monitored ambient air quality data available. If this data showed no
NAAQS violations, then the previous four quarters of monitoring data were to
be examined to assure that the indication of attainment was not the result of
unrepresentative meteorologic conditions. In the absence of sufficient
monitored air quality data, other evaluation methods were used, Including air
quality dispersion modeling.
*In general, a new major source to be located in a rural area with
infrequent short-term violations of the particulate NAAQS will be allowed to
construct after applying the required controls provided that the dust in
question is uncontaminated by pollutants from industrial activity and the
emissions of the source in conjunction with emissions from other sources in
the vicinity (excluding such dust) would not cause a violation of the ap-
plicable increment(s) or the applicable NAAQS, assuming as to the NAAQS an
appropriate "non-urban" background concentration.
-24-
-------
These Section 107(d) designations are subject to revision under Section
107(d)(5) whenever sufficient data is available to warrant a redesignation.22
Both the State and EPA can initiate changes to these designations, but any
State redesignation must be submitted to EPA for concurrence. Private parties
must, therefore, work through the State or EPA to initiate a change in the
designations.
The major significance of the 107 designations is not their impact on new
source preconstruction review, but rather on the SIP revision process.* Spe-
cifically, the formal designations of 107(d)(1)(A)-(E) are incorporated into
the SIP revision mandates of Part C (PSD) and Part D (Nonattainment) of the
Act. Section 161 of Part C requires SIP revisions to prevent significant
deterioration in each area identified pursuant to Section 107(d)(l)(D) or (E).
Section 171(2) in Part D defines the term "nonattainment area" to include any
area identified under subparagraph 107(d)(l)(A) through (C), while giving EPA
authority to add other areas based upon monitoring or modeling data. Designa-
tion as a "nonattainment area" triggers the SIP revision process of Part D and
a general ban on growth for nonattainment areas lacking an adequate SIP after
June 30, 1979. Designation as a "PSD area" triggers the SIP revision process
of Part C.
As noted in the preamble to the March 3rd Federal Register promulgation,
the designation of an area as attainment or nonattainment must be considered
only a point of departure and not as a final end in itself.23 Indeed, the Act
makes it clear that the designations are to be revised as appropriate to
reflect more current or accurate data. It must also be noted that the desig-
nations will have only limited significance for new source revisions since
under both the PSD and nonattainment preconstruction review procedures, major
new and modified sources must undergo a preliminary modeling analysis to
determine their impact upon all nearby areas as well as the area in which they
will be located. Thus, case by case impact analysis is required to determine
the impact on all neighboring areas and also to account for the possibility
that an area with a particular designation may encompass "pockets" which do
not fit that designation.
For designations of total suspended particulates (TSP), the localized
nature of the violations precluded the use of general area-size criteria.
States are therefore advised that designations along political boundaries
such as city or county lines were practical from an air quality management
standpoint. In the case of particulates, however, one very difficult problem
was the appropriate designation of rural areas with significant levels of
fugitive dust. As noted earlier, EPA's Fugitive Dust Policy recognizes the
generally greater health impact and toxicity of urban as opposed to rural
fugitive dust. Therefore, for the purposes of these designations, rural areas
experiencing TSP violations which could be attributed to fugitive dust could
claim attainment of the particulate NAAQS. Rural areas for this purpose are
defined as those which have: (1) A lack of major industrial development or
the absence of significant industrial particulate emissions, and (2) low
urbanized population densities.21*
-25-
-------
The attainment status designations are very important with respect to
major new fugitive dust sources.* Specifically, under EPA's soon to be
revised Offset Ruling, major fugitive dust sources locating in clean portions
of formally designated nonattainment areas or in attainment or unclassified
areas shall be subject only to applicable requirements for preventing signif-
icant deterioration of air quality (see 40 CFR 52.21). Thus, only those
major new fugitive dust sources locating in actual nonattainment areas will
be subject to the Offset Ruling.**
*Under both the Offset Policy and the PSD regulations the term "fugitive
dust" refers to particulate emissions composed of soil which becomes suspended
either by the forces of the wind or man's activity. This would include
emissions from unpaved haul roads, wind erosion of exposed soil surfaces and
soil storage piles and other activities in which soil is either removed,
stored, transported, or redistributed.
**Even where the Offset Ruling is applicable to a major new or modified
fugitive dust source, the soon to be revised Offset Policy exempts these
sources from the requirement of demonstrating a "net air quality benefit"
(Condition 4).
-26-
-------
3.0 SIP REVISIONS FOR NONATTAINMEN't AREAS - NEW FOCUS ON FUGITIVE DUST
3.1 Geographic Consideratleas
In analyzing partlculate pollution, It Is necessary to distinguish
between rural and urban nonattainment areas. The sources and characteristics
of particulate pollution in urban areas are fundamentally different from the
rural counterparts. The most important distinction is the generally greater
human health impact of urban particulate pollution. This is true in both a
quantitative and qualitative sense in that urban particulate not only affects
more people, but also is more hazardous than rural particulate matter.
Indeed, contaminated urban fugitive dust, coupled with potentially hazardous
direct industrial particulate emissions in urbanized areas, is one of the most
serious air pollution problems facing the nation from a public health stand-
point.
By contrast, rural particulates frequently consist of windblown native
soil particles. Although these rural fugitive dust emissions are often the
result of human activities such as agriculture, surface mining and quarrying
operations, the actual particles are typically in the non-respirable size
range and are uncontaminated by industrial pollutants. EPA's recent Fugitive
Dust Policy recognizes the fundamental distinctions between rural and urban
nonattainment problems for suspended particulates. Specifically, the policy
provides that major emphasis for fugitive dust control in nonattainment SIP
revisions should center on urban areas. It should be noted, however, that
this general exemption for rural nonattainment does not extend to rural areas
where traditional industrial particulate sources are causing or contributing
to the NAAQS violations. These areas are treated the same as urban nonattain-
ment areas.
With these geographic considerations as a backdrop, it becomes obvious
that TSP nonattainment strategies will focus primarily on major metropolitan
areas. Nearly 70% of the Nation's urbanized areas with populations exceeding
200,000 have been designated nonattainment for TSP.* Thus, the remainder of
this section will focus on the issues inherent in developing acceptable TSP
strategies for urban nonattainment areas.
3.2 Sources of Urban Particulate
For purposes of regulatory control strategy development, urban particu-
late sources are categorized either as traditional or nontradltional. The
distinction is significant in terms of the types of control strategies re-
quired in the 1979 SIPs. Traditional sources, as the name implies, refer to
those industrial sources which have been regulated under earlier SIPs. Both
stack and fugitive emissions associated with such sources must, to the extent
*Some suprising and noteworthy exceptions to the general nonattainment
status include New York-NE New Jersey, Philadelphia, and Detroit.
-27-
-------
necessary, be controlled through formally-adopted regulations in the 1979 SIP
submittals.25 Nontraditional sources are those which generally have not been
controlled in the past. These include emissions from construction activities,
demolition operations and resuspended street dust. By their nature, nontradi-
tional sources are not amenable to the kind of straightforward emission
limitations or design standards typically applied to traditional sources. The
relationship between these nontraditional emissions and resultant ambient air
quality concentrations has never been adequately defined. The development of
emissions factors, modeling techniques, and effective control strategies for
these complex sources will take time and study.
The SIP revision memorandum of February 4, 1978 recognizes these problems
and authorizes states to submit control program schedules (as opposed to
formal regulations) for control of nontraditional particulate sources.26
These schedules must call for the expeditious development, adoption, submittal
and implementation of the necessary control measures. For nontraditional
particulates, field studies and demonstration projects to define the problem
and investigate alternative controls should begin as soon as possible. If
these studies reveal that major nontraditional control efforts are needed to
meet the 1982 NAAQS deadlines, implementation of the required controls prior
to the deadline will be necessary.
The significance of nontraditi6nal sources to the TSP nonattainment
problem has become apparent as traditional source control programs have
failed to achieve ambient standards. In Greater Pittsburgh, for example,
modeling of inventoried traditional sources is able to account for only 15% of
measured TSP concentrations at certain monitors. Assuming the reliability of
the model estimates and the accuracy of the traditional source emissions
inventory, this means that fully 85% of ambient particulates are of nontradi-
tional origins. One of the most significant findings of the 1976 National
Particulate Assessment is that particulates from nontraditional sources con-
tribute 25 to 30 yg/m3 to citywide TSP levels, thereby preventing many urban
areas from attaining the ambient standards.27
These nontraditional particulate emissions are the product of everyday
urban activities. Five major nontraditional source categories have been
isolated in urban areas. These include reentrained street dust, construction
and demolition operations, exhaust (and tire) particles from automobiles, dust
from unpaved roads and secondary particulates. The impact of these sources
depends upon factors such as meteorology, topography, soil characteristics,
road patterns and many other variables which control the impact of wind or
traffic on surrounding particles. As a general rule, however, their air
quality impact is more localized than stack emissions. As such, SIP revisions
relating to nontraditional sources will have to be rather site and source
specific.
Recent evidence indicates that wind and vehicle-induced reentrainment of
street dust is probably the largest source of particulate matter in many
downtown areas. Depending upon the amount and type of road dust, as well as
the speed and level of traffic and wind conditions, reentrainment may con-
tribute from 14-20 yg/m3 to ambient TSP levels.28 The potential toxicity of
many of these particles, coupled with the high level of human exposure to
them, would seem to demand heightened regulatory attention to this problem.
-28-
-------
Automotive-related particulate emissions from tail pipes and tire wear
are closely related to the reentrainment problem since much reentrained
particulate originates from the automobile. The states' failure to address
this problem creates a serious regulatory deficiency in that such emissions
have potentially harmful health effects. For example, lead emissions from
automotive exhaust contribute as much as 4 pg/m3 in certain cities.29 Such
concentrations are believed to produce nervous system damage in children.30
The air quality impact of construction and demolition activities is also
substantial in terms of total particulate loadings. These emissions are
similar to those from unpaved areas generally in that they result from both
wind erosion over exposed soil surfaces and man's physical activities on the
surface. Heavy construction projects such as roadway construction and resi-
dential, commercial or industrial developments are clearly the largest con-
tributors in this category. Field studies indicate an average dust generation
rate of 1.2 tons of particulate per acre per month.31 Applying this emission
factor to a 150-acre residential construction project, the 100-ton major
source threshold is exceeded in less than 3 weeks.
Another TSP source not addressed in earlier SIP strategies is "secondary
particulate." Secondary particulates are the products of chemical reactions
occurring in the atmosphere through forces of sunlight, temperature and
meteorology in the presence of water vapor and gaseous atmospheric pollutants
such as S02 and NO . Found in both rural and urban areas, secondary particu-
lates are composed primarily of sulfates, organics and nitrates. They are a
prime component of urban smog and are a major contributor to ambient TSP
levels.
3.3 Control Strategy Development
From a conceptual standpoint, development of revised control strategies
for TSP is a seven step process beginning with problem definition and ending
with an adopted strategy demonstrating attainment of the 75 yg/m3 primary
standard before 1983. These seven steps include:
o ANALYSIS OF EXISTING AIR QUALITY DATA
o DEFINITION OF THE CONTROL AREA
o PREPARATION OF AN EMISSION INVENTORY
o ESTABLISHMENT OF SOURCE/AIR QUALITY RELATIONSHIPS
o EVALUATION OF ALTERNATIVE CONTROL STRATEGIES
o ANALYSIS OF ESTIMATED AIR QUALITY IMPROVEMENT
o SUBMITTAL OF AN APPROVABLE SIP REVISION
STEP 1 - Analysis Of Existing Air Quality Data
•
Determination of the attainment status of an area is based upon measured
TSP values at selected monitoring locations. High-volume samplers measure TSP
concentrations at these monitoring sites by drawing ambient air through a
particulate filter at a specified flow rate. The collected particles are then
weighed and the results expressed in terms of micrograms per cubic meter of
sampled air.
-29-
-------
The purpose of the ambient air quality analysis is to determine the
validity and representativeness of measured TSP concentrations. Assessing the
validity of measured TSP levels is essentially a matter of determining that
appropriate measurement techniques and quality control procedures have been
employed. Assuming the data is valid, procedures must then be employed to
determine whether it is representative of either a broad or confined area.
Once the relationship between monitored data and contributing sources is
established, it can then be used in a meaningful way to determine the actual
origins of the nonattainment problem.
STEP 2 - Definition Of The Control Area
Based upon the results of the air quality analysis, the next step in the
SIP revision process is to determine the boundaries of the area to be con-
trolled. This determination is based upon the geographic scope of the nonat-
tainment problem, and the location of the sources causing that problem. Both
.of these factors should have been defined in the air quality analysis.
STEP 3 .- Preparation Of An Emission Inventory
Having defined the origins and geographic scope of the nonattainment
problem, it is necessary to quantify the problem. This is accomplished by an
emission inventory. This inventory allows strategists to assess the relative
contributions to various sources and to determine the emission reduction
impact of alternative control options.
As a practical matter, many states simply will not be able to submit
"comprehensive, accurate and current emission inventories" by the Act's
January 1, 1979 deadline. As an interim measure, therefore, these States will
have to use emission estimates to help supplement their data base until more
reliable emissions information becomes available. For traditional sources,
EPA Document AP-4232 and the 1976 TRC analysis of Emission Rates For Eighty
Processes33 are "the primary references recommended by EPA for making these
estimates. For nontraditional sources, emission estimates are less precise
and reliance upon EPA's Guidelines for Development of Control Strategies in
Area with Fugitive Dust Problems is required.3'*
STEP 4 - Establishment Of Source/Air Quality Relationships
Having defined and quantified the problem, it becomes necessary to
quantify the relationship between area emissions and resultant ambient concen-
trations. Atmospheric diffusion models are employed for this purpose.
The selection of an appropriate diffusion model is a highly technical
determination requiring the services of trained professionals. EPA has pub-
lished numerous guideline documents on atmospheric modeling, the most recent
being its 1978 Guideline on Air Quality Models.35 This guideline specifies
the models to be used in assessing control strategies for TSP and other cri-
teria pollutants. Approval of the Regional EPA Administrator is required
before other modeling approaches may be employed.
-30-
-------
One of the major gaps in current air pollution knowledge concerns the
modeling of fugitive emissions and fugitive dust sources. EPA's modeling
guidelines deal primarily with stack emissions. Modeling techniques are
relatively well developed for stack emissions. Due to the complex configura-
tion and emission characteristics of fugitive sources, however, traditional
modeling techniques are not adequate. At the present time, therefore, model-
ing the impact of fugitive control strategies will be reviewed by EPA on a
case by case basis. EPA is now studying alternative modeling techniques for
both fugitive process emissions and fugitive dust sources. It is doubtful
that these studies will produce satisfactory fugitive modeling techniques in
the near future.
STEP 5 - Evaluation Of Alternative Control Strategies
The prior steps in the strategy development process lead naturally to
the evaluation of control alternatives. The objective of this step is to
identify the most cost-effective means of ensuring timely attainment and
maintenance of the ambient standards.
The first task in this assessment is to identify the sources (traditional
and/or nontraditional) which will receive priority attention for control.
Many considerations guide this determination, including economic costs,
administrative and social impacts, technological difficulties, and the air
quality benefits associated with imposition or tightening of controls on a
particular source. A comprehensive analysis of these policy considerations
is beyond the scope of this paper. Suffice it to say that this task forms
the cornerstone of the control strategy and should be treated accordingly by
State and local air quality strategists. The ultimate objective is to achieve
the greatest level of emissions reduction over time at the lowest per capita
cost. The selection of target sources sets the stage for realizing that
objective.
Once the target sources have been identified, it becomes necessary to
estimate the degree of emission reduction that will occur from full compli-
ance with adopted controls. Such control measures may be in the form of
increased or expedited enforcement of existing regulations, or entirely new
regulations and programs. These emissions reduction estimates comprise the
key variable in the modeling exercise through which reasonable progress
toward timely NAAQS attainment is demonstrated.
For existing sources, control strategies will have to impose reasonably
available controls (RACT) to the extent necessary for NAAQS attainment.
This is particularly true for major industrial centers where traditional
stationary sources have not been subject to such controls under prior SIP's.
RACT may represent a relatively stringent level of control for these sources
and may involve forcing technological innovation.* For all sources, however,
*Various efforts have either been completed or are underway to provide
guidance for determining RACT. Technical support documents for new source per-
formance standards are a valuable information source. In addition, EPA is
currently in the process of issuing documents which describe RACT for numerous
particulate sources covering both stack and fugitive emissions.
-31-
-------
RACT will be determined on a case by case basis taking into account such
factors as retrofit feasbillty, economic and energy costs and environmental
impacts.36
In urban areas where fugitive dust is causing high TSP levels, nontradi-
tional source controls for reentrained dust, construction activities, auto-
motive particulates and secondary particulates will have to be considered.
While the impact of these sources are generally localized in nature, they are
typically found throughout a given area and may, therefore, create widespread
TSP violations. The emission inventory should provide the spatial segregation
and emissions quantification necessary for evaluating these sources.
Control of reentrained street dust is one of the most important nontra-
ditional source programs which should be considered. Unfortunately, it is
also one of the most poorly understood particulate problems. Field studies
indicate that reentrained street dust consists primarily of sand and soil, and
automotive exhaust and debris from automotive wear and tear of brakes, bear-
ings and tires. Findings from these studies suggest two basic methods of
control:37
(1) Reducing the amount of material being deposited onto roadway sur-
faces through curbing, gutter and sidew,alk improvements, enforcement
of visible emission regulations for automotive vehicles, improve-
ments in sanding operations for snow and ice control and increased
controls on street dust loadings from construction vehicles.
(2) Street cleaning through sweeping and/or water flushing operations.
Common sense would indicate that street cleaning should reduce re-
entrained dust, but it is still unproven as an effective method for
reducing ambient TSP concentrations.
Urban roadway improvements anticipated in the next several years are
expected to result in significant reductions of re-entrained dust. As older
vehicles are phased out and automotive particulate contributions decrease, TSP
levels can be expected to decline in high traffic areas. State or local
agencies seeking to accomplish additional control over re-entrained dust
should refer to EPA's recent guideline document on Control of Re-entrained
Dust From Paved Streets.38 This guide provides more detailed information on
street cleaning, construction site control and the costs associated with these
controls. With respect to construction-related emissions, wetting or soil
stabilization of exposed surfaces and access roads appear to be the most
feasible control alternatives. Consideration should also be given to limiting
the length of time that any cleared area may remain exposed before construc-
tion activities commence.
Another nontradltional control which will have to be considered in some
areas concerns secondary particulates. Although the relationships between the
precursors of secondary particulates (SO , NO and organics) are not well
understood, it is suspected that these pollutants contribute from 5-15 pg/m3
of the particulate matter in many urban areas, much of it in the respirable
size range.39 To achieve the primary NAAQS for TSP, efforts will therefore
have to focus on controlling SOX, NO and organic emissions so as to prevent
the formation of sulfate, nitrate and organic aerosal particulates. While
-32-
-------
additional understanding of aerosol formation is needed, urban TSP control
strategies should at least consider reduction of secondary particulate pre-
cursors as a possible alternative.
STEP 6 - Analysis Of Estimated Air Quality Improvement
Having identified alternative control strategies, the next step in the
process is to estimate the air quality impact of each alternative. This is
accomplished through application of the atmospheric diffusion models already
selected under Step 4. This model translates estimated emission reductions
over time into estimated resulting air quality. Alternative strategies are
then manipulated until timely attainment of the ambient standards can be
demonstrated.
STEP 7 - Adoption And Submittal
The SIP Revision Memorandum, as noted earlier, distinguishes between
traditional and nontraditional source controls in terms of how they must be
addressed in the 1979 SIP submissions. All necessary stack and fugitive
emission limitations for traditional sources must be adopted and submitted by
January 1, 1979. Where nontraditional source control is part of the attain-
ment plan, however, formal regulations need not address them. For these
sources the plan must instead contain:
(1) an assessment of the impact of nontraditional sources, and
(2) a commitment by the state to develop, submit and implement the
appropriate procedures and programs. Obviously, these schedules
must include milestones for evaluating progress toward timely
attainment of the standards. Recognizing that the development of
these programs will take considerable time, the memorandum advises
states to initiate the necessary studies and demonstration projects
for controlling nontraditional sources as soon as possible.
3.4 Conclusions and Recommendations
Defining and evaluating control options for TSP nonattainment will
obviously take a great deal of study and effort. The development of an
effective control strategy requires an accurate identification of the sources
of the problem, a quantification of emissions from those sources and an
understanding of the relationship between these emissions and air quality.
Once this information has been assembled, a control strategy can be developed
which reflects the relative contribution of traditional, nontraditional and
background sources to known NAAQS violations.
Particulate control strategies have traditionally been oriented toward
fuel combustion, process emissions and incineration. Control programs for
particulates have resulted in substantial emission reductions and resulting
air quality improvements in many urban areas. Yet, further reduction in
these emissions can still be achieved. Several specific recommendations were
made in the National Particulate Assessment, with regard to traditional
sources. These included:
-33-
-------
o Increased regulatory and enforcement attention directed to the
primary metals and minerals Industries;
o Increased inspection, source sampling and enforcement activity to
obtain compliance by major sources with existing regulations;
o Promulgation and enforcement of numerical or visible "property-
line" fugitive emission standards; operating and maintenance stan-
dards for specific processes such as storage piles and material
handling equipment; and increased paving of industrial roadways and
parking areas;
o Increased regulatory attention to residential and commercial space
heating boilers where they are shown to be a problem;
o Considerable tightening of emission limits for fuel-burning instal-
lations, especially oil-fired operations;
o Regulation of industrial process emissions on an industry-specific
basis. Under existing regulatory practice, general process weight
standards are established for all processes, regardless of how
difficult it is for any particular industry to meet them;
o Consideration of extending regulatory coverage to smaller indus-
trial point sources where the contribution of such sources to NAAQS
violations is found to be significant.
Even with the traditional control measures discussed above, it is appar-
ent that many urban areas will not be able to attain the ambient TSP stan-
dards before 1983. This is the single most important conclusion of the
National Particulate Assessment. It presents difficult planning problems for
urban air quality strategists since emission reduction efficiencies are lower
(and less certain) for nontraditional sources than for traditional sources.
The difficulties lie not so much in the technical feasibility of nontradi-
tional controls as in their cost and political implications.
Still, much can be done to control particulate emission from automo-
biles, and construction activities and from re-entrainment of particulates
from urban streets.
Finally, it is important to remember that there are a number of gaps in
current data and understanding which will have to be overcome before TSP
control strategies can be evaluated with the degree of precision necessary.
Two of the most critical areas of needed investigation are siting criteria
for TSP monitors and the relationship between meteorology/climatology/topog-
raphy and ambient TSP concentrations. Improved emissions/air quality data
bases and expanded research into techniques for modeling and controlling
fugitive emissions and fugitive dust will also be Important to the control
effort.
-34-
-------
FOOTNOTES
1. Clean Air Act (hereafter "CAA") §109.
2. 40 CFR 50.
3. A3 FR 46246.
4. CAA §110(a)(2)(A).
5. 40 CFR §51.13 and 51.18.
•6. 41 FR 55524.
7. CAA §110(a)(2)(I).
8. Sierra Club vs. Ruckelshaus, 4 ERG 1205.
9. CAA, Subchapter I, Part C, Subpart 1.
10. CAA §162.
11. CAA §162.
12. CAA §165 and §169(1).
13. CAA §165(a)(4) and §169(3).
14. CAA §165(c).
15. CAA §166.
16. 42 FR 57459.
17. 43 FR 26380, 40 CFR 51.24, 52.21.
18. 40 CFR 52.21(b)(b).
19. 43 FR 26395.
20. Id.
21. 43 FR 8962.
22. CAA §107(d)(5).
23. 43 FR 8963.
24. Id.
-35-
-------
25. 43 FR 21676.
26. Id.
27. EPA-450/3-76-025, p. 30.
28. Id. at p. 78.
29. Id. at p. 70.
30. 43 FR 46247.
31. EPA Guideline for Development of Control Strategies in Areas With
Fugitive Dust Problems, EPA-450/2-77-029 at p. 3-10.
32. This document establishes emissions factors for many air pollution
sources.
33. EPA Contract Report No. 68-02-1382 T.O. #12.
34. Supra, note 31.
35. EPA-450/2-78/027, May 1978.
36. December 9, 1976 EPA memorandum from Roger Strelow to EPA Regional
Administrators.
37. Supra, note 27 at pi 128.
38. EPA-907/9-77-007.
39. Supra, note 27 at p. 48.
-36-
-------
THE IMPACT OF FUGITIVE EMISSIONS ON
TSP CONCENTRATIONS IN URBAN AREAS
Prepared By
Ronald G. Draftz
IIT Research Institute
10 West 35th Street
Chicago, IL 60616
-37-
-------
ABSTRACT*
The primary cause of total suspended particulate (TSP) violations
in urban areas such as Chicago, Phoenix, Philadelphia, St. Louis, Miami,
Denver and many others, is due to fugitive emissions. These fugitive
emissions, contributing from 40% to 70% of the TSP concentrations, are
principally from a single source - suspension of roadway dust and pavement
aggregate by vehicles. This roadway dust is mainly pavement aggregate
but also includes rubber tire fragments and other deposited dusts such as
soil minerals. The contributions from various entrainment mechanisms such
as direct tire erosion or tire and vehicle turbulence is unknown but have
a direct impact on control strategies.
Microscopical analysis supplemented with chemical analysis has
proven to be the most useful method of determining the sources of fugitive
emissions. This analytical protocol will be described in detail along
with the techniques for source assignments.
* This paper was not available at the time of the Proceedings' publication.
-38-
-------
EMISSIONS AND EFFLUENTS FROM COAL STORAGE PILES
by
T. R. Blackwood
and
R. A. Wachter
MONSANTO RESEARCH CORPORATION
DAYTON LABORATORY
Station B, P.O. Box 8
DAYTON, OHIO 45407
-39-
-------
ABSTRACT
Coal storage piles are sources of atmospheric emissions of fugitive dust and
nonpoint effluent discharges.
Pile geometry, coal erodability (dustiness), wind, humidity, precipitation and
temperature are the major parameters known to contribute to the emissions from
coal storage piles. From the literature and a source-oriented sampling pro-
gram, an equation was derived describing the mass emission rate from a given
coal pile. This program also showed that the variability of the emissions
from a specific coal pile are much greater than that between piles. This is
due to the large influence that wind speed and the precipitation-evaporation
index have on emissions. Emissions were analyzed utilizing x-ray fluorescence
and atomic absorption. Particle size analyses were conducted on two samples.
Over 97 percent of the particles, by number, were smaller than 10 microns.
Eighty-eight percent of the particles from a composite Brink® sampler taken
over two days of sampling were smaller than 5 microns in size.
The water pollution potential of coal stockpiles maintained outdoors at pro-
duction and user sites was also studied. These storage piles are sources of
effluents due to the drainage and runoff of wastewater which occurs during
and after precipitation. The runoff usually flows from the drainage area into
the nearest waterway. In this study, the effluent levels from these sources
were quantified by examining coals (both freshly mined and aged) from six coal
regions of the United States. Coals were placed under a rainfall simulator
and grab samples of the drainage were collected. The samples were analyzed
for organic and inorganic substances and for water quality indicators.
The coal leachate and air samples were also analyzed for selected polycyclic
organic materials, and only small quantities of POM's were detected.
-40-
-------
INTRODUCTION
Water effluents and air emissions exist at coal stockpiles maintained outdoors.
Water effluents form from the drainage and runoff of wastewater which occurs
during and after precipitation. Air emissions form from the action of wind
forces on the exposed coal surface.
Effluents were evaluated from six major coal regions of the United States (1).
Data were obtained by placing several coals beneath a rainfall simulator and
collecting integrated samples of the drainage. The samples were analyzed for
chemical compositions and water quality indicators. Hydrologic relationships
were used to estimate the runoff concentrations of full size stockpiles from
the simulator concentrations.
Emissions were evaluated using field sampling of typical stockpiles and wind
tunnel studies of the erosion process (2). Air emissions were analyzed for
chemical composition and air quality indicators.
SOURCE DESCRIPTION
Three-fourths of all coal produced is consumed at electric utilities. In
1975 there were approximately 950 coal stockpiles containing 124 x 106 metric
tons of coal at user facilities throughout the United States. These stock
piles are maintained outdoors and exposed to a variety of atmospheric condi-
tions. Rainfall leaches pollutants from the stockpile which drain into
waterways. Aquatic life forms in the waterways are thus exposed to the
pollutants. Drainage in coal mining operations produces high sulfate concen-
trations and low pH values in nearby streams. Hence, the potential for this
same problem exists at coal stockpiles because pyrites, the prime factor in
acid mine drainage, exist within the coal pile.
Wind blows the loose and fine coal away from the stockpile. The combined
leaching and drying of the coal causes fracturing which releases more fine
material that can become airborne. Mechanical movement of the stockpile
during usage also releases or fractures more coal.
In addition, coal contains inorganic substances in the "ash" (extraneous
mineral matter) and in the coal structure; these substances are released.
Because coal is primarily organic, organic contaminants are also released.
AIR EMISSIONS
Carbon monoxide (CO) hydrocarbons and particulate matter are the criteria
pollutants emitted. Concentrations of carbon monoxide and hydrocarbons
are barely detectable (<20 ppm CO, and <5 ppm hydrocarbons); these can be
expected to be three orders of magnitude below ambient air quality criteria
at a distance of 50 meters from the pile. The average emission factor for
respirable particulates (those smaller than seven microns in diameter) is
6.4 milligrams/kilogram of coal per year. The composition of the emissions
from coal piles is essentially that of coal dust. There is no significant
increase in the level of any known hazardous species other than the normal
increase found when coal is fractured to produce dust. This increase is
shown in Table 1.
-41-
-------
TABLE 1. COMPOSITION OP COAL AND COAL DUST
Amount in coal, Amount in coal dust,
Element ppm ppm
Arsenic
Cadmium
Copper
Iron
Nickel
Vanadium
0.3
0.2
25
1,600
2.7
12
26
3.8
868
79,200
755
166
Pile geometry, coal erodability (dustiness), wind, humidity, precipitation
and temperature are the major factors known to contribute to the emissions
from coal storage piles. Coal erosion has been studied in a wind tunnel, and
the effects of most of these factors, or their equivalents, have been
evaluated. Equation 1 describes the mass emission rate from a given coal
pile in terms of these factors.
U3p 2S0.345
Q = (336) (1)
(P-E)2
Q = emission rate, mg/s
u = wind speed, m/s
pb = coal density, g/m3
s = surface area, m2
(P-E) = precipitation-evaporation index
During this study, variability of the emissions from a specific coal pile were
determined to be much greater than the variation of emissions from one coal
pile to another. This is due to the large influence that wind speed and
precipitation-evaporation rate have on emissions. The location of a specific
coal pile will have a major influence on the emission rate. One storage
pile sampled one day apart showed emission rates of 14 and 41 milligrams per
second. The emission rate may also be essentially zero, as it was during some
days of the sampling program. One of three sampling tests showed undetectable
contributions due to the storage pile.
Because about 90 percent of the coal stored in the United States is stored in
P-E index regions of 90 to 130 (Figure 1), the effect of the P-E index is
reduced substantially for the average coal pile. In addition, the distribution
of coal piles shows that a tremendous quanity of small piles exist. The most
frequently occurring coal pile size is 49,000 metric tons; the average size is
95,000 metric tons (Figure 2).
-42-
-------
27.2
K.t
25.6
21.1
20.0
II.
I '2'7
S 10.»
7.2
01.061
127.27
PCRCCNIAGES ARf SHOWN IN 1 1
I23.Z9I
16.161
10 » 30 40 50 60 70 W 40 100 110 120 1» 140 150 160 170 110
P-C INDEX
Figure 1. Quantity of coal stored as a function of P-E index.
MEAN = 95,000 metric tons
MODE=49,000 metric tons
200 400 600 800 1,000 1,200 1,400
SIZE OF COAL PILES. 103 metric tons
Figure 2. Distribution of coal storage pile sizes.
-43-
-------
A coal storage pile of 112,000 metric tons was sampled at two seasons of the
year: late March and mid-August. Only small variations in the emissions
were observed between these two sampling sets. The average emission factor
was 6.4 milligrams per kilogram of coal per year; the estimated population
standard deviation was 2.9 milligrams per kilogram of coal per year. The
emissions were analyzed utilizing x-ray fluorescence and atomic absorption;
their compositions were essentially that of coal dust. Particle size analyses
were conducted on two samples. Over 97 percent of the particles, by number,
were smaller than 10 microns. Eighty-eight percent of the particles from a
composite Brink® sampler taken over two days of sampling were smaller than
5 microns in size. Based upon these results, it is concluded that essentially
all of the emissions are in the respirable range. No fibers were detected in
any of the samples. The coal samples were also extracted with pentane and
analyzed for selected polycyclic organic materials using chemical ionization
mass spectroscopy. Although small quantities (<0.5 ppm) of benzo(c)phenan-
threne, benzo(a)pyrene and 3-methylcholanthrene were detected, the quantity
identified was insignificant compared to the quantity of coal dust.
In order to evaluate the impact of coal storage on the environment, a repre-
sentative coal pile was defined from average parameters. This representative
stockpile would contain 95,000 metric tons of bituminous coal piled to an
average height of 5.8 meters. The pile would be located in a P-E region of
91, be exposed to an average annual wind speed of 4.5 meters per second, and
be located 86 meters from the nearest plant boundary. Assuming that this coal
pile is round, Figure 3 shows the approximate relationship between the coal
pile, the plant boundary and a typical emissions radius. The ambient concen-
tration of particulate at different distances is given in Table 2 based on a
particulate emission rate of 19 milligrams per second or 610 kilograms per
year from the average pile. These ground level concentrations were calculated
using dispersion methodology as described by Turner (3). The coal pile which
was sampled in this study had parameters very similar to those of the repre-
sentative stockpile.
TABLE 2. AMBIENT CONCENTRATIONS OF PARTICULATE
Distance from
coal pile,
Ground level concentration, yg/m3
Instant 24-hr average
86
528
18
1.9
6.4
0.67
Figure 3. Emissions radius of coal pile
-44-
-------
WATER EFFLUENTS
Because coal is a complex aggregate capable of discharging a vast range of
compounds, the study (1) was limited to compounds listed on the EPA Toxic
Substances List (4), pollutants with effluent limitations for coal storage
areas, and other water quality criteria used to indicate the presence of
classes of compounds. Due to the diversity of stockpiles and coals, several
samples were collected from each of the six major coal regions of the United
States. Aged coals and fresh coals were collected. Samples were selected to
represent the range of inorganic element content in coal and to obtain coals
that were representative of each coal region.
The apparatus shown in Figure 4 was used to simulate rainfall of representative
intensity and duration. Coal samples were placed under this apparatus for
exposure to simulated rainfall. Drainage seeped through the coal, out the
bottom of the pans, and into collection bottles. A background sample was also
obtained of the rainfall water used. Three simulation runs were completed
over a period of 30 days. The time between runs was varied to observe its
effect. The average effluent concentration found for each coal region is
presented in Table 3.
The effluent levels from a representative stock pile were computed as coal-
production-per-region weighted averages. Table 4 presents these levels at
the source. However, the pollutant concentration levels in the stream that
enters the nearest waterway result from the dilution of pile drainage by runoff
waters in the entire coal storage area.
The representative storage pile was assumed to be located 86 meters from the
receiving waterway. The representative rainfall rate is 0.7 centimeter per
hour over a stockpile area of 18,792 square meters. This rainfall occurs
139 days per year (i.e., every 2.6 days) and lasts for'approximately one hour.
However, only 15 percent of the rainfall volume on the pile appears as direct
runoff used in the rational method of hydrology (5). Volumetric flow from
the pile was computed as 21 cubic meters per hour.
Runoff from the entire coal storage area was obtained from a survey of coal
storage sites; the average runoff was 610 cubic meters per hour (6).
The concentration levels shown in Table 4 are diluted by the drainage area
volumetric flow to obtain the effluent concentrations shown in Table 5.
Thorough mixing of runoff waters with pile drainage is assumed. The coal
aggregate retards the runoff flow for a time period sufficient to enable
mixing of upstream runoff.
CONCLUSION
Pollution from coal stockpiles is a site-specific problem. Large, aged coal
stockpiles located in areas of frequent and light rainfall will generate much
higher effluent concentrations than that of the representative pile used in
this study. The variability of the emissions is also dependent on the site,
although local meterology will be the dominating factor. Air emissions will
not be a major problem because of the small contribution to ambient air quality
at typical exposure distances, even for coal stockpiles of several million tons.
Water effluents from Interior Western (Iowa to Texas and Arkansas to Nebraska)
and possibly the Great Northern Plains (Montana, North and South Dakota) could
-45-
-------
DISTILLED WATER
Figure 4. Rainfall simulation apparatus.
TABLE 3. AVERAGE EFFLUENT DRAINAGE CONCENTRATION FOR EACH COAL REGION
Effluent concentration, g/m3
Effluent parameter
Total suspended solids
Total dissolved solids
Sulfate
Iron
Manganese
Free silica
Cyanide
BOD5
COD
Nitrate
Total phosphate
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Copper
Lead
Nickel
Selenium
Silver
Zinc
Mercury
Thallium
pHb
Chloride
Total organic carbon
Appalachian
1,521
259
66
3.1
0.03
12.3
<0.001
<5.0
1,407
0.12
NDL
2.1
23
NDL
NDL
NDL
0.02
0.05
0.06
23.8
NDL
0.008
<0.001
NDL
6.28
0.33
251.7
Great
Northern
Plains
1,282
430
1,598
1.5
0.14
NDLa
NDL
<7.5
1,324
0.14
NDL
NDL
1.8
HDL
NDL
NDL
NDL
0.05
0.02
NDL
NDL
0.17
0.003
NDL
6.93
NDL
373.2
Interior
Eastern
1,264
1,136
648
9.1
0.44
0.8
0.002
NDL
1,556
0.33
NDL
7.5
4.1
NDL
NDL
NDL
NDL
0.06
0.09
12.5
NDL
0.14
NDL
NDL
7.62
NDL
380.1
Interior
Western
1,853
5,539
4,860
1,131
17.9
86.3
NDL
<1.2
1,053
0.09
NDL
10.3
10.1
NDL
0.05
0.03
2.2
0.33
10.2
25.2
NDL
25.0
0.004
NDL
2.81
2.3
90.5
Western
2,486
1,900
240
8.2
0.4
NDL
NDL
<2.5
1,826
1.8
NDL
14.0
5.6
NDL
0.005
0.04
NDL
0.07
0.05
15.0
NDL
0.15
0.005
NDL
7.24
NDL
318.4
Southwestern
1,538
356
190
5.5
0.04
NDL
NDL
<7.5
769
0.16
NDL
6.5
4.1
NDL
NDL
NDL
0.02
0.05
0.03
21.5
NDL
0.04
0.002
NDL
6.60
NDL
158.7
3No detectable level. Negative logarithm of hydrogen ion concentration.
-46-
-------
TABLE 4. COAL PRODUCTION-WEIGHTED
EFFLUENT CONCENTRATION
Concentration,
Effluent parameter g/m3
Total suspended solids
Total dissolved solids
Sulfate
Iron
Manganese
Free silica
Cyanide
BOD5
COD
Nitrate
Total phosphate
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Copper
Lead
Nickel
Selenium
Silver
Zinc
Mercury
Thallium
pH
Chloride
Total organic carbon
1,551
754
401
39
0.69
10.1
<0.001
<3.8
1,436
0.31
NDL3
4.6
15.7
NDL
0.002
0.004
0.08
0.06
3.1
19.9
NDL
0.80
<0.001
NDL L
6.78
0.27
280
a
No detectable level.
b
Negative logarithm of hydrogen
ion concentration.
-47-
-------
TABLE 5. CALCULATED RUNOFF CONCENTRATIONS FROM
THE REPRESENTATIVE COAL STORAGE FILE
Concentration
entering waterways,
Effluent parameter g/m3
Total suspended solids 7.8
Total dissolved solids 3.8
Sulfate 2.0
Iron 0.2
Manganese 3.4 x 10"1*
Free silica 0.05
Cyanide <5 x 10~6
Nitrate 1.5 x 10~3
Total phosphate NDLa
Antimony 0.02
Arsenic 0.08
Beryllium NDL
Cadmium 1 x 10~5
Chromium 2 x 10~5
Copper 4 x 10~5
Lead 3 x I0~k
Nickel 0.02
Selenium 0.1
Silver NDL
Zinc 4 x lO'4
Mercury 5 x 10~6
Thallium NDL
Chloride 1.3 x 10~3
Total organic carbon 1.4
2-Chloronapthalene 7 x 10~5
Acenapthene 7 x 10~5
Fluorene 7 x 10~5
Fluoranthene 8 x 10~5
Benzidine 7 x 10~5
Benzo(ghi)perylene 2 x 10~^
a ~ ~~~ ~ ~
No detectable level.
-48-
-------
present a major effluent problem which would require neutralization and sedi-
mentation of the runoff (primarily for control of acid).
ACKNOWLEDGMENT
This research was funded by the Industrial Environmental Research Laboratory
of the U.S. Environmental Protection Agency, Cincinnati, Ohio, under contract
68-02-1874. The content of this paper does not necessarily reflect the views
or policies of the U.S. EPA. The authors are indebted to Dr. Lawrence Huggins
and the Agricultural Engineering Department of Purdue University for their
assistance in this research. This acknowledgment does not convey an endorse-
ment by Purdue University of the research results.
REFERENCES
1. Wachter, R. A. and T. R. Blackwood. Source Assessment: Water Pollutants
from Coal Storage Areas, EPA-600/2-78-004m, U.S. Environmental Protection
Agency, Cincinnati, Ohio, May 1978. 123 pp.
2. Blackwood, T. R. and R. A. Wachter. Source Assessment: Coal Storage Piles.
EPA-600/2-78-004k, U.S. Environmental Protection Agency, Cincinnati, Ohio,
May 1978. 98 pp.
3. Turner, D. B. Workbook of Atmospheric Dispersion Estimates. Public Health
Service Publication No. 999-AP-26, U.S. Department of Health, Education,
and Welfare, Cincinnati, Ohio, May 1970. 84 pp.
4. Sampling and Analysis Procedures for Survey of Industrial Effluents for
Priority Pollutants. U.S. Environmental Protection Agency, Cincinnati,
Ohio, March 1977. 180 pp.
5. Handbook of Water Resources and Pollution Control. H. W. Gehm and
J. I. Bregman, eds. Van Nostrand Publishing Company, New York, New York,
1976. pp. 444-447.
6. Development Document for Proposed Effluent Limitation Guidelines and
New Source Performance Standards for the Steam Electric Power Generating
Point Source Category. EPA-440/1-73-029, U.S. Environmental Protection
Agency, Washington, D.C., March 1974. pp. 128-130.
-49-
-------
URANIUM MILL TAILINGS AREA
FUGITIVE EMISSION
BY
P. G. Complin, P.Eng.
and
P. Piersol
of
JAMES F. MacLAREN LIMITED
435 McNicoll Avenue
Willowdale, Ontario
M2H 2R8
-51-
-------
ABSTRACT FOR THE THIRD SYMPOSIUM ON FUGITIVE EMISSIONS;
MEASUREMENT AND CONTROL
Uranium Mill Tailings Area Fugitive Emissions
For the past year and a half, an environmental assessment of the expansion
of Canada's largest uranium mining area has been underway. Investigation
of the air quality aspects has involved an extensive study of tailings
area fugitive dust emissions by field monitoring and emission and dispersion
modelling.
The hi-vol field program resulted in approximately 500 samples of 24-hour
suspended particulate concentrations at locations in the immediate
vicinity of seven tailings areas. Micrometeorological stations located
on the tailings areas, provided wind speeds, wind directions, rainfall and
other meteorological data at the source during some of the dust sampling
periods. There were very large day to day changes in suspended particulate
levels next to tailings areas. The majority of the readings obtained were
less than 60 ng/m3 with an occasional value above 1000 yg/m3 and a number
of values below 20 yg/m3. It was ascertained that suspended particulate
levels are highly dependent on tailings moisture content and surface
windspeed.
A computer model for the emissions and dispersion of dust from the tailings
areas was developed based on the concepts of particle saltation, suspension,
and deposition and atmospheric transport using a Gaussan tilted plume model.
The selection of a suspension factor of 10"7 per meter resulted in good
agreement between the model and the field program results. The ratio
of suspended particulate concentrations calculated from the tailings
to actual measured suspended particulate ranged from 0.07 to 2.91, with
an average ratio of 1.1.
-52-
-------
URANIUM MILL TAILINGS AREA FUGITIVE EMISSIONS
Since the fall of 1976, an environmental assessment of the expansion of
Canada's largest uranium mining area has been underway. Investigation of
the air quality aspects has involved an extensive study of tailings area
fugitive dust emissions by field monitoring and emission and dispersion
modelling, with the intent of predicting the potential impact of tailings
emissions after mine expansions.
In Elliot Lake region of Canada, underground mining of uranium deposits in
quartz-pebble conglomerates has been conducted since the 1950's and is expec-
ted to continue until the turn of the century. In underground mining, the
ore is drilled, blasted, transported to underground primary crushing,
hoisted to the surface, crushed again and conveyed to the mill for uranium
extraction. A "typical" mill illustrated in the schematic shown (Fig. 1)
consists of ten primary steps:
Ore Crushing
Ore Grinding
Leaching
Solid/Liquid .Separation
Solution Upgrading
Uranium Precipitation
Uranium Thickening/Filtration
Drying
Packaging
Tailings Treatment/Discharge
After tailings neutralization, the tailings slurry is disposed of in contain-
ment areas. Within the Elliot Lake region, the tailings areas are usually
natural basins often containing swamps and small lakes with waste rock used
for containment dams. The uranium mining industry has been active in this
area since the 1950's and there are presently both active and inactive mines
and mills. There are a total of seven majcr tailings areas of which two are
-53-
-------
ft n ftf
ORE RECEIVED
AT MILL
MILL
ORE CRUSHING
ORE GRINDING
SOLUTION
UPGRADING
LEACHING
SOLID/LIQUID SEPARATION
TAILINGS
NEUTRALIZATION
URANIUM
PRECIPITATION
Solution
Decoding
Rtogcnts
URANIUM THICKENING/
FILTRATION
DRYING
V
LLMJt
PACNNG,
'SHIPPING
Underground
PRIMARY FLOWS
TAILINGS STREAMS
SECONDARY FLOWS
TAILINGS SETTLING
BASIN
f
PRECIPITATION
BASIN
Troottd
Efflutnt
Oitehorg*
Uranium Mill Process
Schematic
FIG. I
-------
now active and five inactive.
Very little information is presently available about windblown particulate
from tailings areas. Dust emissions from tailings areas share many simi-
larities with the common fugitive sources such as unpaved roads, agricultural
tilling operations, aggregate storage piles and heavy construction operations.
To gain more information, a field programme of measuring suspended particulate
and dustfall was undertaken in the immediate vicinity of the tailings areas.
Suspended particulate was collected with high-volume samplers and particulate
deposition was collected in dustfall jars. One hi-vol sampler and one dust-
fall jar constituted a sampling station.
Wind data for the region showed that although no wind direction predominates,
north and west winds were expected to occur most frequently. Consequently,
upwind-downwind sampling was used in assessing the tailings dust emissions.
In addition to the hi-vol samplers, micro-meteorological stations were
located directly on the tailings areas to obtain data on conditions which
'affect the wind generation of the dust.
This approach was modified to account for topographic constraints, access to
sites and the proximity of interferring sources such as roads and construc-
tion. Within these general siting criteria, the precise locations for the
sampling stations were chosen in areas generally clear of dense bush with a
visible line of sight between the tailings and the sampling station. If
locations were in close proximity to roads, calcium chloride solution was
used to limit road dust generation.
There were very large day-to-day changes in suspended particulate levels
next to the tailings areas. The majority of the readings were less than
60 yg/m3 with an occasional value above 1000 yg/m3 and a number of values
below 20 yg/m3. A summary of the results is presented in Table 1.
Further interpretation of the data confirmed some beliefs on the affect
of wind, rain and snow-cover on the degree of dust generated. Figure 2
illustrates the results from a portion of the 1977 summer particulate
-55-
-------
TABLE 1
I
Ln
Tailings
Monitoring
Site
Stanrock 1
2
3
Long Lake I
2
3
Nordic 1
Lab
Crotch Lake 1
2
3
Panel 1
2
Quirke 1
2
3
Number
of
Readings
31
30
29
13
13
15
73
66
38
35
9
5
2
48
28
8
SUMMARY
Distance
To Dry
Tailings
(m)
100
100
250
60
60
1000
100
200
20
20
60
10
20
120
30
1000
OF TAILINGS
Average
of
Readings
(yg/m3)
115
91
22
44
15
38
94
24
54
43
15
10
44
43
48
67
SUSPENDED
Geometric
Mean of
Readings
(yg/m3)
45
55
19
26
13
19
32
20
27
22
13
7
30
33
23
43
PARTICULATE
Highest
Reading
(yg/m3)
1051
575
60
240
49
262
1965
65
341
300
34
24
75
147
441
276
SURVEY
Lowest
Reading
(yg/m3)
14
9
6
4
6
1
0
4
4
4
5
1
12
7
1
16
% of Readings
Above 120
yg/m3
(%)
16
17
0
8
0
7
10
0
13
9
0
0
0
2
7
13
% of Readings
Above 400
yg/m3
6
7
0
0
0
0
5
0
0
0
0
0
0
0
4
0
Sheriff Lake
61
1000
21
18
48
-------
EFFECTS OF WIND AND RAIN
ON SUSPENDED PARTICULATE
FIG 2
B R ;
H
wind vanes p
y
tailtngs
wet
tailmgs drying pi n
fvt m N
-c.t6 .- •' . •• &
ifttfl
10 ' ' '
JUNE
•
-
r r '*<•
*T °t*
1 ' 15 ' '
h
[
I
1
gh
' ! M
I
If
1 S i
-
lr
llL
1 20 ' ' ' r '
]
|
/\
°'s^f
1965
-i
r
r
21 ' '
JULY
r4'
LO
pit
-9
>
WIND
high
[
25
VEL
!
J
i
fliS
-
h
I i I I
: a B i i
• . wet • «L
i 1
ow
30
OCITY (km/hr)
v max
V
• average
RAINFALL (mm)
TAILINGS
MOISTURE
>V^
1
1
LvJtr
n ii
AUGUST
-100
-50
-10
40
20
0
10
15
0
100
50
5 ' ' ' '
STATION 1
r STATION 2
TIME (days)
SUSPENDED PARTICULATES
(Concentration pg/m3 )
NOTE : lOng/m3 is limil o( detection
-57-
-------
survey selected to demonstrate the effects of tailings moisture content and
windspeed on suspended particulate levels. Between June 12 and June 16, the
tailings were drying out following rain and the suspended dust levels rose
quite steadily day by day from 20 to 70 Ug/m3 at Station 1 and from 13 to 62
yg/m3 at Station 2. During this time wind speeds averaged about 6 kra/h
(4 mi./h). The rainfall on June 16, 17 and 18 caused dust levels to drop to
0 at Station 1 and 13 and Station 2 by June 18, even though the wind speeds
over this period were higher.
The role of high wind speeds is clearly shown by the July 25 data. On July
25, average winds of 18 km/h (11 mi.A) with gusts to 30 km/h (19 mi./h)
combined with dry tailings resulted in the highest dust level recorded at
Station 2. However, on July 31, comparable wind gusts and only a slightly
lower average wind speed showed low dust levels due to an increase in tailings
moisture content.
The principle survey was conducted during the summer of 1977 and a lesser
continuing programme is still underway. Annual data, Table 2, indicates the
effects of seasonal climatic changes on airborne dust released from tailings
areas. The geometric means and highest values recorded during the winter
months are very low. During the spring period, the values increase somewhat,
but it is obvious from the data that summer conditions result in higher
values. Clearly, the tailings are not prone to becoming airborne during the
winter months, and are less likely to drift in the spring than summer.
Sufficient data was available from two active tailings areas to perform a
upwind-downwind analysis. The results do confirm the belief that tailings
dust is carried downwind and will result in elevated suspended particulate
levels immediately adjacent to the tailings area. They do not, however,
identify the effect of tailings fugitive emissions on the areas of concern,
the townsites and residential areas.
For this reason, a theoretical model, based on sand physics, was used to
predict air emissions, and classical atmospheric dispersion techniques were
used to transport and deposit the airborne particulate. The results of the
monitoring program were used to calibrate the model.
-58-
-------
TABLE 2
SEASONAL SUSPENDED PARTICULATE RESULTS
JUNE 1977 - AUGUST 1977
MONITORING
SITE
NORDIC 1
NORDIC LAB
NUMBER
OF
READINGS
73
66
AVERAGE
(ug/m3)
94
24
GEOMETRIC
MEAN
(yg/ta3 )
32
20
HIGHEST
READING
(ug/m3 )
1965
65
NOVEMBER 1977 - MARCH 1978
NORDIC 1
NORDIC LAB
45
44
13
13
6
10
81
33
APRIL 1978 - JUNE 1978
NORDIC 1
NORDIC LAB
27
28
35
31
21
20
236
106
-59-
-------
The air pollution effect of fugitive emission sources depends on the quantity
and drift potential of the dust particles entrained into the air. In addition
to large particles which tend to settle out near the source, fine particles
can be entrained and will disperse from the source.
Studies on sand movement and desert dunes have shown that the transport of
materials similar to tailings is affected by particle size distribution, local
wind velocities (including wind gusts), and other variables. Exposed surface
area and the physical condition of the tailings also have a pronounced effect
on dust entrainment.
As part of the overall programme to expand the data base for parameters which
affect tailings emi-sions and provide information for the model exposed
tailings areas were examined and a series of tailings particle size analyses
were initiated.
Aerial photographs and field observations were combined in an effort to
delineate the extent of wet and dry tailings and develop the dry exposed
surface areas. Size sampling programmes were undertaken at two non-operating
tailings areas and at two operating areas. Samples were taken from tailings
areas which differed in terms of texture and appearance, and these samples
were aggregated and dried prior to sizing analysis.
A model for dispersion of dust from a uranium mill tailings area was developed
by James F. MacLaren Limited for use at the Elliot Lake properties. The model
is based on a concept developed by the Oak Ridge National Laboratory and
includes particle saltation, suspension and deposition and atmospheric trans-
port using a Gaussian plume model.
The suspension of dust at the source is related to the process of saltation
as described by Bagnold and an emplirical estimate 'of the particles which
remain airborne near the tailings surface, referred to as suspension.
The saltation rate equation presented here is influenced by the windspeed,
V, and the theshold wind velocity, Vt, required to initiate particle motion.
-60-
-------
TABLE 3
TAILINGS PARTICLE SIZE DISTRIBUTIONS
PARTICLE
SIZE
TYLER
SCREENS
FOR SAMPLES
PREDOMINANT
EMISSION
PROPERTIES
PREDOMINANT
DISPERSION
PROPERTIES
ym
mesh
t
300-150 to 100 CREEP
AND
150-106 100 to 150 SALTATION
106-80 )
) 150 to 200
80-75 ) J
75-44 )
) 200 to 400 SUSPI
44-38 )
<38 <400
3NSION
f
I
DUSTFALL
SUSPENDED
PARTICULATE
-61-
-------
Saltation Rate, q
The material remaining in suspension per unit area, S, has been shown to be
direct function of the saltation rate:
Suspension, S = kq
where the value of K, the suspension factor, varies from 10 to 10 per
metre dpending upon the particle size and mass.
The dust emission rate or source strength is the material remaining in sus-
pension per unit area multiplied by the source area:
Emission rate, Q = SA
The dispersion and deposition of the suspended materials was based on the
"titled plume" method described by Pasquill:
X (x,0,0) = 2i exp / - (H- v * x/
2 7T 6 6 u -j7
y z \ 26z
This equation assumes that material reaching the ground will not be reflected
or entrained. Some percentage of smaller particles will be reflected at the
ground and so this equation overestimates deposition and underestimates
suspended particulate concentrations.
To ensure that the model predictions resulted in conservative values for
both deposition and suspension for small particles separate models were used
to calculate suspended particulate and deposition.
The previous equation was used to model deposition and modified for suspended
particulate as follows:
X (x,0,0) - Qi exp/- (H-v ' x/u)2 \ + exp / - (H-v ' x/u)2
2 TT 6 6 u ' S
26^
z
-62-
-------
The principal difficulty in using the nodal to quantify tailings dust
emissions and dispersion is the selection of the suspension factor, K. As
noted, K values can range from 10~5 to 10~9 per metre,-which covers a ten
thousand fold variation.
Value of 10~7 per metre was compared with the arithmetic average values for
the tailings hi-vol stations as shown in Table 4. As shown in the table,
the ratio of suspended particulate concentrations calculated from the tailings
to actual measured total suspended particulate matter (Sm/Ss) ranged from
0.07 to 2.91, with an average ratio of 1.1.
At two stations, gravel pit operations add appreciably to the total
particulate loadings. If these stations are not considered, the average
ratio is increased to 1.25.
Since the tailings hi-vol stations were all in close proximity to tailings
and, except where other known sources are also in proximity, most of the
particulate measured is from tailings. However, there will in all cases be
some background contribution due to natural and man-made sources. On the
assumption that this background would not be less than 10 yg/m3 the average
ratio (Sm/Ss) is increased to 1.65.
The selection of the K value at 10~7 per metre appears to result in good
overall agreement between the model and the field results. More detailed
calibration did not appear to be reasonable.
Annual contours of suspended particulate and dustfall were predicted by the
model to determine the extent of the effect of the emissions and to assess
the calibration of the model. The model prediction of the present mining
and milling is presented in Figure 3.
A comparison of the suspended particulate model to the air quality, Table 5,
results indicates that the model is reasonable. The measured overall air
quality at town monitoring stations was approximately 40 yg/m3 (arithmetic
average) while the tailings area contribution predicted by the model is less,
ranging from 8 to 17 yg/m3 at locations in the general vicinity of the
-63-
-------
TABLE 4
COMPARISON OF MODEL
MONITORING
SITE
STANROCK 1
2
3
NORDIC 1
LAB
SHERIFF LAKE
CROTCH LAKE 1
2
3
PANEL 1
2
QUIRKE 1
2
3
LONG LAKE 1
2
3
SURVEY
Number
of
Readings
31
30
29
73
66
51
38
35
9
5*
2*
48
28
8
13
13
15
WITH SUSPENDED PARTICUIATE
RESULTS
Arithmetic
Average
(Ss)
(vig/m3)
115.0
92.8
22.3
93.9
24.3
20.7
53.8
43.1
15.2
(10.4)
(43.5)
42.4
48.2
67.0
44.3
14.8
38.4
MODEL
FIELD STUDY
DISTA]
FROM j
Prediction Evaluation TAILI1
(Sm)
(yg/m3)
62
58
32
35
43
10
67
79
38
55
37
33
25
5
59
43
9
Sm/Ss
0.54
0.62
1.43
0.37
1.77
0.42
1.25
1.83
2.50
0.78
0.52
0.07 (1)
1.33
2.91
0.23 (1)
(m)
90
120
240
200
200
1000
20
20
50
20
20
120
30
1000
20
50
1000
AVERAGE OF ALL STATIONS
*Not enough readings for meaningful results
(1) Stations were ne*ar other known dust sources
1.1
-64-
-------
PRESENT ANNUAL
QUALITY PREDICTIONS
- FIG 3
-IlJ)> Boundary of exposed tailings
Boundary of 2g/m2 + 30d dustfall
and 20ug/m3 suspended
particulate concentrations
O Boundary of 4.6g/m2+30d dustfall
and 30ug/m3 suspended
particulate concentrations
o
L
Kilometres
-65-
-------
tailings and 4 yg/m3 in the more removed town of Elliot Lake. Background
contributions include other fugitive sources such as gravel pits, unpaved
roads, and construction operation which also have variable emission
characteristics. From the results of the particulate survey, these could
make up 20 +_ 10 Vg/m3 at any one location at any one time.
If the model predictions were increased by a factor of two, they would
approach the measured air quality and the amount left for other background
sources would be unreasonably low.
The same approach can be applied to the dustfall results. Table 6, and the
same conclusion reached.
As previously mentioned, the uranium mining industry on Elliot Lake is
undergoing an expension and the effect of the increase in tailings areas
is of concern. The ultimate use of the model in using its ability to
predict trends and extent of the tailings dust on air quality was to deter-
mine the future air quality. Overall air quality patterns only changed
slightly with the isopleths around the sources expanding, Figure 4.
Specific results at the critical areas of the townsites indicates that the
expansion of tailings will not have a significant effect on the existing
air quality, Table 7.
-66-
-------
TABLE 5
COMPARISON OF TAILINGS SUSPENDED PARTICULATE MODEL TO AIR QUALITY RESULTS
FIELD SURVEY RESULTS
Arithmetic
Number of Highest Lowest Geometric Average
MONITORING 24 h Reading Reading Mean (Xs)
STATION Measurements yg/m3 yg/m3 yg/m3 yg/m3
Elliot
Lake 55 136 5 35 42
Nordic
Townsite 29 121 6 30 39
Denison
Townsite 27 108 1 30 40
Quirke
Townsite 29 85 6 32 38
MODEL RESULTS
(TAILINGS ONLY)
Annual
Average
(Xm) Xm/Xs
yg/m3 x 100
3.5 10
17 45
13 35
8 20
-------
TABLE 6
COMPARISON OF TAILINGS DEPOSITION. MODELL TO AIR QUALITY RESULTS
SURVEY RESULTS
MODEL RESULTS
(TAILINGS ONLY)
oo
STATION
Hillside Drive
Kilborn Way
Roman Avenue
imber
of
;o d
•ements
11
10
8
Highest
Reading
g/m2
30 d
7.5
6.5
17.7
Lowest
Reading
g/m2
30 d
0.8
1.0
0.5
Arithmetic
Average
(Ds) g/m2
30 d
3.8
3.6
4.1
Monthly
Average (Dm)
g/m2
30 d
0.2
Dm/Ds
x 100
10%
Denison Townsite
11
24.1
0.9
8.4
1.8
20%
Quirke Townsite
11
4.3
0.4
2.2
0.9
40%
-------
FUTURE ANNUAL
AIR QUALITY PREDICTIONS
FIG 4
—
O
Boundary of exposed tailings
Boundary of 2g/m2 dustfall
and 20ug/nri3 suspended
particulate concentrations
Boundary oh4.6g/m2-t-390d a dustfall
and 30ug/rrt3 suspended
particulate concentrations
Kilometres
-70-
-------
TABLE 7
COMPARISON OF EXPANSION CASE AIR QUALITY ESTIMATES
IN RESIDENTIAL AREAS WITH PRESENT MEASUREMENTS
Station
SUSPENDED PARTICULATE
(pg/m ) (geometric mean)
Existing
Expansion Case
Approximation
Existing
DUSTFALL
(g/m2. 30 d)
Expansion Case
Approximation
vo
i
ELLIOT LAKE
Hillside Drive
Kilborn Way
Roman Avenue
DENISON TOWNSITE
QUIRKE TOWNSITE
35
30
32
37
45
44
3.8
3.6
4.1
8.4
2.2
3.9
3.6
4.2
9.6
3.7
CRITERION
60
60
-------
BIBLIOGRAPHY
R. A. Bagnold, "The Physics of Blown Sand and Desert Dunes" Methuen and Co.
Ltd., 1959.
James F. MacLaren Limited, "Environmental Assessment of the Proposed Elliot
Lake Uranium Mines Expansion"
Volume 1 Background Information, March 1977
Volume 2 Background Information Update, February 1978
Volume 4 Environmental Assessment, April 1978
Addendum 1 September 1978
M. T. Mills, R. C. Dahlman, and J. S. Olsen, "Ground Level Air Concentrations
of Dust Particles Downwind from a Tailings Area During
a Typical Wind Storm", Oak Ridge National Laboratory, 1974.
F. Pasquill, "Atmospheric Diffusion" John Wiley and Sons Ltd., 1974.
G. A. Sehmel and M. M. Orgill, "Resuspension by Wind at Rocky Flats" Batelle
Northwest Laboratories, BNWL-1751, Pt. 1
-71-
-------
IDENTIFYING SOURCES AND QUANTITIES
OF FUGITIVE EMISSIONS IN BALTIMORE
Robert C. Koch
and
John T. Schakenbach
GEOMET, Incorporated
Gaithersburg, Maryland 20760
and
Kathryn G. Severin
IIT Research Institute
Chicago, Illinois 60616
-73-
-------
ABSTRACT
Violations of Federal primary standards for total suspended particulate
(TSP) matter have been observed in the Baltimore area for the past several
years and, based on modeling results, are expected to continue to occur
through 1985. A study was undertaken to estimate fugitive emissions from
nonstandard sources and to analyze hi-vol samples to determine the types
of sources which are contributing to the concentrations which are presently
exceeding standards. The results of a survey of fugitive source character-
istics in the designated TSP nonattainment area in the metropolitan Baltimore
area were used to estimate fugitive dust emissions. Emissions from gravel
and dirt roads were found to account for 74 percent of fugitive emissions.
Fugitive emissions in the nonattaintment area were estimated to be at an
annual average rate of 10,000 Ib/hr. A labortory analysis of 18 hi-vol
filters based on optical microscopy and a battery of additional chemical
and physical tests were used to characterize the filter sample into six
source categories. The analysis was made on half-filters and was repeated
for three samples. The source type attributions for sources consisting
of over 15 percent of the sample were reproduced within 12 percent of the
original estimate. The hi-vol filter analysis showed that 50 percent of
the sample materials are.associated with fugitive sources. Possible fugi-
tive emission control strategies were identified based on these results
which will reduce fugitive emissions by 70 percent. At locations with the
highest measurements, the emissions controls may reduce TSP concentrations
by 50 percent. Future study of the air quality impact of possible control
alternatives is recommended.
-74-
-------
INTRODUCTION
A study of suspended particulate matter in a selected portion of the
Metropolitan Baltimore area has characterized the types of particles which
contribute to exceedances of the National Ambient Air Quality Standards.
A survey of the nonattainment area was made to identify fugitive sources
and estimate emissions. Potential control techniques for reducing fugi-
tive emissions were identified.
The total suspended particulate (TSP) air quality problem in the
Baltimore area is shown by the data in Tables 1 and 2. Monitoring sta-
tions which exceeded Federal primary TSP standards in 1975 or 1976 are
listed for annual geometric means in Table 1 and for second highest
24-hour concentrations in Table 2. The locations of these monitoring
sites are shown in Figure 1.
Table 1. Stations Exceeding Annual Primary TSP Standard*
in the Metropolitan Baltimore AQCR in 1975 or 1976
Station
Fire Department HQ
Fire Department #10
Ft. McHenry
Fire Department #22
Patapsco STP
S.E. Police Station
AAI
Annual Geometric Mean
1975
86
128
86
86
150
78
(not operated)
( g/m3)
1976
74
164
105
82
144
77
90
* Annual Standard = 75 g/m geometric mean.
-75-
-------
Table 2. Stations Exceeding 24-Hour Primary TSP Standard*
in the Metropolitan Baltimore AQCR in 1975 or 1976
Station
Riviera Beach
Fire Department #10
Patapsco STP
Lansdowne High School
Second Highest TSP Concentration (ug/m,)
0
1975
357
358
398
101
1976
174
559
509
271
* Primary Standard = 250 vg/m geometric mean.
Figure 1. Location of TSP monitoring stations
exceeding primary NAAQS in 1975 and 1976.
-76-
-------
A nonattainment area for TSP was designated for Metropolitan Baltimore
as shown in Figure 2. This study was conducted to determine the causes of
the high concentrations of TSP in this area and to suggest alternative
methods of controlling the contributions to the high concentrations from
fugitive emissions, which were suspected to be a major cause of the exceed-
ances.
- City ttmiu
___ County Umttf
Ncaattaiamaat ana
Figure 2. Metropolitan Baltimore nonattayiment area for TSP,
including the three hi-vol stations used in the project.
-77-
-------
The Maryland Bureau of Air Quality and Noise Control (MBAQNC) collected
24-hour hi-vol samples at three sites in the nonattainment area (see Figure 2)
during the 7-month period from June to December 1977. Eighteen filters were
selected for physical and chemical analysis to determine generic types of par-
ticKs collected. The particle types, their relative mass on each filter,
and their frequency of occurrence at each site and for different days helped
determine significant sources or classes of sources.
A ground-area, "windshield" survey was performed by MBAQNC on the loca-
tion and characteristics of fugitive sources of TSP which were not included
in the state emission inventory. Types of surface material, e.g., grass,
asphalt, dirt and gravel, were identified along with quantitative factors
affecting emission rates, including size, amount of vehicle activity, and
types of industrial activity. The survey results and other data relating
to mobile sources (i.e., cars, trucks and trains, etc.) were used to esti-
mate fugitive emissions in each square mile of the city. The estimates will
be used to update the state emission inventory.
A review of control strategies found in the literature was made to recom-
mend methods to implement for each of six types of fugitive particulate sources,
including paved roads, dirt and gravel roads, construction and waste disposal
sites, wind erosion, railroads and storage piles.
FUGITIVE EMISSION ESTIMATES
The data collected by MBAQNC in its survey of fugitive emission sources
covered an area of 83 square miles within the City limits of Baltimore and
284 squares measuring 1,000 feet on a side in parts of the surrounding
Baltimore and Anne Arundel Counties. An inspector visited each area and
characterized each open lot or other area by the type of surface, its use
(e.g., parking, recreation, construction, unused, railroad, storage, etc.),
the type of terrain (e.g., open, enclosed, steep slopes, etc.), the types
of emissions (e.g., vehicles, stockpiles, material handlers, natural occur-
rences, etc.) and location identification data. Special characteristics,
reflecting mobile activities were also noted, including the number of
various vehicles, speed of vehicles, and frequency of stockpile turnover.
The Maryland emission inventory was updated to include fugitive par-
ticulate emissions in two ways: (1) by adding fugitive emissions from new
sources identified in the survey, and (2) by adding fugitive emissions
from sources in the present inventory. Both the amount of fugitive par-
ticulate emissions and information on particle sizes were included in the
inventory data. Fugitive emission factors applicable to each source were
developed from the available literature and from data obtained by direct
contact with investigators whose work had not been published. The factors
were adapted to the meteorological conditions, the materials and the types
of activities found in the Baltimore area.
-78-
-------
Paved Roads
The following emission factor for paved roads was derived from work
done primarily by Bonn and Cowherd (1978):
EF = 0.45 (s/10)(L/5000)(W/3) (1)
where EF = emission factor (Ib/vehicle mile traveled)
s = silt content of road surface material (%)
W = average vehicle weight (tons)
L = surface dust loading on traveled portion of road (Ib/mile).
There was not sufficient information available to assign detailed spatial
and temporal variations to the parameters in this relationship. Values
estimated to be representative for Baltimore were selected. Street runoff
measurements by Sartor and Boyd (1972) were used to estimate that 12 per-
cent is a reasonable estimate of the silt content. The average weight of
vehicles was estimated to be 1.5 tons corresponding primarily to light-duty
vehicles. The dust loading of streets was assumed to be 17.7 Ib/mile based
on informal discussions with investigators in EPA Region III and Midwest
Research Institute. The resulting emission factor for Baltimore paved
roads is 0.43 g(9.6 x 10 Ib) per vehicle mile traveled (VMT).
Unpaved Roads
Emission factors for unpaved roads, including both gravel and dirt
roads, were derived from the EPA publication AP42 (1977). The annual num-
ber of days per year with precipitation greater than or equal to 0.01 inches
was found to be 112 days for Baltimore. The silt content of gravel roads
was estimated to be 12 percent. For dirt roads, information from local
sources was used to estimate the silt content. The following relation-
ship was used to determine emission factors:
EF = 0.6(0.81)s(S/30)(l -(P/365)) (2)
where EF = emission factor (Ib/VMT)
s = silt content (%)
S = average vehicle speed, mph
P = days per year with 0.01 inch or more
precipitation or with reported snow cover.
-79-
-------
To apply equation (2) it is necessary to know the number of vehicle miles
traveled, the average vehicle speed, and (for dirt roads) the silt content
of the roadway. If a better estimate was not available, vehicle speeds
were assumed to be 25 miles per hour. When applied to parking lot traffic,
vehicle speeds were assumed to be 10 miles per hour. The silt content of
dirt roads was estimated for each quadrant of the city and found to vary
from 62 to 75 percent.
In order to be compatible with other emission factors used in this
study, the EPA equation was multiplied by 0.6 to estimate the fraction
of the total particulate emissions with aerodynamic diameters less than
30 urn.
Construction Sites
An emission factor for construction sites was developed from work by
Cowherd and others (1974). Emissions of 1.2 tons/acre/month (269 g/m /month)
are representative of average construction activity and are assumed to repre-
sent solid waste disposal sites which use similar types of operations. Vari-
ations in activity levels may result in emissions varying by a factor of two
over a period of a month. Activity levels are difficult to estimate without
precise information on the nature and intensity of construction or disposal
operations. Since observations from the southwest United States were used
to develop the emission factor, it may be high for east coast areas.
Wind Erosion
For fugitive emissions due to wind erosion, the following emission
factor was used based on an analysis by Bonn and Cowherd (1978):
EF = 381 (e/50) (s/15) (f/25) (PE/50)'2
2
where EF = emission factor, g/m /yr
e = surface credibility, ton/acre/yr
s = silt content of surface material, %
f = frequency that wind exeeds 12 mph, %
PE = Thornthwaite's precipitation-evaporation index.
For the Baltimore area, the wind exceeds 12 miles per hour 21 percent of the
time and the Thornthwaite precipitation-evaporation index is 108. For sur-
faces common in the Baltimore area, the surface credibility varies from 40 to
200 tons/acre/year. For most clay soils the factor is 85 tons/acre/year. A
list of soil credibility factors for various soil textural classes was com-
piled by Cowherd et al. (1974).
-80-
-------
Railroads
An emission factor for railroad traffic was developed to treat the large
amount of railroad volume much of which connects with the Baltimore Harbor.
Based on observations of rail traffic, conversations with railyard managers and
considerations of emission factors for other mobile sources, the emission
factor was set at 10 percent of the emission rate for unpaved roads. The silt
content of railroad loads was estimated at 10 percent and railroad car speeds
were estimated to be 15 miles per hour in the city limits and 40 miles per hour
outside the city limits. Railroad data were used to estimate the railroad
car miles traveled in city and county areas.
Storage Piles
An emission factor for aggregate storage piles developed from work by
Cowherd et al. (1974) is as follows:
where EF = emission factor, g/kg (placed in storage)
PE = Thornthwaite's precipitation-evaporation index.
To apply this factor, it is necessary to know the amount of material moved
into the storage pile over the course of a year. This emission factor was
applied in the Baltimore area to large piles with a material throughput on
the order of 100,000 tons per year or more.
Summary of Fugitive Emissions
There are many uncertainties associated with the emission factors.
Based on a review of the variability reflected in the data used to develop
the factors, considering the uncertainty of other parameter estimates used
to apply the factors, and considering the uncertainty of the applicability
of the factors to the Baltimore area, it is estimated that over the course
of the year actual fugitive emissions will be within a factor of 2 of the
calculated fugitive emissions. On any given day, the emissions may be off
by a greater amount. The validity of the emission factors for determining
maximum 24-hour concentrations for verifying compliance with National
Ambient Air Quality Standards needs further study. A summary of the emis-
sion factors used for the Baltimore area is presented in Table 3.
-81-
-------
Table 3. Summary of Fugitive Emission Factors and Required Source Data
Type of Source
Emission Factor
Required Source Characteristics
Paved road
Gravel road
Dirt road
Construction
sites
Wind erosion
Railroads
Storage piles
0.43 (g/VMI)
0.96 x 10" J (Ib/VMT)
61 S (g/VMT)
0.13 S (Ib/VMT)
5.1 sS (g/VMT)
0.0112 sS (Ib/VMT)
9.0 (g/rrT/day)
79 (lb/acre/day)
2.6 x 10"4 e s
(g/mVday)
2.3 x 10"3 e s
(Ib/acre/day)
6.1 S (g/VMT)
0.013 S (Ib/VMT)
0.14 (g/kg)
0.28 (Ib/ton)
VMT (vehicle miles traveled)
VMT
S (vehicle speed, mph)
VMT
S
s (silt content of road
material, %)
Site area
Area
s
e (surface credibility,
ton/acre/year)
VMT
S
Quantity placed in storage
When the emission factors and the-required source data are applied to the
designated nonattainment area of Metropolitan Baltimore, the fugitive emissions
presented in Table 4 are obtained for the portion within the city limits, the
portion outside the city limits and the total for the nonattainment area. From
these results, it can be seen that, of the seven nontraditional sources, dirt
roads, gravel roads and construction operations contribute the greatest amount
of fugitive particulate emissions. The emissions for unpaved roads (dirt and
gravel) amount to 74 percent of the total. This only includes material kicked
up from the pavement and does not include tire wear or tailpipe emissions.
-82-
-------
Table 4. Baltimore TSP Nonattainment Area Fugitive Participates (Tons/Day)
Type
Paved roads
Gravel roads
Dirt roads
Construction
Wind erosion
Railroads
Storage piles
Total
Within City
.
26
45
10
2
-
4
87
Outside City
12
5
9
4
_
2
32
Total
.
38
50
19
6
-
6
119
HI-VOL FILTER ANALYSIS
The State of Maryland collected hi-vol filters at 3 sites in the non-
attainment area where noncompliance with Federal standards had been reported.
A total of 18 filters collected over a period of 7 months between June and
December 1977 were selected for microscopy and chemical analysis. The location
of the sampling sites is shown in Figure 2. The selected filters represent
different wind directions, wind speeds, and particulate loadings as shown in
Table 5. On 3 days filters were selected from all 3 sites; on another 3 days
filters were obtained from 2 sites; and the remaining 3 filters were selected
on 3 separate days. The wind speeds were slower than normal on the selected
days, varying from 2 to 6.5 as a daily average, comapred to the climatological
average for Baltimore of 10 miles per hour.
Microscopy and chemical analysis were performed on the selected filters
by I IT Research Institute. One-half of each selected filter was submitted for
analysis. At the conclusion of the study, three additional halves, selected
at random from the original 18, were submittted for duplicate analysis.
The microscopy was performed by removing the particles from a piece of
the filter half and viewing the particles optically under a high-powered
microscope which is illuminated by polarized light. The optical properties
observed to classify particles by different types of sources include trans-
parency, color, refractive index, birefringence, reflectance, pleochroism and
fluorescence. The physical properties observed include size, shape, surface
texture, magnetism, solubility, melting point and density. To identify par-
ticle types, the microscopist relies on reference collections of particle
data which include atlases of photomicrographs, handbooks of optical pro-
perties, the microscopist's previous experience and actual source samples.
-83-
-------
Table 5. Hi-Vol Filters Selected for Analysis
Date
6/9
6/21
7/15
8/2
9/19
9/25
10/25
11/24
12/12
Day
of
Week
Th
T
F
T
M
S
T
Th
M
f Sample no
Measured Concentrations (vg/m )
Fort
McHenry
59
136*
114*
154
69*
39
118
41
85
t analyzed
Fire
Dept.#10
63
145*
221
142*
165*
51
306
42
169
Fire
Dept.#22
—
60
149
93*
108
38
-
52
129
Rainfall (in)
Day of
Sample
1-36
0
0
0
0
T
0
0.01
T
Day
before
Sample
0
0.36
0
0.03
0
T
0
0.31
0
Wind
Speed
(mph)
6.2
6.3
2.0
2.8
3.5
6.5
3.2
4.3
4.6
Prevailing
Wind
Direction
S to SSW
WNW
ESE to SE
WNW
SW
ENE
E
W
S
•
A simple particle counting and sizing procedure is used to determine the
weight percentages of specific particle types present in each hi-vol sample.
A mass per unit area of the filter is determined for each type of identified
particle. The sizing and counting procedure was used to calculate mass con-
centrations of silicate, calcite, hematite, rubber tire fragments, flyash, coal
and cornstarch. Other types of particulates (e.g., ammonium sulfate, auto
exhaust, pollen and magnetic fragments) were determined by nonmicroscopical
procedures, including low temperature plasma ashing, plasma emission spectro-
scopy, scanning electron microscopy and sulfate analysis.
Low-temperature ashing determined the concentration of combustible matter
such as starch, pollens, and carbonaceous vehicle exhausts. It generally
divided the samples between organic and inorganic materials. A chemical anal-
ysis procedure was used to identify the sulfate concentration of the filter.
Lead and vanadium concentrations were analyzed by plasma emission spectroscopy
to determine the concentrations of auto exhaust and oil soot particles in the
sample. The auto exhaust content is taken as 1.5 times the lead concentration,
and oil soot is taken as 39 times the vanadium content. The scanning electron
microscope was used to estimate the concentrations of 18 elements. The results
were helpful in substantiating estimates obtained by other methods. Unfortu-
nately, the variation in elemental content of the filters themselves made
the estimates of doubtful value as an independent estimate.
The analysis of hi-vol vilters revealed that six types of particles
accounted for most of the material (from 70 to 96 percent) collected on
18 filters. The percentages of each type, including silicates, sulfates,
rubber, calcite, cornstarch and slag are shown in Table 6. A summary of the
partciles classified by six types of sources using results from both micro-
scopy and other analyses is presented in Table 7. The mineral category
-84-
-------
Table 6. Predominant Materials Found on Hi-Vol Filters
I
00
Ul
Station 6/9/77 6/21 7/1S 8/2 9/19 91
FD *10 Si 41* SI 40* Su
Su 21 Ca 30 SI
Ru 12 So 13 Ru
Ca 13 Ru 7 Ca
Total 87 Total 90 Total
Ft. McHemy SI 44* Co 41% Su
Su 3O. SI 23 SI
Co 11 Su IS Total
Total 85 Total 79
FO I2Z si 32* SI 54* SI 45* Su
Su 24 Su 16 Sn 22 Si
Ru 14 Ru 9 Ru 11 Ru
Total 90 Total 79 Total 78 Total
25 10/25
36% SI 31%
35 Ca 25
11 SI 17
10 Su 13
92 Total 96
SO* SI 44%
24 Su 18
74 Co 8
Total 70
40*
37
12
89
11/24
Su
Si
Ru
Total
Su
SI
Ru
Total
SI
Su
Ru
Total
35*
32
11
78
45*
32
14
91
35*
32
24
91
12
Ca
SI
Su
Ru
Total
SI
Su
Ru
Ca
Total
Si
Su
Ca
Ru
Total
/12
40*
30
13
8
91
33*
25
12
14
84
37*
17
14
10
78
SI =»illcate«
Su = nlfatei
Ru = rubber
Ca = calcitc
Co = conutarch
SI =flag
Includes all material with * by weight MO only.
-------
Table 7. Summary of Particles Classified by Source Type
i
00
ON
Date
6/9/77
6/21/77
7/15/77
8/2/77
9/19/77
9/25/77
10/25/77
11/24/77
12/12/77
Site
FO 110
Ft. Me
FD 122
FO 110
FD 122
Ft. Me
-FD 122
FD 110
FD 122
Ft. Me
FD 110
Ft. Me
FD 110
FD 122
Ft. Me
FD 110
FD 122
Ft. Me
TSP
Concen-
tration
(ug/ra3)
63
59
60
221
149
154
108
51
38
39
306
118
42
52
41
169
129
85
Low
Tempera-
ture
Ashing
(% Removed)
41%
51%
44%
29%
40%
65%
42%
39%
50%
51%
23%
48%
45%
47%
60%
31%
40%
51%
Source Type
Minerals
%
57
49
54
71
58
30
56
46
39
29
44
50
40
40
35
70
56
48
Mobile
Emissions
%
13
3
15
8
10
5
12
12
13
7
4
8
13
26
16
9
11
13
Combus-
tion
Sources
%
3
4
3
4
7
2
6
5
5
10
4
9
8
3
4
2
4
7
Non-
specific
Combus-
tion %
23
31
24
16
20
16
24
36
40
50
16
23
37
33
47
19
22
28
Corn
Starch
5
2
12
1
1
4
43
3
<0.5
2
2
31
9
3
<0.5
1
1
3
5
Blolooicals
%
2
<0.5
4
1
1
1
1
<0.5
2
1
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
-------
includes both soil and road surfaces. Every sample contains a high percen-
tage of mineral content, but the percentage content is generally highest
when the total particle loading (as indicated by the TSP concentration) is
highest.
As a verification of the validity of the microscopy and supporting anal-
yses, 3 filter halves which were matches to the original 18 were submitted
for analysis. The comparative results are shown in Table 8. The mean dif-
ference between independent analyses of the same filter divided by the mean
of all measurements of a given particulate component is shown in the last
column. This ratio varies from 0 to 0.85 for the 17 components listed with
an overall mean of 0.25. These results are very encouraging regarding the
reproducibility of the analysis.
In summary, the results in Tables 6 and 7 show that mineral sources,
such as roadway kickup, constuction operations and wind erosion, are the
major contributors to TSP loadings in the Baltimore nonattainment area.
Another important source is nonspecific combustion sources which are pri-
marily evident by the presence of sulfates. Particulates from mobile source
tailpipes are generally the third highest component of particulate loadings.
A fourth type of particulate, which is generally small, but is occasionally
the largest type, is cornstarch.
SUGGESTED CONTROL STRATEGIES
Based on the inventory of fugitive sources of TSP and the analysis of
hi-vol filters, it appears that dirt and gravel roads and construction opera-
tions are the major local sources which contribute to the highest ambient
TSP concentrations. There is also a unique and highly specific component of
TSP due to cornstarch, which is identified with a few very specific sources.
Controls for cornstarch can be specific to these sources. Alternative con-
trol strategies for dirt and gravel roads and an estimate of their effective-
ness in reducing emissions were developed from various sources of emission
measurements. The following measures are cited as effective, although their
costs have not been evaluated:
1. Paving will reduce emissions by 99 percent.
2. Applying water or a chemical stabilizer will reduce emissions by
50 percent.
3. Oiling and a double chip surface will reduce emissions by
85 percent.
4. Reduce vehicle speeds.
5. Decrease multiwheeled truck activity on dirt and gravel roads.
The following suggested control strategies for construction sites have been
compiled, but useful estimates of their effectiveness have not been determined:
1. Regularly water exposed soil.
2. Keep handled materials at solid waste disposal sites wet.
3. Cover hauling operations.
4. Reduce the free fall of materials handled.
5. Revegetate waste disposal sites.
-87-
-------
Table 8. Results of Duplicate Microscopy Analysis
Component
Silicates
Calcite
Mica
Clays, humui
Hematite
Carbonaceous tailpipe
exhaust
Rubber tire fragment!
Classy ftyash
Coal fragment!
Oil soot
Fine carbonaceous
particles
Recrystallized sulfates
Const arch
Pollens, spores, conidia
Plant parts
Insect parts
Magnetic fragment*
Total
•« by Weight
Set 1
1st
23
5
-------
About 70 percent of the fugitive emissions in the Baltimore area can be
eliminated by local area controls. This would reduce overall TSP levels by
35 percent in areas where fugitive emissions constitute 50 percent of the
total TSP loading. In areas with high TSP loadings, it is likely that fugi-
tive and other locally controllable emissions (e.g., corn starch and slag)
constitute over 70 percent of the TSP levels. Therefore, the highest TSP
levels can be reduced by more than 50 percent.
A relatively large degree of uncertainty exists concerning fugitive
emission factor estimates which are only weakly supported by measurements.
Questions concerning how representative the particulates caught on hi-vols
are of the particulates in the atmosphere necessarily raise some doubts
regarding how representative these samples are of what really makes up
ambient TSP matter. The number of measurements used to derive emission
factors is small. More data representing many variations in roadway emis-
sions are needed to better estimate the impact of this important source of
TSP loading.
REFERENCES
Bonn, R.C.T., Jr. and C. Cowherd, Jr. 1978. "Fugitive Emissions from
Integrated Iron and Steel Plants." MRI, Kansas City, Missouri.
Cowherd, C., Jr., K. Axetell, Jr., C.M. Guenther, and G.A. Jutze. 1974.
"Development of Emission Factors for Fugitive Dust Sources," MRI,
Kansas City, Missouri.
Sartor, J.D., and G.B. Boyd. 1972. "Water Pollution Aspects of Street
Surface Contaminants." Publication No. EPA-R2-72-081. Environ-
mental Protection Agency.
U.S. Environmental Protection Agency. 1977. "Compilation of Air Pollutant
Emission Factors" (Including Supplements 1 through 7). Third Edition.
Research Triangle Park, North Carolina.
-89-
-------
FUGITIVE EMISSIONS FROM THE BY-PRODUCT COKE OVEN PUSHING OPERATION
Robert B. Jacko
Purdue University
West Lafayette, Indiana
ABSTRACT
An intensive sampling program at the by-product coking facility of a
major midwestern steel company has quantified the release of trace metals
and other participate to the atmosphere resulting from the "pushing" operation,
Emission factors and emission rates were obtained from "clean" (completely
coked) and "green" (incompletely coked) pushes. A number of coke oven push
plumes were sampled isokinetically at a rate of approximately 30 SCFM. The
sampling head with associated temperature and velocity sensors was suspended
over the quench car by means of a tower/boom arrangement mounted on a mobile
laboratory. The cross sectional area of the plume in the horizontal plane
of the sampling was determined by measurements on motion picture films
taken from two vantage points. The mass emission rate of total particulate
was calculated from the particulate concentration, plume velocity and cross
sectional area. Trace metal emissions were quantified by atomic absorption
analysis of all particulate samples for Cd, Pb, Zn, and Cu.
-91-
-------
INTRODUCTION
A total of 66 coke oven pushing plumes have been sampled using a specially
designed mobile laboratory for characterizing the plume parameters of tem-
perature, velocity, shape, and particulate emission. An earlier publication
by this authorl reported on the first 15 samples while this work concentrates
on the complete sample set of 66. This paper will be helpful to those invol-
ved in the quantification and control of coke oven pushing emissions and the
design of fugitive emission sampling techniques.
MEASUREMENT METHODOLOGY
The pushing emission can be classified as a "fugitive emission" of sorts
in that it does not emanate from a confined conduit or stack. As such,
quantification of the particulate emissions has a "high degree of difficulty"
and requires a special experimental design. The basic approach in this study
was to place sampling nozzles and sensors for temperature and velocity in a
stationary plane in the plume over the duration of the push. Meandering of
the plume around the sampling nozzle caused by atmospheric motion reduced the
biasing effect of large variations in plume parameters. Figure 1 contains a
cross-section view of the sampling apparatus aligned adjacent to the coke
oven. Note the mobility of the apparatus and freedom of any supports from
the coke oven structure itself. This provided for unrestricted movement
of the sampling laboratory and no interference with the coke oven production
schedule. The sampling nozzle and instrumentation were at the elevation of
the top of the coke guide and positioned approximately over the center of the
quench car.
Temperatures were measured with two iron/constant and thermocouples at
the end of the boom. A cup anemometer provided plume velocities and parti-
culate samples were taken at a rate of approximately 43 scfm through 8 x 10
inch glass fiber filters.
Cross-sectional plume shape was determined by taking 16 mm black and
white motion pictures in two planes over the duration of each push. Subse-
quent stop-action projector analysis yielded plume cross-sectional dimensions.
The camera set-up is shown in Figure 2. Since the motion picture camera
frame rate was calibrated, plume velocities could be checked by use of a
stop watch. In addition, the gas sample volume was easily corrected for
those instances where the sampling nozzle was not in the plume by examina-
tion of the film.
The particulate concentration was determined by drawing a gas sample at
an approximate rate of 43 scfm through an 8 x 10 inch glass fiber filter.
The gas sample volume was determined with an orifice plate and the sampling
rate was controlled from within the vehicle to achieve an isokinetic rate
as a function of the anemometer readout. The reader is referred to the pri-
mary author for specific hardware details. Knowing the weight of coal
charged to each furnace, an emission factor was subsequently computed for
each push.
-92-
-------
RESULTS AND DISCUSSION
Plume Temperature
Fifteen of the 66 pushes sampled were taken in December and the remaining
in April. Since the steel mill was located in Northwest Indiana, ambient
temperatures were significantly different in each of these seasons. The
December ambient temperatures were close to 0°F and the April temperatures
were close to 40°F. This difference in temperature is partially reflected
in average plume temperatures seen in Table 1.
TABLE 1
Coke Oven Push Plume Temperatures °F
No. of Average Range One Standard
Samples Deviation
DECEMBER
Green 7 157 81-309 77
Clean 8 89 77-108 11
Overall 15 121 77-309 62
APRIL
Green 33 232 109-534 101
Clean 18 117 71-167 25
Overall 51 191 71-534 99
Note from Table 1 that the green push plume temperatures are about twice
that of the clean pushes regardless of the season. This is a result of the
incomplete combustion of the remaining volatiles in the green push coke and
the attendant flames. Also, the range of plume temperatures is quite wide
from 71 F to 534 F as is probably expected but it is interesting to note that
the clean pushes have a significantly lower standard deviation value as com-
pared to green pushes. The December and April standard deviations were 11
and 25 F for the clean pushes respectively and 77 and 101 for the green
pushes. Apparently, the clean push plume temperature is more closely re-
lated to the coke temperature which is relatively constant from push to
push as compared to a green push whose temperature is probably more closely
related to the amount of flame in the plume which can vary greatly from one
green push to another.
The temperature-time history characteristic curve would be of interest to
those designing capture hoods or other control devices. Figure 3 contains
a plot of the three temperatures recorded over the 66 pushes sampled. Note
that the maximum temperature was 534 F and corresponds to a green push. The
average rate of increase in temperature for the three temperature curves
in Figure 3 was 4°F/sec with the maximum occurring about 40* of the total
time into the push.
Plume Velocity
The buoyancy of the push plumes result in a vertical velocity which was
measured with a cup anemometer whose cups were located in a vertical plane.
Table 2 contains the average statistics.
-93-
-------
TABLE 2
Push Plume Velocity, ft/sec
No. of Plume Velocity One Standard
Samples Average Range Deviation
Green 40 15.7 6.8-21.3 3.9
Clean 26 14.1 6.2-20.9 3.3
Overall 66 15.0 6.2-21.3 3,7
From Table 2 it is seen that the average and range of plume velocities
are not greatly different for the clean or green pushes. Note also that one
standard deviation about the average value is not exceptionally large and
reflects the relatively consistent plume velocity from push to push. The
overall range, however, is quite wide from 6.2 to 21.3 ft/sec; a 3.5 to 1
change.
Figure 4 shows a typical green push velocity-time history plot. Note
the erratic nature of the trace reflecting the billowing nature of the
plume. Also, the peak velocities occur about 15 seconds in the push, are
sustained for 15 seconds and decay in about one-half the time it took to
reach peak velocity.
Plume Cross-Sectional Shape
Analysis of the two motion picture film records which were shot at approx-
imately 90° to one another (Refer once again to Figure 2) allowed the plume
cross-sectional shape to be defined. The procedure involved the use of a
stop action projector and measuring the plume width frame-by-frame for each
90° view. Table 3 shows the resulting plume dimensions.
TABLE 3
Plume Cross-Sectional Shape
Plume Shape, ft. Range One Standard
Average Deviation
Length "A" 18.2 10-29 3.8
Length "B" 17.8 10-31 5.6
Overall 13.0 10-31 4.7
Ratio of "A"/"B" 1.02 1.0-0.94 0.3
Note that the two diameters are very nearly equal with a ratio of 1.02. This
Indicates that a circular plume cross-sectional area is a reasonable model
to use for estimation of the emission parameters.
From Tables 2 and 3 the plume volumetric flow rate can be estimated on an
average basis. Table 4 shows the plume actual volumetric flow rate as cal-
culated from the statistics in Tables 2 and 3.
-94-
-------
TABLE 4
Plume Actual Volumetric Flow Rate
Average
One Standard Deviation
about the average
Observed overall range
ft3/m1n
229,000
94,000-454i000
61,000-920,000
Table 4 Indicates the average volumetric flow rate to be 229,000 actual
ft3/m1n. Using one standard deviation value to both reduce and increase the
plume diameter and velocity a one standard deviation range on volumetric
flow rate was calculated. This range as seen 1n Table 4 1s 94,000 to
454,000 actual ft3/m1n. The observed minimum and maximum values are also
shown in Table 4.
These volumetric flow rates are of interest to those Involved 1n the
design of control hardware and yield an estimate of gas flow rate that must
be handled. However, keep in mind that these measurements were made 1n the
open atmosphere where gaseous and particulate diffusion 1s relatively un-
limited. If the control hardware confines the push plume, significantly
lower plume volumes will probably result. One reason for this 1s the re-
striction on plume diffusion and dilution. For green pushes, a confined
duct limits the amount of oxygen available to the remaining coal volatiles
thus Inhibiting combustion and therefore temperatures which results in
lower volumetric flow rates.
Particulate Emissions
The concentration of partlculates measured in the push plume is seen
1n Table 5.
TABLE 5
Push Plume Total Particulate Concentration*
One Standard
No. of Range 3 Average- Deviation.
Samples grams/m grams/m grams/m
Green
Clean
Overall
39
25
64
0.22-16.0
0.07-5.0
0.07-16.0
3.3
1.8
2.7
2.9
1.2
2.5
*per standard conditions of 70 F, 1 atm
As expected, the green pushes have a significantly higher particulate 3
concentration relative to a clean push. The green pushes were 3.3 grams/m
as compared to 1.8 for clean. Note the large standard deviation for both
the green and clean pushes. The coefficient of variation (average •» std.
dev. x 100) for green and clean pushes 1s 88% and 67% respectively.
Emission rates and factors of total partlculates are shown in Table 6.
The emission factors are based on steel mill records of coal charged for
each push sampled and Indicated an average oven charge of 30,000 Ib-coal.
-95-
-------
TABLE 6
Summary of Mass Emission Rates and
Emission Factors for Total Participates
No. of Range Average One Standard
Samples Deviation
Mass Emission
Rate (grams/sec)
Green 39 21-1495 382 335
Clean 25 11-377 138 79
Overall 64 11-1495 287 291
Emission Factor
(Ib/ton-coal)
Green 39 0.09-9.0 2.0 19
Clean 25 0.05-2.0 0.7 0*4
Overall 64 0.05-9.0 1.5 l.*6
In an earlier paper , the average mass emission rate of the first 15
December samples was reported. The values were 407 grams/sec for green
pushes and 147 grams/sec for clean pushes. It is interesting to note that
these values compare favorably with those reported for the total sample
set of 64. In other words, the initial 15 samples when reported were
reasonably representative of the larger sample mean composed of many more
samples.
The emission factors expressed as Ib-part./ton-coal charged range from
0.05 to 9.0, a change of 180 to 1. This is a relatively wide variation and
is reflected in the large standard deviation values. The resulting "coeffl
cient of variation" for green and clean pushes is 95% and 57% respectively
This indicates the clean push emission factors are not quite as variable
as the green pushes and reflects the degree of flame and remaining volatile
in green push. Ies
The average emission factors are of interest and were 0.7 Ib/ton for the
clean pushes and 2.0 Ib/ton for the green pushes. The overall value was
1.5 Ib/ton. The emission factor reported in "AP-42" for total participates2
for coke "discharging" is 0.6 Ib-part./ton-coal charged.
The distinguishing feature of the emission factors is the wide variability
However, whether or not one will experience a green or clean push appears to
be directly related to the individual coke oven. In this sampling program
the same ovens were repeatedly sampled. Upon analysis of the data it
appeared that a specific oven would either yield a clean push or a green
push. Table 7 contains a listing of the ovens that were sampled three or
more times and how many of the samples were classified clean or green.
-96-
-------
TABLE 7
Characterization of Coke Push Plumes by Individual Ovens
Oven No. of Times No. of Green No. of Clean
Number Sampled Plumes Plumes
49 5 05
52 3 03
54 4 13
57 5 41
59 5 50
52 4 40
64 4 04
67 3 30
69 5 50
72 4 40
It is interesting to note that ovens #49, 52, 54 and 64 were sampled 16
times and all were clean pushes except one. This contrasts to the record
of ovens #57, 59, 62, 67, 69 and 72 which were sampled a total of 26 times.
Of these, all were green pushes except one. These results suggest that a
clean or green push is a function of the oven from which it came. It may
be related to the condition of the internal heat transfer surfaces and to
the time between rebuilds. It should be stated that no correlation whatso-
ever was found between the coking times and the particulate emission factors.
SUMMARY AND CONCLUSIONS
For the 66 coke pushing plumes sampled the overall average parameters
observed were:
1) temperature 176 F
2) velocity 15 ft/sec
3) shape circular w/ratio front
to side - 1.02
4) volumetric flow rate 229,000 ACFM
5) particulate concentration 2.7 grams/m3
6) particulate emission rate 287 grams/sec
7) particulate emission factor 1.5 Ib/ton coal charged
Specific significant differences were measured for green and clean pushes.
Generally, green pushes were greater in all of the above parameters by at
least 2 to 1 except the velocity which was approximately the same. The one
dominant facit regarding all of the pushes sampled was the wide variability
1n plume parameters not only between green and clean but within each of these
categories. One standard deviation values expressed as a percentage of the
mean value approached 100% in some cases.
The data suggested that a green or clean push is a strong function of the
oven from which it came. Repeated sampling of the same ovens showed a trend
that some ovens produce consistently clean pushes while others yield green
pushes. Oven maintenance and condition may be the important factor here.
-97-
-------
REFERENCES
1. R.B. Jacko, D.W. Neuendorf and J.R. Blandford, "The By-Product Coke Oven
Pushing Operation: Total and Trace Metal Participate Emissions", APCA
paper no. 76-12-2 Portland, Oregon, 1976.
2. "Compilation of Air Pollutant Emission Factors", U.S. EPA, 2nd edition
1973. '
ACKNOWLEDGMENTS
The research described in this paper was supported by the National Science
Foundation Research Applied to National Needs (RANN) under Grant No. GI-
35106 and the Environmental Protection Agency under Grant No. T900196.
-98-
-------
CHARGING
BY-PRODUCT CAR
COLLECTION
MAINS
DOOR
MACHINE
PUSHER
CONTROI
CAB
RAM
PUSHING
MACHINE
COKE
6UIDE SAMPLING
HEADS
BOOM BLOWERS
COKE OVEN
•••••••
RECUPERATOR
QUENCH
CAR
WASTE
HEAT FLUE
Figure 1 VIEW OF COKE OVEN AND SAMPLING VEHICLE PLACEMENT
-99-
-------
Coke Oven
Push Plume
o-'
Camera
B
Quench Car
6
Camera
A
Figure 2 TOP VIEW OF PUSH PLUME MOTION PICTURE CAMERA SETUP
-100-
-------
600
500
400 —
I
300
2OO
100
10
15 20 25
TIME INTO PUSH SECONDS
30 35
40
Figure 3 COKE OVEN PUSH PLUME TEMPERATURE -
TIME HISTORIES FOR THE 3 HIGHEST TEMPERATURES MEASURED
-101-
-------
10
10
15 20 25
TIME INTO PUSH SECONDS
30
35
Figure 4 COKE OVEN PUSH PLUME VELOCITY-
TIME HISTORY FOR A TYPICAL GREEN PUSH
-102-
-------
ESTIMATING DUST PRODUCTION FROM SURFACE MINING
by
Karl F. Zeller, Douglas G. Fox, and William E. Marlatt-
— Bureau of Land Management, Denver, Colo.; USDA Forest Service, Rocky
Mountain Forest and Range Experiment Station, Fort Collins, Colo.; and
Department of Earth Resources, Colorado State University, Fort Collins,
Colo., respectively.
-103-
-------
ABSTRACT
The Federal Land Manager has a responsibility to evaluate the effects of
activities he conducts, or permits to be conducted under his administrative
jurisdiction, on the air resources of the United States. The operation of
surface mining is among these activities. Regulatory classification of surface
mines suggests the need for studies of the emission of particulates from them.
In particular, two aspects of the emission must receive attention; namely,
the chemical/physical characteristics of the particulates, in order to deter-
mine what portion of the total particulate burden is "fugitive dust" and what
portion is industrial pollution, and secondly, the emissions resulting from
specific mining activities such as crushing, screening, dragline overburden
removal, stockpiles, etc., must be determined.
This paper describes a preliminary study designed to characterize these
emission factors for surface coal mines in and around Craig, Colorado. The
study includes consideration of such off-site impacts as deposition downwind
of the site, and effects on visibility. The approach includes a general
characterization of source configuration and the dispersion pattern, which
should lead to an improvement in dispersion modeling for surface mining.
Modeling approaches used to calculate downwind concentrations and surface
deposition will be reviewed, leading to suggestion of a preferred model.
Finally, the main data collection phase of this study will be described.
The result will be an improved set of emission factors for surface mining.
Future plans to investigate the effects of BACT will also be described.
-104-
-------
ESTIMATING DUST PRODUCTION FROM SURFACE MINING
I. INTRODUCTION
Wind blowing over the surface of the earth often generates enough shear
stress to mobilize particles in the first few centimeters of the soil.
Especially during drought, the amount of material introduced into the atmos-
phere by wind can be extremely large. The magnitude of the effect depends
strongly upon soil type, vegetative cover, and soil moisture content. Natural
surfaces, particularly in arid and semiarid regions, develop a "shear-proof"
layer that resists wind erosion. Any activity that disturbs the "shear-proof"
layer will then cause an increase in blowing dust.
Besides surface mining itself, construction and use of roads, blasting,
and overburden removal all increase dust emissions from the surface. How
increased mining in the semiarid regions of the Western United States can be
conducted with minimal effect on the atmospheric environment is particularly
important now that the Clean Air Act Amendments of 1977 require that certain
federal lands in the Western United States be managed so that virtually no
air quality degradation occurs there.
Recent regulatory proposals from the Environmental Protection Agency
(Fed. Reg. ^3., 118, 19 June 1978, p. 20395) clarify the application of the
concept of PSD (prevention of significant deterioration) to surface mining and
discuss the question of fugitive dust. These regulations affirm that surface
mining is a stationary source subject to PSD review, but that fugitive dust
(defined as particles of native soil uncontaminated by pollutants from
industrial activity) will be excluded from such consideration. This exclusion
will be reviewed periodically by EPA pending more information, and when new
sources are planned, "...the burden of showing to what extent emissions from
the proposed source or modification would be made up of fugitive dust rests
with the applicant". Fugitive dust is not excluded from estimates of poten-
tial emissions (a test to determine if PSD review is needed). Any mine that
emits more than 250 tons of particulates per year, including fugitive dust,
must apply BACT (Best Available Control Technology).
This regulatory classification underlines the need for studies of
particulate emission from surface mines. Two aspects of the emission must
receive particular attention, the chemical/physical characteristics of the
particulates (to determine what portion of the total particulate burden is
"fugitive dust" and what portion is industrial pollution) and emissions
resulting from specific mining activities (such as crushing, screening,
dragline overburden removal, and stockpiles). The effectiveness of BACT
must also be evaluated.
The Bureau of Land Management (BLM) is responsible for evaluating the
impact of mining its leased minerals. To evaluate air pollution from surface
coal mining in the area around Craig, Colo., the BLM is supporting a coopera-
tive research effort between the Rocky Mountain Forest and Range Experiment
Station and Colorado State University.
-105-
-------
The major activity in the study will be field measurement to determine
type, size, and chemical separation of particulates at various locations
downwind from the mine source. Meteorological conditions will also be
measured so that modeling can be used to relate the particulate measurements
back to emissions and the amount of activity. The objective will be to
determine emission factors (i.e. emissions as a function of incremental
activities) for surface mining.
To characterize the impact of surface mining on the environment, we will
conduct two additional activities. One will establish a deposition monitoring
site near Craig to provide information about wet and dry removal of pollutants
from the atmosphere. This site will be a component of NC-141, USDA Regional
Atmospheric Deposition Project, which will be collecting similar data
throughout the United States. Data collected at Craig can be evaluated
relative to the entire country.
The second, somewhat speculative activity is related to visibility.
Visibility has been a major issue in public concern about air quality for many
years. The impact of surface mining on visibility is largely unknown. The
assumption that most of the particulates generated from mining are large
enough to cause scattering of visible radiation will be tested in a study of
background visibility in Northwestern Colorado.
I I. CURRENT STATE OF KNOWLEDGE
The latest numbers for emission factors from surface mining, taken from
a recent EPA study (EPA-908/1-78-003, Survey of Fugitive Dust From Coal Mines,
by PEDCo Environmental, Inc.) are shown in Table 1. While these numbers give
some indication of the particulate loading from mining activities, they do
not segregate soil from coal particulates, nor are the data sufficient to
determine site-specific emissions numbers accurately. Some shortcomings of
the numbers in Table 1 are that there is no consideration of such factors as
soil surface conditions, or moisture content, and that the numbers are based
upon application of a Gaussian model to the entire source area.
Recent work has characterized the background dust from the semiarid
Southwestern United States (Gillette et al. 1978). Although the soils are
somewhat different, climate conditions in the Craig vicinity are similar, and
one would expect roughly similar atmospheric loading.
The results of this work that are most applicable to our study concern
particle size distributions. Only small particles (less than 10 \im) are
found at altitudes above 1 km, and they are associated with major dust storms.
Sampling at ground level (1 m) showed a dust size distribution similar to the
aloft aerosol along with a large concentration of particles in the ^0-80 ym
diameter range. Clearly, the fine aerosols are well-mixed through the lower
troposphere, while the large particles remain near the surface. For this
reason, the major long range impact of surface mines is the aerosol produced
in the size range below 20 vim. This research also suggests that accurate
measurement of dust size distribution at one location above the surface may
suffice to characterize the total dust produced by the mining activity.
-106-
-------
However, the PEDCo report determined particulate loading by measuring
concentrations with high volume samplers on or near (1-3 m above) the ground.
Such sampling might overestimate the amount of dust exported from the site,
unless a method of simulating the dispersion and deposition were incorporated
into the analysis.
Modeling was done in the PEDCo study to relate measured concentrations
back to emissions from the source. The model used is the standard Gaussian
distribution,
y z
where x = concentration, g/m
a « vertical standard deviation of plume concentration distribution,
m
a = horizontal (perpendicular to plume) standard deviation, m
u = mean wind speed, m/sec
a and a are determined from the equations
ay = c(x + XQ) (la)
az = a(x = x0)b
where a, b, and c are given in Table 2 for different stabilities, and XQ is
determined by observation of the initial distribution of pollutant mass.
Since it is not appropriate to model dust concentrations without simula-
ting deposition, the PEDCo study used a simplified version of the source
depletion method originally developed by Van der Hoven (1968). The source
depletion model deals with deposition by reducing the effective source
strength. Two other modeling approaches, the tilted plume and a surface
depletion model, are available.
Tilted Plume
The tilted plume model is a straightforward combination of the Gaussian
diffusion formula with a steady particle mass fall velocity. Assuming that
the particles in suspension can be described by a discrete size distribution,
the fall velocity for particles in the j compartment is
X = ZXj and Vd - ZVdj
-107-
-------
(given for a limited set of Reynolds numbers, gdj2/l8v). The trajectory of
the particles is tilted at an angle, a, given by
— i.e. a. - ,a,.
u j u
In order to simulate removal from a plume assumed to be released (or have a
maximum plume rise) at height h, then simply replacing z by
Vdi
h - x Sin a. « h - x
so the area source expression becomes
* = ash- exp [-J(a-)2 - * 1-^)1 (2)
y z L y
The deposition flux is given by
Wj = VdjX (3)
This model assumes a completely absorbtive surface with no provisions for
particle "bounce" or resuspension.
Source Depletion
The source depletion model simply accounts for a removal by replacing
the emission by an effective emission, Q*, where
Q* = £ Q*
in equation (1). Then employing conservation of mass, one can relate the
change in effective emissions to the deposition flux
dQ* /• /•
-di = / Wjdy = - / Vdj X (x,y,o) dy (k)
Since this integral has been evaluated for a number of cases, the depletion
factor Q* can be calculated by ratio.
Surface Depletion
A more sophisticated and physically realistic model allows the removal
of material directly at the encountering surface. Within the Gaussian
approximation, this is accomplished by including negative source distributions
along with the positive source. Horst (1977) has developed this model by
assuming, for the case of an area source, that deposition will occur at a
point (^.m.zj) which causes a reduction In concentration at (x,y,z) downwind,
equal to r , '
j X U,m,zd) d £ dm) J (x - i,y - m,z) (5)
-108-
-------
In order to calculate concentration, one sums the source contribution and the
total deposition (integration of equation (5) over the total domain of deposi-
tion), i.e.
00 V
n f C V
X(x,y,z) = S G(x,y,h) - / / -Hi x(£,«.zd) G(x-£,y-m,z) d I dm (6)
-00 0
This has the advantage of allowing specification of the deposition domain, of
differential deposition rates, and similarly, of a resuspension coefficient
should such a mechanism be needed.
Cole and McVehil (1977) applied these three models to a surface mine for
estimating dustfall rates and particulate concentrations. For the material
they were concerned about they estimated that the dj (distribution of particle
diameters) was approximately kO% with d > 44 ym, and 20% with d < 20 ym. After
comparison with experimental results, the most successful model they employed
proved to be an integrated version of equation (6). Their results further
indicated that some re-entrainment of already deposited dust was necessary to
properly predict dustfall rates. In view of this result, our major effort
will be oriented toward properly defining the parameters for equation (6),
which will provide the most accurate simulation of reality.
Ml. FIELD PROGRAM
In order to assess the impact of surface mining on the atmospheric
environment, we are planning a field exercise at a surface mine in the vicinity
of Craig, Colorado (fig. 1). We plan to initiate both short-term and long-term
studies. The short-term studies will include meteorological characterizations
of the immediate source area and detailed measurements of the mass of particu-
lates suspended in the atmosphere and deposited on the ground surface. Data
will also be collected to determine specific mining activities and amounts of
material processed per unit time. Meteorological data will be collected on
site, using a tethered balloon system and an array of surface stations.
Hopefully the study site will be located near a meteorological data tower
that has long-term data that will help in planning the timing of our intensive
data collection periods, as well as assessing the general applicability of our
results.
Short-term studies will also include a particulate emission study at the
Craig, Colorado Utah International Coal Mine. A brief description of the field
research plan for this study follows.
Delay in starting operation of the Colorado Ute Yampa power plant has
resulted in a large stockpile of coal in open fields south of Craig, Colo.
Currently, some of this stockpiled coal is being crushed and sold by Colorado
Ute. A study of this stockpile area is proposed. Downwind distribution of
fugitive dust around the area will be investigated as a function of wind,
atmospheric stability, particle size, material composition, and coal handling.
The objective is to determine how far downwind significant amounts of coal and
overburden remain airborne under different wind and stability conditions.
-109-
-------
Meteorological Measurements
Wind and air temperature measurements aloft are available from a tower
near he study site operated by Sterns-Rodgers under contract to Colorado Ute.
Three-dimensional wind and turbulence data will be taken using Gill u,v,w
anemometers mounted on towers at 200 and 600 m downwind from the aforementioned
coal stockpile. Also, a tethered balloon sounding system will be available
to provide additional information (wind speed and direction, temperature and
humidity) at any elevation up to 600 m above the surface. As timing and
logistics permit, a monostatic acoustic sounder will be deployed to measure
inversion depth in the area.
Particulate Measurements
Dustfall measurements will be taken, using Petri dish collectors spaced
in the grid pattern illustrated in figure 2, at intervals no larger than 320
m. Close to emission sources, collectors will be located at 160 m intervals.
Petri dishes will be placed at approximately the canopy foliage maximum
elevation (15 cm). They will be exposed long enough to allow collection of
an analyzable sample, but hopefully for a short enough period that stockpile
activity (crushing, etc.) remains constant. The Petri dishes will be fitted
with an artificial turf material to improve their collection efficiency and
reduce loss from particle bounce. In order to provide a cross reference,
selected vegetation samples will be taken to help establish the validity of
the Petri dish collectors. Total mass, size distributions, and, via micro-
scopic observation, percentage of coal/surface soil, will be determined from
the dustfall collectors. Stacked filter samplers (Cahill et al. 1976) will
be used to measure mass and size distribution of suspended particulates.
These stacked filter samplers are sufficiently portable to be operated at
various levels on the meteorological measurement towers. Each tower will also
contain a standard hl-vol sampler at Its base, as well as a precise particle
sampler at one tower. The specific configuration of these ambient particle
samplers will be rearranged according to tracer studies indicating the
general air motions.
The study described will be conducted for 1 or 2 days estimated to
provide maximum emissions (strong winds, dry soils) during fall and/or spring
of 1979.
Long-term concerns include studying the impacts on atmospheric deposition
and visibility. We will establish, in conjunction with the Bureau of Land
Management's Craig District Office, a wet/dry deposition sampler as a part of
the NC-lAl Atmospheric Deposition Project. The NC-141 project Is a national
effort to collect background data describing the chemistry of deposition In a
biologically relevant format. Concerns associated with deposition center
mainly on any changes in the basic atmospheric processes. The atmosphere
provides most of the nutrients required for ecosystem survival (Fox 1977).
The atmospheric supply function must be analyzed and protected, particularly
with regard to revegetation.
Visibility represents a major concern for the federal land manager. The
program of visibility research and monitoring is only now being developed.
-110-
-------
Protection of visibility from deterioration by human caused sources has
been established by Congress as a national goal. The majority of existing
information points at sulfates and nitrogen oxides as the primary cause of
reduced visibility (Latimer and Samuel son 1978) in point source plumes as
well as on a regional scale (Trijonis and Yuan 1978).
Clearly, any suspended particulate matter smaller than 3 ym in diameter
will result In a reduction of visual range, so it Is significant to determine
this size fraction loading to the atmosphere as a result of mine activities.
We are planning to establish a long-term monitoring site at Craig (figure 1)
where visibility can be recorded along with background aerosol. The require-
ments for visibility monitoring include three parameters: visual range,
contrast, and chromatic!ty. The site will include a nephelometer which
measures light scattering at the measuring point, thus giving a local measure
of the visual range. It will also include a telephotometer measuring both
contrast and color alteration of distant targets (George and Zeller 1973).
Hopefully, this will be a long-term study so that visibility data can be
compared before and after the power plant goes on line.
IV. DATA ANALYSIS
Since a vast amount of data will be generated by these studies, it is
important to address the problems of analysis. The objective, as stated, for
the short-term studies will be the determination of improved emission factors
for mining activities. Basically, two types of data will be collected from
the short-term field studies—meteorological and particulate. The meteorol-
ogical data will be used to establish the windspeed and direction, and
stability class for each particulate event which Is measured. The models will
be applied to all the particulate data utilizing a family of particle size as
appropriate. By processing all the dustfall and ambient concentration data
in this fashion, we should be able to improve the emissions factors in Table 1
The long-term studies will provide data for Input into national programs.
Atmospheric deposition data will be analyzed through the NC-141 system.
Various tabulations and graphical comparisons of data from all the 40-plus
sites distributed nationwide will be provided on a routine basis. These data
will allow immediate comparison with other sites.
Visibility data may prove to be the most difficult to analyze. Our hope
is to develop a data base for use in evaluation of all aspects of visibility.
The data will initially be used to determine any inputs on visibility result-
ing from mining operations. Impacts due to the growth In town of Craig and
due to operation of the power plant also will be investigated.
REFERENCES
Cahlll, T. A., L. L. Ashbaugh, J. B. Barone, R. A. Eldred, P. J. Feeney,
R. G. Flocchini, C. Goodart, D. J. Shadoan, and G. W. Wolfe. 1977.
Analysis of respirable fractions in atmospheric particulates via sequential
filtration. APCA Journal 27:675-677.
-Ill-
-------
Cole, C. F., and G. E. McVehll. 1977- Modeling participate concentrations
and dustfall rates at the White Rocks sand and gravel mine. p. 272-275-
JJT_ Proc. Fourth Int. Clean Air Congr., Union of Air Poilut. Contr. Assoc.,
1,088 p. [May 16-22, 1977, Tokyo, Japan.]
Fox, Douglas G. 1977. Precipitation quality and its effects on stream water
quality and the forest in general, p. 103-129. Jn. Proc. "208" Symposium
Non-point sources of pollution from forested land. [Oct. 19~20, 1977,
South. 111. Univ.]
George, David H., and Karl F. Zeller. 1973. Visibility sensors in your air
quality program. Presented at Second Joint Conf. on Sensing of Environ.
Pollutants, Instru. Soc. of Amer., Pittsburgh, Pa. [Dec. 10-12, 1973,
Washington, D.C.]
Gillette, D. A., R. N. Clayton, T. K. Mayeda, M. L. Jackson, and K. Sridhar.
1978. Tropospheric aerosols from some major dust storms of the southwestern
United States. J. Appl. Meteorol. 17:832-845.
Horst, Thomas W. 1977. A surface depletion model for deposition from a
Gaussian plume. Atmos. Environ. 11:41-^6.
Latimer, D. A., and G. S. Samuelson. 1978. Visual impact of plumes from
power plants: A theoretical model. Atmos. Environ. 12:1^55-1^65.
Trijonls, J., and K. Yuan. 1978. Visibility in the southwest: An explora-
tion of the historical data base. U.S. EPA Rep. No. EPA 600/3-78-039
[prepared by Technology Service Corp., April].
Van der Hoven, Isaac. 1968. Deposition of particles and gases. j_n_ Slade,
David H., ed. Meteorology and Atomic Energy. U.S. Atomic Energy Comm.
Pub. No. TID-24190, Oak Ridge, Tenn.
-112-
-------
LIST OF SYMBOLS
X • concentration, gm/m
a = vertical standard deviation of plume concentration, m
a = horizontal (perpendicular to plume) plume concentration, m
u = mean wind speed, m/sec
x,y,z = orthogonal coordinates along wind direction, perpendicular in the
horizontal and vertical, respectively.
£,m - replacement variables for coordinates
a,b,c = coefficients (see Table 2, equation 1)
XQ - a virtual source determined visually as the size of the initial
pollutant mass distribution, m
V. « effective deposition velocity, m/sec
=a gd2/l8v for the Stokes range of particles
d » particle diameter, m
g « 9.8 m2/sec = acceleration of gravity
v = kinematic viscosity of the air - .15 m2/sec @ 20°C
h * effective plume height, m
W » deposition flux, gm/m2-sec
Q » source emission, gm/sec
-113-
-------
Table 1.--Emission factors for individual coal mining operations
Operation
Units
Mine
A
N.W.
Colo.
Dragline lb/yd3 0.0056
Haul roads Ib/veh-mi
w/watering 6.8
no watering
B
S.W.
Wyo.
0.053
13.6
17.0
c
S.E.
Mont.
0.0030b
3.3b
D E
Cent. N.E.
N.D. Wyo.
0.021
4.3
11.2
Shovel/Truck
loading
coal
overburden
Ib/ton
0.011*
0.007 0.002
0.0035
Blasting
coal
overburden
Truck dump
bottom dump
end dump
overburden
Storage pile
Drilling
coal
overburden
Fly-ash dump
Train loading
Topsoil removal
scraping
dumping
Front-end
loader
•«~^^^^^^^M«MIIIBMHM~V_^^B_B^^^^H«
lb/blast
25.1 78.1 72.4
1690 14.2 85.3
Ib/ton
0.01A 0.020 0.005 0.027
0.007.
0.002
Ib/acre-hr 1.6 u, where u is in m/sec
Ib/hole
0.22
1.5
Ib/hr 3.9
Ib/ton 0.0002
lb/yd3
0.35
0.03
Ib/ton
0.12
30n1y veh-mi by haul trucks; travel by other vehicles on haul roads (pickup
trucks, ANFO trucks) is incorporated into these values.
These values were all noted to be somehow atypical and should not be used
without first determining .the limitations to their applicability described
in this paper.
-114-
-------
Table 2.—Coefficients for the application of a line source Gaussian model to
near source dispersion. These values are only appropriate within
10 km of the source.
Stability
class
A 0.183 0.945 0.280
B 0.147 0.932 0.197
C 0.112 0.9'l5 0.132
D 0.0856 0.870 0.086
E 0.0762 0.837 0.065
F 0.0552 0.816 0.042
-115-
-------
POWER PLANT
j UTAH INTERNATIONAL MINE
AND EXPERIMENT SITE
Figure 1.—Map of vicinity of Craig, Colorado. Utah International Mine,
Colorado Ute Power Plant, and proposed experimental site are
located as shown.
-116-
-------
xx x= tower, hi.vol
0.
• • • • •
• • • • •
Figure 2.--Approximate location array for samplers around stockpile.
-117-
-------
OCTOBER 24, 1978
Tuesday Morning - SESSION III
Session Chairman: Lyle D. Randen
Environmental Engineer
AMAX Coal Company
Tuesday Afternoon - SESSION IV
Session Chairman: James A. Dorsey
Chief, Process Measurements Branch
EPA/IERL-RTP
-119-
-------
COMPARISON OF PREDICTED AND OBSERVED
EFFECTS OF FUGITIVE DUST FROM COAL OPERATIONS
by
Alexis W. Lemmon, Jr., Ronald Clark,
and Duane A. Tolle
BATTELLE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
-121-
-------
ABSTRACT
In the vicinity of facilities which prepare, process, or utilize coal,
particulate emissions in the air can be measured. This fugitive dust may be
detrimental to the environment where the quantity of particulates is great
and the concentrations of trace elements in the dust are at levels sufficiently
high to be of concern. In the reported study, particulate concentrations were
measured experimentally in the vicinity of a mine-mouth electric generating
station. Operations at this major power complex included coal mining, handling,
transport, and storage as well as construction and normal activities associated
with power generation. Concentrations of elemental components of the collected
particulates have been used to predict environmental effects; the predictions
are based upon currently available values of estimated permissible concentra-
tions. Finally, the predictions of environmental effects are compared with
results of a brief field reconnaissance of vegetation growing in the vicinity
of the power complex.
-122-
-------
INTRODUCTION
Battelle's Columbus Laboratories has contracted with the U.S. Environ-
mental Protection Agency (U.S. EPA) to perform a comprehensive environmental
assessment of physical and chemical coal cleaning processes. The broad goal
of this program (Contract No. 68-02-2163) is to establish a strong base of
engineering, ecological, pollution control, and cost data which can be used
to determine which coal cleaning processes are most acceptable from the
technological, environmental, and economic viewpoints.
In order to obtain the field data necessary for the overall program,
Battelle has undertaken a sampling and analysis program designed to identify
the combinations of coal cleaning processes and environmental conditions that
are most effective in reducing the total impact of coal use on the environ-
ment. This is being accomplished through the characterization of process
and effluent streams from a variety of coal cleaning facilities and their
associated coal transportation, storage, and refuse disposal areas. As part
of this program, fugitive emissions in the vicinity of these operations were
measured and evaluated in terms of their potential effects on the surrounding
environments.
Specific Objectives
The recent construction of an advanced coal cleaning facility at a power
complex near Homer City, Pennsylvania, provided a unique opportunity to obtain
environmental data before operation of the facility for potential comparison
with similar data to be obtained after operation would begin. Battelle
conducted a series of pre-operational, multimedia, grab-sampling campaigns in
a study area which included this facility, in order to document the abundance
or concentrations of selected key parameters. These data were used to evalu-
ate the air, water, and biological quality in the study area. The pre-
operational environmental studies, while not sufficiently long-term to be a
true baseline analysis, were conducted prior to operation of the cleaning
plant as a reference point for future comparisons.
The objective of this paper is to compare the data obtained with the
values listed in the Multimedia Environmental Goals (MEG) documents prepared
for the U.S. EPA by Research Triangle Institute (Cleland and Kingsbury, 1977a
and b). The MEG methodology was developed to meet the need for a workable
system of evaluating and ranking pollutants for the purpose of environmental
assessment of energy-related processes.
MEG values have been estimated for 216 pollutants by extrapolating various
toxicity data by means of simple models. For most of these pollutants, maximum
values have been estimated for each of the three media (air, water, and land).
For each of the three media, separate maximum values have been estimated which
are considered acceptable in preventing negative effects on (1) human health
and (2) entire ecosystems.
-123-
-------
The MEG values which are particularly appropriate for comparison with
the environmental monitoring data Battelle collected in the area near Homer
City are those termed estimated permissible concentrations (EPC's). EPC's
are the maximum concentration of a pollutant which presents no hazard to man
or biota on a continuous long-term basis. These EPC values are considered
acceptable in the ambient air, water, or soil, as appropriate, and do not
apply to undiluted effluent streams. The ambient application of EPC's
corresponds to the ambient type of sampling conducted by Battelle prior to
operation of the Homer City Coal Cleaning Plant. Specifically, particulate
ambient concentrations were measured and compared with appropriate EPC's.
A second type of MEG values considered in this paper are minimum acute
toxicity effluent (MATE) values. MATE's are concentrations of pollutants in
undiluted effluent streams that will not adversely affect those persons or
ecological systems exposed for short periods of time. These are appropriate
for evaluating the potential hazards of particulates deposited on the ground
and on vegetation in the study area.
Description of the Study Area
Nearly all of Battelle's environmental monitoring was conducted within a
study area that can be approximately bounded by a circle 4 miles (6.4 km) in
diameter. The advanced coal cleaning plant in the center of the study area
is about 2 miles (3.2 km) southwest of Homer City, Pennsylvania.
The six major habitat types within the study area are hardwood forest,
coniferous forest, cropland, grassland, water bodies, and areas of industrial
development. The forested areas are primarily hardwoods, dominated by oak
and hickory. Isolated pockets of pine are present as plantations rather than
naturally-occurring species. Cropland is extensive in the study area, including
contour and strip-cropped fields of corn, wheat, and hay. Grasslands include
those areas which are presently grazed and those areas which were previously
grazed or farmed and are now in a transition stage toward becoming a forest.
Almost the entire study area is on top of deep mines; nearby are abandoned
or active strip mines. Homer City, Pennsylvania, is immediately adjacent to
the study area on the Northeast, and Indiana, Pennsylvania, is only 5 miles
(8.0 km) north of Homer City. During Battelle's sampling campaigns, both the
coal cleaning plant and the refuse disposal area for that facility were under
construction in the study area. Finally, the study area includes the Homer
City Power Station with its associated coal storage, water treatment, and
waste disposal facilities.
The Homer City Station is one part of an integrated power complex which
includes two deep coal mines; coal cleaning, storage, and transport facilities;
power generation facilities; and waste disposal and treatment facilities
(Figure 1). Coal used at the Homer City Station comes from the two dedicated
deep mines in the power complex, as well as coal hauled by truck from other
mines. Solid refuse from power complex activities is deposited in three
different types of disposal areas, including an ash disposal area, mine waste
or "boney" piles, and the cleaning plant refuse disposal area.
-124-
-------
Ash Disposal Area
Mine Drainage Treatment Pond
Helvetia Boney Pile
Coal Cleaning Plant
Coal Storage Pile
Power Plant
Industrial Waste Treatment Plant
Helen Boney Pile
Homer
:;:;*::* City
Coal Cleaning
Refuse Disposal
Area
Kilometer
0 0.5 1.C
FIGURE 1. MAP OF THE STUDY AREA SHOWING COMPONENTS OF THE POWER COMPLEX
-125-
-------
SAMPLING AND ANALYSIS TECHNIQUES
During the period from December 1976 through April 1977, a series of
three pre-operational, grab-sampling campaigns were conducted by Battelle in
the ambient media of the study area which included the Homer City Power Com-
plex. These environmental monitoring studies involved sampling, laboratory
analysis, and/or evaluation of the following components of the environment:
Fugitive dust
Water and stream sediments
Aquatic biota
Terrestrial biota
Raw coal and fly ash
Cleaning plant refuse disposal facility
Groundwater.
The fugitive dust data of concern in this paper were collected and analyzed
for comparison with MEG values. The samples collected during three campaigns
were analyzed for both physical and chemical parameters. Information from
the survey of terrestrial biota conducted during one campaign was utilized,
as later described, in attempts to confirm the results of MEG comparisons.
Fugitive dust monitoring was conducted using high-volume (hi-vol) ambient
air samplers during the following three 48-hour sampling periods:
• Campaign I: 8 p.m. December 17 to 8 p.m. December 19, 1976
• Campaign II: 8 p.m. January 5 to 8 p.m. January 7, 1977
• Campaign III: 8 p.m. April 5 to 8 p.m. April 7, 1977.
The first of these three campaigns was conducted over a weekend when both coal
transfer and construction activities were low.
A multiple-source fugitive-dust dispersion model was used to select and
verify locations for hi-vol samplers (Figure 2). This model takes into
account such factors as wind speed, emission rate, particle size, and distance
from selected potential dust sources located within the Homer City Power
Complex. No dust sources outside of the power complex were incorporated in
the model. On the basis of the computer-generated diffusion-modeling results,
ten monitoring sites were established at distances of 175 to 2200 M downwind
from various local dust sources. One of the ten sites was on private property
downwind of the power complex property and one site was on private property
upwind of the complex.
Several potential dust sources, both local and regional, were not incor-
porated into the diffusion model for sampling site selection. Dust generated
by vehicular traffic, parking lots, construction activities, several storage
silos, and especially the dusty surface of the plant grounds were not included
in the model due to their erratic and non-point-source nature. Data for the
Homer City power plant stack emissions were not available in time to include
in the model. In addition, four other major power stations (Keystone,
-126-
-------
Coal Cleaning
Refuse Disposal
Existing Ash
Disposal Area
Coal Cleaning Plant
Coal Storage Pile
CD
Power Plant
Wind Speed and Direction
Recorder
Hi Vol. Sampler
FIGURE 2. LOCATION OF FUGITIVE DUST SOURCES AND MONITORING SITES
-127-
-------
Conemaugh, Seward, and Shawville) are located in the same Chestnut Ridge sector
of the Allegheny Mountains as Homer City. These utilities are fed from coal
mines located either directly under or near the station sites. The model did
not include fugitive emission data from any of these facilities.
Potential fugitive dust sources at the Homer City Power Complex were
investigated during a pre-sampling site evaluation. Some of the dust sources
included an ash disposal area, boney piles at both deep mines, a coal storage
pile, road dust, three power plant stacks, and construction-generated dust.
The coal cleaning plant with its thermal dryers and the cleaning plant refuse
disposal area were under construction during Battelle's sampling campaigns.
Since these two areas were considered to be future potential sources of fugi-
tive dust, they were considered in the selection of sampling sites.
In order to identify the type and quantity of pollutants being emitted
from fugitive dust sources, a variety of analytical techniques were employed.
Particulate mass was determined by weighing the 8 x 10-inch fiberglass filters
used in the hi-vol samplers before and after each of the 12- or 24-hour
sampling periods. A microscopic analysis was made of particulates to provide
a distinction between components such as coal dust, fly ash, pollen, or
construction dust. An Andersen sampling head was used on one hi-vol to obtain
data on the distribution of particles in five size fractions.
Particulates on the filters from the hi-vol samplers were analyzed for
up to 22 elements. The analytical technique used for most elements was atomic
absorption; but neutron activation, colorimetry, a specific ion meter, a total
organic carbon analyzer, an LDC mercury monitor, and potentiometric titration
were also used. Since large amounts of four of these 22 elements (Na, K, Ca,
and Mg) were found in the blank filters, the values for these four elements
were not reported. Four of the remaining 18 elements (Sb, Ti, V, Se) were
analyzed only in the second or third campaign. In general, the filter
exhibiting the highest percentage of coal or ash from each site was used for
analysis. Data from 15 of the elements analyzed are compared in this paper.
-128-
-------
COMPARISON OF ANALYTICAL DATA
WITH MEG VALUES
Analytical data for fugitive dust, fly ash, and raw coal sampled in the
study area have been converted to the units used in the multimedia environ-
mental goals (MEG) study (Cleland and Kingsbury, 1977a and b). These data
are compared with the estimated permissible concentrations (EPC's) and/or the
minimum acute toxicity effluent (MATE) values defined in the Introduction to
this paper.
Average concentrations of 15 elements analyzed in the fugitive dust from
the study area are compared with the EPC's for air in Table 1. Since most
of the fugitive dust appeared to emanate from the coal storage pile and
decline in concentration within 200-300 m downwind (Figure 3), the data have
been averaged for the sampling sites located between 150-175 m and 400-1,800 m
downwind from the coal pile. The fugitive dust concentrations for the upwind
"control" sampling location are also provided. These field data are followed
by the appropriate maximum EPC's for air recommended for each element to
prevent negative effects to humans or the surrounding environment during
continuous long-term (chronic) exposure. A difficulty in making comparisons
between observed and recommended levels of the 15 elements shown in Table 1
is that three EPC's for human health and 10 EPC's for the environment are
not available.
Average concentrations for three of the elements (As, Cr, and Pb) analyzed
in fugitive dust exceeded the EPC's for human health. These values have been
underlined in Table 1. It is noteworthy that two of these elements (As and
Cr) had concentrations above the health-based EPC even at the upwind "control"
location.
Maximum and minimum concentrations of 15 elements analyzed in fugitive
dust are compared with the appropriate EPC's for soil in Table 2. Again, the
data are grouped to include sampling sites less than 200 m (i.e., 150-175 m)
and greater than 200 m (i.e., 400-1,800 m) downwind of the coal pile. Concen-
trations of the same elements in the raw coal are also shown. EPC's for
protection of human health and the environment are given for 12 elements; no
EPC values for iron, chlorine, and fluorine have been determined.
The majority of the elements analyzed showed maximum and frequently
minimum concentrations in the fugitive dust that were far greater than the
EPC levels suggested for the soil. Ten elements exceeded the EPC's for human
health and 11 elements exceeded the EPC's for the environment. Both the
maximum and minimum concentrations of 8 elements (As, Cd, Cr, Cu, Pb, Mn, Ni,
and Se) in the fugitive dust exceeded the EPC's for both human health and the
environment.
Obviously, the concentrations of toxic trace elements in fugitive coal
dust that has settled to the ground does not mean tfcat these same concentra-
tions occur in the soil. However, studies involving soil contamination by
other types of partlculate deposition have shown that toxic trace elements
-129-
-------
TABLE 1. FUGITIVE DUST COMPARISONS (yg/mj): EPC VALUES FOR AIR
VERSUS HOMER CITY DATA
1
t-1
CO
o
Distance from
Coal Pile
Downwind 150-175 ra(b)
Downwind 400-1,800 m(c)
Upwind Control
EPC Category
Health
Ecology
Trace Element Concentrations, ug/m
As Cd Cr Cu Fe Pb Mn Hg Ni Ti Zn Cl
F V Se
Average Concentration in Fugitive Dust During 3 Campaigns at Homer City (24-hr Sampling Periods)
0.014 0.008 .026 0.292 3.45 0.586 0.076 0.00056 0.015 0.44 0.35^ 1.97^
0.010 0.014 .015 0.119 1.87 0.334 0.093 0.00009 0.013^0. 32 '*' 0.22 0.82
0.009 0.005 .014 0.223 1.65 0.258 0.041 0.00003 0.009 0.17(1) 0.13 1.05
Estimated Permissible Concentrations (EPC's) , yg/m
0.005 0.12(£)0.002(f)0.5 — (h) 0.36 12 16(f) 0.04(f)14 9.5
0.04^ — — — 1^ — 0.01^
5.47 ND^ 0.0049
2.03 ND 0.0026^
1.40 0.02(1) 0.0030(1)
1.2 0.5
0.1 0.03(p)
(a) All data were collected between December 1976 and April 1977.
(b) Average for sampling sites 1 and 3; downwind of coal pile.
(c) Average for sampling sites 4, 8, and 9; downwind of coal pile.
(d) Sampling site 6; upwind of coal pile about 1600 m and off of the power station property.
(e) From Cleland and Kingsbury (1977a and b).
(f) Based on a Toxic Limit Value (TLV) which recognizes the element's carcinogenic potential
(g) Based on teratogenic potential.
(h) Not available.
(i) Concentrations were not available for some sampling sites during all three campaigns.
(j) ND = not detectable.
-------
i
M
CO
I-"
1300
•2 1200
o
I "00
•Cool Pile
•Site 7
Site 4
V
Boney Piles
Site 9
Site 8
Park
Future
Coal Refuse
Site
500 1000
Meters From Coal Storage Pile
1500
300
250
200
150
100 £
50 ^
"k
I
2000
FIGURE 3. FUGITIVE DUST CONCENTRATIONS COMPARED TO A TRANSECT OF THE AREA'S
TOPOGRAPHICAL RELIEF
-------
TABLE 2. FUGITIVE DUST COMPARISONS (pg/g): EPC VALUES FOR SOIL
VERSUS HOMER CITY DATA
(a)
Trace Element Concentration, ug/g
As Cd Cr
Cu Fe
Pb
Mn
Hg
Concentrations in Particulate at Sampling Sites
Maximum
Minimum
Maximum
l_i Minimum
U>
Is)
Health
Ecology
Max i mum
Minimum
154 264 471
3,678 28,736
U_ JJ5 .U5 336 6,223
Concentrations in Particulate at
238 619 667
34 ND 46^
10 0.06(c) 0.
2 0.01 10
4
3
Ni
within 200
264
Ti
Zn
Cl F
m of Homer City Coal Pile (Sites 1 and
8,676
0.2 2J5 626
200 and 2,000 m of Homer
2 545.
ND ND
Concentrations
50(c>
Concentrations Determined by Individual Analysis
31 48,750
20 18,000
17
12
.3 74
35
1.1
0.34
3,563
ND
(a)
Coal Sources *
66
46
0.26 108
0.23 91
V
3)(a)
ND
ND
and 9)(S)
ND
ND
1.4
15.
65
55
Se
33
*•
122
1
2
ND
ND
(a) Data from three sampling campaigns conducted by Battelle in the study area.
(b) From Cleland and Kingsbury (1977a and b); all values were multiplied by 100 based on personal communication with Kingsbury (August, 1978).
(c) Based on carcinogenic potential.
(d) Based on teratogenic potential.
(e) Value not available.
(f) SD - not detectable.
(g) Coal sources include: Helen Mining Company and Helvetia Coal Company (from Upper Freeport Seam); and Trucked-in Coal (from Lover Kittanning Se
-------
in these particulates can cause ecosystem disruption resulting in the loss of
essential nutrients and can also result in increased concentration of these
toxic elements in both plants and animals. These types of effects have been
demonstrated for lead smelter emissions (Jackson and Watson, 1977; Kerin, 1975)
and for fly ash emissions from coal-fired power plants (Furr, et al., 1977).
Dvorak, et al. (1978), have speculated that long-term exposure to uncombusted
coal dust may cause changes in vegetation community structure similar to those
caused by particulates from coal combustion.
Mechanisms for the movement of toxic trace elements from particulate
emissions deposited on the ground to the root zone of the soil are complex
(Vaughan, et al., 1975; Dvorak, et al., 1978). A partial list of the factors
which influence leaching of trace elements from deposited particulates into
the soil solution include (1) the size and type of particulates, (2) the
amount and acidity of precipitation, (3) the concentrations and physico-
chemical properties of the trace elements, (4) the texture, organic content,
pH, and other characteristics of the soil, (5) the solubility of elements into
the soil solution, and (6) the temperature of the air and soil.
The fugitive dust quantity and composition found during monitoring has
probably been accumulating on the ground in a reasonably similar fashion since
the power plant (including the coal storage pile) began operation in 1969.
Thus, mobile elements in the settled dust may have leached into the soil.
The quantity of toxic trace elements available to vegetation, however, needs
to be determined by chemical analysis of the soil. In spite of any leaching
of trace elements that may have increased soil concentration, the vegetation
for some distance from the coal pile has not yet shown any adverse effects
that were readily apparent during Battelle's field reconnaissance. An analysis
of soil biota and plant diversity, however, was not conducted.
Another basis for comparison is also possible; MATE values for components
in solid wastes have also been developed. Inasmuch as the deposited fugitive
dusts are tantamount to being a solid waste and these deposits may contact or
be absorbed or consumed by plants and animals, comparisons with MATE values
for solid wastes would appear to be valid. Such a comparison has been made
in Table 3. The table's structure is similar to that of Tables 1 and 2.
In Table 3, the appropriate MATE values are judged to be the ones related
to ecology limits. In general, these have lower values than those for health;
exceptions are mercury (Hg), chlorine (Cl), and fluorine (F), the latter two
for which there are no ecology values available. Twelve of the fifteen MATE
values for health are exceeded by the maximum values for both the close-in
(<200 m) and the more remote (>200 m) sampling sites. Eleven of the ecology
values are exceeded. Comparisons of solid waste MATE values with the elemen-
tal concentrations in the raw coals are also provided in Table 3. Elemental
concentrations in the raw coal exceed many of the same elemental EPC values
exceeded by elements in fugitive dust. However, the levels of toxic elements
in the raw coal are generally lower than the levels in the fugitive dust.
-133-
-------
TABLE 3. FUGITIVE DUST COMPARISONS (yg/g): MATE VALUES FOR SOLID
WASTED) VERSUS HOMER CITY DATA
Trace Element Concentration, ug/g
As
Cd Cr
Cu
Fe
Pb
Concentrations in Particulate at
Maximum 154
Minimum U_
264 471
18 18
3,678
336
28,736
6.223
Concentrations in Particulate at
Maximum 238
I Minimum 34
W
1
Health 50
Ecology 10
Maximum 48
Minimum 22
619 667
ND 46
10 50
0.2 50
Raw Coal
0.26 35
«0.1 30
6,061
220
1,000
IP.
57.576
11,477
300(C)
50
17,241
501
Mn Hg
Sampling Sites
632 3
65 0.2
Ni
Ti
Zn
within 200 m of Homer City
264
«
Sampling Sites Between 200 and 2,000
12.857
566
Minimum Acute
50
10
5.603 2
6 ND
545
ND
Toxicity Effluent (MATE
50 2
20 50
Concentrations Determined by Individual Analysis
31
20
48,750
18.000
17.3
12
74 1.1
35 0.34
45
2
of Three
16.8
12.8
8,676
626
m of Homer
3.007
1.356
3,563
ND(f)
Cl
Coal Pile
39,081
4,043
F V
(Sites 1 and 3)
55,000 ND
ND ND
Se
33
4
City Coal Pile (Sites, 4, 8, and 9)(a)
5.152
943.
27,278
5,806
45.600 ND
1,818 ND
122.
1
•s) for Solid Waste(b)
18,000
160
Homer City
1,329
1,125
5,000
1°.
260.000(d)
-_(e>
7,50Q(h) 500
— (C) 30
10
1
Coal Sources
66
46
0.26
0.23
108 65
91 55
ND
ND
(a) Data from three sampling campaigns conducted by Battelle in the study area.
(b) From Cleland and Kingsbury (1977a and b); all values were multiplied by 100 based on personal communication with Kingsbury (August, 1978).
(c) MATE values listed are for ferrous (Fe+2)_or ferric (Fe+3).
(d) MATE value listed is for chloride ion (Cl ).
(e) Value not available.
(f) ND » not detectable.
(g) Coal sources include: Helen Mining Company and Helvetia Coal Company (from Upper Freeport Seam); and Trucked-in Coal (from Lower Kittanning Sean).
(h) MATE value listed is for fluoride ion (F~).
-------
CONCLUSIONS AND RECOMMENDATIONS
Elemental concentrations in fugitive dust measured in the study area
exceeded both EPC and MATE values for air, soil quality, and solid wastes.
For example, three out of 15 elements analyzed in fugitive dust had concentra-
tions above the health-based EPC's for air quality. Comparisons with ecology-
based EPC's for air quality, however, were very difficult due to the absence
of ten EPC values.
Although no soil concentrations were determined, comparisons of elemental
concentrations in fugitive dust were made with ecology-based EPC's for soil
due to the potential problem of toxic elements leaching into the soil from
deposited fugitive dust. Eleven of the 15 elements studied had concentrations
in the fugitive dust that were above the ecology-based EPC's for soil.
Twelve solid waste MATE's for health and eleven for ecology were exceeded also.
In spite of the dust (particularly coal dust) present on the ground for some
distance around the coal pile, however, the vegetation has not yet begun to
show any obvious adverse effects.
Although applicability of EPC and MATE values are currently somewhat
limited because of known deficiencies now being corrected, the trend in these
observations appears clear. Many trace element concentrations in the study area
are higher than desirable. The conclusion is not, however, that immediate
corrective action is needed. There is no obvious damage to the vegetation
even though the present conditions have evidently persisted for some time.
Several recommendations can be made based on this study. First, more
field experiments are needed to validate these results. The amount of avail-
able data is small and more extensive sampling and elemental analysis, parti-
cularly of soil and plant and animal tissues, are needed. These steps are
necessary to determine the fate of the trace metals in the fugitive dust.
Additional research also needs to be conducted on EPC and MATE values to
evaluate and rank pollutants for the purpose of environmental assessment.
Much of this work was recommended in the initial MEG document (Cleland and
Kingsbury, 1977a) and is now or will soon be in progress. For example, MEG's
need to be related to the specific compounds or ionic forms of an element
that are most toxic rather than having a single value represent all compounds
and ions which have a common "parent" element. Synergistic and antagonistic
effects need to be considered, because these effects may drastically change
the hazard ranking of a pollutant in a specific situation. MEG's are also
needed for many of the master parameters, such as the "totals" identified
by Cleland and Kingsbury (1977a: 115) (e.g., total particulates).
In another vein, the comparison of trace element concentrations in fugi-
tive dust and MEG values points out the need for laboratory and field research,
particularly in relation to fugitive dust which consists predominantly of
coal particles. First, the rates at which toxic elements leach from coal dust
into a variety of soil types need to be explored. Second, the concentrations
of toxic elements present in the soil around a large, open coal pile need to
be determined where this pile has been in existence for a long period of time.
Third, laboratory bioassay and long-term field studies need to be conducted
on the effects of coal dust on plants and animals.
-135-
-------
REFERENCES
Cleland, J. G., and G. L. Kingsbury. 1977a. Multimedia environmental goals
for environmental assessment, Vol. I. Prepared for U.S. Environmental
Protection Agency, Industrial Environmental Research Laboratory, Research
Triangle Park, North Carolina, EPA-600/7-77-136a.
Cleland, J. G., and G. L. Kingsbury. 1977b. Multimedia environmental goals
for environmental assessment, Vol. II: MEG charts and background informa-
tion. Prepared for U.S. Environmental Protection Agency, Industrial
Environmental Research Laboratory, Research Triangle Park, North Carolina,
EPA-600/7-77-136b.
Dvorak, A. J., B. G. Lewis, P. C. Ghee, E. H. Dettmann, R. F. Freeman III,
R. M. Goldstein, R. R. Hinchman, J. D. Jastrow, F. C. Kornegay, D. L. Mabes,
P. A. Merry, E. D. Pentecost, J. C. Prioleau, L. F. Soholt, W. S. Vinikour,
and E. W. Walbridge. 1978. Impacts of coal-fired power plants on fish,
wildlife, and their habitats. Prepared for Fish and Wildlife Service,
U.S. Department of the Interior, Washington, D.C., FWS/OBS-78/29.
Furr, A. K., T. F. Parkinson, R. A. Hinrichs, D. R. Van Camper, C. A. Bache,
W. H. Gutenmann, L. E. St. John, Jr., I. S. Pakkala, and D. J. Lisk. 1977.
National survey of elements and radioactivity in fly ashes: absorption of
elements by cabbage grown in fly ash-soil mixtures. Environmental Science
and Technology, 11(13): 1194-1201.
Jackson, D. R., and A. P. Watson. 1977. Disruption of nutrient pools and
transport of heavy metals in a forested watershed near a lead smelter.
Journal of Environmental Quality. 6(4): 331-338.
Kerin, Z. 1975. Relationship between lead content in the soil and in the
plants contaminated by industrial emissions of lead aerosols. Pages 487-
502. In T. C. Hutchinson (Chief Ed.), International conference on heavy
metals in the environment. Vol. II, Part 2. Symposium was held in Toronto,
Ontario, Canada, October 27-31, 1975.
Vaughan, B. E., K. H. Abel, D. A. Cataldo, J. M. Hales, C. E. Hane, L. A.
Rancitelli, R. C. Routson, R. E. Wildung, and E. G. Wolf. 1975. Review
of potential impact on health and environmental quality from metals
entering the environment as a result of coal utilization. Battelle Energy
Program Report, Battelle's Pacific Northwest Laboratories, Richland,
Washington.
-136-
-------
INDUSTRIAL NON-POINT SOURCES:
ASSESSMENT OF SURFACE RUNOFF
FROM THE IRON AND STEEL INDUSTRY
Gordon T. Brookman, John A. Ripp, and
Bradford C. Middlesworth
TRC - THE RESEARCH CORPORATION OF NEW ENGLAND
125 SILAS DEANE HIGHWAY
WETHERSFIELD, CONNECTICUT 06109
-137-
-------
ABSTRACT
TRC - THE RESEARCH CORPORATION of New England was retained in April
1976 by the Environmental Protection Agency to assess whether surface runoff
from iron and steel mills is an environmental problem. The program included:
- Identifying and characterizing sources of surface runoff unique to
iron and steel mills.
- Assessing the specific problems associated with surface runoff and
evaluating the contribution made by individual sources to the total
problem.
- Identifying industry control systems which could be used to treat
contaminated stormwater.
TRC researched the limited existing data available, conducted plant
tours, and designed and performed a field survey at two fully integrated mills
on tidal rivers. Data collected at the two sites indicate that the coal and
coke storage piles and the coal and coke handling areas have the highest
potential for contaminating stormwater. The data also indicate that TSS
runoff concentrations are typical of urban runoff concentrations while TDS
values are approximately double the typical urban runoff concentrations.
Stormwater controls which presently exist within the steel industry are
limited. The only system specifically designed for stormwater control exists
at Armco's Houston Works where coal piles have been diked as a control measure
for both fugitive air emissions and stormwater runoff. Some mills collect
stormwater runoff with process wastewater for subsequent treatment at a ter-
minal plant. Those methods which are applicable to the industry include
rainfall detention ponding rings for flat roofs, swirl degritters, and reten-
tion basins or sedimentation ponds.
TRC concluded that, with the exception of runoff from coal and coke
storage areas, stormwater runoff is not a problem when compared to point
source control.
-138-
-------
1.0 INTRODUCTION
Since industries and municipalities are on the way to meeting the point
source standards of the 1977 interim goal of PL 92-500 (Federal Water Pollu-
tion Control Act Amendments of 1972), the effect of non-point source pollution
on water quality is gaining more attention. The National Commission on Water
Quality reported in 1976 that "non-point pollutant sources are significant to
the Commission's study because they may in some instances overwhelm and negate
the reductions achieved through point source effluent limitations."1 Based
on these findings the Commission recommended to Congress that "control or
treatment measures shall be applied to agricultural and non-point discharges
when these measures are cost-effective and will significantly help in achiev-
ing water quality standards."2
Non-point sources are diffuse in nature, usually intermittent, site
specific, not easily monitored at their exact source, related to uncontrol-
lable meteorological events (precipitation, snow melt, drought), and not
usually repetitive in nature from event to event. The primary transport
mechanism for non-point sources is water runoff from meteorological events.
The three basic modes of runoff transport are overland (surface) flow, inter-
flow (also called interstitial flow), i.e., flow through the ground between
the surface and groundwater levels, and groundwater flow. Surface runoff will
usually contain the highest quantity of contaminants and is the most rapid
method of transport of pollutants from non-point sources.
Because of the great quantities of water and raw material used in making
iron and steel, mills are usually located near waterways. Contaminated
stormwater runoff from these mills could rapidly reach these waterways; thus
the potential of causing a detrimental environmental impact is often present.
In April 1976 the Metallurgical Processes Branch of the Industrial
Environmental Research Laboratory (IERL) of the Environmental Protection
Agency (EPA) at Research Triangle Park, North Carolina, retained TRC -
THE RESEARCH CORPORATION of New England to perform an assessment of surface
runoff from iron and steel mills as a task under an existing contract (68-02-
2133). The principal objective of this program was to provide EPA with an
evaluation which it can use in determining if stormwater runoff from iron and
steel mills is an environmental problem and should be included in the Agency's
long-term planning as an area of concern.
The program had the following sub-objectives:
1. To identify sources of surface runoff unique to iron and steel mills
and to characterize runoff streams in terms of quantity and
composition.
-139-
-------
2. To assess the specific problems of surface runoff at iron and
steel mills and evaluate the contribution made by these individual
sources to the overall problem.
3. To identify control systems used by the industry or by other
industries which are or could be used to treat contaminated storm-
water .
This paper presents a general overview of the program with emphasis on
the field sampling portion of the program which was conducted at two iron and
steel mills.
-140-
-------
2.0 EXISTING DATA AND IDENTIFICATION OF SURFACE RUNOFF SOURCES
Before this program, little work had been performed on surveying storm-
water and identifying potential sources of stormwater contamination in the
steel industry. Previously, the most comprehensive studies had been under-
taken by Armco Steel Corporation's Houston Works in Houston, Texas,^ and
Kaiser Steel's Fontana, California plant.4
In the Armco study, the mill was divided into drainage basins, each
characterized according to size, activity, and land cover (i.e., buildings,
paved area, railroad track, undeveloped land, stockpiles, and ponds). Each
basin was sampled for several storms. Parameters measured included total
suspended solids (TSS), oil and grease, biochemical oxygen demand (BOD5),
total organic compound (TOG), and chemical oxygen demand (COD).
Armco found that stormwater quantity and quality varied appreciably
with drainage basin characteristics and location. This limited the validity
of any correlation of parameter concentrations between basins. Furthermore,
the quality of stormwater runoff was found to vary directly with storm dura-
tion and intensity and also with the number of antecedent dry days prior to
storms. As antecedent dry days increase, so does the potential particulate
matter to be scoured. Other significant results of the Armco study were the
absence of a "first flush" effect, and the absence of significant quantities
of organic matter. The "first flush" effect is a condition where the matter
accumulated in a basin since the last runoff event is scoured from the area at
the start of the next storm event. In almost all cases a flow dependent
effect was observed; i.e., peak parameter concentrations occurred at peak
runoff flows.
The Kaiser program involved sampling during the rainy season (February
and March) in 1975. Runoff from twelve storm events was sampled for chloride,
conductivity, and oil and grease. The oil and grease results from the Kaiser
program were much higher than those obtained at Armco.
Since the Armco and Kaiser studies were the only data existing on storm-
water runoff from steel mills, several plants were toured as part of this
program in an effort to combine a number of factors which affect site specific
runoff, such as terrain, climate, mill locations and operations, into an
overall industry-wide assessment of the most probable sources of stormwater
contamination. The following companies were contacted and/or visited:
United States Steel
National Steel
Armco Steel
Republic Steel
Youngstown Sheet and Tube
Inland Steel
Kaiser Steel
CF&I Steel
Alan Wood Steel
-141-
-------
Runoff from the activities and operations of steel mills was segmented
into the following groups:
o Runoff from storage and disposal piles (coal, coke, slag, iron).
o Runoff from adjacent urban areas into the mill.
o Runoff from slag handling and processing facilities.
o Runoff of accumulated materials from roof and ground areas
from several mill operations (blast furnace, sinter plant,
EOF shop, open hearth, coke and by-product plant, coal and coke
handling, and finishing areas).
Because runoff is site specific, it was impossible to compare the con-
taminated stormwater potential of an area in one particular mill to the same
area in another mill. Climate, terrain, operations, maintenance, and the
location of processes relative to each other are unique to each mill.
Therefore, a rating system was devised which ranked the relative potential of
each activity or operation at an individual plant. Based on the assessments
of TRC personnel, the ratings were entirely subjective, except where physical
data were available (e.g. Armco's Houston Works). The following activities
or operations were rated as having the greatest potential for contaminating
stormwater:
o Coal storage piles
o Coke storage piles
o Slag dumps
o Iron ore and pellet storage piles
o Coal and coke handling
-142-
-------
3.0 FIELD PROGRAM
A field survey was performed at two different sites in the spring of
1977. Both sites were fully integrated mills on tidal rivers. The general
characteristics of each site are described in Table 1.
Tables 2 and 3 describe the individual drainage basins sampled at Sites 1
and 2 respectively. At Site 1, the drainage basins had been defined in pre-
vious stormwater runoff programs conducted by plant personnel. Because there
are no well defined drainage basins within Site 2, the delineation of drainage
areas was based upon the storm sewer plans for the plant. The storm sewer
system for Site 2 originates at the roof drains from most of the plant opera-
tions, includes the road and railroad line runoffs, and terminates in either a
canal leading into the tidal river or into the river directly.
The general test plan for the survey was designed to determine:
1. Background conditions at each sampling location prior to a
storm event, i.e., dry weather flow conditions.
2. Volume of stormwater runoff and pollutant concentrations in
the runoff as a function of time for the storm event.
The following additional data were gathered:
1. Rainfall accumulation as a function of time for the storm
event.
2. Dustfall accumulation between storms.
No attempt was made to assess the effects of the stormwater runoff
on the receiving water.
The parameters measured in this survey were:
Runoff Flow Rainfall
Total Iron Total Suspended Solids (TSS)
Dissolved Iron Total Dissolved Solids (TDS)
Phenols Cyanide
Ammonia Sulfates
Oil and grease and organic parameters such as BOD5, COD and TOC were not
measured because previous work performed by TRC showed that these parameters
would not be of sufficient magnitude to be of concern.
-143-
-------
TABLE 1
GENERAL SITE CHARACTERISTICS
Age of Plant
Developed Acreage
Terrain
Runoff Receiving
Body
Plant Operations
Period of Sampling
Number of Sampling
Points
Permanent Flow
Devices
Site 1
37 Years
575
Flat, Semi-Permeable
Tidal River
Coke Plant, Sinter Plant,
Blast Furnaces, Electric
Furnaces, Finishing Oper-
ations
3/77 to 4/77
5
Yes
Site 2
25 Years
3900
Flat, Permeable
Tidal River
Coke Plant, Sinter
Plant, Blast Furnaces,
Open Hearth Furnaces,
Electric Furnaces,
Finishing Operations
5/77 to 6/77
13
No
-144-
-------
TABLE 2
DESCRIPTION OF INDIVIDUAL DRAINAGE BASINS SAMPLED
SITE 1
Basin
005
006
009
010
Oil
Activities and Operations
Wide Flange Mill; Shipping Office; Roundhouse (car,
truck, and railroad car repair facility) ; western
halves of the No. 1 Electric Furnace Shop, No. 2 Plate
Mill, Plate Shipping Building, Heavy Plate Shear
Building, and Plate Heat Treat Building.
Direct Reduction Plant.
West end of the Mold Foundry; area between the Coke
Plant proper and the east end of the Stock House;
Coke Transfer.
Coal transfer; main coal conveyor belt from the dock
area to the coal storage area; Coal Shaker Building;
numerous coal transfer points located in immediate
vicinity of the west end of the Coke Plant area
Mold Preparation Shop; eastern part of the No. 2
Electric Furnace Shop; eastern half of the Coke Ovens;
Coke Oven By-products area; coal pile storage area;
eastern half of the Mold Foundry; employee parking
area.
Acreage
142.8
4.6
6.6
2.6.
60.5
-145-
-------
TABLE 3
DESCRIPTION OF INDIVIDUAL DRAINAGE BASINS SAMPLED
SITE 2
Basin
002
003
004
005
006
007
008
009
010
Oil
012
013
014
Activities and Operations
Diesel Repair Shop.
Slag Filled Borrow Area; Railroad tracks.
Mold Preparation Shop; northeast side of Open Hearth
Furnaces; Railroad tracks.
South end of Open Hearth Shop; Railroad tracks; Slag
Dump Area; Mold Preparation.
Hot mills; Slab cooling area; Slab mill; Billet mill.
Hot strip mills.
Blast furnace; Sinter Plant; Employee Parking; Ore
Conveyors .
Sinter Plant; Ore Conveyors; Roadways.
One half of Open Hearth Plant; Coke Plant; Coke
yards; Numerous Railroad lines; Coke By-Products
Complex.
Coal storage.
Southern end of Coke Ovens - surface runoff.
Southern end of Coke Ovens - surface runoff.
Ladle Repair Shop; Railroad track area.
Acreage
5.5
4.8
18.6
4.0
12.0
5.2
9.1
11.4
144.6
9.9
1.3
1.5
0.6
-146-
-------
Previous work by Armco revealed very high concentrations of COD and TOC which
were concluded to be a result of inert coal and coke fines and not reactive
organics.
The sampling sites were located in the areas of material storage and
disposal and coal and coke handling. No slag dump runoff data were obtained
as neither site had a representative slag dump. In addition, due to tidal
backflow problems, no iron ore or pellet storage pile runoff data were
obtained at Site 1. The sample collection method, flow monitoring technique,
sampling schedule, and parameters to be analyzed at each of the sampling sites
within Site 1 and 2 are summarized in Tables 4 and 5 respectively.
3.1 Field Survey Results
3.1.1 Site 1 Results
Table 6 summarizes the storm event data for Site 1. Of the five storm
events sampled at Site 1, only the storm of March 31 approximated the high
intensity, short duration rainfall typical of this semitropical area. From
historical observations of previous storm events at Site 1, it was expected
that the total rainfall at various locations around the plant would differ
over the course of a storm. This uneven distribution of rainfall was never
observed during the field program. During the sampling program, the rainfall
was typically a steady drizzle with occasional heavy downpours uniformly
distributed over the entire plant. In all five events, rain wedge totals
closely corresponded to the recording rain gages.
Table 7 summarizes the flow data from Site 1. Time-weighted average flow
data plus the range of flow for both dry and wet weather sampling are listed.
The dry weather flows at outfall 010 were not measurable; the water levels
over the weir were essentially zero except for a small trickle which volu-
me trically was negligible. Wet flow data were limited at outfall 010 due to
occasional tidal backflows. At outfalls 005 and Oil wet flows were signifi-
cantly higher than dry flows. Outfall 009 showed the effects of tidal back-
flow from the river and neither samples nor flow measurements could be
obtained at the weir during any of the storm events.
Tables 8 through 12 all refer to the pollutant data measured at Site 1.
The range (Table 8), the mean (Table 9) and the average mass loadings (Tables
10 through 12) of pollutants show the'differences between dry and wet weather
conditions. Average mass loadings of pollutants for dry weather conditions
were calculated by multiplying the mean concentrations value measured during
each storm by the time-weighted average flows from Table 7. Average mass
loadings for wet weather conditions were calculated by multiplying the time
weighted average concentrations by the time-weighted average flows, both
determined from the concentration and flow curves for each rainfall event.
The time weighted average wet weather flows pertain to the time over which
-147-
-------
TABLE 4
SUMMARY OF SAMPLING SITES AND SAMPLING PROGRAM
SITE 1
OUTFALL
005
009
010
Oil
SAMPLE
COLLECTION
METHOD
I SCO Sampler
with weir
I SCO Sampler
with weir
ISCO Sampler
with weir
ISCO Sampler
with weir
FLOW
METHOD
ISCO Flow Meter
and Printer
ISCO Flow Meter
and Printer
ISCO Flow Meter
and Printer
ISCO Flow Meter
and Printer
SAMPLING
SCHEDULE FOR
STORM EVENTS
Every storm
event
Every storm
event
Every storm
event
Every storm
event
PARAMETERS
TO BE
ANALYZED
TSS, TDS
TSS, TDS, Total
Fe, Dissolved Fe,
Phenols, Cyanide,
Ammonia
TSS, TDS, Total
Fe, Dissolved Fe,
Phenols, Cyanide,
Ammonia, Sulfates
TSS, TDS, Total
Fe, Dissolved Fe,
Phenols , Cyanide ,
Ammonia
-148-
-------
TABLE 5
SUMMARY OF SAMPLING SITES AND SAMPLING PROGRAM
SITE 2
OUT? ALL
002
003
004
005
006
007
008
009
010 A
010B
Oil
012
013
014
013
SAMPLE COLLECTION
METHOD
Grab
Grab
ISCO Sampler with
weir
Grab
ISCO Sampler with
weir
ISCO Sampler with
weir
Grab
Grab
ISCO Sampler
Grab
Grab
Grab
Grab
Grab
FLOW
METHOD
Bucket and stop-
wacch
Bucket and scop-
watch
ISCO flow aetar
and printar
Suckat and stop-
wacch
ISCO Clow oacar
and prlncar
ISCO flow oatar
and printer
Nona
None
Gurley meter
Nona
Nona
None
Nona
Nona
SAMPLING
SCHEDULE FOR
STORM EVENTS
Every storm event
(whan possible)
Every storm event
(when possible)—
low priority
Sample 2 of sites
004, 006, & 007 for
each storm
Every storm event
(when possible) —
low priority
Same as 004
Same as 004
Every storm event
(when possible)
Every storm event
(whan possible)
Every storm event
Every storm event
Sample 2 of
012, 013. &
014 for each
storm
Sam* as 012
Sam* as 012
Every storm event
PARAMETERS TO BE
ANALYZED
TSS, TDS
TSS, TDS
TSS, TDS
Total Fa
Dissolved ?e
TSS, TDS
TSS, TDS
TSS, TDS
TSS, TDS
Total F*
Dissolved Fe
TSS, TDS
Total Fe
Dissolved Fe
Metals
TSS, TDS
Total Fe
Dissolved Fe
PhenoLs, Ammonia,
Cyanide
TSS, TDS, Sulfaee,
Phenols, Ammonia,
Total F*, Dis-
solved Fe, Metals
TSS, TDS, Phenols,
Sulfates, Ammonia,
Total Fa, Dis-
solved Fe, Metals,
Cyanide
TSS, TDS, Phenols,
Sulfates, Ammonia.
Total Fe, Dis-
solved Fe, Metals,
Cyanide
TSS, TDS, Total Fe
Dissolved F*
TSS, IDS, Phenols,
Sulfates, Ammonia,
Total F«, Dis-
solved Fa, Cyanide
-149-
-------
Table 6
STORM EVENT DATA
SITE 1
MARCH - APRIL, 1977
Date
3/24/77
3/27-
3/28/77
3/31/77
AM/77
ft/16/77
Storm Beginning
0500
2000 (3/27)
1410
0200
0430
Storm Ending
2130
0200 (3/28)
1430
0500
2000
Total Rainfall
cm
0.84
1.42
0.20
0.36
0.71
(Inches)
(0.33)
(0.56)
(0.08)
(0.14)
(0.28)
Average
Rainfall
Intensity
cm/hr
0.05
0.23
0.61
0.13
0.05
(ln/hr)
(0.02)
(0.09)
(0.24)
(0.05)
(0.02)
Maximum
Rainfall
Intensity
cm/hr
0.13
0.61
1.07
0.41
0.56
(ln/hr)
(.05)
(0.24)
(0.42)
(0.16)
(0.22)
Ol
o
-------
Table 7
DRY vs WET FLOWS
SITE 1
MARCH - APRIL, 1977
OUTFALL
DATE
3/24
3/27 - 28
3/29
3/31
4/4
4/5
4/16
4/18
005
M
Avg. Flow
lp.(gp.)
473
(125)
227
(60)
216
(57)
T
Rang*
lp»(gp.)
435-568
(115-150)
227
(60)
76-254
(20-67)
WET
Avg. Flow
lp«(gp>)
1056
(279)
6083
(1607)
401
(106)
2112
(558)
3456
(913)
Range
1|»(8P«>
227-2233
(60-590)
454-15026
(120-3970)
227-984
(60-260)
568-4542
(150-1200)
228-15900
(60-4200)
010
HET(W
Avg. Flow
Ip-(gp-)
12.5
(3.3)
16.0
(4.2)
HD
ND
13.3
(3.5)
Range
IpB(gjw)
0 - 29.1
(0 - 7.7)
0 - 67.0
(0 - 17.7)
ND
HD(c)
0 - 49.02
(0 - 13.0)
Oil
DR
Avg. Flow
Ipa(gpM)
38
(10)
3.4
(0.9)
87
(23)
Y
Range
lp»(gp»)
27-53
(7-14)
0-13.2
(0-3.5)
76-106
(20-28)
W
Avg. Flow
IpM(gpa)
189
(50)
708
(187)
405
(107)
170
(45)
583
(154)
ET
Range
Ipa(gpB)
45 - 534
(12-141)
38 - 2203
(10 - 582)
95-939
(25-248)
15-167
(4-97)
83-1374
(22-363)
(a)I No flow data were taken at Outfall* 006 and 009, nor at the coal pile drainage ditch.
(t>)' There was no measurable dry (low at 010 during the progran.
(c) NO - Ho flow data were obtained.
(d) Flow values are tlae weighted average* for the entire event.
-------
TABLE 8
RANGE OF POLLUTANT CONCENTRATIONS AT THE
SAMPLING LOCATIONS AT SITE 1
MARCH - APRIL, 1977
Ul
S3
I
Pollutant
Total Suspended Solids
Total Dissolved Solids
Total Iron
Dissolved Iron
Phenols
Cyanide (Total)
Ammonia
Sulfate
. Range of Pollutant Concentration, ng
Outfall 005
Dry
4-31
327-463
Wet
11-113
238-964
Outfall 010
Dry
4-649
2007-5438
1.1-8.3
0.1-0.6
16-34
n.d.-0.99(b)
54-96
400-1580
Wet
10-1272
661-4993
1.2-3.6
0.1-0.2
0.02-1.1
n.d.
3.6-73
180-490
I
Outfall dll
Dry
7-42
668-1049
1.1-2.7
o.io(a)
0.02-0.68
n.d.
0.57-26
Wet
9-151
427-1196
0.96-5.8
0.10-0.30
0.01-0.52
n.d.-0.01(b^
0.65-28
(a)
(b)
Only one value obtained.
n.d. - not detectable - detectable limit - 0.001 mg/t.
-------
TABLE 9
MEAN POLLUTANT CONCENTRATIONS, IN mg/K. AT SITE 1
MARCH - APRIL, 1977
Outfall
Pollutant
TSS
TDS
Total Iron
Dissolved
Iron
Phenols
Cyanide
(Total)
Ammonia
Sulfate
005
Dry
15
396
Vfet
45
541
010
Dry
84
3078
3.3
0.4
25
0.5
73
718
Wet
184
2158
2.4
0.1
0.37
-(a)
43
312
Oil
Dry
18
868
1.9
0.1
0.13
-(a)
9.1
Wet
35
919
2.6
0.2
0.086
0.002
3.4
I
I-1
in
I
(a)
Several non-detectable values were also obtained.
-------
TABLE 10
AVERAGE MASS LOADINGS OF POLLUTANTS
DRY VS. WET WEATHER
MARCH - APRIL, 1977
OUTFALL 005 - SITE 1
Data
Paraaatar
Total Suapandad
Solid*
Total Dlaaolvad
Solids
3/2* (Vtat)
Avg.
Cone.,
«g/l
39
938
Avg.
Flev,
lp«
(n»>
1200
(317)
1200
(317)
Avg.
HIM
Loading.
ki/hr
(Ib/hr)
2.82
(6.2)
67.5
(149)
3/27-28 (Wat)
Avg.
Cone.,
«g/l
48
332
Avg.
Flow,
Ip.
(8P-)
3845
(1016)
3847
(1016)
Avg.
Maaa
Loading,
kg/hr.
(Ib/hr)
11.1
(24.4)
76.6
(168.5)
3/29 (Dry)
Avg.
Cone.,
•8/1
16
353
Avg.
Flow,
lp»
(gpn)
473
(US)
473
(125)
Avg.
Maaa
loading,
kg/hr
(Ib/hr)
0.45
(0;99)
10.0
(22.0)
3/31 (W«e)
Avg.
Cone.,
mg/1
38
581
Avg.
Flow,
Ip.
(8P»)
401
(106)
401
(106)
Avg.
Maaa
Loading,
kg/hr
(Ib/hr)
0.91
(2.0)
14.0
(30.8)
Data
Paramatar
Total Suapandad
Sollda
Total Dlaiolvcd
Solid.
4/4 (U«e)
Avg.
Cone.,
«g/l
37
669
Avg.
Flow,
Ipit
(SI")
2434
(643)
2434
(643)
Avg.
Maaa
Loading,
kg/hr
(Ib/hr)
5.46
(12.0)
97.7
(214.9)
4/5 (Dry)
Avg.
Cone.,
•g/1
17
443
Avg.
Flow,
Ip.
(«P->
227
(60)
227
(60)
Avg.
Maaa
Loading,
kg/hr
(Ib/hr)
0.23
(0.51)
6.0
(13.2)
Avg.
Cone.,
»g/l
7S
371
4/16 (Vat)
Avg.
Flow,
lp«
(gP«)
4205
urn)
4205
(1111)
Avg.
Mesa
Loading,
kg/hr
(Ib/hr)
19.0
(41.8)
93.6
(206.0)
Avg.
Cone.,
•»/!
14
409
4/18 (Dry)
Avg.
Flow,
lp«
UP")
227
(60)
227
(60)
Avg.
Maaa
Loading,
kg/hr
(Ib/hr)
0.19
(0.42)
5.6
(12.3)
-154-
-------
TABLE 11
AVERAGE MASS LOADINGS OF POLLUTANTS
DRY VS. WET WEATHER
MARCH - APRIL, 1977
OUTFALL 010 - SITE 1
Date
Parameter
Total Suspended Solids
Total Dissolved Solids
Total Iron
Dissolved Iron
Phenol
Ammonia
Sulfate
3/24 (Wet)
Avg.
Cone . ,
mg/1
76
2170
2.61
0.46
50
235
Avg.
Flow,
1pm
(gpm)
12.7
(3.36)
12.7
(3.36)
12.7
(3.36)
12.7
(3.36)
4.43
(1.17)
14.7
(3.88)
Avg.
Mass
Loading,
kg/hr
(Ib/hr)
0.06
(0.13)
1.65
(3.64)
0.002
(0.004)
0.0004
(0.0009)
0.013
(0.03)
0.25
(0.55)
3/27-28 (Wet)
Avg.
Cone . ,
mg/1
1717
150
0.119
0.559
41.0
224
Avg.
Flow,
Ipn
(gpm)
34.4
(9.1)
34.4
(9.1)
34.4
(9.1)
34.4
(9.1)
34.4
(9.1)
34.4
(9.1)
Avg.
Mass
Loading,
kg/hr
(Ib/hr)
3.54
(7.80)
0.31
(0.68)
0.0002
(0.0005)
o.ooi
(0.003-)
0.08
(0.19)
0.46
(1.02)
-155-
-------
TABLE 12
AVERAGE MASS LOADINGS OF POLLUTANTS
DRY VS. WET WEATHER
MARCH - APRIL 1977
OUTFALL Oil - SITE 1
Date
Parameter
Total
Suspended
Solids
Total
Dissolved
Solids
Total
Iron
Dissolved
Iron
Phenol
Aa*»nla
1/24 (Uet)
Avg.
.Cone.,
•g/1
11
1343
0.98*
0.1**
0.117
1.49
Avg.
Flow.
IP-
(*p->
217
(57.29)
217
(57.29)
282
(74.4)
275
72.8
262
(69.22)
252
(66.63)
Avg.
Mass
Loading.
kg/he
Ub/hr)
0.14
0.32
17. 5
(38.5)
0.02
(0.04)
0.002
(0.004)
0.002
(0.004)
0.023
(O.OS)
3/27-28 (Wet)
AVI.
Cone . i
-«/l
97
624
5.8**
0.1**
0.038
13.04
Avg.
Flow.
IP-
(CP-)
1764
(466)
1764
(466)
742
(196)
696
(184)
1669
(444)
742
(196)
AVB.
Mass
Loading,
kg/lir
(lb/hr)
10.3
(22.6)
66.0
(145.3)
0.26
(0.57)
0.004
(0.009)
0.004
(0.009)
0.58
(1.28)
3/29 (Dry)
Avg.
Cone . ,
•g/1
14
683
1.3
0.1**
0.055
22
Avg.
Flow.
Ip.
(gP->
38
(10)
38
(10)
38
(10)
38
(10)
38
(10)
38
(10)
Avg.
Haas
Loading.
kg/hr
(lb/hr)
0.03
(0.07)
1.56
(3.43)
0.003
(0.007)
0.0002
(0.0005)
0.0001
(0.0003)
O.OS
(0.11)
4/5 (Dry)
Avg.
Cone . .
•g/1
14
1021
2.6
0.1
0.26
4.9
Avg.
Flow.
IP-
(KP»)
3.4
(0.9)
3.4
(0.9)
3.4
(0.9)
3.4
(0.9)
3.4
(0.9)
3.4
(0.9)
Avg.
Haas
Loading.
kg/be
Ub/hr)
0.003
(0.007)
0.21
(0.46)
0.001
(0.002)
2.0xlO~5
(4.4xlO~5)
5.3xlO~s
(1.2xlO"s)
0.001
(0.002)
4/16 (Wet)
Avg.
Cone.,
-g/1
51
1062
3.63*
0.2*
0.08*
0.912
Avg.
Flow.
IP-
(BP-)
568
(150)
568
(150)
935
(247)
935
(247)
477
(126)
568
(150)
Avg.
Mass
Load Ing .
kg/lir
(lb/hr)
1.74
(3.83)
36.2
(79.6)
0.2
(0.45)
0.01
(0.02)
O.OO2
(0.004)
0.03
(0.07)
4/18 (Dry)
Avg.
Cone . ,
•g/1
30
785
1.9
0.1
0.025
0.97
Avg.
Flow.
IP-
(gpuO
87
(23)
87
(23)
87
(23)
87
(23)
87
(23)
87
(23)
Avg.
Mass
Loading,
kg/hr
(lb/hr)
0.16
(0.34)
4.1
(9.0)
0.01
(0.02)
0.001
(0.002)
O.O001
(0.0003)
0.005
(0.011)
•straight average
**one value only
-------
each parameter was sampled and may vary for the different parameters within
each storm event. In some instances, due to lack of data, straight average
concentrations (or in some cases, one data point) were used to calculate wet
weather average mass loadings. When no flow data were measured, mass loadings
were not calculated.
At all outfalls the mean dissolved solids were higher than the suspended
solids. At outfalls 005, 010, and Oil, where automatic sampling was per-
formed, the dissolved solids were consistently higher than the suspended
solids, often by more than an order of magnitude. The reaction of dissolved
solids varied with each outfall and each storm event. However, after plotting
all the dissolved solids data and comparing these curves to the rainfall
intensity and flow curves, no conclusion can be made concerning the reaction
of dissolved solids to a storm event.
The reaction of total suspended solids to a storm event also varied with
each outfall and event. In a few cases, suspended solids correspond directly
to rainfall intensity and flow, but in most instances, there was a time lag
between the rainfall intensity peaks and suspended solids concentration peaks.
The pollutant data from outfall 010 do show some interesting results. As
indicated in Tables 8 and 9, the dry weather concentrations of total dissolved
solids, total iron, dissolved iron, phenols, cyanide, ammonia, and sulfates
were greater than the wet weather concentrations, indicating that the storm-
water runoff at outfall 010 diluted these pollutants.
This same dilution effect was observed for phenol and ammonia concentra-
tions at outfall Oil, although the levels were much lower than those measured
at outfall 010. The mass loading data indicate that the dry weather loading
is at least one order of magnitude less than the wet weather loading of
phenols. In most cases the same is true for the ammonia loadings.
From the limited data taken at outfalls 010 and Oil, the concentrations
of phenols exhibited a consistent pattern. Increases and decreases in meas-
ured phenols loading in these drainage basins correspond directly to increases
and decreases in rainfall intensity with very little time lag.
In most cases, ammonia showed a general trend of decreasing concentration
over the period of the storm, indicating that the stormwater acted to dilute
the ammonia over the course of the storm rather than cause a "first flush"
effect. In no case was the "first flush" effect observed.
3.1.2 Site 2 Results
Only two storm events occurred during the field program (May-June, 1977)
which were of sufficient magnitude to produce surface runoff. Data from these
events are summarized in Table 13.
-157-
-------
TABLE 13
STORM EVENT DATA
SITE 2
MAY - JUNE, 1977
Date
6/9-
6/10/77
6/20/77
Storm Beginning
0500 (6/9)
0900
Storm Ending
1500 (6/10)
2030
Total Rainfall
cm
4.45
2.59
(inches)
(1.75)
(1.02)
Average
Rainfall
Intensity
cm/hr
0.13
0.23
(in/hr)
(0.05)
(0.09)
Maximum
Rainfall
Intensity
During Storm
cm/hr
1.42
_(a)
(in/hr)
(0.56)
_(a)
oo
(a)
No rainfall intensity data were collected on June 20 due to equipment failure
and manpower constraints.
-------
The first storm event started as a steady downpour which then tapered off
to a drizzle with occasional heavy showers. Surface runoff was evident at all
of the sampling locations. The rainfall intensity curve for this storm event
is shown in Figure 1 along with the flow at outfall 007.
The second storm event was short in comparison to the first, but again
resulted in a considerable amount of surface runoff at all of the sampling
sites. This storm was also a heavy downpour. Due to manpower constraints and
equipment failure, very little data except total rainfall and storm duration
was gathered.
There were also several other small storm events which resulted in 1.3 cm
of rain or less. Because most of the plant area is semi-permeable and level,
surface runoff was not detected during any of these storms.
Table 14 shows the average flows and ranges of flow for several of the
outfalls during dry and wet weather. Complete information exists only for
outfalls 004, 006, and 007.
The flow data for storm events at outfalls 004, 006, and 007 show some
interesting trends. The flow peaks at outfalls 006 and 007 corresponded very
closely to rainfall intensity peaks with almost no time lag. Figure 1 also
shows the hydrograph of June 9-10 for outfall 007. At outfall 004 the time
lag between rainfall intensity peaks and flow peaks ranged from 0.5 to 3.5
hours. The difference was probably due to the type of drainage basin associ-
ated with each outfall. Outfalls 006 and 007 receive stormwater either
directly from roof drains or from paved areas. The basin which drains to
outfall 004 is a mostly unpaved (permeable) area, causing the time lag between
rainfall peaks and runoff peaks.
Tables 15 and 16 show the range of concentrations and the mean concentra-
tions of the pollutants analyzed at each of the outfalls for both dry and wet
weather. Tables 17 through 19 indicate the average mass loadings of the
pollutants analyzed at outfalls 004, 006, and 007 for both dry and wet weather
conditions. Mass loadings were calculated in the same manner as at Site 1.
Total dissolved solids concentrations were much higher than total sus-
pended solids concentrations at all of the outfalls during both dry and wet
weather conditions with two exceptions, those being the wet weather concentra-
tions at outfalls 006 and Oil. A consistent pattern was established for total
suspended solids. A direct relationship exists between TSS concentration and
rainfall intensity corresponded directly to an increase in TSS concentration
with no time lag. This is shown in Figure 2.
Several interesting trends occurred with total and dissolved iron. Six
out of nine outfalls showed that an increase in rainfall intensity also cor-
responded directly to an increase in total iron concentration with no time
lag. This was not true of dissolved iron, since five out of nine outfalls
showed dissolved iron to vary inversely with total iron. As total iron con-
centration decreased, dissolved iron concentration increased and vice versa.
-159-
-------
O
SITE 2
RAINFALL INTENSITY
STORM 6/9 TO 6/10/77
SITE 2
OUTFALL 007 STORH 6/9 TO 6/10/77
TOTAL RAIKFALL 1.75 INCHES
NOON
TIME
Figure 1: Rainfall and Flow in Basin 007 Versus Time, Site 2, 6/9 to 6/10/77. Stora
-------
rim
-90
I 1 ,0 It
WOR
t i i 10 it t
\-iaa
TSS AVCUGC LSMtSS "All
.on
-------
TABLE 14
DRY VS. WET FLOWS
SITE 2
MAY - JUNE, 1977
(a)
Outfall
002 (b)
UUi
nnL^
00*
nnfi(c)
OOo
nn?(c)
007
009 (b)
(j(jy
010A(b)
ijj.vj/t
mnB
Ol'JB
Dace
5/10
5/18
6/9-10
6/20
5/10
5/18
6/9-10
6/20
5/10
5/18
6/9-10
6/20
5/10
5/18
6/9-10
6/20
5/10
5/18
6/9-10
6/20
5/10
5/18
6/9-10
6/20
5/10
5/18
6/9-10
6/20
Sampling
Condition
Dry
Dry
Wet
Wet
Dry
Dry
Wet
Wet
Dry
Dry
Wet
Wet
Dry
Dry
Wet
Wet
Dry
Dry
Wet
Wet
Dry
Dry
Wet
Wet
Dry
Dry
Wet
Wet
Average Flow
1pm (gpm)
—
53 (14)
-
-
163 (43)
413 (109)
549 (145)
382 (101)
1120 (296)
2150 (568)
2498 (660)
3066 (810)
4.5 (1.2)
2.9 (0.8)
45 (12)
291 (77)
-
5344 (1412)
-
-
_
l.OSxlO5 (28570)
-
-
_
5.14x10" (13590)
-
™
Range,
1pm (gpm)
_
23-91 (6-24)
-
-
132-310 (35-82)
223-727 (59-192)
163-988 (43-261)
189-795 (50-210)
655-3410 (173-900)
1540-3293 (407-870)
730-4290 (193-1133)
1692-9463 (447-2500)
4.0-4.9 (1.0-1.3)
2.5-4.0 (0.7-1.0)
1.1-216 (0.3-57)
45-5776 (12-1526)
-
5223-5465 (1380-1444)
-
-
_
1.03xl05-1.12xl05 (27280-29580)
-
-
_
3. 3x10" -6. 6x10" (8640-17480)
-
"*
(a)
(b)
(c)
Flow data were not collected at outfalls 003, 005, 008, Oil, 012,
013, 014, and 015.
Straight averages.
Time-weighted averages.
-162-
-------
TABLE 15
RANGE OF POLLUTANT CONCENTRATIONS AT THE
SAMPLING LOCATIONS AT SITE 2 IN mg/1
MAY - JUNE, 1977
Kl.UITAKt
Total Ju»p.nd'«
Solid*
Total DiM«lva<
Solid!
Total Iron
DUaolvad
Iron'd)
Outfill
002
On
u-29
U-2019
ttt
9-174
112-21*
00}
Dnr
U«e
11-112
93-1*8
00*
m
2-*7
113-20}
0.20-1.2
n.o.-0.2
«at
3-3»
160-339
O.U-2.2
O.l-O.l.
006
DcV
20-416
102-1J9
Vtt
B-2S37
1*5-490
007
Ort
1-40
54-245
Wat
3-119
107-411
OOJ
Orr
22-54
112-172
1.0-2.2
0.1-0.3
w«t
!»-«»
224-265
J.S-7.5
0.1-0.7
009
Drr
*-3l
114-138
0.78-1.*
0.1-0.3
Wat
55-109
151-251
3.0-5.2
0.2-0.*
KHUJTAMT
Total Suipandad
Solid.
Total Dlaaolvad
Sol Ida
Total Iran
(d)
Oliaolvad Iron
Phanel'"
Cyanidartocal)'"
A»K>nla
Sulfaca
Outfall
010A
Dry
3-37
76-123
0.37-1.3
0.1
n.d.-O.Ol
0.01
}.»-». 6
u.t
3-*»
131-253
0.91-5.9
0.1-0.4
n.d.-O.Ol
0.01
0.1-O.7
0108
Dr,
13-28
89-133
(bl
l.J
0.2
n.d.-O.Ol
n.d.-O.Ol
7.1(b)
«.t
12-702
137-239
0.74-223
n.d.-l.l
n.d.-0.06
n.d.-O.Ol
0.07-4.1
Oil
Dry
W«t
36-2694
299-691
11-23
0.2
n. a. -0.02
(c)
-
0.23-
0.43
193-270
012
Orr
Vat
29-363
222-546
1.5-26
0.1-0.6
0.01-
0.19
0.09-0.:
0.41-
1.6
52-128
013
Dry __
«at
12-1380
>33-1690
0.92-28
0.9-1.1
0.01-
0.04
0.38-
0.72
ll-ja
36-190
014
Ory
Vat
24-121
232-32*
0.9S-14
0.2-1.2
(a)
(b)
(O
(d)
(e)
(f)
No dry weather samples collected.
Only one sample analyzed.
Cyanide was not analyzed at this outfall.
n.d.-not detectable. Detectable limit for dissolved iron is 0.02 mg/1.
n.d.-noc detectable. Detectable limit for phenol is 0.001 mg/1.
n.d.-not detectable. Detectable limit for total cyanide is 0.001 mg/1.
. -163-
-------
TABLE 16
MEAN POLLUTANT CONCENTRATIONS IN rag/1 AT SITE 2
MAY - JUNE, 1977
I
I-*
ON
I
r)utt.lll
1102
DO!'"'
DO*
IK*
007
DO*
009
D10A
DID*
BU«'>
MM
»I3<">
•I4*-'
Skiing
Condition
Bry
Wet
Dry
Wet
Dry
Wet
Dry
Met
Dry
Wet
Dry
Wet
Ory
Wet
Dry
Wet
Dry
Wet
Dry
Wet
Dry
Wet
Dry
Wet
Dry
Wet
TSS
20
21
11
11
96
298
15
35
44
45
32
7)
IS
2)
19
60
831
2S7
J»2
64
•ros
749
178
no
157
216
130
223
111
227
149
241
124
201
104
18)
102
1H3
471
360
959
416
Total Iron
0.61
0.51
1.6
5.2
1.1
4.0
0.15
1.7S
-------
TABLE 17
AVERAGE MASS LOADINGS OF POLLUTANTS
DRY VS. WET WEATHER
MAY-JUNE, 1977
OUTFALL 004 - SITE 2
Ln
I
Date
•
Parameter
Total Suspended
Sol Ids
Total Dissolved
Solids
Total Iron
Dissolved Iron
5/10 (Dry)
Avg.
Cone . ,
•g/1
9
155
0.2
(b)
n.d.
Avg.
Flow,
Ip,
163
(43)
163
(43)
163
(43)
163
(43)
Avg.
Haas
Loading,
kg/hr
(Ib/hr)
0.09
(0.2)
1.5
(3.3)
0.002
(0.004)
-
5/18 (Dry)
Avg.
Cone . ,
•g/l
15
160
0.68
0.06
Avg.
Flow,
If"*
413
(109)
413
(109)
413
(109)
413
(109)
Avg.
Mass
Loading,
kg/hr
(Ib/hr)
0.37
(0.81)
4.0
(«.«)
0.02
(0.04)
0.001
(0.002)
6/9-10 (Wet)
AVR.
Cone . ,
mg/1
10
250
0.49
0.11
(0.002)
Avg-
Flow.
Ip.
(8P-)
662
(175)
662
(175)
662
(175)
693
(183)
Avg.
Mass
Loading,
kg/hr
(Ib/hr)
0.4
(0.88)
9.9
(21.8)
0.02
(0.04)
0.005
(0.011)
6/20 (Wet)
Avg.
Cone . ,
•g/1
9
203
0.36
0.04
Avg.
flow,
Ip.
401
(106)
401
(106)
401
(106)
424
(112)
Avg.
Mass
Loading,
kg/hr
(Ib/hr)
0.22
(0.48)
4.9
(10.8)
0.01
(0.02)
0.001
(0.002)
£a'0ne value only.
^ 'n.d.-not detectable. Detectable limit is 0.02 mg/1.
-------
TABLE 18
AVERAGE MASS LOADINGS OF POLLUTANTS
DRY VS. WET WEATHER
MAY-JUNE, 1977
OUTFALL 006 - SITE 2
Date
Parameter
Total Suspended
So J Ids
Total Dissolved
Solids
Avg.
Cone.,
-8/1
41
112
5/10 (Dry)
Avg.
Flow,
IP-
<*P->
1120
(296)
1120
(296)
Avg.
Mas*
Loading,
kg/hr
(Ib/hr)
2.8
(6.2)
7.5
(16.5)
5/18 (Dry)
Avg.
Cone . ,
•g/1
130
148
Avg.
riow.
IP-
(BP«>
2150
(568)
2150
(568)
Avg.
Haas
Loading,
kg/hr
(Ib/hr)
16.8
(37)
19.1
(42)
6/9-10 (Wet)
AvB.
Cone . ,
«B/l
495
244
Avg.
Flow,
IP-
«)
1851
(489)
1851
(489)
Avg.
Masa
Loading,
kg/hr
(Ib/hr)
55
(121)
27.1
(59.6)
6/20 (Wet)
Avg.
Cone. ,
•g/1
32
186
Avg.
Flow,
lp«
(8P">
2824
(746)
2824
(746)
Avg.
Mnsa
Loading,
kg/hr
(lb/l.r)
5.4
(11.9)
31.6
(69.3)
-------
TABLE 19
AVERAGE MASS LOADINGS OF POLLUTANTS
DRY VS. WET WEATHER
MAY-JUNE, 1977
OUTFALL 007 - SITE 2
Date
Parameter
Total Suspended
Solids
Total Dissolved
Solids
5/10 (Dry)
Avg.
Cone.,
•g/t
15
149
Avg.
flow.
lp-
4.5
(1.2)
4.5
(1.2)
Avg.
Haas
Loading,
kg/lir
(Ib/ht)
0.004
(0.009)
0.04
(0.09)
5/18 (Dry)
Avg.
Cone.,
•8/1
24
84
Avg.
riow.
IP-
(gp»>
2.9
(0.8)
2.9
(0.8)
Avg.
Mass
Loading.
kg/hr
Ub/hr)
0.004
(0.009)
0.01
(0.02)
6/9-10 (Wet)
Avg.
Cone.,
•g/1
36
222
Avg.
Flow,
IP-
33.3
(«.«)
33.3
(8.8)
Avg.
Haas
Loading,
kg/hr
(Ib/hr)
0.07
(0.15)
0.44
(0.97)
6/20 (Wet)
Avg.
Cone.,
•B/l
22
244
Avg.
Flow,
IP-
(gP-)
165. 4
(43.7)
165.4
(43.7)
Avg.
Haas
Loading,
kg/hr
Ub/hr)
0.22
(0.48)
2.42
(5.32)
-------
Although there were only limited data for phenols, a pattern was observed
similar to that at Site 1. Phenol concentration peaks were found to corres-
pond to rainfall intensity peaks.
There appears to be no relationship between cyanide or sulfate concentra-
tion and rainfall intensity, although limited data prevent drawing any defin-
ite conclusions. No consistent pattern exists.
As at Site 1, ammonia concentrations tended to decrease over the period
of the storm. Ammonia concentration peaked around the time of the first
rainfall intensity peak and then slowly decreased throughout the remainder of
the storm event. Apparently, the stormwater dilutes the ammonia rather than
causing a "first flush" effect. In no case was the "first flush" effect
observed.
-168-
-------
4.0 SUMMARY OF FIELD RESULTS
Based on the data collected at the two sites, the coal and coke storage
piles, and the coal and coke handling areas have the highest potential for
contaminating stormwater. Table 20 is a summary of average concentrations of
the various pollutants in these areas for the two mills sampled.
In order to determine the potential gross impact of stormwater runoff
from the mills sampled, the stormwater runoff mass loadings were compared to
the point source mass loadings which would exist under proposed BAT control.
Since BAT is EPA's next step in the control process (July, 1984), this com-
parison appears to be valid.
Table 21 compares selected annual and hourly runoff mass loadings to
point source loadings based on proposed Best Available Technology (BAT)
Effluent Guidelines for TSS. This table shows that TSS runoff loadings are
generally higher than point source loadings. In addition to »TSS, the field
data indicate that runoff from coal piles could produce substantial mass
loadings of ammonia, phenols and total iron.
In most cases at both sites, the parameter concentrations were rainfall
intensity dependent (i.e., the concentration increased with increased rainfall
intensity and vice versa). In some cases, the size and characteristics of the
drainage basin had an effect on the time lag between rain intensity and runoff
flow, and the time lag between runoff flow and parameter concentrations.
Finally, the runoff data did not show a "first flush" effect.
-169-
-------
TABLE 20
HIGHLIGHTS OF RESULTS
OF FIELD SAMPLING PROGRAMS
SITES 1 AND 2
MARCH-JUNE, 1977
1
Pollutant
ISS
TBS
TOTAL
IRON
DISSOLVED
IRON
SiCa
No.
2
A
2
1
2
1
2
1
2
1
2
1
2
1
2
1
Pocaocial
Problam
Araaa
Coal Seer.
Coka Scor.
Coka » Coal
•' Handling
Coal Scor.
Coka Scor.
Coka I Coal
Handling
Coal Scor.
Coka Scar.
Coka 4 Coal
Handling
Coal Scor.
Coka Scor.
Coka i Coal
Handling
Avaraca Uac
Concancraclonat
•1/1
853
505
,„(.)
184
471
745
,„(.)
2158
18
32.3
2.4
0.2
0.09
0.12
(a)
(b)
CO.
Thart wara Cue taapllnt polaca aaar cha coka acorafa araa ac Slca 2. Tha avarafa concancratlon*
for only ona (oucfall 013) ara ahovn..
n.d. - nona dacaccad.
.a. - noc analytad.
Pollutant
PHENOL
AMMOSU
ciums
SDLFATE
Slca
No.
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
Pocaaclal
Problam
Araaa
Coal Scor.
Coka Scor.
Coka t Coal
Handling
Coal Scor.
Coka Scor.
Coka t Coal
Handling
Coal Scor.
Coka Scor.
Coka 4 Coal
Handling
Coal Scor.
Coka Star.
Coka i Coal
Handling
Avarigt WaC
Coneaner aciona ,
•f/1
0.01
0.06
0.03<"
0.37
0.33
2.1
».!<•>
43
n.d."»
0.01
0.55(»>
».<.
-------
TABLE 21
COMPARISON OF AVERAGE ANNUAL AND HOURLY POINT SOURCE LOADINGS WITH
AVERAGE ANNUAL AND HOURLY RUNOFF LOADINGS OF TSS FOR SELECTED DRAINAGE BASINS
SITES 1 AND 2
March-June, 1977
•vl
M
Site
1
1
Out f .11
DOT
010
Oil
Oil
(Coal
Pile)
010
Oil
(Coal
Pile)
012
Oil
Average. Animal leading
•aaed on MT
Effluent Guideline! 3-°
Kg/yr(lfc/yr>
-
TSS 1850(4100)
TSS 1150(4100)
~— '
TSS 1.1>10S(4.0>10<>)
—
-
-
Aveiage Annual
ttinoff U»4ln( 7
(g/yrdb/yi)
TSS MOO (MOO)
TSS 80 (HO)
TSS ])1S (7290)
TSS 4.1«10*<9.0»I05)
TSS 1.5. 10 ().1«10 )
TS1 7760 (1.7.10 )
TSS 110 (HO)
TS1 150 (1110)
Average Hourly l.nadlnga
•aaed on HHX|MM< 1 Dnj
KAT Effluent Culdelloea 5
Kg/hrdk/kr)
-
TSS O.t (1.1)
TSS O.t (1.1)
—
TSS t.O (11)
—
-
-
Halnfall Ctentl
Ayeraite Naaa Loadlnga of rnllutanta In Runoff
I./hr(lb/krl
1/24/77
-
TSS 0.06(0. 11)
TSS 0.14(0.12)
—
—
-
-
-
1/27-1/28/JI
-
TSS 3.54 (7.S)
TSS 10. 1 (22.6)
~*
_
—
-
-
4/16/77
—
TSS 1.74 (J.gJ)
~
—
-
-
-
4/9-6/10/77
-
—
—
™"
TSS 21* (462)
—
-
-
4/20/77
—
_
"
TSS |]0 (299)
-
-
-
-------
5.0 IRON AND STEEL INDUSTRY CONTROL SYSTEMS
Stormwater controls which presently exist within the steel industry are
limited. The only system specifically designed for stormwater control exists
at Armco's Houston Works, where coal piles have been diked as a control
measure for both fugitive air emissions and stormwater runoff. Runoff col-
lected within the diked area flows by gravity to an earth pond. In nearly two
years of operation, losses from evaporation and percolation have prevented any
observed overflow from this pond. On dry days, 190,000 liters (50,000 gal-
lons) of water (equivalent to 6 mm of rain) are sprayed on the coal piles to
control fugitive dust emissions. This water is supplied from a separate
concrete pump basin which receives water from the blowdown of a coke plant
cooling tower.
Several mills contacted in this program collect stormwater runoff with
process wastewater from certain mill areas and the water is subsequently
treated at a terminal plant. This necessitates a system of combined sewers
within the plant and in several cases a holding pond is needed prior to treat-
ment to handle high flows from storms.
Many mills store their raw materials (predominantly iron ore) in concrete
bunkers and bins. Some of these bunkers have concrete floors and stormwater
has to be pumped out periodically. These bunkers were not installed for
stormwater control but rather to guard against material loss; however, they
can serve a control purpose by containing runoff which can then be pumped to a
treatment system.
-172-
-------
6.0 CONCLUSIONS AND RECOMMENDATIONS
The results of this program show that there are areas within iron and
steel mills which may pose a problem with respect to contaminated stormwater.
From the field results at the two sites sampled we can conclude the
following:
1. With the exception of runoff from coal and coke storage areas, the
majority of the basins tested in the field survey had pollutant
discharges which, on an annual basis, were less than the proposed
BAT Effluent Guidelines for the point sources located within the
basins. No data were obtained from iron ore and pellet storage
piles and active slag dumps.
2. Runoff samples from the coal storage piles indicated TSS concen-
trations to be typical of urban runoff while TDS values were
approximately twice typical urban runoff concentrations.
3. At both plants, runoff from coal and coke handling areas and the
coke plant area generated higher hourly mass loadings of total
suspended solids than the average hourly loadings for point sources
based on maximum 24-hour loadings in the proposed BAT Effluent
Guidelines.
4. The coal storage areas sampled in this study had much lower runoff
concentrations for TSS, TDS, total iron, and sulfate than those
found in runoff from utility coal containing higher percentages of
sulfur.
Based on the program we recommend the following:
1. Site specificity should be the most important consideration when
evaluating surface runoff problems at individual plants because many
factors affect the runoff loadings, including:
o Plant activities and operations
o Climate
o Soil conditions
o Size of drainage basins
o Location of activities and. operations relative to one another
o Neighboring industries and urban areas
o Proximity of plant to receiving waters
o Plant size
-173-
-------
If stormwater runoff is found to be a problem at a specific site,
more work should be performed to determine the feasibility of cost-
effective controls for mill areas identified as having potential
stormwater contamination problems.
2. At some plants, it may be beneficial to treat stormwater from cer-
tain areas to bring runoff mass loadings down to the same order of
magnitude as point sources based on the proposed BAT Effluent
Guidelines. The most likely area is coal piles where it may be
beneficial to treat runoff for total suspended solids, total dis-
solved solids, and total iron.
3. Future studies should be directed to quantify and qualify stormwater
runoff from the iron ore storage and slag disposal areas.
4. Because many steel plants are built on permeable soils next to
waterways, groundwater contamination from storm events is possible.
Future programs should investigate potential groundwater contami-
nation from the industry.
-174-
-------
REFERENCES
1. Staff Report, National Commission on Water Quality (Washington, D.C.,
April 1976).
2. Report to the Congress. National Commission on Water Quality,
(Washington, D.C., April 1968).
3. Armco Steel Corporation, Houston Works, Storm Water Sampling Program
at Armco Steel Corporation, Houston Works, Houston, Texas, April 1,
1976.
4. Kaiser Steel Corporation, Fontana, California, Annual Summary of
Surface Runoff performed for the California Regional Water Quality
Control Board, 1975.
5. Code of Federal Regulations. Title 40 - Protection of Environment,
Part 420 Iron and Steel Manufacturing Point Source Category, Effluent
Guidelines and Standards as of July 1, 1976.
6. Water Pollution Abatement Technology: Capabilities and Cost, Iron
and Steel Industry, National Commission on Water Quality, November
1975.
-175-
-------
Differential Tracing of Oily Waste and the Associated
Water Mass by Tagging with Rare Earths
by
D. L. MeGown
Energy and Environmental Systems Division
Argonne National Laboratory
-177-
-------
Abstract
A method was developed for tagging oily waste with dysprosium and the
associated fresh water with samarium. Neutron activation analysis permitted
determination of rare earth concentrations as low as 40 ng/L in 15-mL water
samples. Shipboard sampling procedures were developed that allowed measure-
ment of the three-dimensional distribution of the spreading wastes and associ-
ated water. The method was applied in two field experiments that involved
tracing oily wastes and polluted water from the Indiana Harbor Canal (IHC)
into Lake Michigan.
For a summer, floating-plume experiment, about 1400 shipboard samples
were collected. Employment of the dual-tracer technique led to the following
results: (.1) after artificial mixing into the water column by a passing ship,
the tagged oil did not immediately resurface, and (2) there were no distin-
guishable differences between the movement of the oil and water over 4 km of
travel.
During a winter, sinking-plume experiment, 1200 lake-water samples were
collected from a boat and from the raw-water intakes of Chicago's South Water
Filtration Plant (SWFP). These data provided positive evidence of the intake
of IHC effluent and oily waste at the SWFP. The different tracers for the
oily waste and underlying water gave evidence of separate pathways to the
SWFP, reflecting differing transport modes for surface and near-bottom waters.
-178-
-------
INTRODUCTION
It is important to distinguish between the dynamics of waste-receiving
waters and the motions of the associated oily wastes within and upon these
waters. Adequate understanding of these separate but complementary motions
is essential for the development of predictive models. A method was developed
for simultaneously tagging both oily wastes and the underlying water, each
with a unique tracer, and for determining their individual motions in fresh
water.
Traditionally, two classes of substances have been used in tracer
studies to tag water or pollutants in water: fluorescent dyes and radionu-
clides. The tracing approach in the present study uses Rare Earth Element
(REE) tags and neutron-activation analysis (NAA) for tag detection. Among
the advantages of this method are the following:
1. No radiation hazard to the environment exists. ,Q
2. Detection of REE concentrations as low as 6 x 10 g
in a 15-mL sample is possible.
3. Tracer loss to suspended matter and sediments appears
negligible.
4. Rare earth element (REE) tags can be selected that are
non-toxic to the environment.
5. Detection and quantification of several REE tracers
simultaneously is simplified.
Two tags were needed to trace simultaneously oily wastes and the waste
receiving waters. The rare earth elements dysprosium (Dy) and samarium (Sm)
were chosen as the most suitable for the present study because
1. They have high detectability and short half-lives, and
2. Their natural occurrence in the coastal waters of Lake
Michigan is at concentrations below the limits of detect-
ability provided by the methods used in the present study.
Water sample analysis (forty-eight 15-mL raw-water samples)
showed that no natural Dy or Sm could be detected.
Application of the tracing technique by Argonne National Laboratory
(ANL) was focused on a series of experiments conducted on the movement of
oily wastes in the coastal waters of southwestern Lake Michigan from the
Indiana Harbor Canal (IHC)(Fig. 1). Several possible sources of oily pollu-
tants exist in the canal. Effluents from an oil refinery, steel mills, and
municipal sewage treatment plants contain such pollutants (Snow, 1974). Small
spills from oil transfer facilities and runoff from industrial sites also con-
tribute to the oily-waste loading. For this study no individual source was
tagged, but rather a simulated oily waste was prepared and spilled in the
canal.
The source of water in the IHC is Lake Michigan. Industry draws water
-179-
-------
DEPTHS IN FEET
METERS
LAKE
MICHIGAN
SOUTH WATER
FILTRATION PLANT
Fig. 1. Location Map for Study Area and Current-Meter
Positions (I-IV) in Southwestern Lake Michigan,
-180-
-------
from the lake, uses the water for industrial cooling processes, and then re-
turns the water to the canal 5-8 C° warmer than the lake. The canal water
moves into the lake as a thermal plume, sinking during much of the time in
winter, and forming a surface plume during the remainder of the year. Signi-
ficant differences in the oil-transport regime are expected during floating-
and sinking-plume conditions. The most difficult transport/dispersion regime
to observe in the field is that of the sinking plume, because it is difficult
to follow and sample a tagged mass of canal water that forms only a thin layer
(commonly 1- to 5-m thick) as it spreads over the bottom. Nevertheless, the
importance of sinking-plume transport in carrying contaminants from the IHC
to municipal water intakes in southwestern Lake Michigan dictated devoting
significant efforts and resources to the tagging and tracing of IHC water
during sinking-plume conditions.
When the wind is from the southern quadrants, during the winter sinking-
plume conditions, the City of Chicago's South Water Filtration Plant CSWFP)
experiences periods of high hydrocarbon odors that require using expensive
activated charcoal in the treatment process. Prior to this study, it was
thought that the odors came from IHC effluent which had migrated northwestward
along the lake bottom (Vaughan, 1970) passing over the water intakes of the
SWFP; however, it had not been proved. This study provides concrete evidence
that IHC effluent enters the raw water intakes of the SWFP.
EXPERIMENTAL PROCEDURES
Tagging Considerations
Dysprosium (Dy) was the agent chosen to tag the simulated oily waste,
and samarium (Sm) was selected to tag the IHC water. A third agent, the
fluorescent dye rhodamine-WT, was also used to tag the canal water so that,
with the use of a pump and fluorometer, the tagged water mass could be fol-
lowed by the sampling boat.
Waters of the IHC are turbid and highly polluted. Introduction of an
REE tag into the IHC water column in the ionic form would hazard its removal
from solution by the formation of insoluble compounds by reaction with some of
the many species in solution in the IHC or on the bottom of the canal, and
adsorption on suspended solids.
Channell (.1971) showed that the persistence of lanthanum complexed with
EDTA (Ethylenediaminetetraacetic acid) in waters that are in contact with
bottom sediment is almost twice that for lanthanum dissolved in acid. In view
of Channell's findings, chelation of REE tags was considered necessary. The
water tracer, Sm, was complexed with DTPA (Diethylenetriaminepentaacetic acid).
DTPA was used because it was cheaper and was easier to complex with Sm than
EDTA.
Numerous bench tests (>50) were conducted to determine how well the
simulated refinery/steel-mill oily waste retained the Dy tag. The simulated
waste was made up of equal parts of 30-W motor oil, #2 diesel fuel, and engine
drain oil. The results of the bench testing are discussed in detail in
McCown et al. (1978) and McCown et al. (1976). A "loss factor" of
0.4% loss of Dy/mL of tagged oil/mL of receiving water was determined from
the bench tests.
-181-
-------
Sampling Procedures
Documentation of transport and mixing of a plume requires many data
points. This requirement dictates the removal of water samples from several
depths and at many locations in the lake. In addition, the position of the
boat when samples are being taken must be known to a high degree of accuracy.
Therefore, a three-dimensional water-sampling system was designed that inclu-
ded:
1. A Motorola Mini-Ranger microwave ranging system interfaced
with an x-y plotter for real-time positioning,
2. A fluorometer, pump, and hoses to follow the dye patch, and
3. A dynamically depressed faired cable with attached small
tubes that extend to the depths to be sampled.
When tracking a sinking plume, an additional large tube was placed in
the faired cable, which was used to draw bottom water samples. A schematic
of the sampling system, as set up for a sinking plume, is shown in Fig. 2.
The intermediate-depth water samples were pumped onboard, through
4.7-mm I.D. nylon tubes, by a Masterflex multichannel tubing pump. The flow
rates, through =20 m of tube, was 1.67 mL/s. This relatively low flow rate
was sufficient because the volume of each sample was only 15 ml. The inter-
mediate-depth tubes were all the same length, so that sample-removal delay
times were equal.
The bottom sample collected for the sinking-plume study was drawn
through a 9.6-mm I.D. nylon tube, which was connected to the faired cable at
a point 1 m above the depressor fin (Fig. 2). The end of the bottom-sampler
tube was connected to a 7.3 kg (16 Ib) shot that trails the depressor fin by
»12 m so that the shot drags the bottom. The bottom sampler had a relatively
high flow rate of 12.5 mL/s and was driven by a positive-displacement pump.
Water from the bottom was fed directly to the fluorometer. A Valved-tee at
the fluorometer inlet directed a small portion (3.33 mL/s) of water to the
sampling manifold for subsequent NAA at Argonne. All the samples for NAA were
drawn in new, 15-mL lab-grade polyethylene Polyvials.
In addition, a surface-skimming water sampler was built; however, no
acceptable means could be determined for calibrating it. Samples drawn by
the surface skimmer were collected in the field, but the results of these
samples are presented in only one instance (see discussion on Fig. 7 in
section on Sinking-Plume Results) as an indication of tracer presence or
absence.
During the sinking-plume study water samples were.also drawn from the
raw-water streams of the shore and crib intakes at the SWFP.
NAA Procedures
The collected water samples were analyzed at ANL's CP-5 reactor. Com-
plete details of the analytical procedures and an error analysis are discussed
in McCown et al. (1978). The significant features of the NAA procedures,
developed at ANL, are the following:
-182-
-------
MINI-RANGER RECEIVER/TRANSMITTER
FOR REAL TIME POSITIONING ON X'Y
PLOTTER
CABLE FAIRING
)EPTH VARIES
KflTH BOTTOM
TOPOGRAPHY
DEPTH FIXED AT
CONSTANT SPEED
WATER INLET
V-FIN (0.6m WING SPAN)
Fig. 2. Schematic of the Three-Dimensional Underway Water-
Sampling System as Set Up for a Sinking-Plume .
-183-
-------
1. No pre-Irradiation sample preparation is necessary except
evaporation of the liquid phase and cleaning of the outside
of the Polyvial,
2. The samples are irradiated in the same Polyvials in which
they were collected,
3. Seventeen samples per hour can be analyzed by the method,
4. The per sample cost of analysis, including personnel, is
$5.00, and _1 _8
5. Minimum detectability of 6 x 10~ g for Dy and 2 x 10~ g
for Sm is possible with samples of only 15 mL.
Current Measurements
For the sinking-plume experiment, lake-current data were gathered at the
four stations (I-IV) shown on Fig. 1. The moorings were placed in about 10 m
of water, and a Bendix Q-15R current-meter sensor was positioned about 1.5 m
above the bottom at each station; at Station I, a second meter was positioned
about 5 m above the bottom in order to obtain information on the vertical
structure of the currents. Detailed information on current-meter mooring de-
sign, recording methods, and processing is contained in Harrison, et al. (1977),
In-situ current-meters were not placed for the floating-plume experi-
ment because of the relatively short duration and limited horizontal extent of
the experiment. ANL current-meters were moored offshore of the Calumet Area
for a different study so that general nearshore current information was
available.
RESULTS AND DISCUSSION
Floating-Plume Results
On September 26, 1976, the IHC water was tagged in-situ with 4.5 kg
(10 Ib) of samarium that ha" been complexed with DTPA, and with 3 kg of rho-
damine WT dye. The samarium/DPTA/Rh-WT tag was dispersed into the top meter
of IHC water by pumping the tag through a pipe with many small holes drilled
along one side. The tagging process required 10 min and occurred in the cen-
ter of the canal at a point about 500 m from the lakeward entrance (Fig. 3).
The Dy-tagged, simulated oily waste was poured on the surface of the
IHC effluent at the same time and in the same location as the in-situ water
tag (Sm). The oily-waste and tracer consisted of 57 L of the waste mixed with
0.45 kg (1 Ib) of Dy. The Dy had been dissolved in a 50% acetic acid solution.
Shortly after tagging and before sampling had commenced, an empty ore
carrier passed directly through the center of the dye patch. Winds on
September 26, were light from the south. There were no significant waves, and
the lake current was <0.05 m/s flowing to the northwest. The plume that re-
sulted from the release of the tracers and subsequent mixing by a passing ore
carrier was mapped four times in the next 10 hr, and over 1300 15-mL water
samples were collected. Sampling depths were 0.5, 1.0, 1.5, 2.5, and 3.5
meters. About 800 of the collected water samples were analyzed by neutron
activation. The lowest Dy concentrations that were contoured at the 0.5 m
level (80 ng/L = 200 counts) for each of the four plume mappings are plotted
in Fig. 3.
-184-
-------
I
!-•
CO
I
Fig. 3. REE Tagging Location and Contours of Lowest Detectable
Amount of Dy that was Contoured for Plume Mappings 1-4,
Floating-Plume.
-------
Table 1 lists the elapsed time after tagging, the percent of the origi-
nal Dy that was accounted for in each plume, the average concentration inte-
grated over the depth, and the average dilution ratios relative to plume 2.
Table 1. Floating Plumes 1-4 Average Concentration Parameters
Plume
1
2
3
4
Elapsed Time
After Tagging (s)
4.8 x 103
1.1 x 10k
2 x 101*
3 x 104
% Dy
Accounted For
131a
83
77
67
Average Concentration
Over the Depth (Kg/m3)
4.8 x 10~7
4.6 x 10~7
2.2 x 10~7
4.9 x 10~8
Dilution
Compared to 2
-
1:1
2:1
10:1
High percentage due to non-synopticity of measurement. (See text for
explanation.)
Diffusion coefficients were calculated using elapsed time and observed
concentration variances for the last two cloud mappings (3[K-N] and [0-R])
at the 0.5 m level. These two mappings were in Lake Michigan proper and thus
would be subject to lake-type diffusion. The calculated horizontal diffusion
coefficients are shown in Table 2 with appropriate values reported by Murthy
(1976) for Lake Ontario. The values of these diffusion coefficients indicates
that our data are similar to data from the Great Lakes found using dye as the
tracer.
Table 2. Calculated Horizontal Diffusion Coefficients (m2/s)
Time
Dy/Oil
K
K
Sm/Water
K
K
Murthy's Values
K
2.5 x 101* sec
2.2 8.4
0.2 6.9
1.5
'10.5
aLake Ontario (dye)
Longitudinal
CLateral
-186-
-------
The percent of the Dy/oil and Sm found above 1.25 m (based on the
total amounts accounted for), is plotted as a function of time after tracer
release in Fig. 4. Each separate plume mapping is indicated by time in
seconds on the bottom horizontal axis. The transition from the area protected
by the Inland Steel Landfill to the nearshore zone of Lake Michigan is indica-
ted by the vertical dashed line.
Floating-Plume Discussion
The Dy loss factor (0.4% Dy lost/mL oil/mL water) that was determined
above may be used to estimate if significant amounts of Dy dissociated from
the oil during the floating-plume study. The Dy is detected in plume 3 over
a volume of 1.6 x 109 L (400m x 1000m x 4m). That volume would indicate
only a 4% loss of Dy from the oily waste. Therefore, the loss of Dy from the
oily waste into the water through plume 3 is not thought to be significant.
In plume 4, Dy can be detected over a volume of 6.2 x 109 L (,600m
x 1000m x 4m) for a maximum Dy loss from oil of 15%. It is doubtful if even
a 15% loss would significantly affect the trends exhibited by the results.
Note that the Dy tracer loss from the oily waste discussed above is
not necessarily related to the "percent Dy accounted for" in Table 1. If
some of the Dy were lost from the oily waste, it could remain in nearby
waters and thus be sampled and detected.
The "percent Dy accounted for" was determined by integrating the con-
centrations at sampling points over the three-dimensional structure of the
measurable plume and comparing that weight with the original amount of Dy in
the oil spilled. Deviations from 100% are probably due to the spacing of the
grid and the diffusion of Dy to undetectable concentrations at the plume
edges. Plume 1, which shows 131% of the original Dy accounted for appears
strange, but the seeming disparity is due to the non-synopticity of measure-
ment of that plume. Plume 1 was sampled in a region where the flow of the
canal was relatively fast and the direction of movement was changing from
directly out of the canal to a more lakeward heading. Consequently, measure-
ments' of plume 1 were the least synoptic of the four measured plumes.
Figure 4 shows that the amount of Dy and Sm above 1.25 m changes with
time and horizontal position. Both Sm and Dy migrate toward the surface un-
til the plume passes into the nearshore zone of Lake Michigan. After passing
into the nearshore zone, the plume mixes downward.
The initial movement of both tracers seems to have occurred because
warm canal water flows out and rises over the inflowing, colder Lake Michigan
water. However, when the IHC plume enters Lake Michigan proper, it is then
subjected to large-scale, lake-type diffusion and it is more dense than ini-
tially due to cooling. These two factors cause the downward migration of the
plume when it enters the lake proper.
After the oily wastes were mixed into the water column, their movement
did not differ significantly from the underlying Sm-tagged waters. As indi-
cated above, mixing coefficients for the Dy/oil and Sm/water are similar.
-187-
-------
oo
oo
100
90
o
oc.
80
70
60
T I I I
PROTECTED ZONE
T
Dy/OIL
Sm
4,8 x 10'
NEARSHORE
ZONE
I.I x 10*
TIME (SECONDS)
2x|04 3x|04
Fig. 4. Percent of the Total Dy/Oil and Sm Accounted for by Measurement
Found above 1.25 m vs Time after Tracer Release.
-------
Moreover, Fig. 4 indicates no significant differences in the gross vertical
movement of the Dy/oil and Sm/water.
The above results indicate that the oily waste remained suspended in
the water column and did not float back to the surface. The tendency for oils
to remain in suspension and not float back to the surface has also been noticed
and documented by McAuliffe (1977), Shaw (1977), Brown and Huffman (1976),
Brown and Searl (1976), Peake and Hodgson (1966, 1967), and Gordon et al. (1973)
Their results may be summarized as follows:
1. Under agitation, oils break up and disperse into the
underlying waters where most remain suspended and move
with the water, and
2. Colloidal particles (10~ to 10~ m) remain in suspension
for extended periods.
A distinct difference exists between the small, limited spill (57 L)
artificially mixed into the water column discussed in this study and a major
spill involving millions of liters. Extrapolations from the results above to
the case of a major spill may not be valid, and different relative motions
between the oil and underlying water might be expected for a major spill.
This would be particularly true during the initial phases of spreading of a
major spill where gravity, viscosity, and surface tension are dominant.
However, when that type of slick becomes thin and oil mixes into the water
column, behavior similar to that reported in this study might be found.
Sinking-Plume Results
On March 2, 1977, IHC water was tagged in-sitw with 13.6 kg (30 Ib) of
samarium that had been complexed with DTFA, and with 7.5 kg of rhodamine WT
dye. The samarium/DPTA/Rh-WT tag was dispersed in the same manner as for the
floating plume. The tagging process required 10 min and occurred at the "REE
tagging location" (Fig. 5), which was approximately 900 m upflow from the
point where the canal water was sinking. The zone of convergence of canal
and lake water was easily determined with a temperature probe and is indicated
in Fig. 5 as "surface temperature convergence." As the survey vessel NEPTUNE
proceeded lakeward, the surface water temperature dropped from 7.5 to 1.8°C
within a few meters. Canalward of the sinking zone, the water temperature was
vertically isothermal. Just lakeward of the sinking zone, however, the
temperature profile showed warm (=!4.00C) water near the bottom, overlain by
colder (*1.8°C) water.
The Dy-tagged simulated oily waste was poured on the surface of the
IHC effluent at the same time and in the same location that the in-situ water
tag (Sm) was injected. The oily waste and tracer consisted of 170 L of the
waste mixed with 1.4 kg (3 Ib) of Dy dissolved in a 50% acetic acid solution.
A progressive vector diagram of local winds and bottom currents is
shown in Fig. 6 for the period March 1-4 for the winds and March 1-5 for
the bottom currents. The current meters were deployed for this study for the
period January 5 to March 26, 1977. Complete details of the current structure
during the period are published in Harrison et al. (1977). In summary, major
velocity components are well-correlated within the study area for the period
of measurement and suggest the existence of a rather uniform flow field.
-189-
-------
BOTTOM CONTOUR
DEPTHS IN FEET
Fig. 5. REE Tagging Location, Temperature-Convergence Zone, and
Positions of Sampling Transects, Sinking-Plume.
-190-
-------
12"
BOTTOM CURRENTS
(8-MIN VECTORS)
1.0 -•
MIDWAY WINDS
(8-HR VECTORS)
MAR I
MxlO
1.0
Fig. 6. Progressive-Vector Diagrams for March 1-4, 1977,
Midway Airport Winds, and March 1-5, 1977, Lake
Michigan Bottom Currents at Station III (Fig. 1).
About 1050 water samples were collected in Polyvials aboard the
NEPTUNE while tracking the dye that was used to tag the IHC effluent. Of the
1050 samples approximately 200 were collected in Transects A, B, and C (Fig. 5)
in the canal entrance and near entrance areas. All 200 were analyzed by NAA
for their Dy and Sm concentration. Of the remaining 800 samples collected in
the Calumet Harbor area, 350 were analyzed (that is every other or every
third sample along a given transect was analyzed unless a high value was ob-
served, in which case samples on either side were analyzed) and only ten of
those indicated the presence of Sm (canal water); however, many of the samples
from the Calumet Harbor indicated the presence of Dy (simulated oily waste).
All sampling transect locations are plotted on Fig. 5. Each transect
has an arrow indicating the direction of boat travel, and the ends of each
transect correspond to the first or last sampling station. The distance
between sampling stations was about 100 m.
Figure 7 is a plot of the vertical Dy concentration encountered in
Transect A that passed directly through the convergence'zone. The sampling
stations are numbered in ascending order corresponding to the direction of
boat travel. Contour values for Dy in the figure are nanograms per liter
-191-
-------
TO LAKE MICHIGAN
LIGHT TOWER AT INDIANA
HARBOR CANAL ENTRANCE
250
VO
N>
E
CO
z
a.
LU
Q
o
Zi
Q.
CO
10
(BOTTOM)
10
SAMPLING STATIONS
15
-ENTRANCE TO
INDIANA HARBOR
TURNING BASIN
CONCENTRATIONS IN
500 meters
Fig. 7. Transect A: Dy Concentrations. . (All 90 samples collected were analyzed.
The dashed contours indicate how the figure would look if concentration
values for the surface samples could have been considered.)
-------
_
(ng/L); that is, 10 g/L. A region of sinking Dy/oily waste: is clearly shown
between stations 1 and 5. The dashed contours indicate how the data would
have been contoured if data from the surface sampler were considered and
showed that some of the Dy/oil crossed the sinking, zone on the surface.
Plots of all other transects and further explanation of the data are in
McCown et al. (1978). In summary, all of the transects in the Calumet Harbor
area indicated fairly uniform vertical mixing of the Dy tracer. In a number
of instances, a singular point occurred having a very high Dy concentration,
probably due to a large oil particle in the sample.
About 240 water samples were drawn from the raw-water streams for the
shore and crib intakes at the SWFP. These 15-mL samples were drawn every 10
min for almost all the period between 2200 hr on March 3 to 1240 hr on March 4,
1977. From 1240 hr to 1600 hr March 4, 1977, samples were drawn every 20 min.
Water-intake samples were drawn in the same type of new 15-mL Folyvials that
were used for the shipboard samples.
All 240 of the samples drawn at the SWFP were analyzed by NAA and those
that contained detectable Sm and/or Dy are shown on Fig. 8, a plot of relative
amounts of tracer as a function of time for each raw-water intake. Relative
amount is computed as the amount of a given REE found in a Polyvial sample
divided by the quantity of that REE that was released in the IHC effluent
times 100.
Sinking-Plume Discussion
Results of the winter plume-tracking field study show that the effluent
from the IHC enters the City of Chicago's water-purification system at the
SWFP's Dunne Crib and shore intakes. (Any contaminants entering the Chicago
water system are removed at the filtration plant by activated charcoal treat-
ment, but the removal process is expensive.) A detailed picture of the
effluent's path to the crib or shore intake cannot be established from the
tracer data because no measurements of the plume were made in the vicinity
of the intakes. However, two paths may exist for transport of IHC effluent
to the SWFP's water intakes. Each path is governed by the nature of entry of
IHC effluent into the lake, incompletely mixed near-surface water taking one
path and well-mixed, 4°C bottom water taking the other.
During the development of a sinking plume, all of the IHC water does
not sink to the bottom when it mixes with Lake Michigan waters at the subduc-
tion zone. If all of it sank to the bottom, the near-surface temperatures
immediately lakeward of the convergence zone would be the same as ambient .
Lake Michigan water. But temperature profiles immediately lakeward of the
convergence zone show the near-surface water to be 1-1.5°C, not near 0°C as
is typical for ambient Lake Michigan water when ice is on the lake. There-
fore, some portion of IHC water (the portion that does not mix completely)
stays near the surface and moves toward the shore due to the action of winds
blowing from the southerly quadrants. This was the portion of the IHC effluent
that was tracked by the NEPTUNE, and it typically follows a path along the
shoreline.
The other path, followed by the portion of IHC water that sinks to the
-193-
-------
10"
ID'9
10
.-10
icr1
Sm CRIB INTAKE
I
Sm SHORE INTAKE
I I I
(—
Z
cc
UJ
a.
Dy CRIB INTAKE
10*
irr'l
1
M
(
Oy SHORE INTAKE
ia«
in*
jt
•
. *
1
b
2200 2400 0200 0400 0600 0800
1000 1200 1400 1600
4
Fig. 8. Relative Amount of Sm or Dy in the Samples Collected
at the SWFP as a Function of Time. (Dashed line indi-
cates the minimum detectable amount; two dots at a
sample indicate the sample was irradiated twice and
shows the value of each irradiation.)
-194-
-------
bottom, roughly parallels the 9.1-m (30-ft) depth contour (Fig. 1). This iso-
bath coincides with the lakeward terminus of the IHC entrance channel. Such
bottom water would move directly into the lake and then northwestward along
the coast, missing the area enclosed behind the Calumet Harbor Breakwater.
This sinking-plume portion of the IHC effluent would follow the second path to
travel directly to the SWFP. Topographically, the depths of the crib intake
and of the IHC entrance channel are about the same, 9.1 m (30 ft). Effluent
flowing out of the IHC along the bottom would tend to remain near the 9.1-m
(30-ft) contour as it moved northwestward parallel to shore, and this path
would carry it directly to the crib intake. Further evidence of the existence
of a direct lakeward path is found in the high concentrations of Sm detected
in the crib samples and the fact that these samples were drawn at the same
time that the NEPTUNE was drawing water samples in Calumet Harbor, water
samples that contained Dy (oily waste) but virtually no Sm (canal water).
Figure 8 shows the relative Sm amounts in the raw-water samples collect-
ed at the SWFP. Seven of the eleven crib samples contained higher concentra-
tions of Sm than any of the eight shore samples. Two of the crib samples
(March 3 at 2220 hr and March 4 at 0430 hr) contained very high concentrations
of Sm. In contrast, only fourteen of the crib samples contained Dy and all
but two of these were in a three-hour-and-ten-minute span (2210-0120). Thirty-
one of the shore samples contained Dy, however, and these were almost equally
spaced throughout the entire 17-hour sampling interval at the SWFP.
These results indicate that significantly more Sm (the water tracer)
went to the crib intake than to the shore intake. Also, significantly more
Dy (the oil tracer) went to the shore intake than to the crib intake. This
appears logical because a large portion of the oily waste (Dy) would tend to
remain on the surface and thus would be blown toward the shore by the SE
winds that occurred during the experiment. Therefore, the oily waste would
have had a higher probability of being drawn into the shore intake. The water
(Sm tracer), however, would be more likely to sink at the subduction zone and
be carried along the 9.1-m (30-ft) isobath with the prevailing northwesterly
current mentioned above; further evidence of this partitioning of the oily
waste and IHC water is found in the fact that so little Sm (the water tag) was
seen in the Calumet Harbor area.
A simple turbulent diffusion model was used to estimate the mixing of
the water tag (Sm) during the experiment. The limited Sm data at the shore
and crib intakes and the possible partitioning of the tag described above,
however, do not permit validation of even the simple model proposed. Never-
theless, a general comparison between model predictions and measured dilution
ratios at the crib intake may prove instructive.
Table 3 presents dilution ratios for the Sm samples collected during
the experiment at the SWFP's crib and shore intakes. The dilution ratio is
defined as the ratio of the initial concentration in the water at the tagging
location to the concentration in the collected samples. Initially 13.62 kg
of Sm were mixed in the IHC waters over a patch about 24 m x 2 m in area and
1 m in depth, yielding an initial concentration of 0.284 g/L.
The mixing of the tagged portion of the IHC effluent was modeled in two
steps. The dilution ratio due to mixing as the patch moved from the tagging
location out onto the lake was modeled first. As the patch moved to the
-195-
-------
Table 3. Sm Concentrations and Dilution Ratios for Water Samples
Collected at the SWFP
Date
3 Mar.
it
4 Mar.
ti
ii
it
M
ii
ii
it
M
3 Mar.
it
ii
4 Mar.
i>
it
ii
it
Time, Concentration,
CDT yg/L :
Crib- Intake
2220
2310
0030
0300
0430
0450
0720
1040
1200
1220
1520
Shore- Intake
2200
2250
2340
0300
0450
0710
1100
1200
Raw-Water Samples
61. Ob
7.7b
2.8b
6.7b
102. lb
6.1b
2.4b
5.0b
2.1b
5.1b
1.6
Raw-Water Samples
2.9b
1.6
4.0b
1.6
1.6
2.7b
1.8
1.9
% Error
2.0
8.3
23
9.6
1.4
10
24
13
27
12
34
22
42
18
46
41
25
34
32
Dilution Ratio, —
CF
4.7 x 103
3.7 x 10"
1.0 x 105
4.2 x 10"
2.8 x 103
4.7 x 10"
1.2 x 105
5.7 x 10"
1.4 x 105
5.6 x 10"
1.8 x 105
9.8 x 10"
1.8 x 105
7.1 x 10"
1.8 x 10s
1.8 x 105
1.1 x 105
1.6 x I05
1.5 x 105
Minimum detectable amount equals 1.4 yg/L.
Indicates average of two irradiations.
Cj., initial concentration, equals 0.284 g/L. CF is final concentration.
-196-
-------
vicinity of the SWFP's shore and crib intakes, the dilution ratio due to
mixing in the lake was combined with the previous ratio and in addition was
calculated with a model employing 4/3's-law mixing and another employing
linear mixing. The two models produced minimum dilution or center of patch
dilution ratios of 6.1 x 103 and 3.0 x 103 respectively in the vicinity of
the water intakes. The exact location of the patch relative to the intakes
is unknown; however, dilution ratios calculated for the center of the patch
are of about the order of magnitude (Table 3) of those determined from
sampling at the crib and shore intakes. Details of the models and the
coefficients used are discussed in McCown et al. (1978).
SUMMARY
A novel method was developed for simultaneous tagging of both oily
wastes and the underlying water, each with a unique tracer, and for determin-
ing their individual motions in fresh coastal waters.
The overwater sampling aspect of the experiment required the develop-
ment of an underway three-dimensional water sampling system for the collection
of numerous 15-mL water samples. An NAA technique permitted the analysis of
about 17 samples per hour with no pre-irradiation sample preparation other
than cleaning of the outside of the Polyvial. The samples were collected and
subsequently irradiated in the same Polyvial. Tracer detection limits were
6 x 10 for Dy and 2 x 10 g for Sm in the 15-mL samples. The method was
applied to trace simulated oily waste and the underlying, polluted water from
the IHC into Lake Michigan during both the floating- and sinking-plume condi-
tions.
The results of the floating-plume experiment indicated that:
1. When the tagged oil was subjected to severe downward
mixing by a passing ore carrier, it did not resurface
immediately but remained mixed in the water column,
2. Lake diffusion coefficients calculated from the Dy and Sm
data were similar to diffusion coefficients determined by
others in the Great Lakes using other tracer techniques,
and
3. After the oil was mixed into the water column by the ore
carrier, there were no distinguishable differences between
the movement and diffusion of the oil/Dy and the movement
of the water/Sm.
The results of the sinking-plume experiment indicated that:
1. IHC effluent is definitely transported to the raw-water
intakes of Chicago's SWFP under certain lake and meteoro-
logical conditions,
2. A partitioning of the oily wastes and underlying water
(resulting in separate pathways to the SWFP's raw-water
intakes) was made apparent by the employment of the dual
tracer system, and
3. Simple model estimates of IHC plume dilution at the SWFP
were supported by the experimental measurements.
-197-
-------
Acknowledgments
I would like to thank Jack Ditmars, ANL, for his careful review of the
manuscript. The contributions of Wyman Harrison, ANL, co-principal investi-
gator for this study, are acknowledged as are those of the Water Resources
Section, ANL, who helped with the field aspects of the project. , In addition,
the assistance of the City of Chicago's Department of Water and Sewers is
gratefully acknowledged. This program was funded by the U.S. EPA (pass-
through to U.S. ERDA) and the Illinois Institute for Environmental Quality.
REFERENCES
Brown, R.A. and H.L. Huffman, Jr., Hydrocarbons in Open Ocean Waters, Science
191, p. 847-849, 1976.
Brown, R.A. and T.D. Searl, Nonvolative Hydrocarbons Along Tanker Routes of
the Pacific Ocean, Offshore Technology Conference 1, p. 259-274, 1976.
Brown, R.A., et al., Distribution of Heavy Hydrocarbons in Some Atlantic
Ocean Waters. Proceedings Joint Conference on Prevention and Control
of Oil Spills, American Petroleum Institute, Washington, p. 505-519,
1973.
Channel, J.K., Activable Rare Earth Elements as Estuarine Water Tracers,
Ph.D. dissertation, Stanford University, 1971.
Gordon, D.C., et al., Laboratory Studies of the Accommodation of Some Crude
and Residual Fuel Oils in Seawater, J. Fish. Res. Bd. Can., 30, p. 1611-
1618, 1973.
Harrison, W., D.L. McCown, L.A. Raphaelian, and K.D. Saunders, Pollution of
Coastal Waters off Chicago by Sinking Plumes from the Indiana Harbor
Canal, Argonne National Laboratory Report ANL/WR-77-2, 1977.
McAuliffe, C.D., Dispersal and Alteration of Oil Discharged on a Water Surface,
Proc. Symposium, Fate and Effects of Petroleum Hydrocarbons in Marine
Ecosystems and Organisms, Seattle, November 10-12, 1976, D.A. Wolfe, Ed.,
Pergamon Press, New York, p. 19-35, 1977.
McCown, D.L., W. Harrison and W. Orvosh, Transport and Dispersion of Oil-
Refinery Wastes in the Coastal Waters of Southwestern Lake Michigan
(Experimental Design — Sinking Plume Condition), Argonne National
Laboratory Report ANL/WR-76-4, 1976.
McCown, D.L., K.D. Saunders, J.H. Allender and J.D. Ditmars, Transport of Oily
Pollutants in the Coastal Waters of Lake Michigan: An Application of
Rare Earth Tracers, Argonne National Laboratory Report ANL/WR-78-2,
1978 (in press).
Murthy, C.R., Horizontal Diffusion Characteristics in Lake Ontario, Journ.
Phys. Oceanogr., 6(1), p. 76-84, 197.6.
-198-
-------
Peake, E. and G.W. Hodgson, Alkanes in Aqueous Systems, I, Explanatory Investi-
gations on the Accommodation of C -C n-Alkanes in Distilled Water and
Occurrence in Natural Water Systems, j. Am. Oil Chemists' Soc., 43,
p. 215-222, 1966.
Peake, E. and G.W. Hodgson, Alhanee in Aqueous Systems, II, The Accommodation
of C-J-C-. n-Alkanes in Distilled Water, J. Am Oil Chemists' Soc., 44,
p. 696-702, 1967.
Shaw, D.G., Hydrocarbons in the Water Column, Proc. Symposium, Fate and
Effects of Petroleum Hydrocarbons in Marine Ecosystems and Organisms,
Seattle, November 10-12, 1976, D.A. Wolfe, Ed., Pergamon Press, New York,
p. 8-18, 1977.
Snow, R.H., Water Pollution Investigation: Calumet Area of Lake Michigan,
Vols. 1 and 2 of U.S. EPA, Great Lakes Initiative Contract Prog. Kept.
No. EPA-905/9-74-011-A, 306 pp., 1974.
Vaughan, J.C. and P.A. Reed, Progress Report on Water Quality of Lake Michi-
gan Near Chicago, Water Sewage Works, 120(5):73-80, 1973.
-199-
-------
TP 8162
ESTIMATION OF FUGITIVE HYDROCARBON
EMISSIONS FROM AN OIL REFINERY BY
INVERSE MODELING
By
Andrew A. Huang and Sara 3. Head
AeroVironment Inc.
145 Vista Avenue
Pasadena, California 91107
-201-
-------
ABSTRACT
In order to evaluate an ambient air measurement approach for the
definition of fugitive emissions, the hydrocarbon emissions from an oil refinery
were estimated through inverse diffusion modeling. The AVQUAL model was
utilized to predict the hydrocarbon source terms from the measured ambient
hydrocarbon data. The methodology employed in this study is unique in the
following ways: (1) several source areas were handled simultaneously; (2) a tracer
gas was used to define the meteorology on a real-time basis; and, (3) to minimize
experimental errors an over-specified problem was solved to determine the
hydrocarbon emissions.
Five different types of hydrocarbon sources for the refinery were
assumed. The X/Q values resulting from these sources were computed for all
fifteen downwind receptor locations. Subsequently, a linear programming approach
with a non-negative Q constraint was used to solve for the emission rate, Q.
Results showed qualitative agreement with direct source testing data.
-202-
-------
1. INTRODUCTION
Many industrial, commercial, and domestic activities emit gaseous
hydrocarbons and other organic compounds into the atmosphere. These
sources include all kinds of fuel burning, solvent usage, and waste disposal
operations, as well as the more obvious chemical processing and petroleum
refining and marketing. In the South Coast Air Basin (SCAB) of California,
petroleum refineries contribute a significant portion of the stationary
emissions of hydrocarbons.
Refinery hydrocarbon emissions are primarily of the fugitive type.
Leakage from valves, fittings, flanges, pumps, and storage tanks, etc.,
comprises a major portion of the emission, although some stack emissions
result from the operation of process heaters, boilers, incinerators, and
flaring. It is impractical, if not impossible, to test all the fugitive emission
sources in a refinery to arrive at the emission factors. One approach to
inventory the emissions is by means of direct source testing, with emission
factors derived by statistically combining the results of selective test
sources. Another approach is by ambient testing.
An ambient testing approach requires a workable diffusion model and
measurement of ambient hydrocarbon levels in the vicinity of a refinery.
When the atmospheric links between emission strength and ambient
concentrations are established, the hydrocarbon emission factors can be
estimated through inverse diffusion modeling. Theoretically, the ambient
testing approach accounts for every emission source if all emissions are
well-mixed, and thus provides a more accurate' inventory in a rather simple
way. This paper describes the methodology used in testing a refinery in the
SCAB, for the purpose of evaluating a specific ambient testing approach.
Recommendations are also presented for improving the technique.
-203-
-------
2. TEST DESCRIPTION
An ideal test site should be relatively compact, and reasonably distant
from other hydrocarbon emission sources such as gas and oil wells and other
industries so that interference would be insignificant. Also, a heavy traffic
flow area is not desirable. Furthermore, the test refinery should be
representative of the refining operations in the SCAB.
The most favorable meteorological condition for testing would include
a stable atmosphere, and light and uniform wind fields. This kind of
condition usually prevails in the early morning hours when drainage flow
exists. Conducting testing in the early morning hours has the added
advantage of minimum interference from traffic emissions. The test site
further requires enough downwind working space for the sampling.
Extensive site survey and literature research of the past
meteorological data led to the selection of a refinery located in Paramount,
California. Although no direct control over the test conditions (i.e., the
meteorology itself) could be exercised, it was possible to select those days
for testing which exhibited the most favorable conditions in the refinery
area. This was done through the services of a forecaster who could predict
when those days would most likely occur.
2.1 Test Equipment and Approach
To establish links between emissions and ambient concentrations, a
tracer gas was released at the refinery. By knowing the tracer gas releasing
rate and the receptor point concentrations, the source-receptor relation-
ships could be identified through diffusion modeling. Sulfur hexafluoride
(SF6) was used as the tracer. This gas is ideal because it is inert, has very
low background level (<6 ppt), and can be detected down to ppt concentra-
tion.
Both SF6 and hydrocarbons were measured simultaneously so that the
relationship established for a given period of time by the SF6 was clearly
applicable to the hydrocarbon case for the same time span. Real time
monitoring would have been desirable but not practical, owing to the number
of receptors required (15 downwind plus 1 to 2 upwind receptors). As an
alternative, one-hour integrated samples were taken and immediately
analyzed by using a flame ionization detection (FID) hydrocarbon analyzer
and an electron capture detection SF6 analyzer. This was possible by
moving the AeroVironment mobile laboratory (Airlab) to the test site during
the entire test period. The possibilities of sample alteration were thus
minimized. In addition, meteorological parameters needed for diffusion
modeling were monitored by using instruments installed on the Airlab. A
mechanical weather station, secured on the top of one of the refinery
storage tanks, served as a backup.
Some integrated hydrocarbon samples were collected and sent to an
analytical laboratory for analysis by a gas chromatography-mass
-204-
-------
spectrometry (GC-MS). This way the hydrocarbon concentrations
determined by different analytical means, namely FID and GC-MS, could be
correlated, and hydrocarbon speciation data obtained.
o Ambient Air Sampling
Two types of ambient air sampling systems were employed: one for
FID analyses, and one for the GC-MS analyses. To collect samples for the
FID analyses, ambient air was drawn from 6 feet above ground through a
Tygon tubing and a prefilter by a Spectrex Model AS-100 gas sampling pump,
and delivered into a 5-liter Tedlar bag. The delivery rate was pre-set so
that at the end of the one-hour sampling period the Tedlar bag was inflated
enough to provide the analysis needs.
For the GC-MS analyses two samplers were used. One of the samplers
was the same Tedlar bag sampler described above. The other was an
adsorption sampler, in which an adsorption tube filled with activated
charcoal was used to concentrate the ambient hydrocarbons. Ambient air
was drawn from a sample point 6 feet above ground through a Tygon tubing,
the adsorption tube, a rotameter, and a filter by a Nepture Dyna-Pump
Model 2. After sampling, the exposed adsorption tube was immediately
capped and later sent away to the laboratory as a set with the Tedlar bag
sampled during the same time span. These bag samples were analyzed for
low molecular weight hydrocarbons (C6 and below). Adsorption tube
samples were analyzed for higher hydrocarbons (C6 and above). The
overlapping of C4 - C6 from both samplers served as a cross-check.
o SF6 Releasing and Analyzing
The SF6 tracer gas was passed through a copper coil, a liquid trap, a
mass flowmeter, and a rotameter, and released into the atmosphere 28 feet
above ground. Release flowrates were measured by use of a Hastings Linear
Mass Flowmeter and continuously recorded by a strip chart recorder.
Analysis of SF6 content in the ambient air sample was accomplished
by use of an AID Model 511 Portable Gas Chromatograph with electron
capture detector. This instrument, produced by Analytical Instrument
Development Inc., Avondale, Pennsylvania, was capable of detecting down
to the part per trillion (ppt) level. Instrument calibration was performed
before and after the instrument was used on each test day.
o Hydrocarbon Analyzing
A Beckman Model 6800 Air Quality Chromatograph was employed at
the study site to analyze the ambient air sample hydrocarbon contents. This
instrument operates on the FID principle, and was capable of measuring
total hydrocarbon (THC), methane (CH«), and carbon monoxide (CO). The
measurements of CO content gave information pertaining to traffic
interference.
-205-
-------
Instrument calibration was carried out before and after the instrument
use for each test day. The hydrocarbon calibration gas was methane. Thus,
THC data reported were expressed as CH4.
FID instrument response is related not only to the concentration of the
hydrocarbon being measured, but also to the "effective carbon number" of
the hydrocarbon compound. The effective hydrocarbon number varies
depending on the number of carbon atoms in the molecule and on the type of
compound. Thus, an FID instrument calibrated with CH4 generally would
would respond to the heavier hydrocarbons, and the data collected could be
misleading. In order to compensate for the deficiency, some ambient air
samples from the vicinity of the refinery were collected for the GC-MS
analyses. These samples were taken side-by-side with those analyzed by the
on-site FID instrument. This approach constituted the "field calibration" for
the FID results. The GC-MS analytical results also could reveal hydrocarbon
speciation information.
o Meteorological
An anemometer installed 8 meters above in the Airlab was used for
measuring wind speed and wind direction. Sigma meters were used to
measure ow and
-------
Testing was carried out on four days in November - December 1976.
Figure 1 illustrates the typical receptor locations, the SF4 release point,
and the Air lab location for a given test day. The sampler locations for GC-
MS analysis are also indicated.
-207-
-------
ROSECRANS AVE.
Jefferson St.
LEGEND
SFg Release
Airlab
Receptor Location
OSamP'er Location(GC-MS
Scale of Miles
ARTESIA FREEWAY
FIGURE 1. Schematic representation of test site for 30 November 1976.
-208-
-------
3. DIFFUSION MODEL
The diffusion model used for this study was the AVQUAL model.
AVQUAL is a microscale diffusion model developed from Taylor's turbulent
diffusion theory (Taylor 1921), which explicitly incorporates ground
roughness and heat flux. It is simple and accurate. The presence of
inversions is accounted for. Furthermore, this model has the capability of
simulating unsteady effects, and wind shifts can readily be incorporated. A
detailed description of this model has been documented by Lissaman (1973).
In the AVQUAL model, the ratio of ground level ambient concentra-
tion of a given receptor point to the source strength is a function of
meteorological conditions. Parameters involved are wind speed, wind
direction, and vertical and horizontal dispersion speeds etc. This relation-
ship is expressed as follows:
n /TIT ay r2
where x = ambient receptor point concentration at time T
Q = source strength
u = mean wind speed for time span T
x = receptor downwind distance from the source
y = receptor crosswind distance from the source
z = source height
r2 = x2 + y2 + z'2
z' = z/A
A = ratio of vertical dispersion speed to horizontal dispersion
speed; i.e., a /a
As is evident, x and y are dependent on the mean wind direction for time
span T.
For a given test run, X, Q, u, x, y and z are all measurable
parameters. T is equivalent to the time span over which the integrated
samples are collected. Furthermore, A and aw are calculable by equating
a v to aw and A to av /a, where a=u • ae. In the model validation, Q, u, x, y,
z, A, and a are fed into AVQUAL to calculate X. The theoretical A'
values thus calculated are compared with the measured A' values.
Significant discrepancies observed between them would indicate that the
measured meteorological data are insufficient to validate the model. In this
case, it is necessary to carry out a model calibration.
-209-
-------
4. TEST RESULTS AND MODEL VALIDATION/CALIBRATION RESULTS
4.1 Test Results
Testing was conducted on 23, 2k and 30 November, and 1 December
1976. Except for the 24 November testing, the meteorological conditions
during the test hours were favorable. The atmosphere was calm with a
stability class of E or F (Pasquill, 1962). Wind was generally directly from
the north with a wind speed between 3-7 mph, a typical drainage flow
condition in that area. Occasionally, though, the wind would shift greatly
and the wind shift persisted for prolonged periods of time. These were the
runs which were difficult as far as model validation was concerned.
During the 24 November testing the atmospheric conditions were too
calm. There was virtually no wind most of the time. The refinery plume at
that time' would disperse more or less evenly toward all directions.
Consequently, all of the receptors including the "upwind" one recorded about
the same concentrations; THC was around 10 ppm, CH4 around 5 ppm, and
SF6 around 20 ppt
Table 1 shows the results of successful test hours. A receptor location
designated with "A" indicates an upwind site, while that designated with 1
through 15 was downwind site. There were two upwind sites for Run 15 as
shown in Table 1.
It is obvious that the contribution from the refinery plume to a
specific receptor point would be the difference in concentration between
that receptor point and the upwind point. This is true for all species
measured, namely SF6 , THC, CH4, and CO, provided that the incoming wind
is uniform and that no interference is present. These net results will be
used for the diffusion model validation/calibration, and for the hydrocarbon
emission predictions.
Some of the samplers collected air samples for both FID and GC-MS
analyses as mentioned previously. A relationship was established for the
non-methane hydrocarbons from the Douglas refinery air samples:
(NMHC)rr, MQ = 1.36 (NMHC)_IPk
VJ(^-MD rlD
The correlation factor of 1.36 was reasonable considering that the FID
instrument was calibrated by using CH4 standards. This correlation factor
was eventually applied to the final emission prediction data.
4.2 Model Validation/Calibration Results
In the model validation of AVQUAL by use of SF6, Q, u, x, y, z, X,
and av were fed into AVQUAL +o calculate X as mentioned in Section 3.
The theoretical X values thus calculated were compared with the measured
lvalues. S'gnificant discrepancies were observed between them, indicating
the measured meteorological data were insufficient to validate the model.
-210-
-------
TABLE 1. Summary of test results.
Run No: 1
Date: 11/23/76
Time: 0330-0430
Wind Speed: 1.6 m/sec
Wind Direction: 340 deg
0 : 0,07 m/sec
»e: 7.2 deg
Temp: 11 C
Stability Class: E
Q^p : 288,960 ug/sec
ir6
Receptor
A
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
SF6 (ppt)
3.0
6.4
3.0
1525.0
116.0
9.6
_
6.0
.
150.0
420.0
3.0
3.0
_
13.0
21.0
THC (ppm)
5.4
6.6
6.7
7.3
7.9
6.0
6.0
5.6
6.3
5.8
5.4
6.9
6.0
7.1
6.1
7.6
CH^ (ppm)
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
CO (ppm)
3.8
4.0
3.8
3.7
3.7
4.0
3.5
3.7
3.6
3.8
3.9
4.0
3.8
3.9
3.8
3.8
-211-
-------
TABLE 1. (Continued)
Run No: 9
Date 11/30/76
Time: 0330-0430
Wind Speed: 2.5 m/sec
Wind Direction: 335 deg
a : 0.02 m/sec
0g : 4 deg
Temp: 9 C
Stability Class: F
Q<.p, : 299,604 pg/sec
*C6
Receptor
A
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
SF6 (ppt)
19.5
26.0
22.5
4300.0
13.5
16.5
_
20.5
460.0
_
17.5
37.0
740.0
1010.0
16.5
17.5
THC (ppm)
5.1
6.0
6.1
9.4
9.2
5.5
6.0
5.5
7.0
5.2
5.9
5.4
6.0
6.5
5.3
6.0
CH^ (ppm)
2.6
2.6
2.5
2.6
3.2
3.0
2.6
2.5
3.0
3.0
3.1
2.8
2.6
2.8
2.9
3.0
CO (ppm)
2.5
2.8
2.2
2.3
2.5
3.3
2.3
2.1
2.8
3.1
5.2
3.2
2.5
2.5
2.9
2.6
-212-
-------
TABLE 1. (Continued)
Run No: 10
Date: 11/30/76
Time: 0430-0530
Wind Speed: 2.0 m/sec
Wind Direction: 330 deg
wo
0.01 m/sec
5de
Temp: 8VC
Stability Class: F
Q<-F : 299,604 yg/sec
Receptor
A
1
2
3
4
5
6
7
8
9
10
11
12
13
1*
15
SF6 (ppt)
13.5
16.0
17.5
4800.0
13.5
13.0
13.0
24.0
_
13.0
11.8
130.0
500.0
22.5
11.5
13.5
THC (ppm)
4.7
5.3
5.7
10.6
5.5
4.6
6.0
5.3
_
4.6
5.0
5.7
6.3
5.1
4.6
4.9
CH^ (ppm)
2.7
2.5
2.4
2.4
3.1
2.7
2.7
2.4
_
2.8
2.8
2.6
2.7
2.7
2.8
2.9
CO (ppm)
2.3
2.2
1.5
1.5
2.0
2.9
1.8
1.6
-
2.7
2.1
2.8
2.3
2.1
2.7
2.2
-213-
-------
TABLE 1. (Continued)
Run No: 11
Date: 11/30/76
Time: 0530-0630
Wind Speed: 2.2 m/sec
Wind Direction: 3*0 deg
a : 0.01 m/sec
woe: 4deg
Temp: 8°C
Stability Class: F
Qsp : 299,60* pg/sec
Receptor
A
1
2
3
4
5
6
7
8
9
10
11
12
13
15
SF6 (ppt)
5.6
9.8
5.6
4700.0
10.5
6.3
7.7
7.0
455.0
6.3
9.0
62.0
1200.0
425.0
9.4
THC (ppm)
4.7
6.4
5.5
7.6
10.0
4.8
5.6
4,9
7.9
4.7
4.5
5.3
6.1
6.1
4.6
CH^ (ppm)
2.6
2.4
2.3
2.4
3.1
2.6
2.7
2.5
2.4
2.6
2.6
2.6
2.4
2.6
2.6
CO (ppm)
4.1
5.0
1.9
2.4
2.7
6.6
4.5
2.3
1.1
4.8
2.0
2.7
1.8
2.5
3.1
-214-
-------
TABLE 1. (Continued)
Run: 15
Date: 12/1/76
Time: 0800-0900
Wind Speed: 1.6 m/sec
Wind Direction: 330 deg
a : 0.09 m/sec
wae : 7 dee
Temp: 13°X:
Stability Class: F
Qop : 286,220 ug/sec
^6
Receptor
Al
A2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
SF6 (ppt)
76.0
82.0
— t
200.0
1100.0
34.5
96.0
150.0
1400.0
71.0
450.0
480.0
_
_
480.0
500.0
498.0
THC (ppm)
5.2
6.0
7.3
7.8
6.7
5.4
6.4
6.5
7.2
6.0
6.4
6.7
OS
6.9
6.0
5.2
6.8
CH^ (ppm)
2.2
2.3
2.3
2.4
2.4
2.2
2.3
2.2
2.3
2.4
2.3
2.3
2.3
2.3
2.3
2.3
2.3
CO (ppm)
5.0
5.5
4.1
5.2
7.4
4.9
5.5
4.7
5.3
6.0
6.1
4.4
5.0
5.3
5.1
4.1
5.9
-215-
-------
This was most likely caused by the nonrepresentativeness of the Airlab
meteorological measurements. Specifically, the meteorological conditions
at the SF6 release location were different from those observed at the
Airlab. This is understandable because the refinery facilities, such as
storage tanks, cooling towers, and fractionation towers etc., could cause
slight wind direction changes and more horizontal mixing. Perhaps more
important is the fact that the refinery operation's heat output would create
significantly more vertical mixing than what was observed at the Airlab. As
a result, development of diffusion factors empirically by model calibration
was necessary.
In the model calibration, mean wind direction, X , and av were
determined by trial and error to minimize the difference between predicted
and measured values of SF6. The other parameters such as Q, u, and z were
influenced by the refinery operations and thus the measured data were used.
Table 2 shows the comparison of the measured and calibrated meteoro-
logical data for the test runs where satisfactory agreement between
predicted and measured SF6 data could be obtained. Figure 2 illustrates the
typical predicted SF6 isopleths and SF6 concentrations at each receptor
point for each calibrated run. Also given is the measured SF6 concentration
at each receptor point for comparison purposes.
The calibrated mean wind direction was consistently greater than the
measured value by 15 to 25 degrees, as shown in Table 2. The calibrated
values of both av and X were approximately two orders of magnitude larger
ttVan the measured values. Since X is the ratio of av to a, it follows that
only av, the vertical dispersion speed, was significantly impacted by the
refinery operations; whereas a, the horizontal dispersion speed, was in the
same order of magnitude as was observed in the Airlab. Thus, it is possible
that had the meteorological instruments been installed near the SF6 release
point within the refinery complex the AVQUAL model would have validated
satisfactorily using the measured meteorological data.
The model calibration factors listed in Table 2 were used directly for
the hydrocarbon emission predictions in the following section.
-216-
-------
Observed
Predicted
Run: 9
Date: 11/30
Hour: 0330-0430
Wind Speed: 2.5 (rn/sec)
Wind Direction: 355 (deg)
Stability Class: F
Source Strength: 29960"+ (pg/sec)
50 100m
0.10
(O.Oi
0.10
FIGURE 2. Comparison of observed (solid line) and predicted (broken line)
SFg isopleths; at each receptor point observed and predicted
(parenthses) SF, concentrations are also given in wg/m .
-217-
-------
TABLE 2. Comparison of measured and calibrated meteorological data.
Run No.
1
9
10
11
15
Mean Wind Speed
(m/sec)
1.6
2.5
2.0
2.2
1.6
Mean Wind Direction
(deg)
Measured
340
335
330
340
330
Calibrated
360
355
350
355
355
ay(m/sec)
Measured*
0.07
0.02
0.01
0.01
0.09
Calibrated
4.4
2.5
1.8
2.5
2.5
A
Measured**
0.35
0.11
0.06
0.07
0.46
Calibrated
10.0
3.5
3.0
5.5
5.0
I
IS}
M
00
* assuming a = a
** calculated as follows:
w
where aQ in radian
-------
5. HYDROCARBON EMISSION ESTIMATION AND DISCUSION5
One of the initial tasks in determining the method for predicting
hydrocarbon emission rates was to define the sources. The oil refinery
chosen for this study has many fugitive hydrocarbon emission sources. Each
source is characterized by the particular process associated with it.
After the emission source is defined, the next step was to determine
the most practical and accurate way of modeling them. The AVQUAL
model, discussed in Section 3, was then utilized to calculate the relative
inrmacts of the different source types. Finally, a numerical technique was
devised to give a best fit emission rate in accordance with the data. The
method is as follows:
1. For the emission source definition, it was decided to classify the
numerous sources into areas of similar source types. Five source types
were chosen for their dis+inct emission charateristics, namely, process
areas and emulsion plant (Q,), asphalt plant (Q2), loading-unloading
facilities (Q3), fixed roof storage tanks (Q4), and floating roof storage
tanks (Q5). Because not all of the sources of any of the five types
occurred together in a centralized area, smaller areas were defined
which contained only a particular category. The final scheme
consisted of 31 defined areas: five areas of Qi type, two areas of Qa
type, two areas of Q3 type, fourteen areas of Q4 type, and eight areas
of Q 5 type. A diagram of the defined areas and their source category
is depicted in Figure 3.
2. The defined area is further subdivided into a number of point sources
whose optimum spacing was determined. It was found that by reducing
the spacing of point sources the error of estimating downwind
concentrations from an area source was reduced. However, it was also
found that further reduction from a spacing of 25 feet would not
result in a significant improvement in the model output. For example,
reducing the spacing from 25 feet to 12.5 feet would increase the
accuracy by less than 2%. Thus, a spacing of 25 feet was used in the
emission prediction. Within the unit area of 25 feet square, it was
assumed that the source points were distributed evenly. This approach
allowed for equal emphasis of the different source categories on an
area basis.
3. The AVQUAL model can be expressed in a simplified form as follows:
= f (meteorology, coordinates)
Thus, X/Q can be calculated solely from the meteorological data and
the coordinates. For a unit area k of emission type j, its contribution
to the ambient concentration at receptor point i (Xlt k ) is identical
to Qjf ijk . Thus, the total contribution of refinery emissions to
receptor i can be expressed as
-219-
-------
O
Ql: Process Areas, Emulsion Plant 0-5)
Q2: Asphalt Plant (6-7)
Q3: Loading-Unloading Facilities (8-9)
Fixed Roof Storage Tanks (10-23)
Q5: Floating Roof Storage Tanks <2*-31
m mm
' €l»«^^
AIRLAB
RELEASE
DOUGLAS OIL COMPANY
Paramount, California
FIGURE 3. Source definitions used for the refinery emission prediction.
-------
5 k
X = I I
j ijk
=l k=l
In Equation (2), Jj was measured and f j j k was calculated from the
calibrated meteorological data; with Qj 's the only unknowns to be
solved. Therefore, theoretically, only five downwind receptor points
were needed to solve the equation since only five source types were
assumed.
4. Realistically, however, more receptor points were needed for the
accurate prediction. In this study as many as 15 downwind receptors
were used. A numerical analysis technique was therefore necessary to
solve the over-determined set of equations.
5. A linear programming approach with a non-negative Q constraint was
used for solving the over-determined case. This technique, however,
was very sensitive to the variation of receptor concentrations; a noise
of + 5% in receptor concentrations was about its tolerable limit.
Eventually it was necessary to smooth out the receptor concentration
data by fitting the receptor points with the same downwind distance to
a 2nd-order polynomial equation. The hydrocarbon emission prediction
thus obtained gave the most consistent results.
6. The hydrocarbon concentration correlation factor obtained in Section
4.1 was applied to the predicted emission rates.
Table 3 presents the results of the emission rate predictions. Emission
rates from five different source categories as well as the total rate are
tabulated for each run. Run 1 indicates the Q5 source type was the
predominant source, while the rest of the emissions came from the Qi
category; Q2, Q3 and Q4 were negligible sources. Run 9, however, shows
that Q, and Q5 types were approximately equally weighted as far as
contribution to the total emission rate, whereas the contribution from
source types Q2, Q3 and Q4 were still negligible. Runs 10, 11, and 15
predicted that the emissions were predominantly from the Q, type.
One would not expect to observe a change in the predominant source
type from hour to hour, such as the case observed for Runs 9, 10 and 11
which were conducted on November 30. Also, one would not expect the
change from day to day because of the fairly steady refinery operations.
Therefore, the average of the five runs should be more significant as far as
interpretation. As shown in Table 3, the arithmetic average of the source
type emissions indicates that Q, contributed about 60% of total emissions,
while Q5 contributed about 40%; Qa, Q3 and Q4 were essentially negligible.
The average total emission rate from the Douglas refinery was predicted to
be 26 Ib/hr.
-221-
-------
TABLE 3.
Hydrocarbon emission predictions for various source
categories of the refinery in Ib/hr.
Run No.
1
9
10
11
15
Avg.
Q!
8
14
12
18
23
15
Q2
0
0
0
0
0
0
Q3
0
0
0
0
0
0
Q*
0
0
3
1
0
1
Q5
39
10
0
0
0
10
Total
47
24
15
19
23
26
= process areas and emulsion plant
= asphalt plant
= loading-unloading facilities
= fixed roof storage tanks
= floating roof storage tanks
-222-
-------
This predicted value is about one order of magnitude lower than that
in the KVB source testing results, presented in Table k (KVB, 1977). Also
presented is the KVB estimation of the Douglas refinery hydrocarbon
emissions by using the Southern California APCD and AP-42 emission data.
The KVB source testing was conducted at the same time periods as the
ambient testing was performed to formulate a coherent testing program.
Compared to these values, AV's prediction appears to be too low
quantitatively. Qualitatively, however, it is sensible. A major part of the
emissions came from the process areas and the floating roof storage tanks,
The asphalt plant contributed little, if any, to the total emissions. Fixed
roof storage tanks normally should contribute to the source term.
Nevertheless, the ambient testing was conducted during the early morning
hours when the fixed roof storage tanks were "breathing in." Thus, the
prediction of negligible emissions from this particular source type is
reasonable.
Possible causes of the discrepancy between the AV prediction and the
KVB source testing results include hydrocarbon layering under the extremely
stable atmospheric conditions, and heavier hydrocarbon retainment within
the refinery because of the boundary wall. Source testing data clearly
indicated CH4 emissions from the refinery (KVB, 1977). However, AV's
ambient testing data showed little or no difference for the upwind and
downwind CH4 data. It is possible that under the very stable conditions
most of the CH4 emitted would diffuse upward, and thus was not detectable
in the ground level. Heavier hydrocarbons (in reference to air) under the
same stable atmospheric conditions would tend to diffuse downward. They
were, however, more or less contained in the refinery by the existence of
the wall in the downwind side of the refinery. The results of the above
possible causes would be the reduction in the ambient air hydrocarbon
concentrations; and thus, the underestimation of emissions.
-223-
-------
TABLE 4. Comparison of KVB's preliminary hydrocarbon emissions source
testing results with the results calculated by using the SCAPCD
and AP-42 data for the refinery; in Ib/hr.
Stack Emissions
Fugitive Emissions
Tank Storage and Transfer
TOTAL
SC APCD
11
8*
90
185
AP-42
26
123
95
244
KVB/ARB
24-Hr.
13
54
95
162
Av. Night
13
54
67
134
-224-
-------
6. CONCLUSIONS AND RECOMMENDATIONS
The following conclusions may be drawn as a result of the ambient
testing of the refinery:
1. The AVQUAL model can be adequately applied to the Douglas refinery
environs.
2. Emission prediction results for various source types are qualitatively
sound. The predominant sources during the test hours were the
process areas and the emulsion plant (source type Q,) and the floating
roof storage tanks (source types Q5).
3. Quantitatively, the prediction is one order of magnitude lower than
predictions by accepted emissions factors or from source testing. This
is possibly caused by the hydrocarbon layering in the immediate
atmosphere, and also by the existence of a wall at the refinery
boundary in the downwind side.
For future hydrocarbon source emission predictions using the ambient
testing approach, the following are recommended:
1. Meteorological parameters should be measured at a site which is
representative of the source area. The possible heat output from the
area should be taken into consideration in the site selection.
2. Receptor locations should be set up close to the source area in order
to observe maximum impact.
3. Instead of extremely stable atmospheric conditions (Class F), the
testing can be done under less severe conditions, such as Class D, to
provide better air mixing. Alternatively, if the study has to be
performed under Class F conditions, the collection of air samples at
three different heights would be desirable to detect any layering
problem.
k. More receptors should be arrayed crosswind for the same downwind
distance to facilitate the data smoothing technique.
5. Avoid selection of source area with wall perimeters so that no
boundary type problem will exist. In so doing, the source-receptor
relationship would be solely governed by atmospheric diffusion.
In summary, hydrocarbon emission prediction through diffusion
modeling represents a unique approach to the inventory problem. No
statistical analysis is needed, as is with the case of direct source testing.
Also, theoretically, every hydrocarbon source is accounted for since the
mixed refinery plume is collected at the downwind locations.
-225-
-------
The apparent weakness of this approach lies on the fact that the
measurement system might provide too much of an error for the model to
handle. A variation of + 5% for the downwind concentrations proved to be
intolerable. Eventually, it was necessary to smooth out the data to cut
down the noise.
One way to improve the sampling accuracy would be to arrange the
receptor locations closer to the source so that maximum impact can be
observed, and thus the percent error of the receptor point concentrations
can be minimized. Another way would be to utilize real time, continuous
hydrocarbon monitors for all receptor points. This way the handling of air
samples is cut down to minimum, and thus the most accurate results are
achieved. However, this approach is expensive. The methods employed in
this study, i.e., on-site analyses of integrated samples, should be the most
economical and acceptable procedures.
-226-
-------
7. REFERENCES
AeroVironment (1977): Ambient monitoring and source receptor identifica-
tion of hydrocarbon emissions from oil refinery, AV FR 700^.
KVB (1977): Control of hydrocarbon emissions from stationary sources in
the California South Coast Air Basin, KVB 5804-714.
Lissaman, P.B.S. (1973): A simple unsteady concentration model explicitly
incorporating ground roughness and heat flux. Paper presented at the
66th annual meeting of the Air Pollution Control Association, Chicago,
Illinois. (AV TP 311, AeroVironment Inc., Pasadena, Ca.)
Pasquill, F. (1962): Atmospheric diffusion . D. Van Nostrand, New York.
Taylor, G.I. (1921): Diffusion by continuous movements. Proc. London
Mathematical Society 20, 196-212.
-227-
-------
MEASUREMENT OF FUGITIVE EMISSIONS FROM PETROCHEMICAL PLANTS
by
D. R. Tierney
Z. S. Khan
T. W. Hughes
MONSANTO RESEARCH CORPORATION
DAYTON LABORATORY
Station B, P.O. Box 8
DAYTON, OHIO 45407
-229-
-------
ABSTRACT
Fugitive hydrocarbon emissions from petrochemical plants were measured as
part of an emission assessment program being conducted by Monsanto Research
Corporation under contract with the U.S. Environmental Protection Agency (EPA).
The objective of this study was to provide accurate data on fugitive emissions
which would reflect current technology and operating practices in the petro-
chemical industry. Field sampling and analysis was conducted at four plants
producing monochlorobenzene, butadiene, ethylene oxide/glycol, or dimethyl
terephthalate.
Fugitive sources at petrochemical plants include pumps, valves, flanges,
process drains, compressors, agitators, sample valves and relief devices. A
random sample of potential sources at each plant was selected and screened
with an organic vapor analyzer to determine sources having hydrocarbon emission
rates in excess of 0.5 g/hr. These sources were enclosed in a flexible plastic
tent and sampled using a portable sampling train. On-site hydrocarbon sample
analysis was performed using a gas chromatograph with flame ionization detec-
tion to identify and quantify each organic component. Sampling data showed
the average hydrocarbon emission rates from pumps, valves, flanges, and com-
pressors to be 22 g/hr, 4.5 g/hr, 1.4 g/hr and 30 g/hr, respectively. Of the
total number screened, the estimated percentages of pumps, valves, and flanges
having significant (>0.5 g/hr) emission rates were 23%, 7%, and 2%, respec-
tively. The number of such sources at petrochemical plants having signifi-
cant emissions is ^20% lower than the number of such sources at petroleum
refineries.
-230-
-------
INTRODUCTION
BACKGROUND
Hydrocarbon emissions from petrochemical plants can be divided into two broad
categories: stack and fugitive emissions. Stacks and/or vents identified as
principal hydrocarbon emission points are considered to be controlled sources.
Hydrocarbon emission points other than stacks and/or vents are considered to
be fugitive sources. Fugitive emissions may occur due to accidents, inadequate
maintenance, or poor planning, although many fugitive emissions occur even in
the absence of such conditions and are unavoidable characteristics of some
process operations.
The emission factors generally employed for fugitive hydrocarbon emissions are
based on a study conducted to determine miscellaneous hydrocarbon emissions
from petroleum refineries in the late 1950's by Los Angeles County (Ref. 1).
These emission factors are being used by both industry and air pollution
agencies to estimate emissions from fugitive sources, with little or no modi-
fication to reflect current equipment, technology, and operating practices.
Current data on the amount and composition of fugitive hydrocarbon emissions
from petrochemical plants also are not available; they are at best estimated.
Therefore, sampling efforts were necessary to obtain data of the required
quality and accuracy and to identify and quantify emissions of potentially
toxic chemical substances.
The EPA is currently conducting a study on fugitive hydrocarbon emissions from
petroleum refineries to determine the adequacy of established emission factors
in light of current refinery technology and equipment. Information on fugitive
emissions from petrochemical plants generated in this study will be used by the
EPA to compare fugitive emissions from such plants with those from petroleum
refineries.
SCOPE OF STUDY
Fugitive hydrocarbon emissions were measured at plants manufacturing mono-
chlorobenzene, butadiene, ethylene oxide/glycol, and dimethyl terephthalate.
Hydrocarbon emission rates were determined for pumps, valves, flanges, process
drains, compressors, agitators, sample valves, and relief devices. On-site
analysis was performed to identify individual organic components. Sampling
results were extrapolated to evaluate fugitive emissions from all potential
point sources within each process.
Data on fugitive emissions from petrochemical plants and petroleum refineries
were compared. Petroleum refinery data were acquired from the study conducted
by Los Angeles County (Ref. 1).
-231-
-------
SAMPLING AND ANALYSIS
SOURCE SELECTION
Potential fugitive sources at a petrochemical plant are numerous and varied.
Due to budget and time constraints it was not feasible to determine whether
each source type was emitting fugitive hydrocarbons. In order to define the
number of sources required to achieve statistically relevant data three factors
were required. Two of the factors are the desired confidence level and the
corresponding error. On this program , these factors were established as the
95% confidence level and a corresponding error of ± 70%. A third factor
needed to define the number of samples required is the standard deviation of
the measurements. Assuming that sampling data follow a normal distribution
with no bias, the following relationship exists:
S =
x
Ii/n
where:
= estimated relative standard deviation
X,
E = error (difference between true emissions and
mean of measurements, ±70% for this program)
n = number of samples
t = value from statistical tables for "t" distribution
A plot of sx versus n is given in Figure 1 which was used to predict the mini-
mum number of sources required from estimated relative standard deviations to
obtain ±70% uncertainty.
20 40 60 80 100 200 100
ESTIMATED RELATIVE STANDARD DEVIATION. %
Figure 1. Number of samples required for fugitive sampling.
-232-
-------
Fugitive source types which were found at the various petrochemical plants
included pumps, valves, flanges, compressors, relief valves, sample valves,
process drains, and agitators. Each plant site was inspected prior to sampling
to determine the existence of any unusual process configurations. Unusual con-
figurations, such as open vessels containing hydrocarbons, were not found at
the plants sampled.
To guarantee that a sufficient number of samples was obtained, approximately
40 samples of each source type were tested whenever possible. Sources were
selected randomly throughout each plant to insure that fugitive emissions
sampled were representative of and resulted from the process being studied.
Source selection was randomized by one of two methods. The preferred method
was to utilize process instrumentation diagrams which depict all pumps, valves,
compressors, etc. The depicted sources were sequentially numbered on the
diagram, and a random number table was used to determine which numbered sources
were to be tested.
Many process instrumentation diagrams were, however, either not sufficiently
detailed or were outdated, rendering the above method impractical. In this
instance potential fugitive sources were selected by placing a plastic grid
over a diagram of each unit operation or process. Based on the size and shape
of the available diagram, each square millimeter covering the area was consec-
utively numbered. A random number table was then used to select corresponding
squares on the grid. These points were then transferred to the diagram. The
diagram, marked with the number of screening sites selected, was taken into
the field for source selection. Fugitive sources located nearest to the area
marked on the diagram were then tagged for source screening. Figure 2 shows
the selection of sampling sites in a storage tank area.
501
1,001
1,501
2,001
2,501
3,001
3,501
(TO
NOTES
GRIDS ARE NUMBERED SEQUENTIALLY
FROM LEFT TO RIGHT, TOP TO
BOTTOM.
INDICATES LOCATION OF GRID
® CORRESPONDING TO RANDOM
NUMBER
INDICATES STORAGE TANK
LOCATION IN PLANT
710
1,430 2,130 2,WO 3,550
Figure 2. Selection of screening sites in plant storage
tank area using random numbers.
-233-
-------
The selection of sources for screening was supervised by the plant's engineer,
who was responsible for ensuring that all selected sources were a part of the
manufacturing operation being tested.
FUGITIVE SOURCE SCREENING
Fugitive emissions were detected using an Organic Vapor Analyzer (OVA),
Model 128, obtained from Century Systems Corporation. The OVA-128 instrument
utilizes a hydrogen flame ionization detector (FID) sensitive to 1 ppm of
organic vapor and provides a direct meter readout with a 2-second response
time. The instrument is available with a linear readout covering three ranges:
0 to 10 ppm, 0 to 100 ppm, and 0 to 1,000 ppm. Using a calibrated dilution
probe, the instrument's readout capabilities were increased to include ranges
of 0 to 10,000 ppm and 0 to 100,000 ppm.
In prior studies, the EPA has determined that for a fugitive source an OVA
instrument reading of greater than 200 ppm corresponded to an approximate emis-
sion rate greater than or equal to 0.5 g/hr. For this study, it was estab-
lished that only sources with emission rates greater than or equal to 0.5 g/hr
would be sampled. Therefore, if a hydrocarbon concentration greater than
200 ppm was detected by the instrument at any fugitive source, that source was
prepared for sampling.
Pumps, Compressors, Agitators, and Valves
The OVA-128 instrument was used for screening pumps, compressors, agitators,
and valves by placing the instrument probe as close as possible to the inter-
section of the shaft and packing gland at four points, 90° apart. If a
hydrocarbon concentration in excess of 200 ppm was detected at any point, the
source was prepared for sampling.
Flanges
Fugitive emissions from flanges were screened using one of two techniques
depending upon the condition and accessibility of the flange.
The first method, used for screening large, relatively cool vessel and exchanger
flanges, involved taping the outside perimeter of the flange interface, thus
effectively sealing the interface. A small hole was then punctured into the
tape seal. Hydrocarbons leaking from any point around the flange were emitted
through the puncture, and they were screened in a manner similar to that dis-
cussed for screening of valves.
The second method, applicable for smaller and/or hot flanges, involved placing
the OVA-128 instrument at 50-mm intervals all around and against the outside
perimeter of the flange interface. All instrument readings of 10 ppm or
greater were recorded. If the sum of all the readings taken around the perim-
eter exceeded 200 ppm, the flange was prepared for sampling.
Relief Devices, Process Drains, Sample Valves
Relief valve emissions were vented to the atmosphere through a large diameter
pipe, normally located at a high point on the unit which it served. Relief
valves were screened by inserting the OVA-128 instrument probe into the stack
-234-
-------
and measuring the hydrocarbon concentration at this point. The ambient concen-
tration was also determined. If the difference between the stack and ambient
concentration exceeded 200 ppm, the relief valve was prepared for sampling.
Fugitive emissions from process drains occurred at points where the drains are
vented to the atmosphere. Emissions are vented through a perforated steel
plate built into the ground directly above the drain. .Process drains were
screened by placing the OVA-128 instrument probe near the steel plate outlet.
If the instrument detected a hydrocarbon conentration greater than 200 ppm,
the process drain was prepared for sampling.
Process stream samples are collected from sample valves at regular intervals
by plant personnel during normal operation at a petrochemical plant. Sample
valves were screened for fugitive emissions by placing the OVA-128 instrument
probe near the open spout of the sample valve. If a total hydrocarbon concen-
tration greater than 200 ppm was measured, the sample valve was identified for
sampling.
SAMPLING OF FUGITIVE VAPORS
Field sampling of fugitive organic vapors was accomplished using the dilution
or flow-through method. Two separate sampling trains, the vacuum sampling
system and the pressurized sampling system (shown in Figures 3 and 5), were
used. The trains were mounted on a portable cart for easy maneuverability
from source to source.
Vacuum Sampling System
This sampling train, shown in Figure 3, was equipped with a pneumatic vacuum
pump, an orifice meter and a sampling syringe. A cold trap was placed in the
system to condense water and heavier hydrocarbons and to prevent condensation
in downstream lines and equipment. The fugitive emission source identified
through screening was prepared for sampling by tenting. In the tenting proce-
dure, the leaking source (e.g., a valve) was enclosed in a flexible Mylar®
tent as shown in Figure 4. Ambient air was drawn across the emission source
for approximately 20 minutes before sample collection in order to achieve
equilibrium conditions inside the tent. Steady-state conditions were deter-
mined by using the OVA-128 instrument to measure exit gas concentration.
Samples were obtained by drawing a portion of the gas from the main flow using
a 500-ml glass syringe and then injecting the gas sample into an evacuated
Teflon® bag. The vacuum pump was capable of maintaining a flow rate approxi-
mately 0.037 m3/min to 0.034 m3/min during sampling. The source emission rate
was determined from the air flow rate and the hydrocarbon content of the
collected sample. Ambient air samples were obtained in the vicinity of the
leaking source prior to sampling and analyzed for hydrocarbon content to deter-
mine background concentrations. If the background hydrocarbon concentration
was high (>10 ppm) or if the source being sampled was in close proximity to
another fugitive source the pressurized sampling train was used as described
below rather than the vacuum system.
-235-
-------
THIS LINE SHOULD BE
AS SHORT AS POSSIBLE
ORIFICE
METER
TENT
on >
v_^
LEAKING VALVE
rCOLD TRAP
(ICE BATH)
THREE WAY
VALVE
SAMPLE
BAG
EXITGAS
VACUUM PUMP
SYRINGE
Figure 3. Portable sampling train, using a vacuum pump
for sampling or fugitive hydrocarbon emissions.
SUHfR DUCT TAPE
SAG SEAL
MYIAR BAGGING
FLEXIBLE PLASTIC
MESH REINFORCING — \
MATERIAL BENEATH BAG
Figure 4. Tent construction around the seal area of a vertical valve.
DESICCANT
COLUMN
ORIFICE
CONTROL / MEFER
PLANT
AIR
VALVE
C*3=
CHARCOAL
COLUMN
EXITGAS
THREE WAY
VALVE
LEAKING VALVE
Figure 5. Portable sampling train, using plant air for dilution
and sampling of fugitive hydrocarbon emissions.
-236-
-------
Pressurized Sampling System
The pressurized sampling train shown in Figure 5 employed plant air for dilu-
tion and sampling. Plant air under pressure was first passed through a
charcoal filter and desiccant to remove hydrocarbons and moisture, and then
directed through the train and into the tent. After steady-state conditions
were established, a sample bag was connected to the tent and filled by means
of the incoming dilution air.
The vacuum system was the preferred sampling method because filling the sample
bag was not dependent upon air flow through the tent. This permitted less
complicated tenting procedures to be used and resulted in a substantial savings
in sampling time. Also, the vacuum system was equipped with a cold trap to
condense heavy hydrocarbons. If heavy hydrocarbons were to condense inside
the sample bag of the pressurized system prior to GC analysis, a significant
error could be introduced into the data.
The pressurized sampling system was used to sample sources located in areas
where high ambient hydrocarbon concentrations existed or when fugitive sources
were close to each other. In these cases it was felt that the magnitude of
and variations in the background concentration would introduce errors into the
emission rate calculations for the vacuum system. Dilution air in the pres-
surized sampling system was essentially free of hydrocarbons and water since
it was passed through a desiccant and activated carbon prior to mixing with
hydrocarbons from the tented source.
SAMPLING OF LIQUID LEAKS
Liquid leaks were defined as those leaks from which liquid was observed to
escape. If the liquid vaporized rapidly and completely in the vicinity of the
escape point, the emission was treated as a vapor leak and tented.
In order to measure the liquid leak rate at the source, the material was
collected in a graduated container, which was externally cooled with ice and
fitted with covers to contain most of the material. Flow rates were determined
at the leak site by measuring the change in volume with respect to time.
ANALYTICAL PROCEDURES
Analysis for individual Organic Species
After completion of a sampling run, the sample bags were transported to the
field analytical laboratory. Analysis of bag contents was carried out using
either a Hewlett Packard Model 5750 or a Varian 1400 gas chromatograph,
equipped with flame ionization detectors. The gas chromatograph was interfaced
with a Varian GDS-111 microprocessor to convert the chromatograph detector out-
put into a digital integrated output. The system allowed storage of standard
sample data; it provided data output for each sample in the form of retention
times and peak areas; and it identified and quantified each component found
in the samples.
-237-
-------
Analysis for Total Hydrocarbons
If unknown hydrocarbon components were found during sample analysis, the sample
was rerun and the total hydrocarbon content was measured. Total hydrocarbons
were analyzed in the field by bypassing the GC column and injecting the gas
sample directly into the flame ionization chamber.
Laboratory Analytical Procedures
All gaseous samples collected were analyzed in the field. Selected sample bags
were returned to the laboratory for GC/MS analysis. This was done to verify
field measurements and to identify any unknown components.
Liquid samples collected in the field comprised liquid leaks and condensate
collected in the cold trap. These samples were returned to the laboratory for
identification and quantification of each organic component present by gas
chromatography/mass spectroscopy.
RESULTS
Petrochemical plants manufacturing monochlorobenzene, butadiene, dimethyl
terephthalate, or ethylene oxide/glycol were selected for fugitive emission
sampling since they are representative of a variety of operating conditions
and unit operations. Characteristics of each process are described in Table 1.
TABLE 1. CHARACTERIZATION OF CHEMICAL PROCESSES
SAMPLED IN FUGITIVE EMISSION PROGRAM
Process
Monochloro- Dimethyl- Ethylene
Characterization benzene Butadiene terephthalate oxide/glycol
Operating
temperature/pressure
range low/low high/high high/low low/high
Unit Operations Utilized
Distillation x x x x
Absorption x
Scrubbing/washing x x
Extraction x
Evaporation x x
Crystallization x x
Drying x x
Quenching x
Storage x x x x
Unit Processes Utilized
Oxidation x
Hydrogenation x
Pyrolysis x
Esterification x
Hydration x
Chlorination x
-238-
-------
Average hydrocarbon emission rates are given in Table 2 for source types in
each process. These emission rates are based on sources that were found to
have hydrocarbon concentrations in excess of 200 ppm as determined by screen-
ing. The estimated fugitive emission rate for all potential sources is also
given; it is based on the mass of fugitive emissions divided by the total
number of sources. An analysis of variance was performed to determine the
effect of operating temperatures, pressures and fugitive compound vapor pres-
sures on the variability in emission rates from each pump, valve and flange
sampled. Results of this analysis indicated that the factors tested had no
significance in relation to the corresponding emission rates from each sampled
source.
Table 3 gives the estimated annual mass of fugitive emissions from each
process. This information was derived from the number of potential fugitive
sources at each plant and the estimated average emission rates for all poten-
tial sources, as shown in Table 2.
In butadiene production valves were the most significant fugitive source, con-
tributing 90% by weight of the total fugitive emission burden from the plant.
In monochlorobenzene production flanges contributed 91% of the total fugitive
emissions and thus were considered the most significant source of fugitive
losses at the plant. For ethylene oxide/glycol and dimethylterephthalate pro-
duction data on the total number of valves and flanges were not available, and
consequently no emission mass estimates were made.
Factors which determined the magnitude of fugitive emissions from various
source types (i.e., pumps, valves, flanges, etc.) included 1) total number of
potential sources in each process, 2) average emission rate for significant
fugitive sources, and 3) percentage of sources having significant emission
rates.9 The total number of each source type was found to have the most impact
on the overall fugitive emission burden for a petrochemical plant. This was
due to the high variability in the total number of potential sources for dif-
ferent source types. For example, in the butadiene production plant sampled
there were 174 pumps and 6,700 valves. Variations in the average emission
rates and in the percentage of sources having significant emissions for these
two source types were small compared to the differences in number of sources.
Thus the total emissions contributed by valves is very much larger than the
total fugitive emissions from pumps. Since the total number of valves and/or
flanges was large compared to the number of other source types (pumps, com-
pressors, etc.),at each plant sampled, valves and/or flanges are considered
the major sources of fugitive emissions from petrochemical processes.
Fugitive emissions were identified and quantified in the field by gas chromato-
graphic analysis. Table 4 lists the organic compounds detected along with
their weight percent contribution to the total mass of fugitive emissions
determined for each process. On-site analysis was verified by laboratory mass
spectroscopy; it was determined that greater than 98% by weight of the fugi-
tive emissions sampled were identified. Potential human health effects from
exposure to various compounds detected are described in footnotes in Table 4.
Significant fugitive sources are those having a hydrocarbon emission rate
greater than or equal to 0.5 g/hr.
-239-
-------
TABLE 2. HYDROCARBON EMISSION RATES FOR FUGITIVE SOURCES BY PROCESS
Ib/hr (g/hr)
Source type
Pump seals
Average emission
Monochlorobenzene
5.1 x HP2
(23)
Compressor seals
Valves
^ Flanges
O
Relief devices
Process drains
Agitator seals
Sample valves
Note . — Blanks
3.4 x 10-3
(1.5)
1.8 x 10"1
(82)
4.4 x 10"1
(200)
rate for significant fugitive sources
Butadiene
3.7 x KT1
(160)
1.3 x 10-1
(59)
2.5 x 10-1
(120)
0
(0)
3.0 x 10~2
(14)
indicate source type nonexistent
Dimethyl-
terephthalate
4.5 x 10-2
(20)
7.0 x 10~2
(32)
2.4 x 10-1
(110)
0
(0)
4.8 x lO"1
(218)
2.0 x lO"1
(91)
in process.
Ethylene
oxide/glycol
1.8 x 10-1
(82)
2.5 x 10-2
(11)
3.5 x 10-3
(1.6)
2.0 x 10-3
(1.0)
0
(0)
1.5 x 10-1
(68)
Average emission rate for
Monochlorobenzene Butadiene
1.7 x 10~2 1.4 x KT1
(7.7) (63)
1.2 x 10-1
(54)
1.0 x 10-1* 3.7 x 10~2
(0.05) (17)
4.9 x 10~3 0
(2.2) (0)
1.4 x 10~2
(5.0)
4.4 x 10"1
(200)
all potential
Dimethyl-
terephthalate
7.4 x 10~3
(3.3)
3.3 x 10~3
(1.5)
7.6 x 10~3
(3.4)
0
(0)
3.2 x 10-1
(145)
8.6 x 10~2
(40)
b
sources
Ethylene
oxide/glycol
2.9 x 10~2
(13)
1.3 x 10-2
(5.9)
1.6 x 10-"
(0.07)
6.7 x KT5
(0.03)
0
(0)
8.9 x 10"2
(40)
Significant fugitive sources are those having an emission rate greater than or equal to 0.5 g/hr as determined by sampling and analysis.
Emission rates were determined by calculating the mass of fugitive emissions from the emission rates for significant sources. The mass
of emission was divided by the total number of sources screened to arrive at an average fugitive emission rate for all sources.
-------
TABLE 3. MASS OF FUGITIVE EMISSIONS FROM PETROCHEMICAL PROCESSES
Process source
Number of
sources present
a
Estimated mass
of emissions,
metric tons/yr
Total mass of
fugitive emissions,
metric tons/yr
Monochlorobenzene
Pumps
Valves
Flanges
Other
Butadiene
Pumps
Valves
Flanges
Other
Ethylene oxide/glycol
Pumps
Valves
Flanges
Other
Dimethylterephthalate
Pumps
Valves
Flanges
Other
25
640
1,500
_b
174
6,700
26,000
218
69
_b
-b
26
67
_b
_b
37
1.0
0.2
29
1.7
96
1,000
0
16
8.0
_b
_b
7.9
2.0
_b
_b
8.6
31.9
1,112
b
b
Determined from average emission rate from all potential sources as shown in
Table 2.
Data not available.
-241-
-------
TABLE 4. PERCENT COMPOSITION BY WEIGHT OF FUGITIVE
EMISSIONS FROM PETROCHEMICAL PLANTS
Monochloro- Dimethyl Ethylene
benzene, Butadiene, Terephthalate, Oxide/glycol,
Compound % % % %
Acetaldehyde 11
Acroleina 12
Benzene"3 41
i-Butane 0.5
?!-Butane 1.0
1,3-Butadienea 28
1-Butene 11
2-Butene 1.5
Dichlorobenzene3 0.1
Diisobutenea>c 33
Ethane <0.1
Ethene 2
Ethylene oxide& 73
Furfural9 4
Isobutene 21
Methane 2
Methanolc 65
Monochlorobenzenea 59
Xylene 35
Human health effect - irritant
b
Human health effect - suspected carcinogen
C
Human health effect - narcotic
Table 5 provides a comparison of fugitive emission rates for petrochemical
plants and petroleum refineries. Emission rates for petroleum refineries were
derived from emission factors obtained during the sampling of fugitive sources
by Los Angeles County (Ref. 1-4). The percentage of sources having emission
rates in excess of 0.5 g/hr for both petrochemical plants and petroleum refin-
eries is given in Table 6.
For significant fugitive sources, average emission rates at the petrochemical
plants sampled were 3% to 90% lower than corresponding emission rates for
petroleum refineries. Average emission rates for all potential sources were
also found to be lower at the petrochemical plants than the established emis-
sion rates for refineries. Of the total number of sources sampled, the
percentages of significant sources were lower for monochlorobenzene, dimethyl-
terephthalate, and ethylene oxide/glycol production when compared to refin-
eries. For butadiene production, however, the percent of sources having
significant emission rates is essentially the same as that for refineries.
-242-
-------
TABLE 5. COMPARISON OF FUGITIVE EMISSION RATES FOR
PETROCHEMICAL PLANTS AND PETROLEUM REFINERIES
Ib/hr/source (g/hr/source)
a
Significant fugitive sources
Source type
Pump seals
Compressor seals
Valves
Flanges
Process drains
Relief valves
Petrochemical
Plants
1.6 x 10— '
(73)
7.7 x lO-2
(35)
8.2 x ID"2
(37)
1.0 x 10~1
(45)
1.5 x 10-1
(68)
3.1 x ID"2
(14.3)
Petroleum
Refineries
(Ref. 1-4)
4.8 x 10— '
(218)
6.5 x 10~1
(300)
h
8.5 x 10~2
(38)
1.1"
(500)
_C
4.0 x 10— '
(183)
All potential sources
Petrochemical
Plants
4.8 x 10~2
(22)
6.6 x lO"2
(30)
1.0 x ID"2
(4.5)
3.1 x 10~3
(1.4)
8.9 x ID"2
(40)
1.1 X ID"2
(5.1)
Petroleum
Refineries
(Ref. 1-4)
1.7 x 10~1
(78)
3.5 x 10~1
(161)
u
1.1 X ID"2"
(5)
1.1 X 10-2b
(5)
C
1.0 x 10-1
(45)
Significant fugitive sources are those having an emission rate greater than or
equal to 0.5 g/hr.
Data shown arebased on an average emission rate of 0.27 Ib/day/source. EPA
currently uses a value of 0.15 Ib/day/source for valves and flanges; however,
this excludes sources with large leaks.
C
Data not available.
In Tables 5 and 6 petrochemical plant fugitive emissions are shown to be gen-
erally lower in both magnitude and in the number of significant sources when
compared to data on refineries. Two factors which may affect the variations
in fugitive emissions are given below:
• process size - the petrochemical processes sampled were smaller
in size than refinery operations and contained a fewer number of
potential fugitive sources.
• product value - products from petrochemical plants have a higher
value per unit than refinery products. Thus potential losses are
better controlled.
-243-
-------
TABLE 6. COMPARISON OF SIGNIFICANT FUGITIVE SOURCES IN
PETROCHEMICAL PLANTS AND IN PETROLEUM REFINERIES
a,b
Percent of sources having significant emission rates
Ethylene Petroleum
Monochloro- Dimethyl- oxide/ Refineries
benzene Butadiene terephthalate glycol (Ref. 1-4)
Pump seals
Compressor seals
Valves
Flanges
Process drains
Relief valves
20
3
3
39
95
15
0
36
16
5
3
0
16
50
4.5
3
59
0
36
54
13
1
_C
23
a
Significant fugitive emission sources are those having an emission rate
greater than or equal to 0.5 g/hr.
Blanks indicate source type does not exist in process.
'Data not available.
CONCLUSION
This study has identified and quantified fugitive emissions from various petro-
chemical plant processes. Fugitive emissions from petrochemical plants are
generally lower in quantity when compared to EPA data on fugitive emissions
from petroleum refineries.
Physical differences in operating conditions and process materials showed no
relationship on variations in emission rates from individual sources.
ACKNOWLEDGEMENT
This research was funded by the Industrial Environmental Research Laboratory
of the U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina, under contract 68-02-1874. The content of this paper does not
necessarily reflect the views or policies of the U.S. EPA.
-244-
-------
REFERENCES
1. Emissions to the Atmosphere from Eight Miscellaneous sources in Oil
Refineries. Report No. 8 (PB 216 668), Joint District, Federal and State
Project for the Evaluation of Refinery Emissions, Air Pollution Control
District, County of Los Angeles, California, June 1958. 57 pp.
2. Compilation of Air Pollutant Emission Factors. Publication No. AP-42,
U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina, March 1975. pp. 9.1-1 to 9.1-8.
3. Steigerwald, B. J. Emissions of Hydrocarbons to the Atmosphere from Seals
on Pumps and Compressors. Report No. 6, (PB 216 582), Joint District,
Federal and State Project for the Evaluation of Refinery Emissions. Air
Pollution Control District, County of Los Angeles. California, April 1958.
37 pp.
4. Boland, R. F., T. E. Ctvrtnicek, J. L. Delaney, D. E. Earley, and Z. S.
Khan. Screening Study for Miscellaneous Sources of Hydrocarbon Emissions
in Petroleum Refineries. EPA-450/3-76-041, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina, December 1976. 91 pp.
-245-
-------
DEVELOPMENT OF COATINGS TO REDUCE FUGITIVE
EMISSIONS FROM COAL STOCKPILES*
by
R.V. Kromrey, R. Naismith
R.S. Scheffee, R.S. Valentine
ATLANTIC RESEARCH CORPORATION
Alexandria, Virginia
* This work was sponsored by the Department of Energy, Division of Environ-
mental Control Technology under Contract No. EP-78-C-02-4632.
-247-
-------
ABSTRACT
The use of coatings to prevent dust emissions and leachates from coal
stockpiles was studied under sponsorship of the Department of Energy, Division
of Environmental Control Technology. The coatings considered included composi-
tions formed of a filler such as pulverized coal and a binder consisting of a
combination of waxes and plastics. Various latex emulsions, both with and
without fillers, were also evaluated.
The purpose of coal coatings is to seal the surface of the stockpile.
This prevents water penetration into the coal. By sealing the surface of the
pile, dust losses are prevented and leachate formation is minimized. Air cir-
culation through the stockpile is also greatly reduced. This yields the added
benefits of reduced oxidation of the coal and prevention of spontaneous igni-
tion. Cold weather handling characteristics of the coal are also improved.
Physical properties of the coatings were measured. Included were
such characteristics as resistance to water penetration and degradation as a
result of thermal cycling. Application techniques were also evaluated. Both
hot-applied and cold-applied coatings were considered.
Protective coatings appear to be an attractive means of prevention of
fugitive emissions from coal stockpiles. The economic benefits from reduced
oxidation and improved handling more than off-set the cost of application. Use
of coatings on unit trains or barges may also be attractive.
-248-
-------
INTRODUCTION
Coal from mines is transported to its ultimate destination by rail,
truck, barge or pipeline. Once it arrives at its consumption point or a pro-
cessing facility, it is moved over short distances by rubber-tired loaders,
conveyors or rail shuttle cars. During the transportation process, the coal
is often exposed to the elements. This leads to dust emissions and losses
through wind effects. Moisture absorption and oxidation also can occur,
causing deterioration of the coal and loss in fuel value (ERDA, 1976).
Once at the destination, the coal is either used or stored. Most
users maintain stockpiles comprising a 90-day supply at their normal use rate.
.This is done primarily to provide a reserve in the event of temporary loss of
supply. This stockpile may stand for a year or more before it is necessary to
use it.
In the stockpile, the coal is subject to weathering and oxidation.
Heavy dust emissions from stockpiles are common. Absorbed moisture from rain
or snow causes leachate formation and may decrease the fuel value of the coal
by 1 to 8% depending upon initial moisture level and coal rank. Wind and rain
erosion results in coal migration and exposes new coal surface to weather.
Moist coal can also freeze solid in cold weather making normal handling im-
possible. Oxidation may reduce the heating value up to 5% over a one-year
storage period (ERDA, 1976). In addition, internal oxidation can lead to spon-
taneous combustion in low rank fuels (ERDA, 1975; Paulson et al., 1975).
This work presented here addresses the use of protective coatings for
coal to resolve the environmental problems associated with transport and stor-
age. These coatings not only minimize fugitive emissions, but can also be
shown to provide dollar savings substantially greater than the costs of appli-
cation.
TECHNICAL APPROACH
The overall objective of this work was to demonstrate the feasibility
of economical coal coating systems. To accomplish this, the following factors
were considered:
Identification of suitable materials for coal coatings
Development of formulations which are capable of reducing
the adverse environmental effects of weathering of coal stores
Identification of acceptable application techniques
• Tests to determine coating properties and provide data for
economic analysis
Economic analysis of coatings compared to unprotected coal
Cost/benefit analysis
-249-
-------
The first task was intended to provide information to guide the ex-
perimental development of coal coatings. The factors considered are listed
below:
Evaluation of candidate filler-binder materials
Review of current coal protection techniques
Application considerations
User motivation
Filler-Binder Materials Evaluation
A list of materials was generated which could be considered as candi-
dates for inclusion in coal coating formulations. The candidate materials were
screened to establish their characteristics with respect to:
cost
availability
• fuel value
physical properties
environmental pollution potental
The filler materials considered included coal, paper, sawdust, bagasse, rice
hulls, cottonseed hulls, and fly ash. Binder materials considered fall into
two categories. The first included the materials useful for the hot-melt type
of coating such as waxes and plastics. The second type of binder material was
the various latex emulsions. This type of material was considered for use in
filled latex coatings.
Pulverized coal and fly ash emerged as the best filler candidates.
Both of these materials are available at coal user sites in sufficient quan-
tities. They are both available at little or no cost, taking into account
the credit for fuel value of coal. Neither of these materials would cause
additional environmental problems upon combustion, as both are normally present
in coal combustion systems. The other filler material candidates were either
too costly, limited in availability or presented potential environmental
problems.
The best binder materials identified were slack wax, polyethylene,
atactic polypropylene for the hot-melt composition, and polyvinyl acetate
copolymers for the filled latex formulations.
Current Coal Protection Techniques
The most common current means of protecting coal is to carefully
build and compact the stockpile, then monitor it for hot spots. At this time,
it is estimated that less than 5% of the users of coal provide additional
protection in the form of coatings. Coal is shipped by rail car in as-dumped
condition. It is estimated that 5-10% of the shippers use latex crusting
compounds to prevent wind loss and dust emissions in transit. A discussion
of these current protection methods is provided below.
-250-
-------
Stockpile Formation
Large users of coal generally maintain a stockpile sufficient to
provide coal for 90 days in the event of a strike or other loss of supply. A
typical user requires 4,000 tons per day. Thus, a stockpile of 360,000 tons
would be required.
Stockpiles must be properly constructed; otherwise, the risk of spon-
taneous combustion is high. Coal stores have been known to ignite spontaneously
within six days after pile formation (Wilson, 1975). A typical method for con-
structing a coal stockpile is given below:
1. The coal is dumped from the rail cars either by bottom hopper
or car inversion.
2. The coal falls onto a conveyor and passes through a mill where
it is ground to <2-inch size pieces.
3. The coal is transported to the stockpile areas either by
conveyor or by rubber-tired vehicle.
4. A bulldozer spreads the coal and compacts it. Typical piles
are in the form of a truncated cone, 100-200 feet wide at the
base and 25-50 feet high. The length of the pile may be up
to thousands of feet.
2
Coal as dumped has a bulk density of about 50 Ib/ft . After compaction, the
bulk density is about 65 Ib/ft (Paulson, et al., 1975).
The stockpiling technique varies with coal source and rank. Higher
rank coal may be formed into larger piles of greater height and stored longer
than low rank coals (ERDA, 1975). East coast coals may be stored longer than
west coast coals as they are less reactive. The finer the coal, the more reac-
tive it is as a result of greater surface area.
Small coal stockpiles may be stored as dumped for short periods.
Such piles should not exceed 15 feet in height (Wilson, 1975). Longer term
storage of uncompacted coal is risky in terms of spontaneous combustion danger.
Uncompacted coal piles should be used within a few days of dumping.
Stockpile Protection
Stockpiles are formed solely by the action of a bulldozer in shaping
and compacting the pile. Thermal probes are sometimes buried in the coal to
monitor pile temperatures. The probes set off an alarm if the pile temperature
reaches 130-140°F (54-60°C).
Dust emissions from coal stockpiles are a problem of increasing con-
cern. Some users reduce dust formation by washing the coal before piling and
transporting it wet. Other users wet the pile after formation or spray it with
oil to reduce dust emissions.
Leachate formation is also a serious problem except in the case of
very high grade coal. Water runoff from the pile is collected in a trough
-251-
-------
around the pile. It is then sent to a holding tank or pond where it is treated
before discharge to the environment. The treatment generally consists of in-
troducing a flocculant or precipitating agent, then allowing the solids to
settle out before discharge.
Few users currently protect their stockpiles against oxidation and
the resultant fuel value losses. ERDA (1976) reports that losses up to 5% may
occur in the first year of storage. Leonard (1968) found that properly stored
coal will lose only 1-2% per year of its energy value due to oxidation. How-
ever, improper storage was found to result in losses of 3-5% per year.
Latex Crusting Compounds
The only protective coating of significance in current use is the
latex crusting compound. This material is essentially a paint base, and it is
produced by paint manufacturers. Numerous small firms purchase this material
and resell it as a protective coating for coal or as a surface stabilizer for
earth, sand or other materials stored in stockpiles. The latex compounds are
normally applied after dilution by a factor of 3/1 to 20/1. They are sprayed
onto the material to be protected using any equipment capable of spraying water.
Formulation and Evaluation of Coatings
During this work, 119 hot-melt formulations and 39 latex mixes were
produced and evaluated. The types of tests conducted were as follows:
Visual observation Viscosity
cracking tendency Density
surface texture Thermal expansion
adhesion to substrate Water permeability
Compressive strength Rheology
Tensile strength Grindability
Not all tests were run on all formulations. Some tests, such as tensile and
compressive strength, were run on a limited number of specimens to establish
the order of magnitude of the results. No means of converting strength data
into a useful parameter for characterizing the utility of coatings was iden-
tified. Thus, the data are of interest primarily as a rceans of estimating
cracking potential.
Other tests, such as water permeability, effects of thermal cycling
and coefficient of thermal expansion relate directly to the quality of the
coatings. Tables I and II summarize the results of these tests on formula-
tions that were near the optimum compositions for both the hot-melt and filled
latex coatings.
Test Panels
Those formulations which appeared to best meet the above criteria
were cast onto frames, 12" x 12", with about 3-4" of coal as a substrate.
-252-
-------
Table I. Physical Properties of Hot-Melt Formulations
CO
ENTRY NO.
2-16-4
2-17-1
2-17-2
5-2-3
6-1-2
VISCOSITY TENSILE PROPERTIES (-30°F) COMPRESSIVE
Avg. KP At Rupture Tangent STRENGTH, (70°F) STRAIN IN DENSITY
RPM Stress Strain Modulus at Failure COMPRESSION GM/cm
COMPOSITION 1 2.5 5 (PSI) (%) (PSI) (PSI) At Failure (%) (70°F)
85.0%
12.0%
3.0%
85.0%
12.0%
3.0%
85.0%
12.0%
3.0%
77.5%
14.9%
3.8%
3.8%
80.0%
15.2%
2.4%
2.4%
Coal 333 1.9
SW 392 Avg 342 1.3
PE 302 1.6
Coal 248 1.8
SW 278 Avg 258 2.1
PE 249 2.1
Coal 148 j 2.2
SW 179 ! Avg 173 1.9
PE 192 J 2.3
Coal 8 7 4 210 0.4 106,000 1.1
SW 79 0.7 18,200
PP
PE
Coal 222 0.4 84,800 1.08
SW 233 0.6 68,200
PP 241 0.5 83,600
PE
SW = Slack Wax
PE » Polyethylene
PP « Polypropylene
-------
Table II. Characteristics of Hot-Melt and Filled Latex Compositions
LEAK RATE, % OF SPRAY 36" CTE STRIP AVERAGE RESTRAINED
REFERENCE NUMBERS 70°F +160°F -30°F Shrinkage Reheat Shrinkage CTE ENDS GAP GRINDABILITY
5-1-2
80.0% Coal
1 T /,<9 OTJ _r
3.3% PE 3'3 8'1 11'° °'335" - 7.6 x 10 ^ - Fair
3.3% PP
5-2-3
77.5% Coal
, 3.8% PP
to
f 6-1-2, 5-30-1
80.0% Coal
1.4 5.2 8.0 0.305" - 6.6 x 10~5 - Fair
1 5 9? <5W -">
:,"7 JT 14.0 20.0 78.0 0.285" 0.278" 6.2 x 10 D 0.118" Fair
2.4 PP
48
50.0% Coal
23.0% Vinyl ?>? Good
acetate
copolymer
27.0% Water
SW = Slack Wax
PE = Polyethylene
PP = Polypropylene
-------
The frames were constructed with wooden pegs spaced at 3/4" intervals around
the periphery. The pegs served to restrain the coating from shrinkage to en-
hance crack formation and present a more realistic test than would an unre-
strained coating. The coating was applied in thicknesses ranging from 1/8" to
1/2". Figure 1 shows two views of a test frame with a hot-melt coating in
place. Figure 2 shows a cross-section view of the coating on a coal bed.
The test frames were subjected to a "rainfall" test. In this test,
the equivalent of 2" of rain was sprayed on the surface of the test panel.
The quantity of water adhering to the surface or passing through the coating
was measured. For the best coatings, less than 10% of the water was retained
or passed through the coating. The remainder was repelled and ran off the
coating surface. The water permeability tests were generally run 3 times, at
ambient temperature and after exposure to temperatures of +160°F and -30°F.
Thermal Expansion Tests
Thermal expansion or contraction is an important characteristic of
hot-melt coatings. If the degree of contraction upon cooling exceeds the
strain-to-failure of a coating material, cracking will result. Similarly,
if the thermal expansion upon heating exceeds the compressive strain-to-failure,
the coating will shatter.
The coatings listed exhibited Coefficients of Thermal Expansion (CTE)
ranging from 6 x 10~5 to 14 x 10~5 in/in°F. Thus, for a 100°F temperature
change, the degree of expansion or contraction would be about 0.6% to 1.4%.
The tensile strain to failure was measured to be about 0.5%. Thus, most coat-
ings would be expected to crack upon cooling by 1008F or more. Heating does
not appear to be a problem. Compressive strain-to-failure was measured to be
about 2%, and none of the coatings tested expanded that much.
The CTE is not the only factor in cracking of the hot-melt coatings.
The materials were generally cast at 200-250°F. In general, the coatings re-
mained fluid until they cooled to below 200°F. They then start to shrink, if
unrestrained. Total shrinkage upon cooling to 70°F ranged from 0.6% to 1.7%.
The cooling behavior was a function of the composition, however. The CTE
curves for typical compositions are shown in Figures 3 and 4. Formulations
with CTE values similar to that shown in Figure 4 would probably be satisfactory
for stockpile applications.
Latex Formulation Results
A total of 39 latex formulations were prepared and evaluated during
this work. Included in this number were four tests of commercially available
latex crusting compounds for comparison purposes. The remainder of the formu-
lations were filled latexes, in which pulverized coal was used as an extender
to improve the waterproofing character of the latexes.
Some of the latexes tested were not compatible with coal. In the
presence of coal, they coagulated or solidifed rapidly. Daratak SP-1065,
Everflex GT and DLR-H resins yielded satisfactory coating films. All others
tested were incompatible with the coal.
-255-
-------
Figure 1. Hot-Melt Coating on Test Frame
-256-
-------
Figure 2. Cross-sectional View of Hot-melt Coating on Coal Bed
-257-
-------
100
I
ro
u»
oo
o
r-l
X
o
. 10
H
<
W
H
U
FORMULATION 5-1-2
80% Coal
20% Binder
80% Slack Wax
20% Virgin Polyethylene
AVERAGE CTE
7.63xlO-5
PHASE CHANGE
105 125
TEMPERATURE, °F
165
185 205
Figure 3. Coefficient of Thermal Expansion Curve for Formulation 5-1-2
-------
100
10
o
10
tJ
<
v^-
u
1.0
_ — -7
/
/
^ AVER,
.5
/
/
-T/-
d
AGE CTE
6xlO-5
jrT\
^J
Y
__
U.
X
\
— - r
I
FORMULATI01
77.5% Coal
22.5% Blnd<
66 2/
16 2,
16 2;
L. —
1 ^^
^
__ — _
I
¥ 5-2-3
i
»r —
>3% Slack Wax
'3% Polyethylene
'3% Polypropyleng.
. — . — . -
\
r~°
65 85 105 125 145 165 185 2(
TEMPERATURE, PF
5
Figure 4. Coefficient of Thermal Expansion Curve for Formulation 5-2-3
-------
The probable reason for the observed incompatibility is de-emulsifi-
cation of the resin by chemical reaction with ionic components of the coal.
Another possibility is instability resulting from pH change because of acidic
substances in the coal.
Filled latex coatings made with polyvinyl acetate resins were applied
to 1-ft x 1-ft panels similar to those previously described for the hot-melt
coatings. All of the coatings were leakers in the "rainfall" test. The best
of these coatings was an Everflex GT mix with 50% coal, which exhibited only
a 13% leak rate.
To improve on the leakage rate with latex coatings, several test
panels were made using a specially prepared surface finish on the substrate
coal. One-half inch of fines, with a particle size of
-------
MOTOR
PRESSURE
HEATER
STEAM BOILER
DRAIN
JACKETED MIXER
P
STEAM
EJECTOR
I
I
DISPENSER
Figure 5. Schematic Diagram: Coal Cover Application System
-261-
-------
g) service lines (heated)
h) steam boiler
i) dispensing system
A schematic of the system is shown in Figure 5. As shown, a mixing
vessel is included in the current subscale apparatus. This mixer is used as a
batch heater and dispenser. For larger-scale applications, a continuous sys-
tem would be more desirable.
Hot flow application resulted in an excellent layer of coating on an
inclined 4x8 foot panel. The material thickness varied from 1/8- to 1/4-inch
thick. Figure 6 shows a technician in the process of coating a 4' x 8* panel.
ECONOMIC ANALYSIS
The following section is an evaluation of the costs and benefits
which would accrue from the use of protective coatings for coal. Uses of
latex crusting compounds, filled latex coating and hot-melt coatings are
considered. Both stockpiles and rail car applications are evaluated. The
steps involved in forming and protecting a stockpile are shown in Figure 7.
Summary of Stockpile and Rail Car Protection Costs
Table III shows a summary of the costs of building and protecting a
coal stockpile. A 250,000 ton stockpile was used as a basis for comparison.
Table IV is a summary of the costs of various techniques for protecting coal
in rail cars.
Cost/Benefit Analysis - Stockpile Applications
This section illustrates the benefits to be derived from coal pro-
tection and-the estimated return on investment. This analysis is based upon
protection of a 250,000 ton stockpile for a period of one year. The value of
the coal is assumed to be $20 per ton.
The cost factors used in determining treatment costs are listed
below.
Cost, $/Ton
Build and compact pile 0.08
Apply latex crusting compound 0.034
Apply filled latex coating, 0.025" 0.078
Apply hot-melt coating, 1/8" 0.051
The benefit factors used in determining the return on cost of apply-
ing particular coatings are:
Reduce or eliminate dust emissions.
Reduce or eliminate leachate formation.
Maintain or reduce moisture content (fuel value effect).
Minimize freezing of coal into large agglomerates.
Prevent increase in moisture content (grinder operation).
Prevent stockpile migration.
Reduce or eliminate oxidative energy loss.
-262-
-------
% V'
'
^.
f
.V - \
tfl
Figure 6. Application of Hot-Melt Coatings
-------
Dump coal from cars
Grind to <2"
Transport to pile
Unload on pile
Form and compact pile
Current practice*
unprotected piles
Latex Crust
1. Mix with water,
3-10/1
2. Spray apply
4-11 gal/500 ft
Filled Latex
1. Grind coal for surface
preparation *
2. Grind coal for coating
mix
3. Mix coal, water, latex
4. Spray apply 7-31 gal/
500 ft2
Hot Melt
1. Grind coal for
coating mix
2. Prepare wax/plastic
binder
3. Heat and mix
components
4. Spray apply 20-80 gal/
500 ft2
* optional
Figure 7. Coal Stockpile Formation and Protection
-264-
-------
Table III. Summary of Stockpile Protection Costs
Basis: 250,000 ton stockpile
Treatment
Build and Compact
Stockpile
Monitor Stockpile
for 1 year
Latex Crusting
Compound, 1 gal/
200 ft2
Filled Latex on
Surface Fines, .025"
Filled Latex on
Surface Fines, .050"
Filled Latex on
Surface Fines, .100"
Filled Latex on
Normal Surface, 1/8"
Hot-melt on Normal
Surface, 1/8"
Hot-melt on Normal
C,,-*-fn~f> 1 //,"
Time Required
Days
50
365
3.5
13.3
26.6
53.2
11.7
. 11.7
23.4
Total Costs
$
55,000
22,750
8,400
19,500
25,908
42,000
r
33,506
12,788
25,576
Cost/ton
$
0.22
0.091
0.034
0.078
0.107
0.168
0.134
0.051
0.102
-265-
-------
Table IV. Summary of Rail Car Protection Costs
Basis: 1 unit train = 100 rail cars
containing 100 ton/car
treatment rate = 100 cars/day
Treatment
Latex Crusting Compound,
1 gallon/500 ft2
Filled Latex on Surface
Fines, .050"
Filled Latex on as Dumped
Surface, 1/8"
Hot-Melt on as Dumped
Surface, 1/8"
Total Cost
$ '
700
3,844
4,320
1,355
Cost/ton
$
0.07
0.38
0.43
0.14
-266-
-------
Reduce or eliminate spontaneous ignition.
Prevent wind or rain erosion.
Reduce need for snow removal.
Reduce monitoring costs.
Reduce fire prevention and extinguishment requirements .
The basis for determining the dollar value or equivalent for these benefit fac-
tors is described in Kromrey, et al. , (1978). The results are summarized below.
Total Benefits
Treatment Costs, Direct $ Savings, Value, Equivalent
_ Treatment _ _ $/ton _ _ $/ton _ _ $/ton _
Compaction 0.08 0.275 0.61
Latex Crust* 0.114 0.65 1.09
Filled Latex Coating* 0.158 1.04 1.65
Hot-melt Coating* 0.131 1.04 1.90
* includes compaction
From this table, the highest return per unit cost occurs by use of the hot-melt
formulation, i.e., return of benefits valued at $1.90 per ton at a cost of
13. lc per ton.
Cost/Benefit Analysis - Rail Car Applications
This section illustrates the benefits to be derived for protection of
coal in rail cars and the estimated return on investment. The results are
based upon protection of a unit of 100 cars containing 100 tons each of coal.
The value of the coal is assumed to be $20 per ton. The benefit factors con-
sidered are listed below:
Reduce or eliminate dust emission.
Reduce wind losses.
Prevent moisture increase (fuel value effect) .
Prevent moisture increase (grinder operation) .
Prevent freezing of coal into large agglomerates.
Reduce or eliminate spontaneous combustion.
Again, the basis used to convert these factors into an equivalent dollar value
is described in Kromrey, et al. , (1978). The results are summarized in the
following table.
Total
Treatment Costs, Direct $ Savings, Value, Equivalent
Treatment _ $/ton _ _ $/ton _ ___ $/ton _____
Latex Crust 0.07 0.54 0.74
Filled Latex Coating 0.38 0.61 0.81
Hot-melt Coating 0.14 0.61 1.01
This table indicates the benefit return per unit cost to be highest for the
latex crusting compounds with $.74 per ton in benefit value resulting from
costs of $.14 per ton. For rail car applications, the latex crust appears to
be most cost-effective.
-267-
-------
CONCLUSIONS
Based upon the results of this technical effort, the following
conclusions are presented:
1. Commercially available latex resins used as coal crusting
compounds can prevent dust loss and wind erosion, but do
not waterproof the coal surface.
2. Hot-melt formulations consisting of about 77.5% coal,
15% slack wax, 3.75% polyethylene and 3.75% polypropylene
are capable of sealing a coal surface against water
penetration.
3. A filled latex formulation consisting of 50% coal,
23% Everflex GT resin and 27% water is capable of sealing
a coal surface against water penetration provided the
surface is coated with fines of less than about 1/8"
particle size.
4. Hot-melt coatings appear to be most cost effective for
application to coal stockpiles.
5. Latex crusting compounds appear to be the most cost
effective means of protecting coal in rail cars.
6. The anticipated return in dollars and intangible benefits
by use of the coal protection methods described herein is
in the range of 10rl5 times the cost of such protection.
-268-
-------
REFERENCES
ERDA (1975), Report ERDA-TR-120, "The Storage of Solid Fuel at Dumps of
Thermoelectric Power Stations of the USSR, Methods of Preventing
Dusting and Spontaneous Combustion of Fuel in Piles."
ERDA (1976), "Energy from Coal," ERDA-TR-76-67, prepared by Tetra-Tech, Inc.
under Contract No. E(49-18)-2225.
Kromrey, R.V., Scheffee, R.S., DePasquale, J.A., and Valentine, R.S. (1978),
"Development of Coatings for Protection of Coal During Transport
and Storage," Final Report on Contract EP-78-C-02-4632., U.S.
Department of Energy. Prepared by Atlantic Research Corporation.
Paulson, L.E., Axeley, S.A., Wegert, C., and Ellman, R.C., (1975),
"Experience in Transportation of Dried Low-Rank Western Coals,"
Society of Mining Engineers of AIME, Prepoint No. 1 75-B-337.
Wilson, H.S., (1975), "An Update on Coal Storage Technology," Combustion 47,
33-36 (August).
-269-
-------
NEW CONCEPTS FOR CONTROL OF FUGITIVE DUST
by
Dennis C. Drehmel
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina
and
Thomas Blackwood
Monsanto Research Corporation
Dayton, Ohio
and
S. Calvert, R. 6. Patterson, S. C. Yung
A1r Pollution Technology, Inc.
San Diego, California
-271-
-------
ABSTRACT
NEW CONCEPTS FOR CONTROL OF FUGITIVE DUST
by
Drehmel, Blackwood, Calvert, Patterson, and Yung
In the vast majority (92%) of air quality control regions not
meeting total suspended particulate (TSP) standards, fugitive sources
exceed point sources. In rural areas, unpaved roads and agricultural
activities are important sources; in urban areas, paved roads and construction
sites. In industrial areas, both process and open sources of fugitive
particles must be controlled not only for TSP but also for lead air
quality standards. Hence, the need for control of fugitive dust requires
technology to span a variety of source types and control requirements.
Current technology involves improvement in gathering fugitive
process emissions and prevention of open source fugitive emissions. One
new concept under study by EPA does neither of these but rather attempts
to capture particles without hooding and evacuation to a control device.
Instead the particles are caused to settle out by introduction of another
aerosol with appropriate charge and chemical properties. Another new
concept under study uses a different approach for prevention of emissions
from unpaved roads. Rather than using wetting, oiling, or chemical
treatment, the nature of the roadbed is modified. Testing of these
concepts is in an early stage but preliminary results show 50 to 93%
control is to be expected.
-272-
-------
NEW CONCEPTS FOR CONTROL OF FUGITIVE DUST
Introduction
Among the standards used by the Environmental Protection Agency
(EPA) to prevent adverse health and ecological impacts of air pollution,
is the National Ambient Air Quality Standard (NAAQS). The NAAQS is
currently expressed in a number of ways involving the limitation of
total suspended particulate (TSP). Unfortunately, the TSP standard is
not being met in many air quality control regions. Fugitive dusts
account for most of this problem. In more than nine out of ten control
regions, fugitive emissions are greater than ducted point source emissions.
In one out of three control regions, fugitive emissions are more than 10
times greater than ducted point source emissions. Within cities the
contribution of point sources is expected to be smaller than fugitive
emissions. In Philadelphia, more than one-third of the TSP was attributed
to reentrained street dust and vehicular traffic while point sources
were responsible for less than one-fourth of the TSP.
Because of the importance of fugitive sources in meeting NAAQS, EPA
is involved in a program of control technology research and development.
Re-entrained dust emissions and fugitive process emissions, both described
below, are two ongoing projects which have been labeled "new concepts."
This term does not imply that such technology has not been previously
investigated but rather that the technology has not been fully utilized
and will remain new until it is accepted into conventional practice.
Re-entrained Dust Emissions
Fugitive emissions come from uncontained process emissions or from
the re-entrainment of dust. For the latter, the problem can be solved
if the surface of the dust can be stabilized and kept free of unscabilized
dust. The surface may be stabilized by the application of water, oil,
or chemical surface agents. In the case of unpaved roads, application
of such materials is less cost effective than paving the roads.
Advances in fabric development now provide another possibility to stabilize
a surface; especially, unpaved and haul roads. Using this approach, an
unpaved road would be covered with fabric and then coarse aggregate.
The fundamental concept behind use of a fabric roadbed stabilizer,
or "road carpet," for control of fine particle emissions from unpaved
roads is prevention of vortex entrainment by separation of fine roadbed
materials from the coarse aggregate where the traffic movement occurs.
Large aggregate is held from settling, while newly deposited fines
(< 70 ym) are filtered by gravitation and hydraulic action down to a
zone away from vortex entrainment. Road carpet can be made from spunbondedi
thin-film polypropylene on nylon sheet (Celanese), continuous filament
polyester fibers needled to form a highly permeable fabric (Monsanto),
or other spun or needle-punched synthetic materials. The mechanical
interlocking of fibers makes a formed fabric with the required durability
-273-
-------
and toughness. Designed for road construction use, this fabric is laid
over poor loadbearing soils to help support and contain the overburden
aggregate. It spreads concentrated stress from heavy-wheeled traffic
over a wider area, siphons away ground water, and contains fine soil
particles in the roadbed that could otherwise contaminate ballast or
road overburden.
Capital and operating costs differ for various control options.
Although initial costs for some control alternatives may be lower than
for road carpet, maintenance costs for these alternatives may be quite
high. Road carpets can add as much as 20% to the cost of a road construction,
but the material will pay for itself in less than a year in reduced
maintenance costs and driver safety. In some cases, a new road can be
built for as low as 20% of the cost of conventional roads.
Treating the road surface with oil once a month is an efficient
method of controlling unpaved road dust emissions. The estimated cost
of such applications (reported in 1972) is $0.10 per square meter treated
per year. However, 70% to 75% of the oil applied moves from the surface
of the road by runoff and dust transport, resulting in ecological harm
caused by the oil or its heavy metal constituents. Furthermore, surface
oiling does not prevent potholes, a major part of road maintenance.
Roads constructed using road carpet, on the other hand, require
significantly less maintenance because of the increased road stabilization
imparted by the carpet. The most effective method of reducing particulate
emissions is paving the road surface; but, owing to high initial cost
and subsequent maintenance and repair costs, paving these temporary
roads is impractical. Road carpet can be installed to give a virtually
permanent, maintenance-free road.
Use of road carpet fabric results in no health or safety hazards or
any other unfavorable environmental impact. In the development of these
fabrics, various synthetic polymers (including nylon, polypropylene, and
polyester) were evaluated. Fabrics made from any of these products
generally are resistant to mildew, mild acids, and alkali, and are rot
and vermin proof. Polyester was chosen by Monsanto because of the
following distinct advantages:
resistance to chemicals, including those found in soils
constant properties over a wide range of temperatures
high melting point
little change in wet or dry properties
low moisture absorption
high abrasion resistance
high modulus of elasticity and excellent resilience
excellent creep resistance
Thus, use of road carpet precludes any environmental damage due to
leaching of hazardous chemicals or heavy metals.
-274-
-------
Fugitive Process Emissions (FPE)
Control of fugitive process emissions often has to overcome the
disadvantages of a dispersed, low concentration emission in a large
volume. If the particulate were not dispersed it could be collected in
a high efficiency hood and ducted to a conventional control device. If
the particulate were of a high concentration or in a building of small
volume, the process building could be evacuated to a conventional control
device. Assuming that the FPE cannot be hooded or evacuated to a control
device, EPA is investigating the use of chemically and electrostatically
treated water spray droplets. This method applies to fugitive dust
emissions which cannot be controlled by stabilizing the surface of the
source. Examples are transfer points, conveyor belts, charging operations,
molding lines, and roof monitors. As always, water with or without
chemical additives can be sprayed to stabilize exposed sources, such as
stock piles which are not being processed.
Fugitive dust particles entrained in the gas stream may be collected
with charged or uncharged water sprays by mechanisms such as diffusion,
inertial impaction, interception, and electrophoresis. The larger size
of the water drops would allow easy separation from the gas stream by
methods such as gravitational settling or entrainment separators.
Figure 1 is a functional diagram for the process anticipated for
controlling fugitive emissions. The functional phenomena represented in
this diagram could occur concurrently or separately in several types of
equipment.
Several types of chemicals have been found effective in reducing
fugitive dust emissions when applied to the source. Over 100 chemical
products are presently marketed or under development specifically as
dust control agents. These chemicals act by several different means and
are generally categorized by their composition: bituminous, polymer,
resin, enzymatic, emulsion, surface-active agent, ligninsulfonate, 5
latex, etc. Dust stabilizing agents have been reported in the literature
and are not being considered in this investigation.
In addition to the use of surface active agents, the concept of
using electrostatic sprays involves the controlled disposal of fugitive
dust that is entrained in a gas stream. This approach is the most
permanent way to control FPE because it precludes the re-dispersal of
dust which can occur if stabilization is used or if the collected particles
are deposited on the ground. After the air containing the FPE is conveyed
to the apparatus, it is necessary to bring the spray into contact with
the fugitive dust. The water sprays would be produced with single- or
two-fluid nozzles. Nozzle systems are currently available on the market
which produce charged water drops. Contacting may be obtained by projecting
the water drops into the dusty gas or moving the dusty gas into the
vicinity of the water drops. The water drops will be formed by hydraulic
pressure sprays. An ejector type of spray system is a likely candidate
for contacting with minimum energy usage.
-275-
-------
Containing fugitive emissions with a series of barriers and/or
electric curtains may prove most cost effective. The electric curtains
would be used to deflect the particulate, thus minimizing the volume of
air to be cleaned. Because of the great variety of FPE source situations,
it is not possible to use a single approach to this part of the system.
After spraying in the charged water drops, the next process step
would be to remove the water spray drops after sufficient contacting
time to effect capture of the initial fine particles present in the gas.
At this stage the large size and mass of the water drops is utilized to
separate them from the gas. An entrainment separator may be used for
this step, depending on the mist elimination and pressure drop requirements.
The cleaned gas stream leaves the entrainment separator at this point.
The water from the entrainment separator is passed through a separation
process, such as a filter, to remove the collected dust particles. The
water may then be recycled and the dust may be disposed of in such a way
as to prevent its re-dispersion.
The concept proposed here has some similarity to liquid scrubbers.
Recent studies have shown increased scrubber efficiency can be obtained
by electrostatically charging particle and collectors to opposite polarity. ''
Hoenig has shown that naturally charged fugitive emissions may be
collected more efficiently with charged drops than with uncharged drops.
The results of preliminary analysis show that this technique is most
effective for small particles (less than 5 urn diameter).
Development of Concepts
EPA has planned an extensive research, test, and evaluation program
for both of these concepts. The research portion will identify parameters
affecting the theoretical and economic limitations to utilization of the
technology. Preliminary studies of these parameters will provide an
understanding for optimization. In the case of road carpet, the vortex
entrainment, comminution, and saltation of material on the coarse aggregate
must be modeled. Size and thickness of the aggregate may affect performance.
For the charged water sprays, methods of partial containment of the FPE
must be studied as well as fundamentals of droplet/particle Interactions.
The test and evaluation portion of the program will involve studying
a prototype of each new concept. Using the road carpet, a 30 to 40
meter long section of haul road will be constructed at a quarry. A
second prototype road will be constructed at a different site to evaluate
subsoil differences.
The emissions from the prototype road and a comparison conventional
road will be monitored at several Intervals of time. Sampling will be
conducted when wind conditions are favorable at two different vehicle
speeds and over the course of 3 months. During this time, any changes
in the mass reduction or particle size distribution will be noted and
correlated with observations on the condition of the prototype road. The
durability and other positive attributes of this type of road construction
will be evaluated in order to reassess the economics of this method of
-276-
-------
emissions reduction. Sampling will be conducted using the quick-reaction
sampler at distances of from 10 to 12 meters from the road. Since the
sampling can be conducted over short periods of time (4 to 10 minutes),
desirable wind direction, speed, and atmospheric stability can be chosen
for sampling (i.e., mid-day and approximately perpendicular to the
road).
The test and evaluation program for the charged water sprays will
consider the following parameters:
1. Air flow rate--three levels.
2. Air temperature--two levels.
3. Particle type--two levels.
4. Particle size distribution--two levels.
5. Water-to-air ratio.
6. Water pressure.
7. Nozzle type and arrangement.
8. Electrostatic charge level.
9. Surfactant concentration.
10. Confinement type.
11. Wind conditions (windy and calm).
The spray contactor will be 1.8 m high and 2.5 m wide and attempt to
collect metal oxide or liquid aerosols. Sampling will be by total
filters, cascade impactors, and light scattering devices.
Conclusions
Two concepts for control of fugitive emissions have been identified
and will be investigated under an EPA program. One concept employs road
carpet to stabilize unpaved roads by separating traffic from entrainable
particles. The second employs charged water sprays for direct collection
of fugitive process emissions.
References
1. Carpenter, B. H. and Weant, G. E., Fugitive Dust Emission and
Control, Symposium on the Transfer and Utilization of Particulate Control,
July 24-28, 1978, Denver, Colorado.
2. Bradway, R. M., Record, F. A. and Belanger, W. E., A Study of
Philadelphia Particulates Using Modeling and Measurement Techniques,
Symposium on the Transfer and Utilization of Particulate Control,
July 24-28, 1978, Denver, Colorado.
3. Evans, J. S., Schneider, M., Cooper, D. W., and Quinn, M.,
Setting Priorities for the Control of Particulate Emissions from Open
Sources, Symposium on the Transfer and Utilization of Particulate Control,
July 24-28, 1978, Denver, Colorado.
4. Ochsner, 0. C., Chalekode, P. K. and Blackwood, T. R., Source
Assessment: Transport of Sand and Gravel, EPA-600/2-78-004y (in print),
USEPA, Cincinnati, Ohio, October 1978.
-277-
-------
5. PEDCo, Investigation of Fugitive Dust-Sources, Emissions, and
Control: Volume I. EPA-450/3-74-036a, June 1974.
6. P1lat, M. J. et al.. ES&T, 8: 360-362, 1974.
7. P1lat, M. J. et al., JAPCA, 25: 176-178, 1975.
8. Hoenlg, S. A., Use of Electrostatically Charged Fog for Control
of Fugitive Dust Emissions, EPA-600/7-77-131 (NTIS No. PB 276645),
November 1977.
-278-
-------
WATER
DROPS
ro
FUGITIVE
PARTICLE
EMISSIONS
CONTACT OUST
WITH SPRAYS
COLLECT DUST
ON DROPS
SEPARATE DROPS
FROM AIR
Controlled Disposal of Particles
DISPOSE
OP
DUST
CLEAN
AIR
DISPOSE OF
OR RECYCLE
WATER
Figure 1.
Functional diagram for the major process steps
involved in controlling fugitive particle emissions
with a charged drop system.
-------
BEST AVAILABLE CONTROL TECHNOLOGY (BACT) FOR
FUGITIVE EMISSIONS CONTROL IN THE STEEL INDUSTRY
AUTHOR: Arthur G. Nicola
Manager, Air Pollution Control Systems
PENNSYLVANIA ENGINEERING CORPORATION
PITTSBURGH, PENNSYLVANIA
ABSTRACT
Technology for controlling fugitive emissions related to the various
steps in iron and steelmaking have been given low priorities in the past
due to light concentrations as well as the difficulty in collecting these
emissions with the result that the state-of-the-art for controlling fugitive
emissions in the steel industry has lagged behind the technology developed
in other areas. In some cases, such as blast furnace cast house emission
control, this technology is virtually non-existent in the U.S.
Since the technology for emission reduction through gas cleaning
already exists and can be accomplished by any number of air pollution
control devices, the limited scope of this paper will concentrate on the
best available technology for collecting fugitive emissions from the major
sources which include:
1) Blast furnace - cast house
2) Oxygen steelmaking - hot metal charging, tapping
ladle alloy additions, reladling
3) Electric Furnace Shop - charging and tapping
The ideal solution appears to be collection of the fugitive emissions
at their source preventing their escape into the atmosphere, which allows
emission control with minimum volumes at lowest capital investment and
operating cost.
Although collecting the emissions at their source has not always been
the most practical method in the past, due to interference with normal
operations, present day technology is gradually overcoming these obstacles,
allowing systems that collect the emissions at their source to be installed
in operating shops.
-281-
-------
BEST AVAILABLE CONTROL TECHNOLOGY (BACT)
FOR FUGITIVE EMISSIONS CONTROL IN THE STEEL INDUSTRY
INTRODUCTION
BLAST FURNACE CAST HOUSE EMISSION CONTROL
Tap Hole
Trough
Runners
Hot Metal Cars
Slag Pots
FUGITIVE EMISSION CONTROL DURING THE OXYGEN STEELMAKING PROCESS
Hot Metal Charging
Scrap Charging
Tapping
Slagging
Puffing During Oxygen Blow
HOT METAL RELADLING
FLUX HANDLING
ELECTRIC FURNACE SHOP FUGITIVE EMISSION CONTROL
Charging
Tapping
CONCLUSION
-282-
-------
Figure 1 Cast House Arrangement
Figure 2 Multiple Tap Hole Cast House Arrangement
Figure 3 Retractable Curtains In The Cast House
Figure 4 Hot Metal Car
Figure 5 Hot Metal Charging For Basic Oxygen Process
Figure 6 Schematic Arrangement Of EOF Furnace Enclosure
Figure 7 Components Effecting Fugitive Emission Control With
Furnace Enclosure
Figure 8 Recent EOF Installation With Fugitive Emission Control
Figure 9 Hot Metal Charging Emission Control With Furnace Enclosure
Figure 10 Schematic Arrangement - Ladle Hood For Reladling Emission
Control
Figure 11 Reladling Emission Control - Pouring Through A Vertical
Slot In The Ladle Hood
Figure 12 Schematic Flow Diagram - Pollution Control System For
Electric Furnace Shop
Figure 13 Fugitive Emissions During The Tapping Operation
Figure 14 Typical Furnace Enclosure For Electric Furnace
Figure 15 Charging Operation With Furnace Enclosure
Figure 16 Tapping Operation With Furnace Enclosure
-283-
-------
."BEST AVAILABLE CONTROL TECHNOLOGY (BACT)
FOR FUGITIVE EMISSIONS CONTROL IN THE STEEL INDUSTRY"
Introduction
Current air pollution control regulations dictate that the Best
Available Control Technology (BACT) be utilized for controlling
fugitive emissions in the Iron and Steel Industry. These emissions
can be generally divided into two (2) categories - open dust sources
and process fugitive emissions. Open dust source fugitive emissions
include those sources from which emissions are generated by the
forces of wind and machinery acting on exposed aggregate materials,
while process fugitive emissions include uncaptured particulates and
gases that are related to the various steps in the iron and steelmaking
process.
Since air pollution is associated with practically all steps of iron
and steel production, the related air pollution control equipment is
an important factor in all of these operations. A production process
may become obsolete if it is not capable of meeting today's stringent
air pollution control requirements.
Technology for controlling fugitive emissions related to the various
steps in iron and steelmaking have been given low priorities in the
past due to light concentrations, as well as the difficulty in
collecting these emissions with the result that the state-of-the-art
for controlling fugitive emissions in the steel industry has lagged
behind the technology developed in other areas. In some cases, such
as blast furnace cast house emission control, the technology is
virtually non-existent in the United States.
Since the technology for emission reduction through gas cleaning
already exists and can be accomplished by any number of air pollution
control devices, the limited scope of this paper will concentrate on
the Best Available Control Technology (BACT) for collecting the process
fugitive emissions generated by the main metallurgical production
processes which include Blast Furnaces, Basic Oxygen Steelmaking and
Electric Furnaces.
The alternatives for collecting fugitive emissions generated during
the main metallurgical processes are generally limited and usually
have many disadvantages.
The ideal solution of collecting fugitive emissions at their source
to prevent their escape into the atmosphere, allowing emission control
with minimum volumes at lowest capital investment and operating cost,
is normally the most difficult system to install.
-284-
-------
Collecting fugitive emissions at their source has not always been the
most practical method in the past due to possible interference with
normal operations; present day technology is gradually overcoming these
obstacles, allowing systems that collect the emissions at their source
to be installed in operating shops.
BLAST FURNACE CAST HOUSE EMISSION CONTROL
The cast house structure surrounding the blast furnace, encloses the
runners and operating area and provides weather protection for the
operators and equipment (Figure 1). At the same time, this enclosure
also contains the fumes generated during the cast. Since environmental
control efforts have been concentrated in other areas, blast furnace
cast house emission control has been given a low priority with the
result that little has been done to date to control these cast house
emissions in the U. S.
In the U. S. today, there are 184 blast furnaces producing basic iron
with 3 new blast furnaces scheduled to start up in the near future.
However, none of the operating blast furnaces are capable of meeting
current air pollution control regulations. It is anticipated that in
the future, all new blast furnaces will be required to incorporate cast
house emission control and existing blast furnace cast houses will be
required to install some degree of cast house emission control.
Virtually all blast furnace cast house emission control technology to
date has been developed by the Japanese. During the past ten (10)
years, they have developed their systems to the point where they now
have integrated their ironmaking and emission control. At the present
time, one-hundred percent (100%) of the blast furnaces in Japan have
some degree of cast house emission control.
Primarily, the Japanese approach is to capture the fumes at their source,
preventing their escape into the atmosphere. This is accomplished by
close fitting hoods and covers wherever the hot metal is exposed to the
atmosphere in the cast house. At the same time, the Japanese often
employ a separate secondary system consisting of canopy or monitor hoods
for scavenging and building evacuation.
Particulate emissions from cast houses in the U. S. have been found to
be in the range of 0.2 to 0.6 Ibs./ton of hot metal cast. However, in
Japan, particulate emissions have been found to exceed 1.0 Ibs./ton of
hot metal cast. In general, these emissions are approximately 751 iron
oxide with small percentages of manganese, silicon oxides and sulphites.
Differences in the levels of fume generated from cast house to cast
house can be attributed to the variations in operating practices and
materials used in the blast furnace cast house.
-285-
-------
Sources of fugitive emissions in the blast furnace cast house are the
tap hole, trough, runners, hot metal cars and slag pots. These areas
generate practically all of the fugitive emissions associated with cast
house operation. Existing cast house installations in the U. S. offer
many problems in adopting Japanese technology. However, the new blast
furnaces with multiple tap holes scheduled to start up in the U. S. in
the near future are incorporating Japanese technology for cast house
emission control (Figure 2).
Tap Hole
The primary source of fugitive emissions in the cast house is from the
hot metal as it exits the blast furnace at the tap hole. Approximately
30% to 401 of the total fugitive emissions can be attributed to the tap
hole. With the Japanese technology, these fumes are normally controlled
with local hoods or retractable curtains (Figure 3).
Trough
The hot metal pool adjacent to the tap hole normally extending to the
dam and skimmer (25' to 50' long x 3' to 4' wide x 2' deep) serving as
a holding pit to separate the hot metal and slag, is also a major source
of emissions in the cast house. The fumes here are effected by the area
of hot metal exposed to the atmosphere, temperature of the hot metal,
type of refractory used, and other factors. Operating systems collect
these emissions with the tap hole fumes utilizing a retractable curtain.
Runners
Emissions from the runners are also dependent on the pool areas exposed
to the atmosphere and the metal temperature. As the metal cools, carbon
emerges from the saturated solution as "Kish", a form of graphitic
carbon that is light and flaky. These emissions can be controlled by
replaceable refractory lined runner covers.
Hot Metal Cars
Molten metal from the runner spouts is poured into the hot metal cars
located outside of the cast house or in arcade under the cast house
floor (Figure 4). The density of these fumes is dependent on the rate
of cooling of the hot metal. Local hooding can normally be utilized to
control these fumes since the hot metal cars are always in the same
relative position during this operation.
-286-
-------
Slag Pots
Fumes generated here are the result of running the slag into refractory
lined slag pots or a pit adjacent to the cast house and can be readily
controlled with local hoods.
Since the installation of cast house emission control systems must have
a minimal impact on established operating practices and at the same time
maintain high safety standards on the cast house floor, transfer of
Japanese technology for blast furnace cast house emission control in
the U.S. is not a simple matter due to the difference in operating
practice as well as cast house geometry. Tap hole restrictions, number
of pouring stations, runner layout and pouring station geometry are all
limiting factors on the majority of older blast furnaces with rated
capacities of 2,000 tons/day or less.
The alternative of building evacuation requiring 45-80 air changes per
hour is a costly solution to the cast house emission control problem.
FUGITIVE EMISSION CONTROL
BASIC OXYGEN STEELMAKING SHOP
In the EOF Shop, a number of sources exist which generate fugitive
emissions that are not captured by the main gas cleaning system during
the oxygen blowing process. Today, collection of these fugitive
emissions has become more of a concern than the primary gas cleaning
itself. This concern for fugitive emission control is based on the
difficulty of efficient fume collection at the emission source, since
once the fume escapes into the building, it is almost impossible to
control.
In an effort to comply with current air pollution control regulations,
many systems of different design have been installed and are in
operation today with varying degrees of success. However, to success-
fully control fugitive emissions with minimum capital expenditure, they
must be collected at the source and not allowed to escape into the
building.
Major sources of fugitive emissions in the basic oxygen steelmaking
process are scrap charging, hot metal charging, tapping and ladle alloy
additions, slagging, puffing during the oxygen blow and flux handling.
Other sources such as ladle transporting and teeming also contribute
to fugitive emissions, but these are low in volume and dissipate readily.
They normally do not contribute to monitor emissions from the EOF Shop.
In general, fugitive emissions from the EOF Shop are effected by a
number of factors and can be substantial. However, the particulate
loading is relatively low.
-287-
-------
Hot Metal Charging
The worst condition occurs during charging of hot metal into the
furnace which already contains scrap. The hot metal and the effect of
the hot metal on the scrap both contribute to emissions during this
period (Figure 5). Test results have shown that fumes generated at
this time are composed of approximately 35% iron oxide, 30% kish and
others, with participate size less than 100 microns and an approximate
emission rate of 0.3 to 0.4 Ibs. per ton of hot metal poured.
Since the prime factors controlling fume generation during this period
are the condition of scrap and the rate of hot metal poured, operating
practice is a big factor in the amount of fumes generated.
Scrap Charging
Scrap charging itself is not a primary source of pollution, however,
when the scrap contains a foreign substance such as dirt, paint or
oil, it becomes a major pollution source when the hot metal is added.
The quantities of pollution generated during scrap charging itself
are minor compared to the fumes generated during hot metal charging and
can be minimized by scrap selection.
Tapping
A dense fume normally results from the tapping operation itself and
with ladle additions such as ferro silicon and ferro manganese, the
magnitude of the problem increases. Fume composition here is dependent
on the alloys employed but generally would consist of approximately
75% iron oxide with particle size less than 10 microns and an average
emission rate of 0.15 to 0.20 Ibs. per ton.
Slagging
Emissions from the furnace slagging operation can be a major problem
depending on the type of steel being produced since it influences the
slag volume and emission rate at this time. The method of killing the
slag is also a factor which influences the emissions during the slagging
operation. Fumes are generated inside the furnace at this time and
continue with the slag as it falls inside the slag pots below the
charging floor. These emissions normally have a particle size less than
100 microns.
Slagging fumes are relatively cold and do not have the necessary thermal
energy to cause the fumes to rise into an overhead canopy hood making
them difficult to capture.
-288-
-------
Puffing
Puffing during the oxygen blow occurs intermittently. When this occurs,
generally small quantities of fume are emitted around the mouth of the
furnace. This is not a major source of fugitive emissions, however, it
does contribute to the total emission problem in the shop.
Methods of Collection
With present technology, the alternatives available for collecting the
oxygen steelmaking fugitive emissions are:
1. Complete or partial building evacuation.
2. Local hoods and dampers.
3. Furnace enclosures.
Canopy hoods located in the building trusses for partial or complete
building evacuation systems are operating in a number of shops with
varying degrees of success. In most cases, this method is normally
considered as a last resort due to the high capital expenditure and
operating cost involved. However, these systems usually are given
favorable consideration by maintenance and operating personnel since
they require minimum maintenance and do not restrict operating practices.
Volume requirements to achieve the necessary collection efficiency with
a canopy hood system, would normally be in excess of 1,000,000 CFM for
complete building evacuation.
Local hoods appear to be more effective than canopy hoods in collecting
these emissions, since they are in closer proximity to the emission
source. However, relatively high volumes are still required to
effectively collect fugitive emissions in this manner. Local hoods are
also undesirable from the maintenance and operating point of view.
Dampers utilized to direct charging emissions into the main gas cleaning
hood have also experienced only limited success to date.
Collection of fugitive emissions by means of a furnace enclosure has
proven to be the most economical solution, as well as the Best Available
Control Technology (BACT) since it allows collection of emissions at the
source and prevents their escape into the atmosphere. With properly
designed furnace enclosures, it has been demonstrated that it is possible
to effectively control scrap charging, hot metal charging, furnace
tapping, ladle alloy additions, furnace slagging and puffing emissions
with low volumes. At the present time, systems of this type are doing
an effective job of fugitive emission control with volumes of
approximately 350,000 ACFM.
-289-
-------
The EOF furnace enclosure (Figure 6) essentially forms a gas tight seal
when the bi-parting doors are closed. Since the furnace enclosure
extends below the charging floor, the only openings are for the ladle
car. If desired, the ladle car openings can be effectively reduced by
air curtains or the addition of a vertical shield on the end of the
ladle car, as a means to increase the efficiency of the furnace enclosure
during tapping.
The furnace enclosure design is based on sound engineering principles
and actual testing, taking into consideration the volume of fumes
generated inside the furnace due to the reaction of pouring hot metal
into the furnace, as well as the velocity and temperature of the fumes
leaving the mouth of the furnace (Figure 7).
During charging, the bi-parting doors are opened while charging scrap
or pouring hot metal and the fumes are collected through the secondary
hood located inside the enclosure directly above the furnace mouth.
For controlling emissions during all other phases of operation, the
enclosure doors are closed, while the fumes are evacuated from the
enclosure through the main or secondary hood.
With a properly designed furnace enclosure,' it is possible to collect
secondary emissions generated by the basic oxygen process with
approximately 901 efficiency, provided the charging of hot metal into
the furnace is done at a controlled rate and the scrap is relatively
clean.
One of the more recent EOF installations with effective fugitive
emission control facilities (Figure 8) incorporates individual furnace
enclosures over two (2) 350 ton vessels, with the main gas cleaning
systems being utilized for secondary emission control (Figure 9).
Practically all new BOF/Q-BOP vessels that have been installed in the
U. S. in the past seven (7) years have included a partial or full
furnace enclosure for fugitive emission control. Since the original
enclosure designs had many deficiencies, these systems are operating
today with varying degrees of success. At the present time, however,
there are approximately ten (10) installations operating or in the
construction stage which incorporate the secondary hood inside the
furnace enclosure with sufficient volume for effective fugitive emission
control being provided by the main gas cleaning or an auxiliary system.
-290-
-------
The first retrofit fugitive emission control system incorporating
furnace enclosures for controlling emissions has been installed in an
operating EOF Shop without interrupting normal production, demonstrating
that it is possible to install such equipment without disrupting normal
production schedules. In this case, an existing baghouse system was
utilized for gas cleaning.
Hot Metal Reladling
The hot metal reladling operation is another primary source of fugitive
emissions in the oxygen steelmaking process. The fumes generated
during hot metal reladling consist of approximately 55% iron oxide
less than 3 microns, 42% graphite greater than 75 microns and 3% others
with an approximate emission rate of 0.25 Ibs. per ton of hot metal.
With today's technology, collection of reladling emissions is not a
problem since it is possible to utilize local hoods or close fitting
ladle hoods depending on the arrangement of the reladling facility.
The main factors effecting the collection of reladling emissions is
the distance of the hood from the ladle mouth, as well as the rate of
pouring hot metal. In actual practice, since the handling time remains
almost constant regardless of the amount of hot metal handled, the
volume required to control reladling emissions with a canopy or local
hood is proportional to the volume of hot metal.
In contrast, utilization of a close fitting refractory lined ladle hood
for collecting the reladling emissions allows fume collection at the
source. The close fitting refractory lined ladle hood (Figure 10)
utilizes the "air seal" principle where the fumes are drawn through a
pouring slot in the hood and the gap between the hood and the ladle with
sufficient velocity to prevent fumes from escaping into the atmosphere.
Actual efficiency of the ladle hood has been demonstrated by a recent
installation (Figure 11) where the reladling emissions for a 350 ton
furnace are effectively collected with 125,000 CFM. In this design, the
hot metal is poured through a vertical slot in the movable ladle hood
which serves two (2) reladling stations. It is estimated that canopy
or local hoods would require volumes in excess of 300,000 CFM to
effectively collect these emissions.
Flux Handling
Flux handling also contributes to total fugitive emissions in the EOF
Shop, but these emissions are minor in nature and easily controlled.
-291-
-------
FUGITIVE EMISSION CONTROL
ELECTRIC FURNACE SHOP
Fugitive emissions associated with the electric furnace shop are mainly
charging and tapping emissions which are usually heavy and difficult to
capture. Today, a typical emission control system for an electric
furnace would normally consist O-F Direct evacuation or side draft hoods
for the primary gas cleaning with canopy or local hoods for collecting
charging and tapping emissions (Figure 12). Total emissions from
electric furnaces producing carbon steel are approximately 30 Ibs./ton
of steel produced with approximately 3 to 5 Ibs./ton associated with
charging and tapping emissions, while the particulate loading for
electric furnaces melting alloy steel is approximately one-half
of the emissions generated by carbon steel furnaces.
Charging Emissions
Charging emissions resulting from charging a hot electric arc furnace
with scrap are usually heavy and difficult to capture. The intensity
of the charging emissions is a function of scrap cleanliness. Scrap
containing heavy rust, oil, grease or dirt produces heavy emissions
during charging. These emissions are highly carbonaceous consisting
primarily of smoke and soot.
Tapping Emissions
Tapping emissions generated while pouring steel from the furnace into
the ladle are primarily composed of metallic oxides resulting from
contact with the air and from bath agitation. Tapping fumes, like
charging emissions, are also difficult to capture (Figure 13).
Methods of Collection
With present technology, the alternatives available for collecting
charging and tapping emissions are complete or partial building
evacuation and local hoods, in addition to partial or full furnace
enclosures.
Today, most electric furnace shops utilize canopy hoods for charging
and tapping emission control. Although they offer little operating
restrictions, they have the disadvantages of high volume requirements,
high capital and operating costs and result in poor working conditions
due to the shop atmosphere.
-292-
-------
Local hoods have achieved some degree of success for collecting tapping
fumes. However, operating practices do not allow the installation of
local hoods for collecting charging emissions.
Based on the success in collecting fugitive emissions in the EOF Shop
with furnace enclosures, the alternative of collecting Electric Furnace
Shop fugitive emissions by means of a furnace enclosure offers a
practical solution. Although the furnace enclosure offers an efficient
and economical solution to the problem, objections from operating
personnel appear to be a big obstacle.
While the furnace enclosure technology applied to electric furnaces
is still in its infancy, an electric furnace shop consisting of two (2)
60 ton arc furnaces, utilizing complete furnace enclosures for total
emission control, has been operating since 1976 without adversely
effecting their operating practices.
The typical furnace enclosure (Figure 14) essentially forms a gas
tight enclosure from which the fumes are evacuated.
During the charging operation, the bi-parting doors are opened to
allow entry of the scrap bucket (Figure 15). After positioning the
scrap bucket, the bi-parting doors are closed and the scrap is charged
into the open furnace. The fumes generated at this time are evacuated
from the enclosure for gas cleaning.
During the tapping operation, the bi-parting doors are closed while
the fumes are evacuated from the enclosure (Figure 16).
In addition to offering the Best Available Control Technology for
electric furnace shop emission control, the furnace enclosure offers
many other advantages, including effective emission control with low
volume requirements, collection of emissions at their source, lowest
capital investment and operating costs, improved working conditions
and a practical solution for sound control in electric furnace shops.
The Texas Air Control Board recently made an analysis of electric
furnace shop emission control systems with total emissions control
in the State of Texas since 1972. Utilizing a ventilation index
developed by the Texas Air Control Board to grade the various systems,
the furnace enclosure had the lowest index number proving to be the
most efficient installation.
-293-
-------
CONCLUSION
Technology for fugitive emission control has lagged behind the
technology developed for main gas cleaning systems in the Iron and
Steel Industry because the fugitive emissions were considered minor
in comparison to the emissions generated during the actual melting
periods, and the fact that regulatory authorities were willing to
tolerate these emissions while concentrating on other fnore important
areas. As a result, at the present time, system applications for
fugitive emission control have not reached the same degree of
reliability.
This gap in technology is gradually closing due to current pressure
by the regulatory authorities to control all fugitive emissions. This
is evidenced by the evolution of the total furnace enclosure for
controlling fugitive emissions from oxygen steelmaking facilities and
electric furnace shops and development of the close fitting ladle hood
for controlling reladling emissions.
With present day air pollution control regulations, new technology
will undoubtedly be developed for blast furnace cast house emission
control. Proven Japanese technology with modifications to accommodate
U. S. operating practices and cast house geometry appears to be a
practical solution. Two (2) new blast furnaces scheduled to start up
in the near future in the U. S. will incorporate Japanese technology
for cast house emission control.
The Best Available Control Technology (RACT) for fugitive emission
control that will be acceptable to the Iron and Steel Industry
requires a technology that offers the most effective and economical
type of fugitive emission control from the standpoint of working
environment which will allow these emissions to be collected at their
source and prevent their escape into the atmosphere.
-294-
-------
References
1. Blast Furnace Cast House Emission Control Technology Assessment,
EPA-600/2-77-231 (Nov. 1977).
2. Inspection Manual For Enforcement Of New Source Standards For
Producing Electric Arc Furnaces. EPA 540/1-77-007 (May 1977).
3. A Regulatory View Of Air Pollution Controls On Electric Arc
Furnaces In Foundries And Steel Mills In Texas by James C. Caraway,
Texas Air Control Board, Austin, Texas.
-295-
-------
Typical open cast house
with crane parallel to iron trough
Iron Trough
& Skimmer
Hot Metal
Runners
Hot Metal
Cars
PLAN
Section A-A
Figure 1 - Cast House Arrangement
Slag
Runner
Slag Pits
Crane
Slag Pits
-296-
-------
I
NJ
VO
Blast Furnace
Slag Pits
Iron
Runners
Slag Runners
Tilting Spout'
Enclosure
Typical cast house 10,000 ton day furnace capacity
Figure 2 - Multiple Tap Hole Cast House Arrangement
-------
Iron
Runner
to
10
oo
Runner
Retractable enclosure
over tap hole,iron
trough & skimmer
O
Enclosure take-off
duct to control
device
Plan view single tap hole furnace
partial emission control concept
Figure 3 - Retractable Curtains in the Cast House
-------
Figure 4 - Hot Metal Car
-------
I
OJ
o
o
Figure 5
Hot Metal Charging For
Basic Oxygen Process
-------
BUMPER
SECONDARY HOOD
HOT METAL CHARGING LADLE
FURNACE CHABONG DOORS
(Retractable)
SLAG POT
I
o
t-1
WATER COOLED HOOD
HOOD TRANSFER CAR
ADJUSTABLE SKIRT
TAPPING EMISSIONS DUCT
SEAL RING
FURNACE ENCLOSURE
OPERATING FLOOR
TEEMMG LADLE
SHOP MR INDRAFT
SLAGGMG&
TAPPING
^••^•^••^^^••••^^^^^•^^^^•iH^^HBBMH^^HMMHMMi^^HB^^HMVMHMB^^^HBHHM^^HBMH^HBMHHHBBMHaHHBiHBHH
Figure 6 - Schematic Arrangement For BOF Furnace Enclosure
-------
FURNACE ENCLOSURE
Figure 7 - Components Effecting Fugitive Emission Control
With Furnace Enclosure
-------
Figure 8 - Recent EOF Installation
With Fugitive Emission
Control
-------
Figure 9 - Hot Metal Charging
Emission Control With
Furnace Enclosure
-304-
-------
TORPEDO CAR
LADLE HOOD
i
u>
o
VJ1
HOT METAL LADLE
Figure 10 - Schematic Arrangement - Ladle Hood For Reladling Fjnission Control
-------
Figure 11 - Reladlin
KeJ.aaj.ine: hmi
Pouring Throu
Slot In The i
hmission Control
A Vertical
le Hood
-------
SCHEMATIC FLOW DIAGRAM
GAS CLEANING SYSTEM FOR
ELECTRIC ARC FURNACE
ROOF CANOPY HOOD
i
u>
o
WATER COOLED
ELBOW
SCRAP CHARGING
TAPPING
ELECTRIC ARC FURNACE
REVERSE AIR FAN
BAG FILTER (PRESSURE TYPE BAG HOUSE)
Figure 12 - Schematic Flow Diagram - Pollution Control System For Electric Furnace Shop
-------
higure 13 - Fugitive Emissions
During The Tapping
Operation
-308-
-------
Figure 14 • Typical Furnace
Enclosure For
Electric Furnace
-------
Figure 15 • Charging Operation
With Furnace Enclosure
-------
ure Hi - T;i| Iperal ion
With 1'iirnace l!n<
-311-
-------
CONTROL OF FUGITIVE EMISSIONS
AT REVERBERATORY FURNACES
AND CONVERTERS
by
Alfred B. Craig, Jr., IERL
U.S. Environmental Protection Agency
and
L.V. Yerino, R.T. Price, and T.K. Corwin
PEDCo Environmental, Inc.
October 24, 1978
-313-
-------
DISCLAIMER
This paper was prepared by PEDCo Environmental, Inc., for presentation at
the Third Symposium on Fugitive Emissions in San Francisco, California, on
October 24, 1978. Information was collected under Industrial Environmental
Research Laboratory's program on fugitive control (U.S. EPA Contract No.
6802-2535, Task 1). The opinions, findings, and conclusions expressed are
those of PEDCo Environmental and not necessarily those of the U.S. Environ-
mental Protection Agency.
ABSTRACT
The following are discussed in this paper: the functions of the rever-
beratory and converter furnaces; methods of charging, operations, slagging,
tapping of each type of furnace, and their related emissions during these
operations; methods of controlling the fugitive emissions during the afore-
mentioned operations; relative merits and efficiencies of the different
methods of controlling fugitive emissions and costs; types of furnaces avail-
able and their relative fugitive emissions in comparison with each other; and
possible future types of furnaces to minimize emissions.
-314-
-------
CONTROL OF FUGITIVE EMISSIONS AT REVERBERATORY
FURNACES AND CONVERTERS
Summary
Introduction
A. Copper Smelting Furnaces
1. The Smelting Process
a. The Reverberatory Furnace
b. The Electric Furnace
c. The Flash Furnace
d. The Noranda Furnace
2. Fugitive.Emissions - Smelting Furnaces
B. Copper Converters
1. The Converting Process
a. The Pierce-Smith Converter
b. The Hoboken Converter
2. Fugitive Emissions - Converters
C. Discussion of Fugitive Emission Controls/Practices
1. Smelting Furnaces
D. Relative Merits and Efficiencies of Different Methods of
Controlling Fugitive Emissions - Converters
E. Costs of Fugitive Emission Control Systems
1. Cost Parameters - Secondary Hooding
2. Capital Costs
3. Operating Costs
F. Current and Possible Future Methods/Processes For Emission Control
-315-
-------
SUMMARY
In terms of domestic consumption, copper is the largest of the three
primary nonferrous metals (lead, zinc, and copper). In 1976, about 1.33
million metric tons of copper were produced in the 16 U.S. smelters from
domestic ores, and 1.8 million metric tons from domestic and foreign ores.
Vast amounts of particulate and sulfur dixoide (S02) emissions are generated
in producing this tonnage of copper. Copper smelters are a major source of
S02, emitting approximately 80 percent of the total amount of S02 emitted from
the copper, lead, and zinc industries. The industry is implementing control
methods to recover some of the S02 as a marketable product. All of the
smelters have installed pollution control equipment to reduce emissions into
the ambient air, for example, electrostatic precipitators to remove particu-
lates from dust-laden air out of roasters, dryers, reverberatory furnaces, and
converters; and many of the smelters have acid plants to reduce S02 emissions
primarily from converters. Even with these emission controls, it is estimated
that 52,000 tons/yr of particulates, an indeterminate amount of fumes that
contain noncondensables, and 3.5 million tons/yr of S02 are still emitted into
the air through the stack. (This is based on 1974 reports and does not con-
sider new modifications to the smelters since then.) Sulfur dioxide quanti-
ties are especially high because SOX from multiple-hearth roasters and rever-
beratory furnaces is dilute, and only a portion is sent to acid plants and the
larger portion is emitted without control. Sulfur dioxide from converters can
be sent to acid plants only when the converters are in the blowing operation
since the 502 concentration is high. The emissions discussed so far are
called primary emissions because they are either sent to control equipment or
are at least ducted to a stack for dispersion into the ambient air.
Emissions of concern in this paper are secondary or fugitive emissions.
These fugitive emissions are those that escape through leaks in dryers,
roasters, smelting furnaces, converters, and anode furnaces or escape from
primary hooding systems not designed with great enough capture velocity, or
they escape because of the position of the process equipment in certain
phases of operation. Particulate fugitive emissions may be high around
multiple-hearth roasters where concentrate is being handled and roasted, old
reverberatory furnaces, leaky waste heat boilers, and converters when the
converter is not in stack or in-line with the main hood position. Sulfur
dioxide fugitive emissions can be considerable from roasters, smelting fur-
naces, and converters. Various reports state that 2 to 25 percent of the SO?
generated by the copper industry is fugitive emissions. The amount of fugi-
tives varies considerably within a plant from day to day and hour to hour.
Fugitive emissions are a problem at all smelters, however, and in some
plant configurations they are a difficult problem to handle efficiently. The
following is a summary of the copper smelters and their process equipment:
-316-
-------
No. of plants No. of units
Copper smelters
Green feed
Multiple-hearth roaster
Rotary dryer
Fluid bed roaster
Reverberatory furnace
Electric furnace
Flash furnace
Noranda furnace
Peirce-Smith converter
Hoboken converter
Anode furnace
Fire refining furnace
Acid plant
Liquid S0£
Dryer or roaster gas to
acid plant
Furnace gas to acid plant
Converter gas to acid plant
Total concentrate feed rate
of 16 plants, tons/day
Total copper production at
16 plants, tons/day
16
5
4
7
4
11
3
1
1
15
1
12
1
13
3
4
6
13
50
(includes dryers at concentrator
plants)
25
3
63
5
28
20
4
-20,000
* 5,700
This paper will briefly describe the smelter operations, address the
effectiveness and costs of fugitive emission controls, and also discuss some
ideas on possible future plant designs to reduce fugitive emissions.
-317-
-------
INTRODUCTION
Copper produced by the domestic primary copper industry is mainly from
sulfide ores comprised of a variety of minerals. Small amounts of copper ore
also are recovered from oxide ores, low-grade waste, and imported ores.
Because most of the domestic ore is mined in the southwestern states, the
majority of the plants are located in that area1. Figure 1 shows the locations
of the 16 primary copper smelters in the United States. Because most copper
ores are sulfides, copper recovery processes have been developed to treat
these ores. These processes recover copper while removing most of the impuri-
ties present in copper ore. Remaining impurities are removed by refining.
Copper metal primarily is recovered from copper ores by pyrometallurgical
processing with the remainder consisting of hydrometallurgical processes. The
pyrometallurgical processes convert ore concentrate into an impure copper
called blister copper. The processes consist of roasting or drying, smelting,
converting, and refining. The anode copper product (as high as 99.8 percent
copper) is then sent to a refinery for final purification.
Figure 2 represents a general flow sheet of the copper industry in the
United States. The figure shows the pyrometallurgical steps and the refining
processes.
This paper deals with the state-of-the-art in the control of fugitive
emissions created in the smelting and converting departments, which will be
discussed in the following paragraphs. Before beginning a more detailed
discussion of the smelting process, the following brief description of the
traditional and still most commonly used smelter operations will be presented.
Ore concentrate, containing about 25 percent copper, is usually railed to
the smelter and stored. From the storage area, the ore is generally conveyed
to a dryer, then to subsequent operations, or to hoppers above a bank of
multiple-hearth roasters. Feed from the hoppers is discharged on the top
hearth of a roaster where rotating arms push the concentrate to the center of
the hearth where it cascades to the hearth below. From this hearth, the
concentrate is directed to the periphery where it cascades to the next hearth
and then is directed through several additional hearths. The multiple-hearths
are gas- or oil-fired. This roasting process dries the concentrate and
removes some of the sulfur in the ore so as to improve the matte grade in the
following operation. (Since burning of sulfur is exothermic, the temperature
in all of the hearths can be maintained at adequate temperatures with supple-
mental heating as required.)
The dried or partially roasted (calcined) ore concentrate is usually
transferred from the roasters in larry cars to a reverberatory furnace(s).
-318-
-------
Figure 1. Locations of primary copper smelters in the United States.
-------
?1
MULTIPLE
ROASTING
FLUIDIZATION
r\
~ u
3 i
o /
vy
/-\
= .
-
•
-»
-»
RtVtKBtKfllUKl
SMELTING
1
SLAG TO
DUMP
ELECTRIC
SMELTING
|
SLAG TC
£
w
A
HOBOKEN
CONVERTER
1
SLAG TO ••—'
ELEC. FURN.
SLAG TO
SMELTING FURN.
ROASTING
DUMP
I
OJ
N)
O
x^
A
a'
nnv 1 ur ^3?
1 • DRYING • "X — «J
go
w
-» FLASH SMELTING -J Cl ..
*. (OUTOKUMPU)
1, RECOVERABLE COPPER
A
_ 7? i.i
"" NORANDA " °|T
(* S3
w
vy
ELECTRIC SLAG
A 1
FLOTATION
"~* SLAG TREATMENT
RECOVERABLE COPPER
"SLAG TO DUMP
•SLAG TO DUMP
Figure 2. General flow sheet of the copper industry in the United States.
-------
The concentrate is charged into the reverb(s) through pipes and/or hoppers
located at the top or along the side walls of the reverb. The concentrate
melts by reverberatory heating. The melt with the addition of a silica flux-
ing agent forms a copper-bearing matte layer and a waste slag layer. The
matte is tapped near the bottom of the reverb. The lighter slag is tapped at
a higher elevation than the matte. The slag is collected in a slag pot or
ladle and carted to, and dumped at, a disposal area. The ladle, in which the
matte is collected, is usually transferred by crane to a Peirce-Smith con-
verter (horizontal cylindrical furnace). This furnace converts the matte into
blister copper and slag by means of air blown into the converter below the top
of the bath line. The reaction is exothermic. The slag, containing recover-
able copper, is poured from the mouth of the converter into a ladle, and the
slag is then returned to the reverb furnace. After final blowing, the blister
copper (about 98 to 99 percent copper) is poured into a ladle and then trans-
ferred to an anode furnace. This furnace is usually fired with natural gas
and completes the smelter refinement of copper (99.5+ percent copper). The
anode copper is poured from the furnace into molds or a continuous casting
wheel.
These anodes, varying from 460 to 1000 Ib are cooled by quenching and
then stored and/or loaded onto rail cars for delivery to a refinery. Figure 2
shows other alternatives to the traditional method of copper smelting.
-321-
-------
A. COPPER SMELTING FURNACES
1. The Smelting Process. Copper smelting is the process of heating a
roasted or dried sulfide ore concentrate to its melting point to separate much
of the iron and undesirable impurities from the copper sulfides and other
valuable metals. In the smelting furnace, hot roasted calcines or raw (green)
concentrates are melted with siliceous and/or limestone flux, as well as
materials with recoverable copper values such as converter slag, flue dusts,
oxide ores, and copper scrap. As the temperature in the furnace increases, a
complex series of reactions takes place and the charge separates into frac-
tions. The sulfur in the charge combines preferentially with copper to form
stable copper sulfides, which are mutually soluble with the molten copper
metal. This fraction, known as matte, typically contains from 20 to 45 per-
cent copper (using the reverberatory practice) as well as other impurities
such as sulfur, antimony, arsenic, iron, and precious metals. Since copper
has a weak chemical affinity for oxygen, very little copper oxide forms. Iron
combines preferentially with oxygen to form iron oxides, which in turn react
with the flux to form iron silicates. These compounds and any calcium, mag-
nesium, and aluminum minerals that were present in the charge form a slag of
lower specific gravity that is insoluble in the matte and floats on top of it.
Any sulfur left over from the matte- and slag-forming reactions reacts with
additional oxygen to form SOX gas (primarily S02, some S03), which mixes with
the combustion off-gases. The energy necessary for smelting can be provided
by fossil fuel combustion, electrical power, and partially by heat of reaction
from the oxidation of iron sulfide to iron oxide.
a. The Reverberatory Furnace. The workhorse of the U.S. copper indus-
try is the reverberatory furnace or reverb (Figure 3), which was first in-
troduced in 1879 and is still used in modified form at 11 of the 16 domestic
smelters. The modern reverb is an arch-roofed or suspended-roof horizontal
chamber, generally about 35 m in length and 10 m in width. Heat is supplied
by fossil-fuel-fired burners located at one end of the furnace. A reverbera-
tory furnace operates on the basis of the heat from the flame radiating from
the roof onto the charge. Thermal efficiency of the reverb is low. However,
the reverb is flexible with respect to concentrate composition and is capable
of accepting as much as 1800 metric tons of material per day.
Although specific methods may vary considerably, reverbs are generally
charged either through the furnace top or along the top portion of the side
walls. Belt slingers (high speed conveyors), hoppers, and Wagstaff guns
(inclined chutes) may be used to better distribute the charge over the molten
bath. Drag chains and screw conveyors have also been used for charging.
Continuous charging is not practiced because of the problem of creating local
thermal imbalances that could reduce the furnace temperature below the effici-
ent level needed for the smelting reactions. Slag is drained periodically
from skimming bays at one end of the furnace and conveyed by short launders to
slag pots. The slag can be cooled, solidified, granulated, or dumped molten.
Matte is withdrawn periodically through tap holes in the lower furnace wall.
The matte flows down launders and into ladles, which are conveyed to the
converter by overhead cranes.
-322-
-------
A. BALLOON FLUE
B. CONVERTER UPTAKE HOOD
C. ADDITIVE SYSTEM
D. MONITOR
E. E.O.T. CRANE 60/60
F. PEIRCE-SMITH CONVERTER
G. LADLE
H. CONVERTER AISLE
I. REVERBERATORY FURNACE
J. WASTE HEAT BOILER
K. REVERB. FLUE
i
u>
NJ
Figure 3. Typical reverberatory furnace and converter
(Adapted from Kirk-Othmer Encyclopedia of Chemical Technology, Second Edition,
Volume 6, page 148.)
-------
The outlet gases from the reverb, which may run upward of 120,000 m /h,
usually contain between 0.2 percent and 1.5 percent S02- The specific S02
content is dependent upon the sulfur content of the ore and whether or not the
concentrate was roasted prior to smelting. Although there is considerable air
infiltration, the principal reason for the low-S02 content is the substantial
quantity of combustion air that must be provided to the furnace. The outlet
gases contain vapors (i.e., water vapor and some sulfurous and sulfuric acid)
as well as considerable dust and fume, including trace metal compounds of
arsenic, antimony, lead, and zinc. These gases are generally passed through
waste heat boilers to recover much of the heat of the combustion gas, and are
finally cleaned of particulate by means of hot electrostatic precipitators or
baghouses and vented to the atmosphere.
b. The Electric Furnace. The electric copper smelting furnace (Figure
4) has traditionally been used in Scandinavian areas where hydroelectric power
is cheap and fossil fuels are expensive. The first such furnace in the United
States started operation in 1972, and two more smelters have since adopted
this technology. The electric furnace is rectangular in cross-section with a
firebrick sprung-arch roof. The largest are about 35 m in length and 10 m in
width. Carbon electrodes are placed in contact with the molten slag, and the
heat required for smelting is generated by electrical resistance of the slag
to the submerged arc between electrode pairs. Electrical ratings vary up to
51,000 kVA. The chemical and physical changes that occur in the molten bath
are similar to those that occur in a reverb, although the thermal efficiency
of the electric furnace is much higher.
The charge of concentrate and fluxes is delivered to the roof of the
furnace by drag conveyors and then fed to the molten bath through multiple-
feed spouts near the electrodes and/or between the electrodes and sidewalls.
As the charge materials melt, they settle into the bath and form additional
matte and slag. Separate launders or chutes on the furnace end wall may be
used to charge converter slag and reverts. Matte is tapped into ladles from
tap holes placed in the hearth area near one end wall. Slag is skimmed from
tap hole(s) in the opposite end wall and delivered by launder(s) into slag
pot(s), which are usually hauled to the dump by trucks.
Although originally designed as an alternative to the use of expensive
fuels in the Scandinavian countries, the electric furnace also facilitates air
pollution control. Large amounts of combustion air are not required, re-
sulting in an outlet gas about an order of magnitude less than a reverb, but
containing the same quantity of sulfur oxide emissions. S02 concentrations of
2 to 4 percent can be expected, and particulate emissions should be less than
a reverb because of the lower gas volume and more uniform gas flow. The
electric furnace off-gas at all three domestic smelters is combined with other
high-S02 gas streams and used as feed to contact sulfuric acid plants.
c. The Flash Furnace. A more recent development in copper metallurgy
is the continuous flash furnace, which is more efficient in terms of energy
consumption and also produces a more easily controlled stream of flue gas than
the reverb or electric furnaces. Flash furnaces are of two types, the
Outokumpu Oy and the Inco, which differ primarily in their use of either
preheated air or commercial-grade oxygen to sustain the smelting reaction.
-324-
-------
ELECTRIC
POWER
CONVERTER
SLAG
LAUNDER
CALCINE
FEED
MATTE
SLAG
Figure 4. Electric smelting furnace.
CONCENTRATE
PREHEATED
AIR
CONCENTRATE BURNER
—-OIL
SLAG
SLAG MATTE
Figure 5. Outokumpu flash smelting furnace.
-325-
-------
The flash furnace is in widespread use throughout the world, although only one
is operating in this country (under license from Outokumpu).
The Outokumpu furnace (Figure 5) combines the functions of roasting,
smelting, and partial converting in a single furnace with three sections--
reaction shaft, settler, and uptake shaft. Dried ore concentrates are in-
jected continuously along with flux and preheated air into the reaction shaft
through concentrate burners. Oil may also be injected into the shaft. The
finely divided concentrate burns in a "flash" combustion as the particles fall
down through the shaft, and the heat released from the combustion of the oil
and sulfur sustains the smelting reaction. The process is similar to the
combustion of pulverized coal. The molten particles fall into the settler
part of the furnace and separate into matte and slag layers. The matte, which
contains 45 to 75 percent copper, is tapped from the settler and transferred
to converters for further processing. The slag, which contains too much
copper to discard, is also processed further. The flash furnace in the U.S.
does not have the flexibility to recover copper from converter slags, and as a
result these also require slag treatment facilities. The furnace outlet gases
are withdrawn from the uptake shaft. They contain 10 to 20 percent S02 and
considerable quantities of entrained molten or semimolten particulate matter.
The gas is cooled in a waste heat boiler, cleaned of dust in an electrostatic
precipitator, and used for sulfuric acid production. The Inco flash furnace
(Figure 6), which requires a low-cost source of oxygen to be economically
competitive, produces gases containing 75 to 80 percent S02» ideally suited
for liquid S02 manufacture.
d. The Noranda Furnace. A newer type of smelting furnace used in the
United States is the Noranda continuous furnace (Figure 7), in which roasting,
smelting, and partial converting reactions are combined in a vessel similar to
a lengthened Pe1rce-Sm1th converter. One U.S. plant has started operating
Norandas within the past year. The reactor is a horizontal cylindrical fur-
nace about 21 m long. It is fired from both end walls, and oxygen-enriched
air is blown into the matte layer through side-mounted tuyeres. The furnace
can be rotated on its horizontal axis, bringing the tuyeres out of the bath
and stopping the smelting process. The compact design facilitates process
control, and the domestic Noranda smelter is highly instrumented. The Noranda
was originally developed as a one-step process that would eliminate the con-
verter, thus significantly reducing capital costs and eliminating the need for
a crane aisle. In both of its commercial applications to date, however, the
use of a converter has been retained to allow better trace element control,
increased production, and longer reactor-lining life.
Concentrate and fluxes are fed to the Noranda by a slinger at one end
that spreads the feed over the molten bath. The high-grade matte, which
typically contains approximately 70 percent copper, is periodically tapped
from the side of the furnace and then transported by ladles to standard con-
verters where it is batch-treated to remove additional sulfur and iron prior
to fire refining. Slag, which contains 6 to 8 percent copper, is periodically
tapped from the end of the vessel opposite the slinger. Slag is upgraded by
milling, producing a concentrate (which is returned to the reactor) and a
tailing (which is discarded). The off-gases leave the Noranda furnace through
its mouth where they are captured by water-cooled hoods and ducted to a waste
-326-
-------
CHALCOPYRITE
SAND CONCENTRATE
CONSTANT
WEIGHT FEEDERS
OXYGEN-*
OXYGEN
PYRRHOTITE, CHALCOPYRITE
CONCENTRATES, AND SAND
SLAG MATTE
Figure 6. Inco flash smelting furnace.
CONCENTRATES
AND FLUX
TAPHOLES
COPPER
or
MATTE
21.3m
Figure 7. Noranda process reactor.
-327-
-------
heat boiler. The gases are passed through cyclones and electrostatic precipi-
tators to remove participate matter, and then used as feed to a sulfuric acid
plant. With 30 percent oxygen enrichment, the off-gases to the acid plant
contain 6 to 7 percent S02- The S02 concentration will be only about 4 per-
cent if oxygen enrichment is not used.
2. Fugitive Emissions - Smelting Furnaces. Fugitive emissions are
those that escape from processes because of the nature of the operation in-
volved or escape from primary control systems for various reasons (i.e., less
than 100 percent efficiency; inadequate capture velocity). The emissions of
concern in the copper industry consist of particulates, trace metal fumes, and
sulfur dioxide (S02). Some fugitive emissions (fugitives) occur with all
types of smelting furnaces. The emissions are present at all charging and
tapping locations, as well as through leaks in the furnaces and auxiliary
equipment. During the operation of smelting furnaces, the amounts of fugitive
emissions will depend on:
0 Charging technique
0 Furnace design and age
0 Amount of time required to charge feed and converter slag
0 Location and efficiency of any existent secondary hooding
0 Furnace maintenance and operations.
Fugitive emissions will be less variable over time with the continuous flash
and Noranda furnaces than the reverb or electric furnaces. The release of
materials through cracks and leaks is most serious with the reverb because of
both its increased gas volume and the fact that these furnaces are generally
older and in poorer repair than the other designs. The electric furnace
presents the additional problem of leakage from the furnace at the roof line
around the electrodes, as well as burning of the electrode paste in this area
at the openings in the band containers.
With the Noranda furnace, the primary hood cannot maintain a perfectly
tight seal with the reactor during smelting, and some fugitive emissions will
be released. There will also be dilute gases discharged through the mouth
when the furnace is rotated out from under the hooding system, even though the
smelting reactions will have largely stopped.
A final fugitives source near all smelting furnaces is the ladle, which
is used to transport the molten matte to the converter aisle.
B. COPPER CONVERTERS
1. The Converting Process. Copper converting is the batch-type process
by which blister copper is produced from the copper/iron/sulfur matte formed
in a smelting furnace. It is essentially an adaptation of the Bessemer
process developed by the steel industry. Molten matte and silica flux are
charged to the converter, and air or oxygen-enriched air is blown through
-328-
-------
tuyeres into the melt to oxidize the iron sulfides. This results in the
formation of an iron silicate slag, which floats on top of the matte and can
be skimmed from the converter. Further blowing oxidizes the copper sulfides
to blister copper, which is approximately 98 percent pure. Remaining impuri-
ties can be removed by fire refining at the smelter and electrolysis at a
refinery. The slag contains sufficient copper to be returned to the smelting
furnace. Almost all of the sulfur remaining in the matte is removed in the
converter, and the S02 content in the off-gas is sufficiently high that many
smelters produce sulfuric acid from this feed stream. However, sulfur re-
covery is complicated by the cyclic nature of the converting operation, which
results in major and frequent fluctuations in off-gas volumes and S02 con-
centrations. Converting is exothermic and no fuel is needed to maintain the
bath in a molten state. In fact, smelter reverts, copper scrap, and, in some
cases, copper concentrate may be charged to reduce its temperature and prevent
damage to the refractory lining.
a. The Peirce-Smith Converter. The Peirce-Smith converter (Figure 8)
is a horizontal, refractory-lined, cylindrical furnace with an opening in the
horizontal side that serves as a mouth for charging feed materials and pouring
slag and blister copper. It can rotate through an arc of about 120° from the
vertical for operational purposes. Compressed air is supplied through a
horizontal row of tuyeres along its back. The standard unit is about 4 m in
diameter and 9 m in length. First developed in 1909, the Peirce-Smith con-
verter is now used at 15 of the 16 domestic copper smelters, with as many as
nine units installed at one plant. Two or three converters are generally
associated with each smelting furnace. The Peirce-Smith converter is a rela-
tively efficient furnace whose high air flows permit both large copper through-
puts and the charging of bulky materials and copper scrap.
With the converter rotated partly forward, molten matte that has been
transported from the smelting furnace by the overhead crane is charged through
the converter mouth. The compressed air blast is put on, then the converted
returned to its operating position. Silica flux can be added either before or
after the converter is rotated to the upright position. The initial blow is
for the formation of slag during which the iron sulfides in the matte are
preferentially oxidized and sulfur is removed as SOz with the off-gas.
Periodically, slag is skimmed from the converter and fresh matte and scrap are
added. Blowing is discontinued during the skimming operation. The process is
continued until a sufficient quantity of copper sulfide (white metal) has
accumulated in the converter and the iron has been removed as art iron silicate
slag. The copper blow then begins, and it continues until the white metal has
oxidized, forming mainly S0£ and blister copper. Upon completion of the
copper blow, the blister copper is poured into ladles for transfer to anode
furnace(s). A complete converter cycle lasts about 10 to 12 hours for a 40
percent matte grade.
The outlet gases from the Peirce-Smith are released through the same
mouth used for charging and pouring. When the converter is in its upright
operating position, the gases are drawn into a fixed hooding system situated
above the converter. The hood has a retractable gate that can be lowered to
reduce leakage, and the off-take is controlled by regulation of the draft on
the flue system. By stolchiometric calculation, the S02 content in the con-
verter off-gas can be as high as 20 percent. However, the actual S02
-329-
-------
(end v«w)
(side view)
CO
O
I
(1) Shell; (2) Hood; (3) Air baffle; (4) Stack fluxing belt and chute; (5) Radiation
pyrometer; (6) Converter mouth; (7) Turning mechanism; (8) Turning motor and speed
reducer; (9) Rollers; (10) Turning rungs; (11) Air-supply duct; (12) Air-distribution
ducts; (13) Automatic mechanical tuyere; (14) Tuyere; (15) Puncher protection shields.
Figure 8. Peirce-Smlth converter.
(Adapted from Kirk-Othmer Encyclopedia of Chemical Technology, Second Edition
volume 6.)
-------
concentration is in the range of 2 to 10 percent primarily because of excess
air infiltration into the hood, which cannot be physically tight with the
converter. Oxygen enrichment of the compressed blast air can be used to
partially reduce the dilution effect of this air infiltration. Although a
reduction in S02 content is not desirable, the dilution air does serve to cool
the hot gases somewhat, which is necessary to prevent damage to downstream gas
cleaning equipment such as electrostatic precipitators or baghouses. Waste
heat boilers are also often employed on the converter gas. The cleaned gas is
used for acid production at most smelters; however, three domestic plants
release it without specific S02 control. To minimize the fluctuations in gas
flow and S02 content that occur throughout the operating cycle, and ensure a
relatively steady feed to the acid plant, the operation of two or more con-
verters must be properly scheduled to ensure that one furnace is always
charged with matte and blowing.
b. The Hoboken Converter. The Hoboken converter (Figure 9) is an
alternative to the Peirce-Smith that was designed to largely eliminate the
problem of excess air infiltration into the flue gas off-take system. First
developed in the early 1930's in Belgium, a Hoboken has been operated at a
single domestic primary smelter for about 4 years; it is also installed in a
number of foreign copper plants. The design of the furnace itself is similar
to the Peirce-Smith. However, instead of withdrawing the off-gas through the
converter mouth into a hood, the Hoboken is equipped with an integral side
flue located at one end of the furnace. Shaped like an inverted "U," this
flue, or siphon, rotates with the converter, as does the cylindrical duct to
which it is connected. A counterweight balances the siphon. The cylindrical
duct is connected by an airtight joint to a fixed vertical duct that leads to
the gas cleaning system. The Hoboken thus provides a direct link at all times
between the converter and the gas off-take, irrespective of its operating
position.
2. Fugitive Emissions - Converters. Fugitive emissions from converters
are always present during charging and tapping, and during slag and copper
blowing. Fugitives also occur during the rolling in and out of the converter
with blast air on. During the operation of a Peirce-Smith converter (10 to 12
hours to produce blister copper), the fugitive emissions that occur are depen-
dent on:
0 Matte grade - affects blowing time
0 Charging technique
0 Equipment design
0 Material additions of anode slag, cold material (i.e., skulls,
clean.ings, etc.) and flux.
The Peirce-Smith converter is a principal source of fugitive emissions in
a copper smelter. Basically, the furnace operates under a negative draft, but
due to the clearance required between the mouth of the converter and main
hood/gate, ambient air infiltrates into the main hood. This infiltration
disrupts the flow and causes some of the discharge gases from the converter to
-331-
-------
Figure 9. Hoboken converter.
-332-
-------
escape through this opening. Fugitives increase significantly during the
times in the operating cycle when compressed air is blown through the tuyeres.
Substantial fugitive emissions also occur during charging and pouring opera-
tions when the converter is rolled out and the hood is no longer positioned
above its mouth.
The Hoboken converter's operating cycle and charging and pouring opera-
tions are similar to the Peirce-Smith. The principal difference is that the
link to the gas cleaning system is never broken, as is the case when the
Peirce-Smith rotates forward from under its hood. With this in mind, no
fugitive emissions theoretically should escape the Hoboken. In actual prac-
tice, under good controlled operations, the Hoboken can eliminate fugitives.
A constant zero or slightly negative draft can be maintained; thus minimizing
fugitive emissions at the converter mouth. Maintenance of the proper draft is
complex. If excessive fan draft occurs, this will cause entrainment of the
material in the molten bath, which carries over into the siphon system and
forms accretions that restrict the gas flow and would disrupt the draft at the
mouth of the converter and would eventually cause emissions.
Additional sources of fugitives common to both types of converters can be
leaks in the off-gas ductwork and the ladles that carry matte to the converter
and ladles that transport blister copper to the anode furnaces.
C. DISCUSSION OF FUGITIVE EMISSION CONTROLS/PRACTICES
1. Smelting Furnaces. Methods to decrease fugitive emissions around a
reverberatory and electric arc furnace could include the following:
0 Good, tight fitting refractory/maintenance of the furnace and its
auxiliaries to reduce fugitive emissions through leaks in the brick-
work, and around electrodes and inspection ports
0 Hooding over slag and matte launders to pick up fugitive emissions
during tapping and slagging operations
0 Hooding over slag and matte ladles to pick up fugitive emissions
during tapping
0 Having adequate draft to maintain furnace under slightly negative
pressure
The Noranda furnace, although a smelting furnace, is similar to the
Peirce-Smith converter in design and operation. Thus, fugitive emissions can
be treated in the same manner as a converter.
D. RELATIVE MERITS AND EFFICIENCIES OF DIFFERENT METHODS OF
CONTROLLING FUGITIVE EMISSIONS - CONVERTERS
Control of fugitive emissions in the vicinity of Peirce-Smith converters
could reduce or nearly eliminate these emissions and help improve working
conditions for smelter plant personnel. Control can be achieved with varying
degrees of success by the addition of fixed or movable secondary hoods or by
-333-
-------
enclosure of the converter building. Efficiency of fugitives control ranges
from 30 to 100 percent, depending upon hooding configurations as outlined in
Table 1. Following are the major types of secondary hoods:
0 Fixed type -- made of structural steel with an eliptical cross
section. It is attached to the primary or uptake hood.
0 Fixed and movable -- consist of a movable intermediate hood and a
hood fastened to the gate. Both hoods are made of structural steel
with eliptical cross sections so that they telescope in the re-
tracted mode.
0 Swing-away type with a fixed overhead hood — made of structural
steel, refractory lined, and supported by pivot arms with a motor-
ized drive to permit positioning during blowing and pouring opera-
tions.
0 Combination of fixed and movable and swing-away type.
0 Controlled building evacuation with adequate number of air changes
discharging to a fabric .filter (baghouse).
The positions of the movable hoods for the various converter operations
are outlined in Table 1.
Estimated fugitive emission control efficiencies are presented in Table
2.
A descriptive summary of the secondary hooding systems is shown in Table
3. This table points out some of the operational and maintenance problems
involved, and factors that affect control efficiency.
E. COSTS OF FUGITIVE EMISSION CONTROL SYSTEMS
1. Cost Parameters - Secondary Hooding^. This cost study includes
various items that must be installed or modified to achieve control of fugi-
tive emissions from the Peirce-Smith converter (this furnace being the worst
source of secondary emissions). It does not include costs of certain opera-
ting procedures that would minimize fugitive emissions; e.g., allowing minimal
clearance between the primary uptake hood and the apron of the converter, or
maintaining proper matte charges to provide for direct flow of gases from the
mouth of the converter to the centerline of the primary uptake hood.
The following items are considered in this cost study:
0 A fixed hood having an elliptical cross section of 17 ft 6 in. on
the major axis and 7 ft on the minor axis, and 9 ft 6 in. long. The
plate is 3/8-in. carbon steel, with stiffeners of 7-in. channels.
The fixed hood is bolted to the primary hood and to the smoke plenum
of the secondary duct system.
-334-
-------
Table 1. POSITIONS8 OF MOVABLE HOODS
i
U)
u>
Ul
Type
Movable
Gate hood
Swing- away
Matte
addition
Retracted
Retracted
Retracted
Blowing
or
holding
Extended
Extended
Retracted
or in oper-
ating posi-
tion
Skimming
Extended
Extended
Operating
position
Rabbling
Partially or
fully extended
Partially ex-
tended
Retracted
Collar
pulling
Retracted
Retracted
Retracted
Pouring
Extended
Extended
Operating
position
* Following are definitions of hood positions:
Retracted - hood is in its highest or extreme position away from the
converter.
Extended - the hood is in its lowest or operating position.
Partially extended - hood is extended as far as practical to maximize secondary
emissions control.
-------
I
CO
co
Table 2. PEDCo'S ESTIMATED HOODING EFFICIENCIES'
(Values In percent)
Hood type
Fixed
Fixed and movable
Fixed and swing-away
Fixed, movable, and swing-away
Enclosed building
Matte
or
hot metal
addition
30-50
30-50
30-50
30-50
95b-100
niister
or
hot metal
pouring
30-50
40-70
80-90
BO-90
95b-100
Skimming
or
rabbling
30-50
40-70
50-70°
60-«0C
95b-100
Blowing
60-70
70-90
80-90
80-90
95b-100
Host system efficiencies would be higher if air motion* (i.e., open doors, man-cooling
fans, monitors, etc.) could be eliminated. Skimming in removal of slag from the con-
verter by tilting of the converter. Rabbling is removal of slag from the converter by
tilting of the converter and manual use of a rake to work the molten bath.
Reduced efficiency due to air motions; if doors are left open this efficiency could drop to 50%.
c Efficiency during skimming would be similar to blister pouring.
-------
Table 3. SUMMARY OF CURRENT FUGITIVE EMISSION CONTROL SYSTEMS
Type
Monitor. Natural
U.S.
Monitor, puwered
I
U>
OJ
Fixed hood with
secondary emission
ducting - U.S.
Enclosed converter
hood-swing away type
with fixed hood
Enclosed building
Design and operation
Simple design. Relies on outside
»1r moven«nt for removal of emis-
sions from the area.
Sirple design. Large air move-
ment required at the fans. Re-
moval rate is constant.
Clearance problems for crane hook
and cables during collar pulling or
matte additions. Retrofit diffi-
culties for ducting, fans, breech-
ing, and dust bins. Operational at
ill tines that converters are on
line. Good face and capture
velocities required.
Clearance problems for crane hook
and cables during collar pullino or
irstte additions (fixed hood). Clear-
ance problem for floor space re-
lationship tc the fixed hnod; rugged
drive mechanism needed for swing-
away converter hood.
Operational and maintenance problems
Haze in building during emissions; outside air movement
affects time required to clear the area. Crane oper-
ator and rainter.ance personnel working above the con-
verter may be required to wear face aspirators. Visi-
ble emissions in the monitor area. Maintenance in the
converter area, electric overhead traveling crane and
roof trusses for removal of the settled emiscions other
than gases.
Blind pockets or short-circuited flows could cause haze
and emission buildup in the roof line area. Grant
operator may require the use of a face aspirator at
times. Maintenance of fans and drives. Visible emis-
sions at each powered monitor. Maintenance in the
converter area, electric overhead traveling crane and
roof trusses for removal of settled emissions other
than gases.
Operational damage to hood by swinging or uncontrolled
electric overhead traveling crane action during matte
addition or collar pulling. Maintenance is less in the
converter area, electric overhead traveling crane and
roof trusses due to particle buildup.
Requires careful consideration of
all openings (personnel, t*-ock,
rail, materials) to minimize air
notion. Roof monitor must handle
all ventilation air for workers.
emissions and in leakages. In-
creases building costs because of
wind load design, tightness, and
close-fitting openings.
Space occupied in the aisle by the swing-auay converter
hood when adding matte, rabtlini. or skimming could
hamper crane movements. Crane must deposit the ladles
for pnuring or skinning and at the completion of the
operation nust await retraction of the hood before en-
gaging the ladle. Maintenance of the swing-away
mechanism and minimal maintenance for removal of par-
ticulate buildup that occurs during matte additions
and rabbling.
All openings must be maintained constantly against ex-
cessive air infiltration. Tight siding and roofing
required. Air circulation within tne building for
the workers and process must te carefully controlled.
Intoke and exhaust fans reajfre preventive raintenance.
Cleanup mainter.ar.ee for settled particulates in the
converter area is similar to that for a monitored
system, either natural or powered.
Efficiency
Dependent on outside air
currents and inside air
cation.
Dependent on number of
monitors, fan size, build-
ing design above tne con-
verter proper; air motions.
Dependent on the distance
of th? niouth of the fixed
hood from the emission
source; also on the capture
and face velocity created
by the fan at the mouth of
the fixed hood.
Dependent on operational
cycle. When pcjrir.j, blow-
ing or slocgir.; the emis-
sion ccntrol hould be cood;
when addino rcstte, or rab-
bling it wojld be similar
to the fixed hoed. Air no-
tions influences efficiency
1n all operations.
| Dependent on oulldinj
tightness, air ration con-
trol, monitor exr.aust capa-
oi 11 ties.
-------
0 The movable hood in the retracted position is parked above the fixed
hood. It has an independent track system with a five-speed, double-
grooved hoist unit and slack cable limit switch. The movable head
is 9 ft long and is elliptical, with a major axis of 18 ft 6 in. and
minor axis of 7 ft 6 in. These dimensions provide a 3-1/3-in.
clearance between the movable and fixed hoods. There are mating
plates on the lower end of the fixed hood and on the top of the
movable hood. The lower end of the movable hood is fitted with a
12-in. asbestos-type curtain that follows the elliptical perimeter
to form a seal with the gate hood. The movable hood is constructed
of 3/8-in. carbon steel with stiffeners of 7-in. channels.
0 The gate hood is elliptical in cross section with a major axis of 16
ft 6 in. and a minor axis of 6 ft 6 in. It is 9 ft long. Clearance
between the fixed hood and the gate is thus 3-1/2 in. The hood will
be bolted to the gate. The plate is 3/8-in. carbon steel reinforced
with 7-in. channels.
The dimensions listed above would be modified for each converter layout
to provide the required clearances. Design considerations may dictate that
the fixed hood is the largest unit, with the movable hood under it and the
gate hood under the movable hood.
0 If crane runway rail height is a problem, the smoke plenum of the
secondary hooding duct system could be fitted in as shown. The
plenum would span the primary uptake hood and would have a secondary
hood dust bin affixed on each end. The dust bins would be equipped
with pneumatic dust valves and discharge pipes. No provision is
made in this study for removal of dust in the dust bins. The smoke
plenum for this cost study is 4 ft by 4 ft 8 in. by 21 ft. It is
constructed of 3/8-in. steel with 6-in. channel stiffeners.
0 The secondary hooding duct system would have an uptake from each
dust bin adjacent to the converter and then pass to its main ducting
header for fugitive emissions. The damper valve shown would be
adjacent to the main ducting header and would be closed when the
converter is out of service. Existing facilities would determine
the path of retrofit. The gases go to a dust bin ahead of the fans
and from there to the breeching into the main converter gas duct and
to the stack (Figure 10). For this study, the length of the main
duct runs is 600 ft.
0 The fans considered in this estimate are Buffalo Forge type 1320 BL,
single inlet, arrangement 1, class 3 with 145 bhp, 785 rpm, 80,000
ft3/min at 200°F. There would be one fan for each converter in the
plant; as many fans as are required would be tied into the system.
0 Support items for the system include piping, wiring, foundations,
supports for ducting every 20 ft, expansion joints, miscellaneous
platforms, and walkways. Valves, fans, dust bins, and similar items
are flanged for. ease of maintenance.
-338-
-------
U)
LJ
SECONDARY
HOOD
DUCTING
DROP OUT
DUST BIN
BAGHOUSE
i-c
\ /
;
\ /
;
\ /
i "j ' N r ) /
j
/
BUTTERFLY
VALVE
MAIN
CONVERTER
GASES
DUCT
Figure 10. Fugitive dust ducting system (with baghouse)
from converter building to stack.
-------
Estimates of total installed costs are based on current (1978) costs for
major components of specified sizes, as provided by equipment suppliers.
Estimates of fabrication costs and installation charges are based on general
accepted engineering practices.
A 5 percent contingency factor to allow for emissions testing, equipment
changes, etc., is applied to the total of the direct and indirect costs. An
escalation factor of 7-1/2 percent per year is used to account for increases
in cost of equipment, labor, and services before and during construction.
Direct capital costs include equipment, instrumentation, piping, electrical,
structural, foundations, site work, insulation, and painting. Indirect capi-
tal costs include the following: engineering costs, contractor's fee and
expenses, interest (accrued during construction on borrowed capital-estimated
at 9 percent per year), freight, off-site expenditures, taxes (sales tax of 4
percent of equipment cost), start-up or shake-down, and spares.
Annualized costs include both operating costs and fixed capital charges
and consist of: utilities; labor and fringe benefits, maintenance, plant
overhead, and total fixed costs (amounting to 19.97 percent of total installed
costs and consisting of depreciation over 15 years at 6.67 percent unless
otherwise indicated, property insurance at 0.3 percent, property taxes at 4
percent, and interest on borrowed capital at 9 percent).
2. Capital Costs. Table 4 shows the direct cots, indirect costs
Including 1 year of contingency and escalation, and total capital costs for
plants containing one to nine converters.
3. Operating Costs. Operating costs include the following: operating
labor at $8 per man-hour, supervision at 15 percent of labor, maintenance for
labor and supplies at 2 percent of total capital cost, maintenance materials
at 15 percent of maintenance labor and supplies, electricity at 35 mills/kWh,
plant overhead at 50 percent of operations, and payroll at 20 percent of the
operating labor costs. The fixed costs include a straight-line 15 years
depreciation, 0.3 percent for insurance, 4 percent for taxes, and 9 percent
for capital costs. Table 5 lists the operating costs for a multiconverter
plant.
Additional handling of slag and blister ladles may cause delays in opera-
tion of the movable and swing-away converter hoods. It is estimated that a
delay of 5 to 15 seconds may occur with each ladle movement, equivalent to a
total delay of 3 to 10 minutes per day or a 0.23 to 0.7 percent production
slowdown. This loss is calculated on an annual basis as part of the operating
cost since it is negligible in comparison with other delays that are encoun-
tered such as delay because of lack of matte or because the anode furnace
cannot accept more blister copper or maintenance of converters or reverbera-
tory or electric arc furnace, etc.
Each of these systems is connected to a main discharge duct and an ex-
hauster type fan, which exhausts to a stack.
Relative capital costs of hooding for a five-converter smelter increase
in the following order: fixed, fixed with movable, fixed with swing-away,
-340-
-------
Table 4. ESTIMATED
HOODING FOR
CAPITAL COSTS OF SECONDARY
MULTICONVERTER PLANT
(dollars)
No. of
converters
1
2
3
4
5
6
7
8
9
Direct
costs
760,000
1,216,000
1,532,000
1,771,000
2,211,000
3,219,000
3,601,000
3,880,000
4,402,000
Indirect
costs
532,000
785,000
963,000
1,255,000
1,463,000
2,122,000
2,383,000
2,545,000
2,850,000
Total
costs
1,292,000
2,001,000
2,495,000
3,026,000
3,674,000
5,341,000
5,984,000
6,425,000
7,252,000
-341-
-------
Table 5. ESTIMATED ANNUAL OPERATING COSTS FOR SECONDARY HOODING
IN A MULTICONVERTER PLANT
(dollars)
No. of
converters
1
2
3
4
5
6
7
8
9
Labor
supervision
10,000
21,000
31,000
41,000
51,000
62,000
72,000
82,000
93,000
Maintenance
labor, supplies,
and materials
30,000
46,000
57,000
70,000
85,000
124,000
138,000
148,000
167,000
Overhead
plant and
payroll
22,000
38,000
50,000
64,000
78,000
105,000
119,000
131,000
149,000
Utilities
1,000
3,000
4,000
5,000
6,000
8,000
9,000
10,000
11,000
Fixed
costs
258,000
400,000
498,000
604,000
734,000
1,077,000
1,195,000
1,283,000
1,448,000
Total
annual
costs
321,000
508,000
640,000
784,000
954,000
1,376,000
1,533,000
1,654,000
1,868,000
ro
I
-------
fixed with movable and swing-away, and controlled building evacuation. Fol-
lowing are estimated costs to retrofit a smelter with five converters:
Fixed - $2.98 million
Fixed/swing-away - $3.04 million
Fixed/movable - $3.39 million
Fixed/movable/swing-away - $3.67 million
Enclosed building - $5.14 million
The enclosed building with baghouse, fans, and ducting is estimated at $11.78
million.
As shown in preceding tables, the costs of retrofitting fugitive emission
control systems are high. The costs shown are for so-called "typical" smel-
ters. Retrofit factors have been included in the costs and are estimated
based on the "normal" problems encountered with retrofitting a converter shop
that was not originally designed to have equipment installed. Some existing
shops are so constricted that retrofitting ductwork that will not interfere
with operations (i.e., crane, etc.) and is practically impossible. No smelter
is really typical. They are all different and have different problems associ-
ated in retrofitting the equipment. Therefore, it must be pointed out that
fugitive emission control costs must be analyzed on a site-specific basis and
much thought given to all related operations near a hooding system before
retrofit factors are assigned to a given plant.
F. CURRENT AND POSSIBLE FUTURE METHODS/PROCESSES FOR EMISSION CONTROL
The industry has remained the same in technology and design since 1909
with the reverberatory furnace and Peirce-Smith converter. Some methods have
been utilized and many methods could possibly be adopted in order to control
fugitive emissions in smelting and converting operations at a copper smelter.
The ideas discussed here would involve design of equipment new to the copper
industry, but not new to other metallurgical industries (i.e., steel indus-
try).
0 Charging floor technique
0 Q-BOF type furnace
0 Cascading system from calcined ore to matte to blister copper to
anodes
0 Covered ladles/stopper rods/autopour units
0 Evacuated building with each furnace in its own enclosure
-343-
-------
The following opinions have been formed with respect to control of fugi-
tive emissions from personal observations of various copper smelters in the
country:
Roasters - Nearly all of the multihearth roasters observed have observ-
able particulate fugitive emissions and detectable st-Tfur dioxide emissions.
As far as can be determined, these fugitive emissions have been greatly
reduced by the use of fluid-bed roasters with cyclone collectors. These
roasters appear to be operating with minimal problems. Besides nearly elimi-
nating fugitives due to leakage, the concentrate is fed in a slurry form to
the fluid-bed roaster, thus minimizing the escape of fugitives during con-
centrate transport. (Sulfur dioxide in the fluid-bed roaster gas stream,
unlike multihearths, is rich enough to be sent to an acid plant. Thus, fluid-
bed roasters, besides reducing fugitive emissions, also improve ambient air.)
Smelting furnaces - The reverberatory furnaces (reverbs) observed usually
had leaks in the brickwork in which fugitive emissions could be seen. Sulfur
dioxide emissions can usually be detected, particularly when standing at a
location even with or above the reverb roof.
The electric furnaces observed have had a minimal amount of fugitives
around them. The furnaces have had much tighter brickwork than the reverbs.
At one smelter, no fugitive emissions could be seen coming from the electric
furnace. At another smelter, only some emissions around the electrodes con-
sisting of paste that was ignited at the furnace roof line could be seen.
The Noranda furnaces (reactors) observed showed no fugitive leakage
during the smelting period.
Converting furnaces - The most common converter is the Peirce-Smith. All
of these converters generate varying amounts of fugitive emissions, from
minimal (barely visible) to considerable (dense cloud) depending on the mode
of operations. Some Peirce-Smith converters emit minimal amounts of fugitives
during blowing operations because the hooding system and the sliding gate form
a close fit with the mouth of the converter. If wide gaps are present between
the converter and the main hood and gate, or if leaks are present in the
hooding system, then fugitives are heavy and are very noticeable in the con-
verter aisle during blowing. The main point, however, is that fugitives can
be minimal during slag or copper blowing if a converter has good hooding and
good maintenance. The problem with all Peirce-Smith converters observed
with respect to fugitive emissions occurs when the converters are "out of
stack," meaning that the gate hood is up and the converter mouth is rotated
out from under the hood (i.e., receiving a charge or is pouring slag or
blister copper, etc.). When a converter containing matte is rolled out, air
is blown through the tuyeres (which are approximately 8 in. below the bath
level during blowing) for a short time to prevent plugging as the tuyeres
emerge; also when the matte is added to the converter, air is blown through
the tuyers for a short time to prevent plugging. Fugitive emissions are
usually great during this period. Fugitives continue to escape the main
hooding system even when the blow air has stopped, but in lesser amounts.
Secondary hooding can help capture some of these fugitive emissions. In
summary, fugitive emissions from Peirce-Smith converters can be controlled
-344-
-------
effectively when they are in the blowing mode by means of tight hoods. When
out of stack, however, only a portion of the fugitives from a Peirce-Smith can
be controlled by secondary hooding. One smelter has a modified enclosed
converter building in which 60 percent of the roof truss area is under control
of a fugitive exhaust system to a baghouse. This system will not be commented
on because it still is being modified and is not a totally enclosed and con-
trolled evacuation type.
Escape of fugitive emissions could be minimal from the Peirce-Smith
converter if it could be charged from a ground-operated machine (discussed
later) rather than by an overhead-type crane. Secondary hooding would be
effective if it could be kept in position directly over the fugitive sources.
Retrofitted secondary hoods, now at some smelters, must be positioned to be
clear of the crane operations for charging the converter, etc. This makes
fugitive emission control at Peirce-Smith converters a difficult problem with
the current building design.
Hoboken converters - It may not be fair to compare Hobokens and Peirce-
Smith converters (with secondary hooding) on an equal basis. The Hoboken unit
is designed to operate with a zero or slight negative pressure. The Hoboken-
Overpelt unit in Belgium operates with little or no emissions during blowing
operations. Some fugitive emissions are released during tapping, slagging,
and charging operations at the ladle. A telescopic secondary hood operating
in conjunction with a Hoboken converter would eliminate emissions during
slagging and tapping blister copper and greatly reduce the small amounts
of emissions during charging.
A Peirce-Smith converter with retrofitted secondary hooding cannot really
be compared with new equipment designed to eliminate fugitive emissions. A
better comparison would be the installation of a Peirce-Smith converter in a
new shop which would be designed with hoods specifically to capture fugitive
emissions.
The Hoboken converter can control fugitive emissions much more readily
than the Peirce-Smith type.
Other observed equipment currently in use to control fugitives are as
follows:
Hoods over matte and reverb slag tapping holes can be very effective.
Hoods over a filling slag pot can also control emissions to the point where
none can be observed. Swing-away hoods that can be lowered over ladles
receiving matte can be effective. The hoods observed appear to pull most of
the fugitives effectively, although some escape the outer periphery of the
ladle or hood because of the large clearance between the hood and the top of
the ladle.
Even newly designed shops installing secondary ductwork on a Peirce-Smith
converter that would minimize fugitives and not be in the way of the overhead
cranes etc., create many design problems.
-345-
-------
A possible answer to the Peirce-Smith fugitive emission problem would be
a charging floor technique. A charging machine(s), illustrated in Figure 11,
could replace the overhead cranes that transport ladles of matte or slag. A
charging machine, for example, could lift a ladle of matte, move back and
pivot around 180 degrees, then move to the converter. The arms would lift the
ladle to the desired height and tilt the ladle for discharging the matte into
the converter. The converter would be contained or encapsulated with a hood
to capture the fugitive emissions. A portion of the hood for the converter
would be retractable during the times an overhead crane is used for main-
tenance. A Hoboken converter without a main hood would be better than a
Peirce-Smith converter in this mode of operation.
With the elimination of overhead cranes for production in the converting
aisle, the proper ducting could be installed above the ladle and mouth area.
The ducting could be located in positions most suitable for when the converter
would be in the rolled out position. A charging machine would perform all of
the duties of overhead cranes, except for maintenance.
Whether using overhead cranes or charging machines, the transport of open
ladles from a matte-producing furnace to a blister-producing furnace means
escape of emissions from the ladles. Fugitive emissions during ladle trans-
port are not a major source, but they do add to the fugitives problem. Because
the ladles are often near the converter aisle floor, the fugitives that are
generated are near the operations people. A possible method of reducing these
"ladle" emissions would be to have covers on the ladles that could be placed
on or taken off with a minimum amount of time and trouble. (This may be
difficult in existing shops.) In conjunction with covers, specially designed
bottom-pour ladles with stopper rods (Figure 12} could possibly be used at
smelters for molten metal transport. Such a ladle with an autopour unit is
used in the steel industry where it is attached to the operation of the stopper
rod mechanism (Figure 12). This would allow the crane operator to control the
flow of matte directly into a converter or blister copper into an anode fur-
nace. A bottom-pour ladle could even be seated on a ladle support, and the
ladle tapped to a specially designed opening in the end of a converter through
a refractory-lined air tight joint/cover. (In this latter case, the mouth of
the converter could be made smaller, if so desired, and would be used solely
for pouring and not receiving hot metal charge.)
The electric overhead traveling crane (EOT) could be adapted to minimize
fugitive emissions from ladles in transport, ladles being filled, or ladles
pouring molten metal into a converter or anode furnace, by using a capture
hood fixed to the spreader beam (Figure 13). The capture hood would be fixed
to a sectionalized telescopic column (similar to that used in the steel indus-
try for handling ingots going to and returning from a soaking pit). The
telescopic column would be attached to the trolley, where a transfer duct
would discharge to a fixed split rubber covered duct fixed to the EOT's girder.
This duct would discharge to another split rubber covered duct adjacent to the'
building columns on the crane runway at the walkway level or at the roof truss
line. This second duct would discharge to a baghouse. An I.D. fan, posi-
tioned on the trolley, walkway, or girder, would supply the required capture
velocity. Another technique would be to discharge from the trolley to the
roof truss and use the building monitor for discharging to a baghouse.
-346-
-------
I
to
MODIFIED
CHARGING
MACHINE
Figure 11. Modified charging machine.
-------
Figure 12. Hydraulic cylinder mounted on barrel of ladle rigging raises
and lowers stopper rod to control flow of molten steel from ladle to
ingot mold.
-348-
-------
Figure 13. Sketch of E.O.T. crane with capture hood.
-349-
-------
Another fugitive-emission control scheme could be an evacuated building
with each furnace in its own enclosure with an independent pendant-operated
crane for handling ladles, etc., within the enclosure. Ladles would be removed
or brought into the enclosure through a movable rail-mounted car that would
pass through a sliding door. Emissions would be exhausted through a duct
system to a central point.
In summary, as far as fugitive emission control is concerned, there are
methods currently used that are reducing fugitives, but many ideas need to be
developed and good ideas tested so that fugitive emissions can be eliminated
in a feasible manner. Elimination of fugitives would improve working condi-
tions in a smelter. Besides that, new ideas and equipment that have been
implemented in industry to solve a problem have often lead to a technical
breakthrough that was not originally foreseen.
-350-
-------
APPENDIX A
Figure A-l, Matte charging operation.
-351-
-------
Figure A-2. Peirce-Smith converter - retracted
hooding, pictorial view.
-352-
-------
T.O. RAIL
SMOKE
HOOD
PLENUM
TO SECONDARY
HOODING
MAIN DUCT
Figure A-3. Secondary converter hood configuration
-353-
-------
.JICOHMIT WCT
/CUM lid
Figure A-4. Fixed, movable and gate hoods - front view.
-354-
-------
TO CttMM WWi>
wr*
SUMMIT MCI
INSTALL LIMIT SWITCH TO
TRAVEL OF TROLLY FOR
CHARGING OR COLLAR PULLING
TO PROTECT HOODING: FOR
OTHER MAINT. WORK, ETC.
REQUIRING THE HOOK TO WORK
IN THIS AREA THE LIMIT
SWITCH WILL ENERGIZE A GONG
ft FLASHING LIGHT TO ALERT
THE CRANEMAN t CONVERTER
OPERATORS THAT THE LOAD ON
THE CRANE OR HOOK CAN
INTERFERE OR DAMAGE THE
HOOD UNLESS THE HOODING
IS RETRACTED.
TUt MM
Figure A-5. Peirce-Smith converter - side view,
blister pouring operation, hooding extended.
-355-
-------
CSMW.
(UCATIOH VVt> It
« Itt CWUIt MMUkl)
TO inwt,!i mini
!U (»ll
ULOWM.T MCI »
wtunt
NP09 MOIST ICCMUia
"•>,.,
111 M TW ma
Figure A-6.
Peirce-Smith converter
hooding in position.
- side view,
-356-
-------
Photo A: Morenci - converter with gate (upper left) and
apron (center left), converter operating platform (lower
right), and typical ladle with lifting bale (lower left).
Photo B: Morenci - fixed hood, smoke plenum and duct (upper
center and left), and gate (lower left).
-357-
-------
Photo C: Morenci - fixed hood, smoke plenum and duct to dust
bin (upper right, upper and middle center), and gate (center
and lower right).
Photo D: Morenci - view of converter uptakes, fixed hood,
smoke plenum, dust and gates.
-358-
-------
Photo E: Morenci - view of converter aisle. Converters to
the left and reverberatory furnaces are on the right.
r: Morenci - view of secondary hood ducting (center
left to center right).
-------
Photo G: Morenci - view of secondary hood ducting leading to
thp ctarlc
Photo H: Ajo - ladle (center) pouring matte into the con-
verter by means of the overhead crane. Gate in raised
position (upper center). Apron on converter is seen clearly.
-360-
-------
Photo I: Ajo - secondary emissions (at center) are visible
with the gate in a raised position (center) and fixed hood
(upper center).
Photo J: Ajo - secondary emissions (at center) are visible
with the gate in a raised position (upper center), the fixed
hood, smoke plenum and dust bin (upper center and left)
-361-
-------
REFERENCES
1. Environmental Assessment of the Domestic Primary Copper, Lead, and Zinc
Industries (prepublication copy). PEDCo Environmental, Inc. EPA Con-
tract No. 68-03-2537, Work Directive No. 1. October 1978.
2. Background Information for New Source Performance Standards: Primary
Copper, Lead, and Zinc Smelters. EPA-450/2-74-002a. October 1974.
3. S02 Control for the Primary Copper Smelter Reverberatory Furnace (Draft),
Pacific Environmental Services, Inc. EPA Contract No. 68-03-2398.
August 1977.
4. Copper - State-of-the-Art. Donald G. Treilhard. Engineering/Mining
Journal. April 1973.
5. Flash Smelting - A World Beating Finnish Process. World Mining. March
1978.
6. Kennecott Plans Mid-year Smelter Startup. Mining Engineering. May 1978.
7. The Hoboken Copper Converter. Mechanism, Metallurgie Hoboken-Oberpelt.
8. Secondary Hooding for Peirce-Smith Converters. PEDCo Environmental, Inc.
EPA Contract No. 68-02-1321, Task No. 47. December 1976.
-362-
-------
Core Room Emissions in Foundries
William D, Scott
Southern Research Institute
Birmingham, AL
Robert C. Cummisford
Krause Milling Co.
Milwaukee, WI
Presentation to Third EPA
Symposium on Fugitive Emissions
October 23-25, 1978
San Francisco, CA
-363-
-------
ABSTRACT
Cores are the preformed sand inserts placed in foundry molds to produce the
internal cavities in the castings used by our society. Traditionally the
sand used for coremaking has been bonded with drying oils that were thermally
set by the heat of electric, coke, oil or gas-fired ovens. Thirty years ago
synthetic thermosetting resins, including furan and phenolic-based products,
were formulated for coremaking. More recently other chemical binders, in-
cluding some that do not require baking, have been developed that give faster
cure times. For the most part, technical developments in foundry binders
have been independent of environmental considerations.
It is estimated that the nation's foundries annually produce 3.3x106 pounds
of emissions during coremaking. These emissions range from simple hydrocar-
bons such as methane through unsaturated aliphatics, various solvents, the
carbon oxides, and others most noted for their odors. These emissions have
two basic sources: The binder (and its subsequent polymerization reaction
products) and the oven emissions associated with fuel combustion.
Recent research has developed a new binder system based on a starch product
that satisfactorily bonds core sands which is inherently less prone to produce
objectionable emissions. These low emissions levels have been demonstrated
in the laboratory. The binder requires a drying step to develop maximum
strength. Microwave energy has sucessfully been used with this system,
effectively eliminating oven emissions at the foundry.
-364-
-------
CORE ROOM EMISSIONS IN FOUNDRIES
The casting process is the basis of all manufacture involving metal.
In the casting process, molten metal is poured into a mold which contains
a cavity with the desired shape of the final product. The metal solidifies
and cools in this configuration. Foundry molds are typically made from sand
and have two mated parts: the cope (the upper half) and drag (the lower half).
Sand molds are prepared by ramming or blowing sand around a pattern. Binders
are added to the sand prior to molding so that the patterns may be withdrawn
to leave the desired mold cavity. These binders may be as simple as the wet
day in green sand molds or as complex as organic resins.
When the configuration of the casting requires internal voids or hidden
cavities, it is most economical to cast them in place. Preformed sand in-
serts, called cores, are used for this purpose. A core must be rigid and
self-supporting, and therefore the binding agent must give it more strength
than is usually provided by mold binders. Traditionally, the sand is mixed
with a drying oil which polymerizes or cures, to bond the sand when heated.
Cores are prepared in a core box, such as illustrated in Figure 1. Cores
are often made with the aid of mechanical equipment such as squeeze molders,
core blowers, or core shooters, which inject the sand mixture quickly and
automatically into the core boxes. The boxes are stripped away from the sand
and the core is laid on a flat or contoured support, called a core drier,
which prevents deformation during the oven baking cycle.
One of the most significant advances in foundry technology in the last
thirty years has been the development of new binders for sand molds and cores.
These binders make it practical to design more complicated cores and they
also allow the foundryman to achieve better dimensional accuracy in the cast
product. Table I lists some of these processes and the binders employed.
The development and description of specific binders for these processes have
recently been reviewed. (1,2)*
In general, binder development has been directed toward the use of com-
plex organic resins, and the use of catalysts that allow curing without heat.
Many of these "no-bake" binders were developed with no specific concern for
their environmental impact. However, in current research and development
of binders, more attention is being paid to reducing the level of emissions
both in the mixing and curing, and during the pyrolysis of these binders.
(3,4,5) Inorganic systems based on phosphates and silicates have been touted
as solutions to environmental problems in the foundry.(2) Poor collapsibility
of the cores has been reported, which makes core removal from the casting
difficult and prevents their widespread use.(5,6) The addition of organic
additives to these systems to improve collapsibility produces pyrolysis pro-
ducts of similar composition and magnitude to those from organic resins, (7,8),
but the molding and mixing emissions are low.
* Numbers in parentheses refer to items in "References."
-365-
-------
CAN BE METAL
OR PLASTIC
CORE BOX
PATTERN
CORE
MOLD
MOLTEN METAL
CORE
FLASK
POURING CUP
SPRUE
COPE
DRAG
FINISHED CASTING
Figure 1. CASTING NOMENCLA TURE
-366-
-------
process
Oil
Shell
Hot-Box
No-Bake [Acid Cured]
No-Bake
Cold-Box
Spirit
Silicate
TABLE I
Common Core Binders
Binder
Drying oils
Phenolic Novalac Resins
Phenolic or Furan Resins
Phenolic or Furan Resins
Urethane Resins
Phenolic, Furan or Urethane Resins
Carbohydrates (Starches)
Sodium Silicate
-367-
-------
The widely accepted organic binders that require heat for curing — speci-
fically those for the shell, warm box, hot box and core oil processes - are
declining in usage as shown in Figure 2. (9) Oil-sand mixtures must be heated
in a core oven. These ovens may be heated by coal, oil, gas, or electricity.
Oven designs have been proposed which heat cores with hot exhaust gases from
metal melting furnaces. Whatever the fuel, core ovens require extensive floor
space. Vertical ovens which allow sufficient time at temperature (dwell time)
to cure the core binder are available which conserve floor space.
Recently, core ovens have been developed which employ microwave energy
to cure the binder in a more controlled manner. Design requirements are
radically different for this type of oven, in that wave reflection and shield-
ing rather than insulation and burner design are critical to the oven's ef-
ficiency. (10,11,12) These ovens are particulary effective in drying core
coatings.
Core Room Emissions
There are two sources of emissions in the core room: the emissions from
the core ovens and those released during the curing of the sand binders.
The area where the cores are made is often separated from the remainder
of the foundry, sometimes even detached from the foundry building. Core rooms
are ventilated to remove the chemical emissions, and particularly objectionable
odors, from the work area. However, these emissions are usually exhausted
directly to the atmosphere outside the foundry, and may contribute signifi-
cantly to foundry fugitive emissions.
Foundries using thermally-cured resins with core ovens can control in-
plant emissions if the ovens are properly vented, but unless they are equipped
with afterburners or scrubbers, pollutants will still be discharged to the
atmosphere outside the plant. Such controls are rare. Table II lists the
emissions expected during the mixing, molding, and curing of cores prepared
with thermal setting binders.
Another approach has been to use "no-bake" binders, which require no
heat to cure, but may still release objectionable emissions as indicated in
Table III.
Some core binder systems inherently release objectionable emissions during
curing. Hexamethylamine (Hexa) is added to "shell" sand to provide formalde-
hyde to cure the phenolic resole resin, but it also introduces ammonia into
the atmosphere. (14,15) The urethane-base cold-box process uses odorous ter-
tiary amines such as dimethylethylamine (DMEA) as a catylst, which must be
controlled with a phosphoric acid scrubber. (16,17,18) Furan-phenol formalde-
hyde binders employing gaseous sulfur dioxide as a catalyst are also being
marketed as cold-box systems. (19) For both of these processes, the catalytic
gases must be contained and scrubbed to avoid contamination of the work area.
(18,19) These controls also are effective in reducing fugitive emissions or
these catalysts.
As shown in Tables II and III, hydrocarbons are emitted during the curing
of the binders. The most common hydrocarbon emissions are solvents, which
are added to the binders to lower the viscosity. Some of these are aromatic
-368-
-------
300
200 -
2
u.
o
Cured No-Bike
Box
^, podium Silicate
•Alkyd Urathine
52 54 56 58 60
Phenolic Uretlwm
48
70 72 74 76
Figure 2. A NNUAL BINDER CONSUMPTION (US).
-369-
-------
TABLE II
AIRBORNE EMISSIONS PROM CHEMICALLY BONDED THERMO-SETTING
SYSTEMS DURING MIXING, MOLDING AND COREMAKING (13)
CHEMICAL SPECIES
/
/ f
/ / /
CORE OIL
FURAN HOT-BOX
PHENOLIC HOT-BOX
SHELL
CARBOHYDRATE
7
7
7
7
7
0
7
X
X
7
0
0
0
7
0
0
0
0
X
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
?
0
7
7
0
7
7
7
7
0
7
7
7
7
0
0
?
?
0
0
0
7
7
0
7
ASUMES NORMAL VENTILATION, OPTIMIZED BINDER USAGE AND PROPER HANDLING
OF BINDER SYSTEM.
0 NOT EXPECTED TO BE PRESENT IN SUFFICIENT QUANTITIES TO BE CONSIDERED
A HEALTH HAZARD
? POSSIBLY PRESENT IN WORKING ENVIRONMENT DEPENDING UPON SPECIFIC
FORMULATION AND SAND QUALITY
X PRESENT IN SUFFICIENT QUANTITIES TO BE CONSIDERED A POSSIBLE
HEALTH HAZARD
-370-
-------
TABLE III
AIRBORNE EMISSIONS FROM CHEMICALLY BONDED
"NO-BAKE" SYSTEMS DURING MIXING, MOLDING & COREMAKING (13)
CHEMICAL SPECIES
•* * /
FURAN NO-BAKE
(H3P04)
FURAN NO-BAKE
(ISA)
PHENOLIC NO-BAKE
(TSA)
ALKYD URETHANE
PHENOLIC URETHANE
SODIUM SILICATE
(C02 OR ESTER)
0
0
0
0
0
?
7
X
0
?
Q
0
0
0
0
0
0
0
0
0
0
0
0
?
?
0
0
0
0
0
0
0
?
0
0
0
0
?
0
0
0
?
X
0
?
0
0
?
0
0
0
0
?
0
?
0
?
0
0
0
0
0
?
0
0
0
ASSUMES NORMAL VENTILATION, OPTIMIZED BINDER USAGE AND PROPER HANDLING
OF BINDER SYSTEM
0 NOT EXPECTED TO BE PRESENT IN SUFFICIENT QUANTITIES TO BE CONSIDERED
A HEALTH HAZARD
? POSSIBLY PRESENT IN WORKING ENVIRONMENT - DEPENDING UPON SPECIFIC
FORMULATION AND SAND QUALITY
X PRESENT IN SUFFICIENT QUANTITIES TO BE CONSIDERED A POSSIBLE
HEALTH HAZARD
-371-
-------
compounds, which are readily released into the atmosphere. (1) Thermally
cured binders are likely to give off methane and aliphatic hydrocarbons.
In core rooms with inadequate ventilation, an oil mist is sometimes observed
in the area. Some CO may be produced when curing thermal setting binders,
especially if coated sand is exposed to higher temperatures. Traces of other
organic compounds in the core room atmosphere, such as amines, can be objec-
tionable because of their strong odor. (20)
The other source of core room emissions is the oven. Depending upon
the oven type and configuration, there are combustion products from the fuel.
Natural gas or propane is preferred for a number of applications, including
the hot-box process and the shell process which was heated core boxes, because
the combustion/temperature is readily controlled. Drying ovens with natural
gas are also available, but with increasing industrial unavailability of this
fuel, more conversions are seen to either oil or coke. In any case, conven-
tional oven emissions such as methane, carbon monoxide, carbon dioxide, and
water will be produced. More complex species may be formed, depending upon
oven efficiency. Often all these combustion products are usually ventilated
directly to the outside.
The use of electric heat essentially defers emission of these combus-
tion products to a power plant, where control may be better effected. A com-
parison of oven emissions during core baking between an electric oven and
a conventionally fired oven will define the proportions of the emissions due
to the combustion of the fuel and that produced by binder decomposition.
During the work on organic binder decomposition emissions, a promising
new system based on carbohydrates was identified. It is marketed as a core
oil substitute and expected to produce fewer emissions than the conventionally
used core oils since it is in a water medium. This binder is based upon the
controlled formation of starch polymers. (21)
The Spirit^ binder system is based on starch. Starch is predominantly
otl-4 anhydro glucose polymer with some al-6 linkages. This polymer is quite
large in its natural form. In the Spirit system, starch is chemically treated
give a controlled distribution of lower molecular weights that are dispersible
in water. Addition of aldehyde groups with at least two aldehyde groups on
each molecule completes the formulation. At the time of use, an inorganic
salt catalyst is added. This gives a resin system that is basically carbohy-
drate. Crosslinking is presumed to proceed through a reaction of the aldehydes
with alcohol groups present in the carbohydrate to form hemiacetal and then
acetal configuration. Theoretically, low temperature pyrolysis, such as
drying temperatures in the range of 120-150°C, should result in principally
water and carbon with possibly some small quantities of formaldehyde.
t registered Trade mark of Krause Milling Co., Milwaukee, WI
-372-
-------
SPIRIT BINDER EMISSIONS
Experiments were designed to evaluate the emissions during the baking
of cores bonded with the carbohydrate-based Spirit binder. Cores were baked
in a conventional electric resistance oven and also in a microwave oven.
Table IV reports the results of these experiments. Analysis of samples of
the gas taken in the oven cavity showed that the carbon monoxide content was
less than 25 ppm. The total hydrocarbon concentration was approximately 25
ppm, but the aromatic portion, as reflected in the concentrations of benzene/
toluene, xylene, and naphthalene, was extremely low. Trace amounts of formal-
dehyde were found. As the binder contains no nitrogen in its formulation,
nitrogen-containing compounds were not expected nor analyzed for. These
emissions are relatively low, approximately of the same order of magnitude
as for the sodium silicate process.
It is anticipated that a foundry would be more satisfied with the shake-
out characteristics of the Spirt binder than with sodium silicate. The general
surface finish of a gray iron automotive casting produced with a Spirit bonded
core (no mold wash) was equivalent to that of an oil-bonded core with a re-
fractory core wash, as illustrated in Figure 3. (22) Both of these are com-
mercially acceptable.
Shot blasting and grinding, which produce unhealthy dust emissions, are
often necessary to remove adhering sand from castings produced with sodium
silicate; and internal cores such as in Figure 3 might be difficult to produce
with sodium silicate bonded sand because of the problems with core removal.
Microwave Ovens
As the starch-based binder must be dried to cure, it is particularly
well adapted to microwave curing. The literature contains several references
to this new application of microwave energy.(12,21,23,24) Essentially the
microwave oven offers the foundryman a controlled, emission-free oven that
readily dries sand. The sand aggregate is only indirectly heated in the oven
cavity, and therefore energy is not wasted. Currently the largest technical
problem in the use of microwave ovens to produce foundry cores lies in the
development of suitable materials for core boxes and drying racks. Ideally,
such materials should be readily formed, transparent to microwave energy,
solvent-resistant, and resistant to temperatures up to approximately 300°C.
(20,24) Many materials, including aluminum, transite, and plastics of several
types, have been tried, but at this time the optimum material has not been
developed. Once this problem has been overcome, it is anticipated that many
foundries will consider conversion and subsequently may reduce their core
room emissions.
Pouring Emissions
Finally, it is important to note that the emissions from the core material
during the pouring are only a small part of the total pouring emissions. The
core is normally surrounded by metal, such that the effective sand-to-metal
ratio is very low and the pyrolysis of the core binder is nearly complete.
Table V shows the results of the evaluation of organic core materials on
-373-
-------
Table IV
Average Oven Emissions From
Baking of Starch Bonded Molds
Carbon Monoxide
Total Hydrocarbons*
Benzene
Toluene
M-Xylene
O-Xylene
Naphthalene
Formaldehyde .
Total Aldehydes*
Propanol +
Ethanol/Methanol
Microwave Oven
(ppm)
0.05
0.01
0.08
<0.01
<0.01
1.04
<0.5
0.05
0.31
<25
35
(0.02)
(<0.0
(0.02)
(0.8)
(0.02)
(0.23)
Electric Re-
sistance Oven
(ppm)
<25
18
0.05 (0.02)
0.09 (0.02)
0.09 (0.02)
<0.01
<0.01
0.60
1.5
0.11
0.18
(0.5)
(0.04)
(0.14)
* As methane
f As acetaldehyde
+ As ethanol
-374-
-------
Figure 3. COMPARATIVE CASTINGS. Core for casting on left made with
Spirit binder; Core on right made with core oil.
-375-
-------
Table V
Results
Green Sand - Cored
Max.
I
LO
Core Sand
Green
Green
Fur an
Furan
Alkyd
Alkyd
sand
sand
- TSA
- TSA
isocyanate
isocyanate
Phenolic urethane
Phenolic urethane
Shell
Shell
Max.
(ppm)
2400
2400
1930
1770
1850
2070
2220
2500
2010
1690
CO
(Time)
(3 min)
(3 min)
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
(5 min)
(3 min)
Max. CO 2
(ppm) (Time)
18,100
11,400
7,990
6,750
7,750
5,540
11,500
10,880
9,720
9,890
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
Max.
(ppm)
680
540
380
430
320
460
240
310
330
330
CH,,
(Time)
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
(1 min)
(4 min)
(5 min)
(4 min)
HCT
(ppm)
980
1460
910
840
710
860
510
610
850
570
(Time)
CH,,
(2 min)
(6 min)
(3 min)
(3 min)
(10 min)
(1 min)
(20 min)
(2 min)
(2 min)
(7 min)
NH3
ug/1
1.0
<0.20
0.48
<0.20
0.32
0.45
0.21
0.17
<0.20
0.20
HCN
wg/i
0.63
0.91
0.64
0.85
<0.21
2.70
<0.21
<0.21
:
Phenol formaldehyde
1850 (5 min)
7,800 (1 min)
310 (5 min)
550 (5 min)
0.30
-------
overall emissions during the pouring of a simple cylindrical casting, such
as illustrated in Figure 1. These results clearly indicate that the molding
sand decomposition (as for green sand) predominates emissions produced from
thermal decomposition of the core. These emissions may also become fugitive
emissions, (25) so it is important to bear in mind that whereas a new core
binder, such as the carbohydrate-based spirit binder, may reduce core room
emissions, it may have little effect in the pouring area. The relative chemi-
cal simplicity of this binder indicates that a reduction in emissions in pour-
ing and shakeout areas might be possible if the molds were bonded with it.
Further development along these lines is expected. These efforts are expected
to include quantification of the advantages which might be gained in reducing
pouring room and shakeout emissions if the carbohydrate binder is used for
molding.
CONCLUSIONS
Typical core room emissions include hydrocarbons ranging from methane
to the light oils as well as aromatic compounds, including benzene. Other
emissions may include carbon monoxide, carbon dioxide and other simple organic
compounds such as formaldehyde. Some processes may release amines, ammonia,
and sulfurous gases, depending upon the specific core binder.
There are essentially two sources of these emissions: the curing oven
and the sand binder. In either case, these emissions are most often directly
vented to the atmosphere. They present an undesired pollution and may have
an irritating odor.
One engineering solution to the problem of controlling the fugitive emis-
sions from the core room of the foundry may lie in binder substitution. A
new type of binder based upon starch has been identified as a possible core
oil substitute, and it promises to greatly reduce binder emissions in the
core making and baking processes.
A new method of drying these carbohydrate-bonded cores employing micro-
wave energy may also improve, or essentially eliminate, core room emissions
in the foundry.
-377-
-------
REFERENCES
1. Scott, W.D. and Feazel, C.E., "A Review of Organic Sand Binder Chemistry,"
Modern Casting Tech Report, No. 765, American Foundrymen's Society, June
1976.
2. Plummer, J.F., "Bonding Mechanisms of Inorganic Foundry Binders," address
to Am. Chemical Society's Symposium on New Developments in Foundry Binders,
Sept. 13, 1978, Miami Beach, Fla.
3. Schafer, R.J. and Toeniskoetter, R.H., "Industrial Hygiene Aspects of
the Use of Sand Binders and Additives in Molds and Cores," BCIRA Con-
ference on the Working Environment in Iron Foundries, University of War-
wick, March 1977.
4. Eftax, D.S.P. and Brechter, M., "Environmental Considerations on the
Use of No-Bake Foundry Binders," BCIRA Conference on the Working Environ-
ment in Iron Foundries, University of Warwick, March 1977.
5. Wallace, R.A., "Inorganic Binders for Mold and Core Preparation," Steel
Castings Institute of Canada, Research Report 25, March 1978, pp 127.
6. Nicholas, K.E.L., The CO,-Silicate Process in Foundries, British Cast
Iron Research Association, Birmingham, U.K., pp. 223.
7. Scott, W.D., Bates, C.E., and James, R.H., "Chemical Emissions from Foun-
dry Molds," AFS Transactions, Vol. 85, 1977, pp 203-208.
*
8. Scott, W.D., James, R.H., and Bates, C.E., "Point Source Emissions from
Foundry Molds," Proceedings of BCIRA Conference on Working Environment
in Iron Foundries, March 1977, University of Warwick, U.S.
9. Dorfmuller, A. and Schafer, R.J., "Cold Box Process-Research to Reality,"
AFS Transactions, Vol. 8.1, 1973, pp 404-409.
10. Crowley, T. and Apelbaum, J., "Microwave Ovens in the Foundry," Elec-
tronic Progress, Vol. 18, 1976, pp 8-18.
11. Tarasescu, M. and Dragoi, F., "General Consideration of the Use of Micro-
wave Energy in Foundries," Metalurgia, Vol. j!9(5) , pp 246-251.
12. Schroeder, R.E. and Hackett, W.S., "Microwave Energy in the Foundry,"
AFS Transactions, Vol. TT_, 1969, pp 141-145.
13. Data drafted for Cured Sand Committees, American Foundrymen's Society,
Des Plaines, 111., April 1976.
14. "Phenolic Resins," Foundry Health and Safety Guide, No. 12, American
Foundrymen's Society, Des Plaines, 111., April 1976.
-378-
-------
15. Wile, L.E., "Progress in Coremaking: Shell," Foundry, March 1970, pp
50-57.
16. Toeniskoetter, R.H. and Gardikes, J.J., "Chemistry of the Cold Box Core-
making Process," address to Am. Chemical Society's Symposium on New
Developments in Foundry Binders, September 13, 1978, Miami Beach, Fla.
17. Toriello, L.I. and Robins, J., "Cold Box Processes - Sand and Metal Per-
formance," AFS Transactions, Vol. 76, 1968, pp 212-221.
18. Sahlin, L.B., "Cold Box Process - Process Engineering Studies," Modern
Castings, February 1969, pp 17-22.
19. Stewart, H., "Insta-Draw Process," Tech. Bui. Core-Lube, Inc., Danville,
111., June 1977.
20. Kiesel, R.F. and Van Oene, H., "A Method for Odor Analysis of Foundry
Emissions," address to Am. Chemical Society's Symposium on New Develop-
ments in Foundry Binders, Sept. 13, 1978, Miami Beach, Fla.
21. Cummisford, B.C., "Curing Spirit with Microwave," Ibid.
22. Scott, W.D. and Bates, C.E., "Decomposition of Resin Binders and the
Relationship Between Gases Formed and the Casting Surface Quality, AFS
Transactions, Vol. 83, 1975, pp 519-524.
23. Smith, W.A., "Microwaves and Materials Science in Foundry Applications,"
address to Am. Chemical Society's Symposium on New Developments in Foun-
dry Binders, September 13, 1978, Miami Beach, Fla.
24. Cole, G.S., "The Effects of Microwave Heating on Core Processes," AFS
Transactions, Vol. 86, 1978, in print.
25. Scott, W.D. and Bates, C.E., "Measurement of Iron Foundry Fugitive Emis-
sions," to be published in Proceedings of EPA Symposium on Fugitive Emis-
sions, Hartford, Conn. May 18, 1976.
-379-
-------
OCTOBER 25. 1978
Wednesday Morning - SESSION V
Session Chairman: Henry J. Kolnsberg
Senior Project Manager
TRC - The Research Corporation
of New England
-381-
-------
FUGITIVE EMISSIONS PROBLEMS AND CONTROL AT A SURFACE COAL MINE
By: Lyle D. Randen, Environmental Engineer
AMAX Coal Company
ABSTRACT
Fugitive dust control is not an exacting science and varies from region
to region depending mainly upon climate, soils, and vegetative cover. Many
new and Innovative control techniques are infeasible due to their cost,
effect upon equipment, or other forms of environmental degradation, i.e.
water pollution. After studying various forms of fugitive dust control, I
feel that implementation of conventional control techniques are usually the
best.
' Specifically, wind erosion of barren areas can be best controlled by
application of hay mulch and/or seed depending upon the length of time an
area is to be left unmined. When using electric drive trucks, i.e. Electro-
Haul trucks, the most feasible control of pit haul road dust is through the
use of water trucks. Chemical additives oftentimes can have a detrimental
effect upon the electrical or mechanical systems of a haulage vehicle when
used over a period of time. Many other forms of fugitive dust control have
been recommended, i.e. irrigation sprinklers on roads, blasting mats, spray
nozzles on shovels, water bags in blasting holes, etc. AMAX will continue
to evaluate these types of control for application at their Wyoming mines.
However, most of the tests so far have shown new techniques not feasible
due to weather extremes, costs, effect on equipment and general unworkable
situations.
-383-
-------
Fugitive dust takes on many different meanings when defined in the eyes
of the regulator, the environmentalist, the industrial representative, the
agriculturalist and so on. Its definition as published in the "Prevention of
Significant Deterioration Regulations," developed by the United States
Environmental Protection Agency will ultimately be litigated. Regardless of
the outcome of all of the interpretations or adaptations of the definition of
fugitive dust, it will continue to cause problems for all concerned. My
remarks today are limited to those problems associated with a large surface
coal mine operating in the State of Wyoming, namely the AMAX Belle Ayr Mine.
The Belle Ayr Mine produced some 13.3 million tons of coal in 1977,
giving it the distinction of being the largest producing coal mine in the
United States. We will exceed this production level in 1978. Along with the
prestige of being the largest mine comes the requests to do environmental
impact studies on the mine, i.e., air, water, wildlife, etc., inspections
from curious and serious regulators and environmental groups, salespersons,
interested citizens and so on. Each of these groups have their own ideas
on how we should control fugitive emissions. These range from doing nothing
to enclosing the pit and putting on a baghouse to shutting down the entire
operation. The latter two, however, would not be economically feasible for
AMAX nor in the best interests of our Nation's energy goals. The AMAX Belle
Ayr Mine is regulated by U.S.G.S., BLM, EPA, OSHA, MSHA, OSM, Wyoming
Department of Environmental Quality which has an Air Quality, Land Quality,
and Water Quality Division and the list appears to be broadening during each
session of Congress. This serves only to further distort already cloudy
issues.
Fugitive emissions sources are numerous in a surface coal mine. They
include:
1. Haul road traffic - coal and overburden
2. Shovel-truck loading operations - coal and overburden
3. Topsoil removal
4. Blasting - coal and overburden
5. Truck dumping - coal and overburden
6. Open storage piles - not usually allowed for coal
7. Exposed areas including topsoil piles and grading and redistribution
of topsoil and overburden
8. Haul road construction and maintenance
9. Train loading - coal
A few of the above sources are amenable to conventional treatment and
are easily controlled. The other sources of fugitive dust are not so easily
controlled both economically and technically. Control devices or
methodologies must change with geographical locations due to climate, soil
types and conditions and differing types of equipment used to mine coal.
-384-
-------
Electric wheel motors, 100 degree temperatures, 10-20% relative humidities,
wind chill factors of -50 degrees and strong winds, make control of fugitive
dust seem impossible. Many natives of the State of Wyoming also feel it is
impractical given existing dust levels.
There are several approaches to control of fugitive dust. AMAX has
experimented with a number of these with limited success. When a new control
is suggested, before any hands-on experimentation, one must go through a
rigorous review of the control and potential problems. We review new control
practices and techniques for their economical impact, i.e., cost per ton of
coal produced, technical feasibility such as the applicability of water spray
system operating during freezing conditions, safety considerations, effect on
operating equipment, secondary environmental effects, area resource depletion
and product availability. Many of the control techniques that I criticize
may well be techniques or methods that others have found suitable. I am only
reporting findings as they apply to our mine, and those that generally would
apply to surface coal mines operating in the "Powder River Basin" of Wyoming.
Let us take each of the sources I have previously listed and discuss
fugitive dust problems and control of them:
1. Haul Road Traffic. For those of you not directly familiar with our
Belle Ayr Mine, we have two types of haul roads, i.e., overburden and coal.
They sometimes cross each other and sometimes are one and the same. The life
of a haul road can vary from several days to several years. As such, haul
roads cannot be considered permanent as can the entrance roads leading into
the mine. Most mines are being required by the Wyoming Department of
Environmental Quality, Air Quality Division, to "hard surface" all entrance
roads leading into the mine as part of their "permit-to'construct" application.
AMAX is experimenting with soil binders and a double chip and seal surface or
a two inch asphalt mat to meet these requirements. There are a number of
different fugitive dust control strategies for haul road traffic. Most
regulatory agencies recognize road watering, silt content of the soil, tire
size and vehicle speed in calculating emissions from haul roads (U.S. EPA,
1973). The U.S. EPA publication AP-42 also establishes a climatic factor for
rainfall using 0.01 inches of rain for 100% control of fugitive dust. In
Wyoming the number of days per year receiving more than 0.01 inches of rain
is 100 (U.S. EPA, 1975). Therefore, our climatic factor is 265. or rainfall
controls 100% of fugitive dust 28% of the time. Our pro- 365 blem is then
controlling fugitive dust the other 72% of the time. This could be done with
water trucks, water trucks with chemical additives, chemically treating the
roads, hard surfacing the haul road or eliminating the haul road and replacing
it with some other means of transporting the coal or overburden in a dust
free manner.
-385-
-------
My first choice would be to operate water trucks In a manner that
simulates a rainfall event on the haul roads comparable to the rainfall event
noted in AP-42. This control could also be altered by using something other
than a water truck to apply the water, i.e., an irrigation system. This
system, however, cannot be used to control fugitive dust during the freezing
conditions unless an antifreeze is used in conjunction with the water.
Unfortunately, the water and antifreeze make the road slippery and create a
severe safety hazard. Also the height of the spray necessary to cover a 100'
wide roadway, creates a mechanical nightmare with water being sprinkled almost
as high as the truck cab.
Water trucks with chemical additives have proven to be very costly. If
we use the chemical for 120 days per year we have spent a lot of money;
potentially enough to buy another water truck that can be depreciated over
several years. Other chemicals with which I am familiar cost large sums of
money at their intital applications, but are reduced in cost on subsequent
applications but require continuous maintenance. Some are corrosive, some
have a potential impact on the electric wheel motors, and others cake the
underside of the equipment prohibiting routine maintenance without high
pressure steam or gasoline cleaning treatment. The main factor in making
this analysis is that the water truck and operator are still essential to
putting on the chemical. We must also be careful that any chemical used will
not enter a water course and result in water pollution or a fish kill.
Chemical treatment and hard surfacing of haul roads can be treated as if
they were the same because the constraints that limit their use within a sur-
face coal mine are the same. Chemical treatment or hard surfacing of a haul
road is very expensive due to the depth the application would have to be made
to in order to withstand the weights of the haul trucks. An unloaded over-
burden truck weighs approximately 35 ^ons and a loaded truck will weigh in
the 145 ton range. The chemically treated road would pot hole out as would
a hard surfaced road. Spillage onto the road would eventually cover the
chemically treated surface as well as a hard surface. The road is probably
not in existence long enough to warrant this type of expenditure. Additionally
most mines use a motor grader to rip the road and then apply aggregate for
traction during winter, i.e., snow and freezing conditions.
The movement of coal and overburden in other "dust free" manners is
being explored by AMAX and other coal companies. I am talking about the use
of enclosed conveyors for long haul situations. This alternative is not only
being explored from a dust reduction standpoint, but also as a method to
achieve greater economic efficiency. However, there are still some unresolved
problems concerning conveyor system usage in AMAX.
Given the above brief explanation of money, technical shortcomings,
mechanical hazards and unworkable situations, I believe that AMAX will
-386-
-------
continue to utilize water trucks for fugitive dust control on haul roads.
We will work with vehicle speeds and other readily available, proven techniques
to control fugitive dust. We will also continue to explore the use of
chemicals in water trucks, chemical treatment of roads and hard surfacing
roads, but probably would not expend any large sums of money on implementation
unless frequent violations of the ambient air quality standards for the State
of Wyoming are noticed. We will continue to look at ways to increase water
truck efficiency. This includes the installation of a surge tank at Belle
Ayr for reduced cycle times and larger water truck tanks. In areas where
water is not so abundant the use of chemicals may be quite advantageous in
controlling fugitive dust. My discussion related only to our situations.
2. Shovel-Truck Loading Operations. Both overburden and coal are re-
moved utilizing a truck-shovel operation at the AMAX Wyoming mines. Typical
shovel bucket sizes are 24 cubic yards on the overburden shovels and 40 cubic
yards on the coal shovels. Fugitive dust control on the shovels is a potential
problem. A great deal of the time the material loaded is moist enough or
frozen to the point that it is not a problem. Problems occur during the hot,
dry, windy days of the summer months. I have worked with the idea of a spray
system mounted on the shovel but feel that a potential safety hazard exists
with water being sprayed around a shovel that utilizes a very high voltage
line for power, i.e., 4160 volts. The spray system would not penetrate the
coal or overburden surface enough to adequately control the dust. Other
potential problems include alteration of the heating value of the coal by
adding too much moisture, which is probably insignificant, coal sticking in
the trucks and loadout silos, overburden sticking to the truck beds,
limitation of the shovel operators' visibility, and overwatering of the load-
ing area making it slick for truck and sho.vel operation. Presently, during
dusty conditions, our water trucks will back up above the shovel and shoot
water onto the bench of coal or overburden as a short-term mitigating measure.
3. Topsoil removal. Surface mine operations are required to save top-
soil for later use. Topsoil removal is accomplished with pan scrapers. Dust
eminates from working areas, haul roads and storage piles. To eliminate
this fugitive dust problem and avoid potential soils contamination by using
chemicals, we require all topsoil removal operations to be accompanied by a
water truck. If the water truck breaks down during the topsoil removal
operation and conditions are dusty, the operation will be shutdown until the
water truck is back in service. All of AMAX topsoil removal operations at
Belle Ayr are on a contract basis. This requires the water truck or trucks
be included in the bid. We are presently experimenting with pre-wetting areas
to be stripped of topsoil with an irrigation system and staged stripping of
topsoil to take advantage of spring rains.
4. Blasting - Coal and Overburden. AMAX normally will blast 75% of the
overburden and 100% of the coal removed. This is done to increase shovel
-387-
-------
efficiency. I have helped with experiments incorporating the use of water
sprayed over the area to be shot, water bags put down the hole on top of the
explosive, differing types of energizing systems to actuate the shot (nonel,
prima cord) with no appreciable reduction in dust levels from the shots.
Some people have suggested the use of blasting mats to control dust. Blasting
mats were developed to control flyrock, not dust. With all of the safety
(MSHA) regulations, OSM regulations and other common sense controls, I am of
the opinion that there is very little or nothing we can do to control fugitive
dust from blasting operations. Maybe blasting operations should not be
considered as a problem as it is short-term, has limited control possibilities,
and most of the dust settles in a short distance from the shot.
5. Truck Dumping - Coal and Overburden. Truck dumping of coal into a
hopper can be easily controlled. There are two approaches: a baghouse
system drawing a negative air pressure on the hopper, or using a spray bar
system creating a fine water mist over and in the truck dump. Such systems
are commercially available and as such do not present an unsolveable problem.
Overburden truck dumping is a source of fugitive dust and I am not familiar
with any types of control other than reducing drop heights from the trucks,
limiting the heights of piles within the pit to keep it out of the wind,
orienting dump areas at right angles to prevailing winds, and watering turn
around and backup areas next to the overburden dump.
6. Open Storage Piles. Open coal storage piles are usually not
allowed by the State of Wyoming Air Quality Division. Open storage of topsoll
is controlled in the same way that we control exposed areas as discussed be-
low. There are a number of chemical binders, mechanical methods and other
control strategies that can be implemented to control fugitive dust from coal
storage piles. As I do not have any actual experience with these, I will not
discuss their application.
7. Exposed Areas Including Topsoil Piles and Grading and Redistribution
of Topsoil and Overburden. I have developed a wind and water erosion control
strategy for the AMAX Belle Ayr Mine as per the OSM requirements. Briefly,
AMAX relies upon SCS established practices for controlling wind erosion of
exposed areas. Included within controlling fugitive dust from exposed areas
is limiting vehicular traffic to established roads or staking off non-travel
areas. The area of soil exposed and length of time that it is exposed should
be kept to a minimum. There are several practices that can be utilized to
control fugitive dust from exposed areas. Usually, the lowest cost practices
are the most effective in controlling erosion. Straw mulch or hay mulch at
an application rate of two tons per acre can control up to 90% of the wind
erosion. The mulch should be anchored to be effective in controlling erosion.
-388-
-------
Practices to reduce fugitive dust;
Practice
1. Disturbed Area Stabilization
(with mulching only)
2. Disturbed Area Stabilization
(temporary seeding potential
fertilization)
3. Disturbed Area Stabilization
(permanent seeding potential
fertilization)
Condition Where Applicable
On areas to be bare less than 6
months or where seedings cannot
be made.
Areas which would remain bare for
one year or less before
permanent grading and seeding.
On bare areas where permanent
vegetation is needed.
Fugitive dust from exposed areas can be effectively controlled if mulch-
ing, seeding and fertilization operations are conducted as soon as possible
after the disturbance. Of course, final and permanent reclamation is the
ultimate answer to controlling fugitive dust from exposed areas.
8. Haul Road Construction and Maintenance. Haul road construction is
usually done by backfill operations utilizing dozers and front end loaders
while maintenance is done by motor graders. The only realistic control again
dictates the use of water trucks to control fugitive dust. Please note
discussion of chemical usage in Part I, Haul Road Traffic.
9. Train Loading - Coal. Actual monitoring tests have shown this source
to be of little significance when left uncontrolled. Coal is loaded by
gravity flow from a 4 silo arrangement into a train driven underneath. The
moisture content of the stored coal probably reduces emissions more than any-
thing. There are spray bar systems with or without chemical additives or
binders available for use. Our experiments have shown to date that these
systems are not necessary. Future tests may require further scrutinization
of these controls.
Other sources of fugitive dust contained within the pit of a surface
coal mine produce insignificant emissions when compared to those discussed
already. I have limited the scope of this paper to those fugitive dust pro-
blem areas and potential control techniques that I am familiar with. I am
sure that there are a number of other control techniques that merit
consideration and experimentation by the AMAX Coal Company. AMAX is dedicated
to continued evaluation of these types of new and innovative techniques for
application at the Belle Ayr Mine. We have found by actual monitoring that
county gravel roads probably produce as much if not more fugitive dust than
the mines. That is why we are hard surfacing light duty access road utilized
-389-
-------
by our employees going to and from work.
A lot of the issues concerning "fugitive emissions problems and control"
will eventually be resolved by the courts. We must not lose sight of the
Nation's energy goals, and the fact that the cost of controlling fugitive
emissions will ultimately be borne by the consumer in the form of increased
electrical rates and product values. Fugitive emissions control should not
neglect public health considerations as related to particle size distribution.
It is my contention that a large percentage of fugitive dust falls out of the
air before it ever hits the pit boundaries, let alone before it reaches
property lines. Property line concentrations should be regulated levels, not
pit concentrations. Pit concentrations are already being monitored by the use
of dust pumps as required by MSHA and compliance with these standards is
ongoing at Belle Ayr. These tests relate to the respirable fraction of the
dust.
We need to take a realistic approach to fugitive emissions control. New
technology must be given time to be proven both technically and economically.
Thank you for the opportunity to speak at this symposium. I hope you find
my remarks enlightening and challenging for without challenge there would be
no change.
-390-
-------
AIRCRAFT TURBINE ENGINE PARTICULATE
EMISSION CHARACTERIZATION
John D. Stockham, Erdmann H. Luebcke
IIT Research Institute
Chicago, Illinois, U.S.A.
and
Larry Taubenkibel
DOT - Federal Aviation Administration
Washington, D.C., U.S.A.
-391-
-------
ABSTRACT
A particulate matter sampler is described along with operational expe-
rience and the data obtained using the sampler on the JT8D and JT9D aircraft
turbine engines. The sampler acquires a sample of the exhaust particles
emitted during the ground operation of the engine and permits the characteri-
zation of the mass concentration, elemental composition, particle size dis-
tribution, and particle morphology. Mass concentration is obtained
gravimetrically. Elemental composition is obtained by energy dispersive X-ray
and particle size and shape by electron microscopy and image analysis. Par-
ticle size data is also obtained with the Electrical Aerosol Analyzer.
The sample is extracted from the exit plane of the engine. The sampling
probe geometry corresponds to the FAA's recommended methodology for gaseous
emission sampling. The extracted sample is conditioned through dilution to
minimize local condensation, particle deposition, and sample bias before col-
lection on substrates suitable for analysis.
-392-
-------
THE PHYSICAL AND CHEMICAL CHARACTERIZATION of particles emitted by air-
craft turbine engines has not received comprehensive investigation. There-
fore, a sampling system was designed and constructed to collect particles
from aircraft turbine engines (1, 2).* The sampler is specifically designed
to collect representative samples of particulate matter from the exhaust at
the exit plane of low bypass ratio, mixed flow, and high byp'ass ratio aircraft
turbofan engines operating in an engine test cell. These engine types are
represented in the current commercial fleet by the JT3D, JT8D, and JT9D
series of engines. The sampler uses a mixing type sampling rake, a special-
ized aerosol transport and conditioning system,'and sample collection near
the engine to minimize particle deposition, re-entrainment, and agglomeration
during transport. A feature of the sampler is the bifurcation of the ex-
tracted exhaust sample - a large portion for bulk analysis and a smaller
portion for detailed particle characterization.
SAMPLER DESCRIPTION
A flow diagram of the sampler is shown in Figure 1. A "diamond" shaped
sampling rake mounts directly behind the engine exhaust plane. An extraction
nozzle is located at the midpoint of each leg of the diamond. The four
nozzles are evenly spaced on an arc whose radius is 62% of the engine's exit
plane radius. The leading edge of the nozzle extends about ten diameters
upstream of the support structure. The nozzle diameters are 0.170 ± 0.008 cm.
The geometry of the rake corresponds to that recommended by the Federal
Aviation Administration (FAA) for gaseous emission sampling and has been
shown to give comparable results to a 24 sampling nozzle rake system (3).
The four sampling nozzles are manifolded to provide a composite sample. The
manifold is connected to the primary diluter via a flexible coupling that
accommodates movement of the engine during testing.
The function of the primary diluter is to condition the extracted partic-
ulate matter sample. The sample is cooled without passing through a dew
point and diluted to reduce particle agglomeration effects. The unique
design of the diluter reduces wall losses and aids in the preservation of
particle mass and particle size distribution (4, 5). The diluter consists of
a porous tube housed within a larger tube. Clean, dry, metered, compressed
air is delivered to the annular space at four locations and flows through the
porous inner tube providing a clean boundary layer that reduces particle
deposition and subsequently mixes with the sample to provide dilution and
temperature control.
^Numbers in parentheses designate References at end of paper.
-393-
-------
After conditioning, the sample is bifurcated. The majority of the
sample is transported to the mass sampling turret where the particulate
matter is removed by filtration onto tared Gelman, type AE, glass fiber
filters, 142 mm in diameter. A smaller portion of the sample is adjusted in
concentration by a secondary diluter similar in design to the primary diluter
and transported either to the Electrical Aerosol Analyzer (EAA) for real-time
particle size distribution measurement or for collection onto membrane filters
and electron microscope grids for subsequent analysis of particle size, shape,
and elemental composition. The membrane filters used are 0.03 pm pore size
Nuclepore filters, 47 mm diameter.
With the exception of the EAA, the sampler is designed for operation in
close proximity to an operating turbine engine. All functions of the sampler,
therefore, are remotely controlled from a centralized control panel that can
be located up to 25 feet from the filtration units in a safe location such as
the test cell control room. The EAA continuously malfunctioned in the test
cell environment despite efforts to provide a suitable sound protection enclo-
sure. As a result, the EAA is operated in the test cell control room.
Flow control valves for the sampler are located downstream of the parti-
cle collection sites to eliminate particle deposition in the valve assembly.
This requirement imposed operational constraints in the system—each sample
has to be acquired sequentially and flow switching procedures must be care-
fully performed to prevent rupture of the filters. Also, because the ex-
tracted sample flow from the engine is not measured directly, the primary
diluter flow and the total sample flow have to be accurately determined. The
dilution flow rates are measured by precision drilled orifices; total flow is
measured with a hot-wire anemometer. The uncertainty in total flow measure-
ment is ±0.09 std. m3/hr. while the dilution flow is repeatable to approxi-
mately ±0.1 std. m3/hr. The maximum error in the extracted sample flow
measurement occurs at idle power and is on the order of 10%.
The turret assemblies for sample collection provide positions for six
filters. Samples, thus, can be acquired at five engine power settings —
idle, approach, cruise, climb-out, and take-off — before the engine needs to
be shut down and the filters recovered. The sixth turret position is a dummy
used during start-up, shut-down, and while power settings are being changed
and engine operation stabilized.
The sampler is designed for isokinetic sampling of the exhaust from the
JT8D and JT9D engines at power settings from idle through cruise. At climb-
out and take-off power, sonic conditions are approached at the engine's exit
plane and sampling at this rate would cause severe flow complications re-
sulting in modification of the extracted exhaust particles. The flow through
the extraction nozzles was, therefore, limited to a local Mach number of 0.8.
Extracted sample flow ranged from 0.9 std. m3/hr. to 4.8 std. m3/hr. for
engine power settings from idle to take-off (6).
-394-
-------
TEST SITE AND DATA ANALYSIS PROCEDURES
The particulate matter sampler was operated at United Airlines' San
Francisco, Ca., Maintenance Facility on JT8D and JT9D aircraft turbine en-
gines. Figure 2 shows the sampler set-up in a JT9D test cell. Due to Unit-
ed1 8 fleet needs, sampling time was at a premium and several different engines
were used during the test series. The engines tested are given in Tables 1
and 2 along with data on the mass emission rate and the particle size distri-
bution of the exhaust particles as determined by electron microscopy and image
analysis. In addition, particle sized distribution measurements were made
with the EAA and particle shape was determined by image analysis. Two fuels
were used during the test series, pearl kerosene (PK), a high hydrogen, low
aromatic content fuel, and JET A, a standard commercial aircraft turbine
engine fuel with a lower hydrogen and higher aromatic content. The use of two
fuels permits the determination of the effect of fuel composition on emis-
sions. A single batch of PK fuel was purchased and stored at United's facil-
ity. Its composition, therefore, was constant throughout the tests and
repeatability measurements were possible. JET A fuel composition varied
during the tests. All fuels were sampled and analyzed; significant data is
given in Table 3.
Mass emission rates were determined by differential gravimetric analysis.
A Gelman, type AE, glass fiber filter was used to collect the particulate
matter for mass determination. The filters were weighed on an analytical
balance before and after sample collection and corrected for shifts in weight
due to hygroscopic effects with control filters.
Particle size data was obtained by two methods. One method used, the
Electrical Aerosol Analyzer (EAA), provided a real-time measure of the parti-
cle size distribution. The EAA used was Model 3030, Thermo-Systerns, Inc., St.
Paul, Mn. The instrument was operated in accordance with the operating and
service manual provided with the instrument.
The second procedure used for particle size analysis was electron micro-
scopy and image analysis. The samples, collected on Nuclepore membrane fil-
ters or carbon coated electron microscope grids attached to the membrane
filter, were used in the analysis. The electron microscope used was the JEOL-
100C analytical scanning/transmission microscope equipped with an energy
dispersive X-ray system for elemental analysis. After applying a thin film of
gold to the filter surface with a Hummer II plasma specimen coater, the sam-
ples were observed and photographed in the scanning microscope at several
magnifications ranging from 10,000 to 100.000X. A magnification of 30,OOOX
was found to be optimal for particle size and shape analysis. Particle size
and shape analysis were performed by interrogating the electronphotomicro-
graphs. with an Imanco Quantimet 720 Image Analyzer. Particle size was com-
puted from the area measurement of each particle. The particle size reported,
thus, represents the diameter of a circle of area equal to that of the parti-
cle.
-395-
-------
Particle shape was determined using the image analyzer and interrogating
the particles for their area and perimeter. A two-dimensional shape factor
was then calculated from the ratio A/P2. This ratio was plotted as a function
of the equivalent area particle diameter.
Elemental composition of the particles was determined while the particles
were under observation in the electron microscope. A Kevex energy dispersive
X-ray spectrometer coupled with the Tracor/Northern 880 X-ray analysis system
was used.
RESULTS AND DISCUSSION
MASS EMISSIONS — Mass emission data are reported in Tables 1 and 2 and
plotted for the JT8D engine as a function of primary air flow in Figure 3.
The data show an increasing emission rate with power. The emission rate
ranged from 0.15 kg/hr at idle to a range of 0.25 to 4.62 kg/hr at climb-out
and 0.76 to 1.79 kg/hr at take-off. Data for the JT9D engine show a similar
trend and the same order of magnitude. However, primary air flows for the
JT9D engine are significantly greater. Primary air flows as a function of
engine power is given in Table 4. With the restricted sampling time avail-
able, 12-30 minutes depending on the operating mode, and the low emission
rates for the JT8D and JT9D engines, the gain in filter weights were only a
few milligrams. Additional testing with longer sampling times are needed for
a more accurate determination of mass emission rates. Earlier data (2) col-
lected from a TF-30 engine suggest the sampler can reproduce mass concentra-
tion data within ±2% at the cruise power setting for this engine. Effects of
engine and fuel variability are discernible, however. The JT8D-7 engine,
648735, gave lower emission rates than the other engines and the JT8D-15
engine tested with both JET A and PK fuel gave a lower emission rate with the
PK fuel.
PARTICLE SIZE DATA — Particle size data obtained by electron microscopy
and image analysis are reported in Tables 1 and 2 and summarized in Tables 5
and 6. For the JT8D_engines, operating on JET A fuel, the average geometric
mean particle size, dg increased with engine power. At idle, dg was 0.045 urn,
at approach dg was 0.068 ym, at cruise and climb-out dg was 0.078 ym, and at
take-off dg was 0.084 ym. Particle size is influenced by fuel type; PK fuel
gave smaller particle sizes. This is consistent with the lower mass emission
rate found for the engine when operated on PK fuel. Also consistent is the
performance of JT8D engine 648735. This engine, which produced the lowest
emission rates, also produced the smallest particle sizes. The geometric
standard deviations for the particle size distributions ranged from 1.6 to
2.3. Engine 648735 gave the widest distribution of particle sizes. The JT9D
engines gave similar trends to the JT8D engines, but indications are that the
particles are smaller.
-396-
-------
Particle size distribution data were also obtained with the EAA. Typical
data for the JT8D and JT9D engines are plotted in Figures 4 and 5. These data
support the results obtained by the electron microscope and image analysis
measurements. Higher power settings produce a greater concentration of parti-
cles in all sizes ranges above about 0.02 ym. Below 0.02 \mt data is con-
flicting but the performance of the EAA is known to be erratic in this
range (7). Particle concentration for the JT8D engine is about one order of
magnitude greater than for the JT9D engine. The influence of fuel composition
on particle size is illustrated in Figures 6 and 7 for the JT9D engine at ap-
proach and take-off power. At all power settings, PK fuel produced few parti-
cles.
PARTICLE SHAPE — Particle shape data shows the particle structure is
more complex at climb-out and take-off than at the lower power settings.
Figures 8 through 11 illustrate this effect. The effect is evident from the
photographs shown in Figures 12 and 13. As an aid in interpreting the shape
factor:, A/P2, Table 7 is presented. The shape factor for a disc, the two-
dimensional representation of a sphere, is 1/4 ir or 0.0795. A 5-chain agglom-
erate has a shape factor of 0.0159; a 10-chain, 0.0078; and a 50-chain,
0.00016. Thus, as the shape factor decreases the particles are less spherical
and more highly agglomerated. Generally, particles with shape factors above
about 0.03 can be considered single particles trending toward a spherical
shape- As shown in Figures 9 and 11, the very large exhaust particles have
small shape factors and are, thus, highly agglomerated. Idle power produces
relatively small particles of simple shape while climb-out and take-off power
produces a wider range of particle sizes of more complex shape.
ELEMENTAL COMPOSITION — The elemental composition of the individual
exhaust particles was determined by energy dispersive X-ray techniques. No
spectrum was obtained indicating the particle composition is essentially
carbon.
CONCLUSIONS
The particulate matter emitted during the test cell operation of the JT8D
and JT9D aircraft turbofan engines was characterized using a sampler designed
expressly for this purpose. The particles are principally carbon and less
than 0.1 ym in size. Particle size increases with engine power. At idle, the
average geometric mean particle size was 0.045 ym and at take-off, 0.097 ym
for the JT8D engine operating on JET A fuel. Particle size is influenced by
the fuel composition. PK fuel with its higher hydrogen, lower aromatic, and
higher smoke point content gave smaller exhaust particles. Particle shapes
are tnore uniform and trend toward sphericity at idle; at climb-out and take-
off the particle structures are more varied and complex. The mass emission
rate for the JT8D engine was 0.15 kg/hr at idle and ranged from 0.76 to 1.79
lcg/hr at take-off. PK fuel produced lower emissions than JET A fuel. Vari-
ability among the engines tested was noted. The limited data available indi-
cate the JT9D produces less particulate matter than the JT8D. Particle
concentration data show approximately an order of magnitude difference between
the JT8D and JT9D engines.
-397-
-------
The results indicate that the sampler provides a valid sample of the
particles present at the exhaust plane of aircraft turbine engines.
ACKNOWLEDGEMENT
The work reported was sponsored by the United States Department of
Transportation -Federal Aviation Administration under contract DOT-FA75-WA-
3722. The assistance of Messrs. Paul Campbell and Ralph Johnson of United
Airlines and Dr. Donald Fenton of New Mexico State University is gratefully
acknowledged.
REFERENCES
1. D.L. Fenton, "Turbine Engine Particulate Emission Characterization,"
Phase I Report to Federal Aviation Administration under Contract DOT-
FA? 5-WA-3722, Report No. FAA-RD-76-141, September 1976, National Tech-
nical Information Service, Springfield, Va. 22161.
2. D.L. Fenton, E.W. Nordstrom, E.H. Luebcke, "Turbine Engine Particulate
Emission Characterization," Phase II Report to Federal Aviation Admin-
istration under Contract DOT-FA75-WA-3722, Report No. FAA-RD-77-165,
October 1977, National Technical Information Service, Springfield, Va.
22161.
3. A.J. Fiorentino, et al., "Evaluation of Federal Aviation Administration
Engine Exhaust Sampling Rake," Final Report, PWA-5534, United Technolo-
gies Corp., East Hartford, Conn., NASA Contract No. NAS3-19447, June
1977.
4. M.B. Ranade, "Sampling Interface for Quantitative Transport of Aerosol,"
Report to Environmental Protection Agency under Contract 68-02-0597,
December 1973, Environmental Protection Agency, Research Triangle Park,
N.C. 27711.
5. M.B. Ranade, "Sampling Interface for Quantitative Transport of Aerosol,"
Report to Environmental Protection Agency under Contract 68-02-1295, July
1976, Environmental Protection Agency, Research Triangle Park, N.C.
27711.
6. D.L. Fenton, E.H. Luebcke, E. Nordstrom, "Physical Characterization of
Particulate Material from a Turbine Engine," Department of Mechanical
Engineering, New Mexico State University, Las Cruces, New Mexico. Paper
to be presented at the ASME Gas Turbine Conference and Product Show, San
Diego, Ca., March 12-15, 1979.
7. G.J. Sem, "Design and Application of an Electrical Size Analyzer for
Submicron Aerosol Particles," Thermo-Systems, Inc., 2500 N. Cleveland
Ave., St. Paul, Mn. 55113. Presented at the 21st Annual Instrument
Society of America Analysis Instrumentation Symposium, Philadelphia, Pa.,
May 6-8, 1975.
-398-
-------
Table 1 — Summary of Turbine Engine Sampling Data - JT8D
Engine
Number
Fuel
Operating
Mode
Idle
Approach
Cruise
Climb-out
Take-off
Idle
Approach
Cruise
Climb-out
Take-off
Idle
Approach
Cruise
Climb-out
Take-off
Idle
Approach
Cruise
Climb-out
Take-off
Idle
Approach
Cruise
Climb -out
Take-off
Primary
Air
(kg/sec)
16
33
50
61
68
17
33
SO
62
69
17
33
49
61
66
17
33
50
62
69
17
33
50
62
69
Particle
Emission
(kg/hr)
0.15
1.00
1.78
2.97
0.95
__
0.30
1.01.
4.62
—
__
0.34
0.25
0.25
0.76
__
0.24
1.12
0.74
1.38
__
0.08
0.24
1.27
1.79
Particle Size
Distribution1
(dg, um)2
0.052
0.086
0.081
0.096
0.099
0.042
0.067
0.092
0.088
0.097
0.043
0.054
0.069
0.049
0.060
0.042
0.066
0.071
0.079
0.080
0.032
0.043
0.079
0.075
0.077
(Og)3
1.7
1.7
1.7
1.6
1.7
1.7
1.8
1.8
1.9
2.0
1.6
2.1
2.0
2.3
2.2
1.7
1.7
1.9
1.6
1.6
1.7
1.8
1.9
1.8
1.9
1/10/78 JT8D-7 654957 JET A
4/24/78 JT8D-15 648796 JET A
4/25/78 JT8D-7 648735 JET A
6/20/78 JTfln-15 696572 JET A
4/24/78 JT8D-15 648796 PK
'Obtained by electron microscopy and image analysis
2Geometric mean particle diameter
'Geometric standard deviation
Table 2 — Summary of Turbine Engine Sampling Data - JT9D-3A
Sample
Date
1/9/78
4/21/78
6/21/78
6/22/78
4/21/78
Engine
Number
662734
663031
663082
662794
663031
Fuel
JET A
JET A
JET A
JET A
PK
Operating
Mode
Climb-out
Take-off
Approach
Take-off
Idle
Approach
Cruise
Climb-out
Cruise
Climb-out
Approach
Take-off
Primary
Air
(kg/sec)
99
111
62
111
30
62
82
99
82
99
62
111
Particle
Emission
(kg/hr)
_
0.60
2.87
0.32
0.22
1.33
Particle Size
Distribution1
(dR. urn)2
0.084
0.117
0.054
0.077
0.041
0.047
0.040
0.041
0.045
0.054
0.050
(Qg)}
1.6
1.6
1.8
l.S
1.5
1.7
1.8
1.7
1.6
1.6
1.9
'Obtained by electron microscopy and image analysis
'Geometric mean particle diameter
'Geometric standard deviation
-399'
-------
Table 3 — Significant Fuel Analytical Data
Analysis JET A PK
Hydrogen Content, % 13.05-13.65 14.25
Aromatic Content, % 17-20 1
Naphthalene Content, % 1.6-2.5 0.1
Olefins Content, % 1-2 1
Smoke Point, mm. 20-21 35
Table 4 — Primary Air Flow as a Function of
Engine Power
Primary Air Flow (kg/sec)
Power Setting JT8D JT9D
Idle 16 30
Approach 33 62
Cruise 50 82
Climb-out 62 99
Take-off 69 111
Table 5 — Particle Size of JT8D Emission as Obtained by
Electron Microscopy
Effect of Fuel on
Geometric Mean Geometric Mean
Particle Size, ym1 Particle Size, um
Operating
Mode Range Aver. JET A PK
Idle 0.042-0.052 0.045 0.042 0.032
Approach 0.054-0.086 0.068 0.067 0.043
Cruise 0.069-0.092 0.078 0.092 0.079
Climb-out 0.049-0.096 0.078 0.088 0.075
Take-off 0.060-0.099 0.084 0.097 0.077
A fuel
-400-
-------
Table 6 — Particle Size of JT9D Emission as Obtained by
Electron Microscopy
Operating
Mode
Idle
Approach
Cruise
Climb-out
Take-out
A fuel
Geometric Mean
Particle Size, van1
Range
0.047-0.054
0.040-0.084
0.077-0.117
Aver.
0.041
0.051
0.041
0.056
0.097
Effect of Fuel on
Geometric Mean
Particle Size, un
JET A
PK
0.054 0.054
0.077 0.050
Table 7 — Particle Shape Characterization
No. of Particles in
Agglomerate Chain
1
3
5
10
50
Aspect Ratio
1/d
1:1
3:1
5:1
10:1
50:1
Shape
Factor
A/P2
0.0795
0.0265
0.0159
0.0078
0.00016
-401-
-------
Monitold
/ 3/4" Flexible Coupling
Sampling
Probe Assembly
2 Stainless Stetl
l" Vac. Hose Bypass Line
Orifice
Indexer
Pressure
Regulotor
\ 30psig
Porticulote
Filter
Dryer
Thermocouple
(4) Electrical Heaters
Vorioc Controlled
•*- Compressed Air
lOOpsig
.Secondary Diluter Orifices
Pressure Regulotor
40psig
Moss- Sampling Turret
Sound Enclosure
I" Vac Hose
l_
Absolute
Filter
—a
Thermocouples
Sample Line
j To EAA
Hot-WIre
Anemometer
3-Way I"
Control Valve
Flowrneter
Roots Rotary
Vacuum Pump
I
3-Woy 3/4"
Control Valve
|/2" Motorized
Gast Vacuum Pump
Flow Dampers pi
2" Motorized
Bleed Valve
Fig. I - Schematic diagram of the engine exhaust participate sampler
-402-
-------
Flexible Coupling
ENGINE TEST CELL
ENGINE CONTROL ROOM
Fig. 2 - Typical layout for FAA - IITR1 sampler in an
engine test cell
-403-
-------
I0c
£
O
'5»
.52
UJ
0.
II IT
Data
Envelope
10 50 100
Primary Air Flow, kg/sec
Fig. 3 - Mass emissions for JT8D
JET A fuel
-404-
-------
9
3
JT8D-I5 Engint 69572
JET A Fuel
• Tokt-off
O Approoch
x Idlt
.001
Porticle Sizt,
Fig. 4 - EAA particle size data for the JT8D engine
-405-
-------
I01
I06
IOB
o»
e
z
<
I0f
I I I
i i i M i r
.001
JT90 Engine 663082
JET A Fuel
• Take-off
O Approach
x Idle
I I I I 11 ll I I I I I 11 ll I III I I
.01
O.I
1.0
Particle Size,
Fig. 5 - EAA particle size data for the JT9D engine
-406-
-------
I07
I06
I0a
10'
I01
T I
II I
••^
1}
<<
JT9D Engine Approach
O JET A Fuel
• PK Fuel
<*
\ —
t i i i i i ii il
.01 O.I
Particle Size,
.001
Fig. 6 - Effect of fuel on particle size of emissions
l.O
-407-
-------
I01
- I I I I I I 11
I0e
10?
to4
10s
I I I I II I I I I I I INI];
JT9D Engine - Take -Off
o JET A Fuel
• PK Fuel
I02
.001
i i i i i i i il i l i i i 11 il l iii
.01 0.1
Particle Size,
1.0
Fig. 7 - Effects of fuel on particle size of emissions
-408-
-------
•I
0.
<
fc
.10
.08
.06
.04
.03
.02
.008
.006
.004
.003
.002
.001
• »•
I 1 I I 1
.01 .02 .03 .04 .05 .06 .07 .08 .09 .10 .11 .12 .13 .14
Porticl* Silt,
Fig. 8 - Porticle shops foctor, JT8D engine, idle, JET A fuel.
-409-
-------
o
-IU
O8
O6
04
03
M
CL O2
N
««
£
•
E
0
i i 1 1 1 1 1 1 1 r 1 1 1 1 T 1 1 1 1 1 1 ) 1 1 1 1 1
-
• .
• •
•
- . • .
• • .
• • ' ' .
•
•
•".'•
• * •
. •
•
0 •
• •
•
I o,J-
i 008 -
* t
.OO6 |-
.005 j-
I
.OO4 |
OO3
002
.OOI
i i '
Ol .02 .03 .04 05 06 O7 08 09 10 II 12 .13 .14 .15 16 17 .18 13 2O 21 22
Particle Site, fj.ir<
Fig. 9 - Particle shape factor, JT80 engine, take off, JET A fuel
-------
.10
.06
.06
.04
.03
.02
0.
x
S.
.008
.006
.003
.004
.003
.OO2
.00
• •
t •
I I L
.Ol .02 .03 .04 .05 .06 .07 .08 .09 .10 .1! .12 .13
Particle Diameter,
Fig. 10 • Particle shape factor, JT9D engine, idle, JET A fuel
-411-
-------
Q.
X.
o
s.
I
.10
.08
.06
.04
.03
.02
.01
.008
.006
.004
.00 3
.OO2
.001
i i i i I I I I 1 1 1 1 L
.Ol .02 .03 .04 .05 .06 .07 .08 .09 .10 .11 .12 .13 .14
Porticl* Size, [Mm
Fig. II - Particle shape factor, JT9D engine, climb out, JET A
-412-
-------
Idle iru i
Fin. 12 1'articulatc Emissions, JT8D iMitaru-, .11:1 A I IK I
Idle { iinh-ou:
Ki>i. 13 Piirt ii-ulati- Emissic^ns, .ITW onciiK', .IKT A I tu 1
-413-
-------
MEASUREMENT OF
FUGITIVE DUST EMISSIONS FROM HAUL ROADS
Chatten Cowherd, Jr.
Head, Air Quality Assessment Section
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
-415-
-------
ABSTRACT
This paper presents the results of a field testing program used to
develop emission factors for dust generated by heavy-duty truck traffic on
unpaved roads. Emission tests were performed on roads traveled by haul trucks
within an open pit taconite mine and within two integrated iron and steel
plants. Both chemically treated and untreated roads were tested. The basic
field measurements consisted of fugitive dust exposure profiles and particle
size distributions using vertically oriented arrays of isokinetic samplers
such that the total passage of airborne dust downwind of the test road seg-
ment could be determined.
Test results a~e used to extend the applicability of the emission factor
equation previously enveloped for light- and medium-duty truck traffic on un-
paved roads associated with agriculture and industry. Correction terms which
appear in the equation account for the dependence of emissions on vehicle
weight, vehicle speed, and the fraction of suspendable fines (silt) in the
road surface material. Based on an expanded data set of 24 tests, including
vehicle weights up to 157 tons, the unmodified equation predicts measured
emission factors with a relative standard deviation of predictive error equal
to 26.1%. Test results suggest that inclusion of an additional correction
term involving number of wheels per vehicle would lower the predictive error.
Particle size data indicate that the fine particle (< 5 ,um) mass fraction of
the suspended particulate (< 30 ^m) emissions is approximately 35%, indepen-
dent of vehicle weight and road surface composition. Finally, limited testing
of chemical dust suppressants shows high initial control efficiency (exceed-
ing 90%) which decays with road usage.
-416-
-------
INTRODUCTION
Until recently, the national effort to control industrial sources of air
pollution has focused on emissions discharged from stacks, ducts, or flues
and carried to the point of discharge in confined flow streams. Control strat-
egies have been based on the assumption that the primary air quality impact
of industrial operations results from the discharge of air pollution from
conventional ducted sources.
However, failure to achieve the air quality improvements anticipated
from the control of ducted emissions has spurred a detailed reexamination of
the industrial air pollution problem. Evidence is mounting which indicates
that fugitive (nonducted) emissions contribute substantially to the air qual-
ity impact of industrial operations and, in certain industries, may exceed
the effects of stack emissions.
Industrial sources of fugitive particulate emissions may be divided into
two classes--process sources and open dust sources. Process sources are fully
or partially enclosed operations from which emissions escape into the work-
place environment and/or the ambient air. Examples of process sources are
crushers, sintering machines, and metallurgical furnaces. Open dust sources
include those sources, such as raw material storage piles, from which emis-
sions are generated by the forces of wind and machinery acting on exposed
aggregate materials.
The travel of heavy vehicles on unpaved roads is the category of open
dust sources to which this paper is addressed. Truck traffic on haul roads is
a major source of fugitive particulate emissions associated with the surface
mining and processing of both metallic and nonmetallic minerals. Quantifica-
tion and characterization of dust emissions from haul roads are necessary in
order to assess control needs and to develop cost-effective control measures,
as necessary.
QUANTIFICATION OF OPEN DUST SOURCES
Fugitive emissions from open dust sources are especially difficult to
characterize for the following reasons:
1. Emission rates have a high degree of temporal variability.
2. Emissions are discharged from a wide variety of source configura-
tions.
-417-
-------
3. Emissions are comprised of a wide range of particle sizes including
coarse particles which deposit immediately adjacent to the source.
The si-berne for quantification of emission factors for open dust sources must
effectively deal with these complications.
In 1972, Midwest Research Institute (MRI) (1) initiated a field testing
program to develop emission factors for four major categories of fugitive dust
sources: unpaved roads, agricultural tilling, aggregate storage piles, and
heavy construction operations. Because the emission factors were to be appli-
cable on a national basis, an analysis of the physical principles of fugitive
dust generation was performed to ascertain the parameters which would cause
emissions to vary from one location to another. These parameters were found
to be grouped into three categories:
1. Measures of source activity or energy expended (for example, the
speed and weight of a vehicle traveling on an unpaved road).
2. Properties of the material being disturbed (for example, the content
of silt in the surface material on an unpaved road).
3. Climatic parameters (for example, number of precipitation-free days
per year on which emissions tend to be at a maximum).
By constructing the emission factors as mathematical equations with multipli-
cative correction terms, the factors became applicable to a range of source
conditions limited only by the extent of the program of experimental verifica-
tion.
The use of the silt content as a measure of the dust generation potential
of a material acted on by the forces of wind and/or machinery was an important
step in extending the applicability of the emission factor equations to the
wide variety of aggregate materials of industrial importance. The upper size
limit of silt particles (75 ^m in diameter) is the smallest particle size for
which size analysis by dry sieving is practical; and this particle size is
also a reasonable upper limit for particulates which can become airborne.
Analysis of atmospheric samples of fugitive dust indicates a consistency in
size distribution so that particles in specific size ranges exhibit fairly
constant mass ratios (1,2).
In order to quantify source-specific emission factors, MRI developed the
"exposure profiling" technique (3), utilizing the isokinetic profiling concept
which is the basis for conventional source testing. Exposure profiling con-
sists of the direct measurement of the passage of airborne pollutant immedi-
ately downwind of the source by means of simultaneous multipoint sampling over
the effective cross section of the fugitive emissions plume. This technique
uses a mass-balance calculation scheme similar to Environmental Protection
-418-
-------
Agency (EPA) Method 5 stack testing rather than requiring indirect calculation
through the application of a generalized atmospheric dispersion model.
Specifically, the exposure profiling method was used to develop emission
factors for: (a) light-duty vehicular traffic on unpaved (dirt and gravel)
roads (1); (b) agricultural tilling utilizing a one-way disk plow and a sweep-
type plow (1); (c) load-out of crushed limestone utilizing a 1.75 cu yard
loader (1); and (d) vehicular traffic on paved urban roadways (2). In order
to extend the applicability of the emissions factor equations to sources in-
volving larger scale materials handling equipment, the following sources were
tested at two integrated iron and steel plants (Test Series A and E) (4):
(a) light-duty vehicular traffic on industrial unpaved roads; (b) heavy-duty
vehicular traffic on unpaved roads; (c) mixed vehicular traffic on industrial
paved roads; (d) mobile stacking of lump iron ore; (e) mobile stacking of pel-
letized iron ore; and (f) loadout of processed slag into a truck with a front-
end loader. All sources were tested under dry road conditions (i.e., daytime
periods at least 3 days subsequent to a precipitation occurrence) so that
worst case emissions could be determined and used as a starting point for
projecting annual emissions.
EXPANDED TESTING OF UNPAVED ROAD DUST EMISSIONS
Figure 1 presents the emission factor equation previously developed for
dust entrainment by vehicles traveling on unpaved roads. Although this equa-
tion was found to have high predictive reliability for dry road conditions
(relative standard deviation of prediction error equal to 11.2% of the mean
measured value), the data base was limited to 11 tests. Moreover, only two
tests had been conducted for vehicle weights representative of haul trucks
used in mining and processing of minerals.
Therefore, in order to strengthen the data base for the predictive emis-
sion factor equation, 16 additional tests of uncontrolled emissions from un-
paved roads were performed during the summer of 1978. Seven of these tests
(Series I) were conducted at an open pit taconite mine and the remaining
tests (Series F and G) at two integrated iron and steel plants. In addition,
five tests of traffic on roads treated with chemical dusts were performed.
As in the previous tests, the primary tool for measuring fugitive dust
generated from vehicular traffic on unpaved roads was the MRI exposure pro-
filer. A vertical line grid of samplers (see Figure 2) was used for measure-
ment of dust emissions. At all times the MRI exposure profiler was posi-
tioned within 5 m of the downwind edge of the test road with air samplers
covering the effective cross section of the fugitive dust plume.
Other equipment utilized in the testing included: (a) cascade impactors
with cyclone preseparators for particle sizing, (b) high-volume air samplers
for determining upwind particulate concentrations, and (c) recording wind
-419-
-------
EF=5.9
(IT) W(T) (30?) |b/veh-mi
Determined by profiling
of emissions from light-
duty vehicles on gravel
and dirt roads under
dry conditions.
Estimated factor to
account for mitigating
effects of precipitation
over period of one
year.
Determined by profiling of emissions from
medium- and heavy-duty vehicles on gravel
and dirt roads under dry conditions.
where: EF = suspended particulate emissions (Ib/veh-mi)
s = silt content of road surface material (%)
S = average vehicle speed (mph)
W = average vehicle weight (tons)
d = dry days per year
Figure 1. Predictive emission factor equation for vehicular
traffic on unpaved roads•
-420-
-------
Figure 2. MRI exposure profiler.
-421-
-------
instruments utilized to determine mean wind speed and direction for adjusting
the MRI exposure profiler to isokinetic sampling conditions. A detailed de-
scription of the testing methodology is provided elsewhere (4).
In order to determine the properties of road surface aggregate being
disturbed by the action of moving vehicles, representative samples of the
surface materials were obtained for analysis in the laboratory. Unpaved road
test segments were sampled by removing loose material (by means of broom
sweeping) from lateral strips of road surface extending across the traveled
portion. Moisture contents of samples were determined in the laboratory by
weight loss after oven drying at 110°C, and texture was determined by stan-
dard dry sieving techniques.
In addition to the silt content of the road surface material, the emis-
sion factor equation (Figure 1) requires data on vehicle speed and weight,
averaged over the vehicle passes (approximately 50) accumulated during a
test. During each test, the speeds of vehicles passing the sampling station
were estimated by timing over a known travel distance. Estimates of vehicle
weights were obtained from plant personnel. In some tests, the vehicle passes
sampled were dominated by controlled test vehicles traveling at preselected
speeds.
TEST RESULTS
The cumulative results of the field testing of vehicular traffic on un-
paved roads are provided in Table 1, which presents the emission factors for
suspended particulate (smaller than 30 /im in Stokes diameter) and for fine
particulate (smaller than 5 ^m in Stokes diameter), along with surface mate-
rial characteristics. The upper size limit of 30 fim for suspended particulate
is the approximate effective cutoff diameter for capture of fugitive dust by
a standard high-volume particulate sampler (based on a typical pirticle den-
sity of 2 to 2.5 g/cu cm) (1). Both Table 1 and Figure 3 compare actual emis-
sion factors with predicted values.
Excluding run Nos. 1-1, 1-7, and 1-8 for the reasons given in the foot-
notes to Table 1, the relative standard deviation of prediction errors is
26.1% of the mean measured emission factor. For Test Series E and G, the emis-
sion factor equation consistently underpredicts the measured factors. This
appears to be due to the effect of 10- and 18-wheel trucks, which comprised
a substantial number of the passes in those tests. In all other test series,
the vehicle mix was dominated by four- and/or six-wheel vehicles.
As indicated in Figure 4, there is no apparent relation between the
fraction of the emissions consisting of fine particles and the average
vehicle weight or the road surface composition. The average value is ap-
proximately 35% by weight.
-422-
-------
TABLE 1. PREDICTED VERSUS ACTUAL EMISSIONS—UNPAVED ROADS
Road surface
Emission factor^/
Test Silt Mean vehicle Mean vehicle
No. Type (%) speed (mph) weight (tons)
R-l ) Crushed 12 30
R-2 , limestone 13 30
S-3 ' 13 40
R-8 j 20 30
R-10> Dirt 5 40
R-13) 68 30
A-14 i Crushed 4.8 30
A-15 f slag 4.8 30
E-l ) 8.7 14
E-2 • Dirt 8.7 16
E-3 ' 8.7 16
F-21i Dirt/ 9.0 15
F-22 crushed 9.0 15
F-23' slag 9.0 15
F-24 | Dirt/slag 0.03 15
F-25 !' (Cohered 0.02 15
G-27 . 5.3 22
G-28 J 5.3 23
G-29 ( Crushed 5.3 24
C-30 / slag 4.3 25
G-31 1 4.3 29
3-32 ' 4.3 22
l-\ ) Crushed 4'7 «
•-1 1 rock and 4'7 «
1 glacial *'7 c
'-* till 4'7 15
1-5 4.7 15
1-7 | Crushed 6.1 13.5
1-8 f rock 5.1 13.5
(taconite/
waste)
I-? j Crushed 1.3 13
1-10 / rock l.il/ 13
I- 11 ' (TREX) 1.3 14
a/ Particles smaller than 30 jjm in diameter
b_/ Based on MRI emission factor.
c/ 100 x (predicted-actuaD/actual.
d/ Test Series 1-1 through 1-5 performed on
£/ Tests performed on day following 2 days
fj Assumed value.
£/ Equation not applicable.
3
3
3
3
3
3
70
70
34
34
23
3
3
4
3
3
17
12
9
14
3
30
b7
67
67
157
157
118
117
110
112
127
previously
lib/vehicle
Predicted^/
5.9
6.4
8.5
9.3
3.3
33
29
29
14
16
12
2.2
2. 2
2.8
if
7.7
6.1
5.0
6.0
4.5
9.8
13.9
13.9
13.9
27.4
27.4
25.5
25.3
.
£/
a/
inactive road.
roila)
Actual
6.0
6.8
7.9
8.1
3.9
32
27
29
17
16
19
3.2
1.7
2.4
0.071
0.39
13.7
9.0
6.8
11.6
6.6
21.8
, .d/
6.4—
9.9
20.3
24.3
30. r
15.1*.'
U.&S/
1.3
2.6
3.3
Percent
difference^'
_2
-6
8
21
-15
3
7
0
-17
0
-37
-31
29
17
-
-44
-32
-26
-48
-32
-55
117
40
-32
10
-11
53
73
.
_
-
Predicted
Actual
0.98
0.94
1.08
l.:i
0.85
1.03
1.07
1.00
0.32
1.00
0.63
0.69
1.29
1.17
-
0.-6
0.68
0.74
0.52
0.68
0.45
2.17
1.40
O.o8
1.10
0.89
1.38
1.73
—
-
—
of rain totaling 1.13 in.
-423-
-------
I
IS)
I
o
ON FACTOR
o
ACTUAL
o
1 Inch Rain on Previous Evening
Tests on Inactive Road
TEST SITE - ROAD SURFACE MATERIAL
A Rural Kansas - Crushed Limestone
A Rural Kansas - Dirt
O Iron and Steel Plant - Crushed Slag
t> Iron and Steel Plant - Dirt/Crushed Slag
• Iron and Steel Plant - Dirt
0 Taconite Mine - Crushed Rock (Taconite/Waste)
• Taconite Mine - Crushed Rock/Glacial Till
1 1
1 1
1 1
1 1
1 1
1 1
1 1
J I
10
20 30
PREDICTED EMISSION FACTOR (U(/VMT)
40
Figure 3. Comparison of predicted and actual emissions—untreated roads*
-------
I.UU
'5& °-80
Z
P= 0.60
(J
oe.
1 "-
6 £
r 2
^ 0.40
u
fe
0.
LU
Z 0.20
u_
0
ROAD SURFACE MATERIAL
A Crushed Limestone
A Dirt
D Crushed Rock (Taconite/Waste)
O Crushed Slag
• Dirt/Crushed Slag
• Crushed Rock/Glacial Till
-
> 0
* ° A * °
A O
~A • •
•
A
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
20
40 60 80 100
AVERAGE VEHICLE WEIGHT (TONS)
120
140
160
Figure 4. Fine particle fractions of TSP emissions.
-------
Analysis of parameters affecting the atmospheric transport of fugitive dust
indicates that only the portion smaller than about 5 urn in diameter will be
transported over distances greater than a few kilometers from the source (5).
As stated above, limited testing of the effects of chemical dust sup-
pressants was also conducted. At the taconite mine site, TREX (ammonium
lignin sulfonate--a water soluble by-product of papermaking) was applied to
the waste rock aggregate comprising the surface of a haul road. A 20 to 25%
solution of TREX in water was sprayed on the road at a rate of 0.08 gal/sq
yard of road surface. The other chemical dust suppressant tested was
Coherex® (a petroleum-based emulsion), which was used to treat a dirt/slag
surfaced service road traveled by light- and medium-duty vehicles at an inte-
grated iron and steel plant. Coherex® was applied at 10% strength in water.
Figure 5 shows a plot of measured dust control efficiency as a function
of the number of vehicle passes following application of the road dust sup-
pressant. Control efficiency was calculated by comparing controlled emissions
with uncontrolled emissions measured prior to road surface treatment. As
indicated, the effectiveness of road dust suppressants is initially high but
begins to decay with road usage. According to taconite mine personnel, the
binding effect of TREX can be partially restored by the addition of water to
the road surface. It should also be noted that the apparent performance of
Coherex®was negatively affected by tracking of material from the untreated
road surface connected to the 200-ft treated segment.
With regard to the effects of natural mitigation of road dust emissions,
the final term in the emission factor equation for traffic on unpaved roads
(Figure 1) is used to reduce emissions from dry conditions to annual average
conditions. The simple assumption is made that emissions are negligible on
days with measurable precipitation and are at a maximum on the rest of the
days. Obviously, neither assumption is defendable alone; but there is a rea-
sonable balancing effect. On the one hand, 0.01 in. of rain would have a
negligible effect in reducing emissions on an otherwise dry, sunny day. On
the other hand, even on dry days, emissions during early morning hours are
reduced because of overnight condensation and upward migration of subsurface
moisture; and on cloudy, humid days, road surface material tends to retain
moisture. Further natural mitigation occurs because of snow cover and frozen
surface conditions. In any case, further experimentation is needed to verify
and/or refine this factor.
CONCLUSIONS
Based on an expanded data set of 24 tests, the MRI emission factor
equation for traffic-entrained dust from unpaved roads predicts measured
emission factors with a relative standard deviation of the prediction error
equal to 26.1% of the mean measured value. It appears that an additional ad-
justment term related to the average number of wheels per vehicle would
-426-
-------
*>
ro
vj
100
90
G so
z
UJ
u
t 70
o
on
\-
z
o
u
60
50
40
VEHICLE TYPE
DUST SUPPRESSANT
O Haul Truck
A Light Duty Vehicles
Lignin Sulfonate
Coherex
100 120 140 160 180 200 220 240 260
VEHICLE PASSES FOLLOWING TREATMENT
280 300
Figure 5. Effectiveness of road dust suppressants.
-------
reduce the mean prediction error because there is a clear tendency to under-
predict measured emission factors when the road is traveled by a substantial
portion of 10- and/or 18-wheel vehicles rather than 4- and 6-wheel vehicles.
Approximately 35% of measured road dust emissions in the suspended par-
ticulate size range (particles smaller than 30 /m in diameter) consist of fine
particles (particles smaller than 5 urn in diameter), which have the potential
for transport over distances greater than a few kilometers from the source*
This fraction appears to be independent of average vehicle weight and road
surface composition.
Limited testing of chemical dust suppressants for industrial unpaved
roads indicates a high initial control efficiency (exceeding 90%), which de-
cays by more than 10% with the passage of 200 to 300 vehicles. Consistent with
the emission factor equation, the lowering of emissions is reflected by the
reduced silt content of the road surface material after the application of
chemical dust suppressants. Additional testing is needed to better quantify
the performance of road dust suppressants. Testing is also needed to verify
and/or refine the emission factor adjustment term which accounts for climatic
mitigation.
ACKNOWLEDGEMENT
The work upon which this paper is based was performed pursuant to
Contract No. 68-02-2609, Task 3, with the U.S. Environmental Protection
Agency.
FACTORS FOR CONVERSION TO METRIC UNITS
1 kilogram =2.2 Ib
1 kilometer = 0.62 miles
1 metric ton =1.1 short tons
-428-
-------
REFERENCES
1. G. Cowherd* Jr., K. Axetell, Jr., C« M. Guenther, and G. Jutze, "Devel-
opment of emission factors for fugitive dust sources," Final Report,
Midwest Research Institute for U.S. Environmental Protection Agency,
Publication No. EPA-450/3-74-037 (NTIS No. PB 238262/AS) (June 1974).
2. C. Cowherd, Jr., C. M. Maxwell, and D. W. Nelson, 'Quantification of
Dust Entrainment from Paved Roadways," Final Report, Midwest Research
Institute for U.S. Environmental Protection Agency, Publication No.
EPA-450/3-77-027 (July 1977).
3. C. Cowherd, Jr., "Measurement of fugitive partlculate," Second Symposium
on Fugitive Emissions Measurement and Control, Houston, Texas, May 1977,
Publication No. EPA-600/7-77-148 (December 1977).
4. R. Bohn, T. Cuscino, Jr., and C. Cowherd, Jr., "Fugitive emissions from
integrated iron and steel plants," Final Report, Midwest Research
Institute for U.S. Environmental Protection Agency, Publication No.
EPA-600/2-78-050 (March 1978).
5. C. Cowherd, Jr., and C. M. Guenther, "Development of a methodology and
emission inventory for fugitive dust for the regional air pollution
study," Final Report, Midwest Research Institute for U.S. Environmental
Protection Agency, Publication No. EPA-450/3-76-003 (January 1976).
-429-
-------
DEVELOPMENT OF MEASUREMENT METHODOLOGY
FOR EVALUATING FUGITIVE PARTICIPATE EMISSIONS
Edward E. Uthe, Charles E. Lapple, Clyde L. Wltham, and Robert L. Mancuso
SRI International
Menlo Park, California 94025
-431-
-------
ABSTRACT
A research project Is under way to develop and demonstrate a method for
evaluating fugitive particulate emissions from an industrial plant, by measur-
ing the flow of total particulate pollution in the atmosphere downwind of it.
Using both surface sampling and remote-sensing techniques for sampling above
the surface level, the method is to be capable of measuring the spatial distri-
bution, absolute concentration, and velocity of the particulate cloud issuing
from the plant.
The objective of the current study is to demonstrate the proposed method
under field conditions, but in circumstances that simplify and control the
variables of the general problem. A field test was recently completed that
used an aerosol generator to emit a controlled aerosol stream into the atmo-
sphere from a point source with known particulate feed rates. A mobile lidar
system was used to make a series of remote observations of the cross-plume
particulate distribution 300 to 500 m downwind of the source.
Data presented show that the cross-plume integrated backscatter is respon-
sive to the particulate feed rate. The problem of deriving an absolute mass
emission rate from such backscatter traverses of a plume is discussed.
-432-
-------
I INTRODUCTION
The total source strength of pollution emitted by industrial plants is the
aggregate of all diffuse and minor specific emissions as well as major identi-
fiable point sources. Therefore, for many plants, the measurement of individual
emissions from the multiplicity of sources is neither economical nor practical.
The only feasible approach is to measure, as accurately as possible, the
concentration throughout a cross section of the downwind "plume" of the com-
bined fugitive emissions, to integrate these, and, from a measurement of the
integrated wind velocity through the plane of the cross section, to calculate
the pollutant mass flow. For particulate fugitive emissions, any deposition or
transformation in the intervening distance must also be considered, to derive
an estimate of the total source strength.
The problems of accomplishing such measurements with existing technology
are many. Specifically, with in-situ samplers it is virtually impossible to
characterize adequately the concentration of particles throughout the total
cross section of the plume, or to relate any measurements made to the total
envelope of the plume, or to determine its extent. This is especially the case
above the surface, because of the extended and variable nature of the multiple
sources of fugitive emissions.
Lidar (laser radar) observation fulfills, as no other method does, the
requirement for delineating the spatial distribution of elevated particulate
pollution plumes, and also for readily distinguishing between background pol-
lution and pollution from the plant being studied. While there are limitations
and difficulties in using lidar backscatter measurements for determining abso-
lute particulate concentrations, it is possible to evaluate, with useful ac-
curacy, the near-instantaneous mass distribution of particulate material within
a selected cross section or envelope. It is thus possible to derive series of
such cross sections in time, and to relate these to a measurement of mean wind
velocity, to derive an estimate of mass flow and hence of source strength.
(Several lidar techniques have been demonstrated for wind measurement, but
these are not currently at an operational stage and simpler, more conventional
methods of wind measurement are still indicated.) Finally, eye safety from
laser systems used over industrial sites must be considered. In principle,
however, it is apparent that lidar offers a unique potential for eventually
providing quantitative measurement of fugitive particulate emissions.
This paper presents results from an initial attempt to demonstrate, using
presently available research equipment, the capabilities of lidar for measuring
particulate fugitive emissions, and recommends several research areas requiring
further consideration to fully exploit the lidar technique for this purpose.
-433-
-------
II THE LIDAR TECHNIQUE
Lidar (Light Detection and Ranging) uses laser energy in radar fashion to
effectively observe remote atmospheric constituents. The lidar system being
used to develop methodologies for fugitive dust measurement is SRI Interna-
tional's Mark IX mobile system, designed for use with atmospheric research
programs. Figure 1 presents views of the lidar system, a block diagram, and
a list of specifications. The lidar is installed within an 18-ft van complete
with real-time data processing and display capabilities as well as with power-
generating equipment, enabling operation at any site.
Figure 2 is an example of an intensity-modulated TV display depicting the
cross-plume aerosol structure observed by scanning the lidar in elevation. The
brightness of this display is proportional to the logarithm of the backscattered
light detected by the lidar system. Clearly, lidar can be used to map relative
aerosol density distributions over remote distances. Computer-generated
vertical-concentration profiles are plotted on the cross section for locations
indicated by the cursor marks drawn above the plume return. Similarly, the
backscattered data can be spatially integrated to determine a relative cross-
plume density. Light that is elastically backscattered from suspended atmo-
spheric particulate matter, as a function of pulse-transmission time, is quan-
titatively related to (clear air and plume) optical parameters along the ob-
served path according to the lidar (radar) equation
P = K E 0 R~2 T2 (1)
where
P = received power from a scattering volume defined by the laser pulse
K = lidar calibration constant
E = transmitted power
8 = backscatter coefficient of the scattering volume
R » range from lidar of the scattering volume
T = transmission to the scattering volume
For relatively clear air, the attenuation term can normally be ignored (T2 « 1)
and P provides information on the distribution of particulate material in
the atmosphere, as shown in Figure 2. However, the evaluation of physical-
density terms (such as mass concentration) requires relating the optical param-
eter 3 to the density. Unfortunately, backseatter-to-concentration ratios are
dependent on the distributions of particle size, composition, and shape. Before
these factors, and the ways in which they can limit the accuracy of quantitative
evaluations of plume density, can be considered, the lidar observations must be
shown to respond in a predictable manner to increases of plume density for a
given aerosol type. An experimental field program using presently available
research equipment has been conducted for this purpose. The design of the field
program is discussed below, followed by some of the preliminary results.
-434-
-------
MARK IX LIOAR VAN
(b)
ANALOG DATA AND FIRE
CONTROL ELECTRONICS
(c)
DIGITAL DATA ELECTRONICS
AND TV DISPLAY
EXTERIOR AND INTERIOR VIEWS OF THE LIDAR VAN
BLOCK DIAGRAM OF THE LIDAR SYSTEM
LIDAR SPECIFICATIONS
TRANSMITTER
6943A Wavelength
0.5 mrad Beamwidth
1.0 J Pulse Energy
30 ns Pulse Length
60 ppm Maximum PRF
RECEIVER
6 inch Newtonian
1 to 5 mrad Field of View
5A Predetection Filter
RCA 7265 PMT Detector
4-decade, 35-MHz Log-
arithmic Amplifier, Inverse-
range-squared or step-
function PMT modulation.
DATA SYSTEMS
Analog video disc recording
(4.5 MHz) with A-scopeand
Z-scope real -time displays.
Digital magnetic tape (data
and programs) recording
(25 MHz) with computer
processing and real-time TV
display (512 x 256 x 4 bit)
of processed data.
MOUNT
Automatic azimuth and elevation
fire and scan with 0.1 minimum
resolution. Automatic reset.
Mechanical safety stops.
FIGURE 1 SRI INTERNATIONAL'S MARK IX MOBILE LIDAR SYSTEM
Ul
5
CO
H
o
UJ
I
FIGURE 2
HORIZONTAL DISTANCE FROM LIDAR
EXAMPLE OF COMPUTER-GENERATED PROFILES OF VERTICAL PLUME DENSITY
Lidar is located at lower left corner. The height and distance
scale is 75 m/div. Vertical concentrations of the plume (rela-
tive to clear air, with a scale of 10 dB/div) are plotted at the
lower left and the horizontal position associated with each
profile is plotted in the upper right.
-435-
-------
Ill FIELD STUDY
The experiments were conducted under actual field conditions but simplified
to the extent that the observed particulate plume would have better-known prop-
erties than could be expected in the general case. Dust was emitted into the
atmosphere in a controlled manner. The intent was to show that observed back-
scatter spatially integrated across the aerosol plume downwind of the source is
well correlated with the source strength. In addition to lidar observations,
conventional wind measurements were made to determine downward dilution of the
particulate plume; near-surface particulate sampling was conducted to investi-
gate possible approaches for measuring actual fugitive emissions with unknown
particulate properties.
An aerosol source was designed and constructed to generate a point source
of dust by feeding a fine powder of known properties into a sonic-velocity air
stream pointing downwind. The source is capable of emitting aerosols at three
levels—1, 3, and 10 m above ground level. The powder is fed at a uniform rate
by means of a grooved-disk feeder of proprietary design. The disk rotation rate
can be continuously varied to control the particle feed rate. A compressed air
stream of 40 ft3/min at 52 psi was used to disperse the dust metered out by the
feeder.
For the dust, Min-U-Sil (a commercially available* ground silica) was
chosen because (1) the particles are predominantly below 5 um in diameter;
(2) the material is available in quantity; and (3) the dust is a realisic
example of material encountered in practice. Figure 3 shows an aerosol plume
generated by dust emissions from the bottom and top nozzles of the generator.
The Mark IX lidar van was positioned 300 to 500 m from the source so that
the laser beam would intersect the plume perpendicularly at a downwind distance
of about 300 to 500 m. The lidar was scanned in elevation to observe the cross-
plume aerosol distribution, similar to that shown in Figure 2. An anemometer
was installed on the Mark IX lidar van and its signal output was input to the
lidar digital system so that both wind speed and direction were sampled for
each lidar observation (firing). In addition, the output of an integrating
nephelometer was similarly sampled. These data and the digital record of the
backscatter signature were stored on magnetic tape for use during the data
analysis program.
An initial data collection program was conducted during May 1978 and a
second data collection period using slightly improved experimental techniques
was conducted in September 1978. For each experimental run, aerosol plumes
were generated at three different particulate feed rates. For some meteorolog-
ical conditions, the cross-plume aerosol density at a given distance from the
source was quite variable. Therefore, a series of about ten vertical lidar
cross sections were collected for each particulate feed rate, so that the value
of the time-integrated cross-plume backscatter could be determined. Preliminary
results obtained from the field program are presented in the next section.
Pennsylvania Glass Sand Corp., Pittsburgh, PA.
-436-
-------
FIGURE 3 CONTROLLED DUST PLUMES GENERATED TO REPRESENT FUGITIVE EMISSION
SOURCES
(a) Emission 1 m above surface
(b) Emission 10 m above surface
-437-
-------
IV PRELIMINARY RESULTS
The range-dependent backscatter signatures were collected while the lidar
was scanned in elevation, to observe the total cross-plume aerosol (Figure 2).
While the resulting data can readily be integrated in polar coordinate form,
an objective method was applied to evaluate an array of backscatter values in
Cartesian coordinates. The advantage of an x-y grid is that the values can
be displayed with commonly used computer techniques, to facilitate indentifica-
tion of plume and clear air regions. The integrated cross-plume backscatter
was determined from the expression
/
'p dA -<* + dA '
where 3 = plume backscatter coefficient [see Equation (1)]
3 = clear-air backscatter coefficient
c
A - area that includes all plume particulates
A' = area that excludes plume particulates.
The first term represents plume and clear-air backscatter integrated over an
area of the vertical cross section containing the plume, and the second term
is an estimate of the clear-air backscatter integrated over the same area.
The primary objective of this study was to demonstrate experimentally that
the integrated cross-plume backscatter responds in a linear manner to changes
in the total mass concentration of particulate material released into the
atmosphere.
The total rate of fugitive dust emission, w, can be calculated from the
distribution of particulate concentration in a given vertical plane downwind
of the source by
W=/CU
sinG dA (3)
where c is the concentration in the area segment dA at which the wind velocity
is u and is oriented at an angle 0 to the vertical plane. Assuming c is pro-
portional to the lidar backscatter coefficient (c = K3), and assuming a constant
wind speed and direction at all parts of the vertical plane observed by the
lidar, we have
w = K u sinG / $ dA (4)
For each observation (vertical cross section), the backscatter records
were processed in terms of an array of backscatter values for identifying appro-
priate clear-air and plume regions, to evaluate Equation (2). The integrated
-438-
-------
cross-plume backscatter values were computed and statistically analyzed for the
average and standard deviation for each set of observations made at a fixed
particulate feed rate.
Figure 4(a) presents experimental results for two observational periods,
using the integrated cross-plume backscatter as an estimate of the rate of
fugitive dust emission. For each case, three emission rates were used, and
for each rate, from 8 to 20 vertical cross sections were made. The data show
that the integrated backscatter generally increases linearly with the particu-
late emission rate. However, relatively large scatter of the data points
occur—i.e., large standard deviations. This can be explained partly on the
basis of variable wind conditions, which were ignored in this estimate of
emission rate.
Figure 4(b) presents the data shown in Figure 4(a) after wind-speed (but not
direction) corrections were applied; these were based on anemometer measurements
made at the lidar site during the data-collection period. The wind corrections
improve the linear relationship between the lidar observations and the particu-
late feed rates (especially for the second case). However, the standard devia-
tions remain about the same. This variability is thought to result partly from
small-scale, uncontrolled variations in the particle feed rates and turbulent
conditions resulting in a nonuniform aerosol concentration downwind of the
source. Both these factors were visually observed to cause backscatter varia-
tions, but an estimate of their relative importance could not be made. Never-
theless, these data illustrate that the lidar technique provides a downwind
measurement that responds in a predictable (linear) manner to the total par-
ticulate emission rate as evaluated from feeder calibration data.
V ABSOLUTE PLUME DENSITY MEASUREMENTS
The data presented in the previous section of this paper indicate that for
a given type of particulate material, the lidar can be calibrated to provide
an absolute measurement of aerosol density of a downwind pollution plume. How-
ever, fugitive dust emissions from a multiplicity of source types may consist
of a heterogeneous complex of particle size, shape, and composition distribu-
tions. The lidar wavelength(s) or observational techniques must be carefully
selected to minimize errors of density measurement associated with uncertainties
of particle characteristics.
Earlier experiments used a large-scale chamber that was especially designed
for making remote lidar observations of generated aerosols of known particle
size, shape, composition, and concentration (Lapple and Uthe, 1976). Observa-
tions of aerosols generated from fly-ash particulates of different size cate-
gories showed that the backscatter-to-mass concentration ratio is less dependent
on particle size at a lidar wavelength of 1.06 urn than at 0.7 urn. Recent trans-
missometer experiments (Uthe, 1978) have shown that extinction measurements in
a wavelength interval of 3 to 4 ym provide a good indicator of aerosol volume
concentration regardless of particle size, shape, or composition (for the range
of typical pollutants). While backscatter certainly is sensitive to changes in
particle shape or composition, measurements in the 3- to 4-ym region may result
in a higher correlation between backscatter and aerosol mass concentration when
-439-
-------
40
30
20
10
8
6
4
n
\ 2
i
i
E
c
]
u
I 40
5 30
|
i 20
10
8
6
4
2
1
1 1 1 1 1 1 1 1
DUST RELEASE
— 1 M ABOVE SURFACE ~
iO = ONE STANDARD
DEVIATION
k
— T X
I ft*
- / J
)8
—
—
- - la —
X 1
T X
X -
V ?13 = Number of observations
16 MAY 1978
1 T 1 1 1 1 1 1 1
_ 1 1 1 1 1 II 1 1
DUST RELEASE
-10 M ABOVE SURFACE ~
X* 10 -
4g)\Q
/ -L - 10
— ^/ —2
- X O 20 —
— / —
X
22 SEP 1978
1 1 1 1 1 1 I 1 1
2 468 10 20 30 2 468 10 2030
EMISSION RATE (FEEDER) - kg/hr
(a) LIDAR ESTIMATE BASED ON INTEGRATED CROSS-PLUME BACKSCATTER, //? dA
1 1 1 1 1 1 • 1 I
DUST RELEASE
~ 1 M ABOVE SURFACE
A
T X
— tl'
E x
T xX
-+ 10 —
)8
__
—
•-10 ~
\ / • -\
V13
-xT -
16 MAY 1978
I I I I I I I I I
2 4 6 8 10 20 30
1 I 1 1 1 1 | 1 1
DUST RELEASE
~ 10 M ABOVE SURFACE ~
~~ -r + 10 ""
^r10.
_ /L - «_
> ~
- X!)^ ~
X 1
_ X
T x
/ / «
_ >fl9
22 SEP 1978
1 1 1 1 1 III 1 1
2 4 6 8 10 20 30
FIGURE 4
EMISSION RATE (FEEDER) - kg/hr
(b) LIDAR ESTIMATE BASED ON PRODUCT OF INTEGRATED CROSS-PLUME
BACKSCATTER, AND MEAN WIND SPEED, u/0 dA
PARTICULATE EMISSION RATE EVALUATED FROM LIDAR PLOTTED AS A FUNCTION
OF EMISSION RATE EVALUATED FROM PARTICLE FEEDER
Dashed line is best-fit linear fetation.
-440-
-------
the particle size is variable. Clearly, additional experiments are needed to
define optimum wavelength regions and to establish uncertainties in plume
density measurements because of unknown particle characteristics.
A downwind in-situ measurement of the backscatter-to-mass concentration
ratio may provide the basis for converting backscatter observations to fugitive
particulate concentrations. One possible approach is to sample the aerosol
mass concentration at a location near the path of the laser beam. For example,
a filter sample could be taken over an extended time period while repetitive
lidar observations are made of nearly the same atmospheric volume. The time-
integrated backscatter from this volume and the mass concentration measurement
would provide an estimate of the backscatter-to-mass concentration ratio re-
quired to quantitatively evaluate the integrated cross-plume density from the
lidar-observed, above-surface backscatter distributions. While this method
may provide the necessary information to interpret backscatter data quantita-
tively in terms of aerosol density, an especially designed instrument for mea-
suring the backscatter-to-mass concentration ratio is desired. Of course, if
an observational technique can be developed to reduce to within acceptable
standards the dependence of the density measurement on the uncertainties of
particle characteristics, these supporting observations will not be necessary.
VI CONCLUSIONS AND RECOMMENDATIONS
Using an existing mobile lidar system, a field program has been conducted
to demonstrate that for a single particulate type, the cross-plume Integrated
backscatter is a good indicator of source strength of particulate emissions.
The positive results obtained from this first field program encourage the
further development of this methodology for measuring fugitive dust concen-
trat ions.
We plan to test the lidar technique using sources of different particle
size, shape, and composition distributions. For example, the next test may be
of the emissions from a liquid-particulate generator capable of producing both
black and white oil smokes. Other areas to be investigated include:
• Selection of optimum wavelength(s) regions to minimize mass measurement
errors associated with uncertainties in particle size, shape, and com-
position distributions.
• Development of an eye-safe lidar technique.
• Evaluation of wind-measuring lidar techniques to evaluate actual flow
of observed pollutants.
• Development of an instrument to measure backscatter-to-mass concentra-
tion in situ.
• Further experimental evaluations of proposed methodologies using sources
of known particulate properties.
-441-
-------
ACKNOWLEDGMENT
SRI International's work was sponsored under EPA Contract No. 68-02-2752,
The authors are grateful to William D. Conner of EPA for his part in the
program and his review of the manuscript.
REFERENCES
Lapple, C. E., and E. E. Uthe, 1976: "Remote Sensing of Particulate Stack
Emissions," AICHE Symposium Series, Vol. 72, No. 156, pp. 181-202.
Uthe, E. E., 1978: "Remote Sensing of Aerosol Properties Using Lidar (Laser
Radar) Techniques," Proceedings of SPIE Seminar on Optical Properties of
the Atmosphere, Washington, D.C., 30 March 1978.
-442-
-------
DEVELOPMENT OF A FUGITIVE
ASSESSMENT SAMPLING TRAIN
FOR PARTICIPATE AND ORGANIC EMISSIONS
Henry J. Kolnsberg, Senior Project Manager
Roland L. Severance, Principal Engineer
TRC-THE RESEARCH CORPORATION OF NEW ENGLAND
125 Silas Deane Highway
Wethersfield, Connecticut 06093
-443-
-------
Abstract
The measurement of fugitive emissions from sources that preclude the cap-
ture of emissions before their diffusion into the ambient atmosphere poses some
unique problems. Devices designed to obtain ambient samples for area studies
generally do not provide samples of particulate matter large enough for meaning-
ful quantification and qualification analyses in a reasonable sampling period,
and are usually completely non-directional in their sampling.
This paper describes a program for the development of a prototype portable
Fugitive Assessment Sampling Train (FAST) designed to obtain a 500 milligram
particulate matter sample in an 8 hour sampling period downwind of most indus-
trial sources. The development of the design criteria, establishment of oper-
ating parameters, selection and design of hardware components, fabrication and
initial testing of the FAST are described in detail, and the program for the
qualification testing of the completed unit is outlined.
-444-
-------
A considerable portion of the air-polluting particulate matter and
organic vapor emissions from industrial and energy-related processes is
generated by sources that do not permit the capture of their emissions for
measurement purposes before their diffusion into the ambient atmosphere.
Obtaining samples of such fugitive emissions of sufficient size to perform
statistically significant analyses of their concentration, particle size
distribution, physical characteristics, chemical composition or biological
activity presents a problem not readily solved using existing devices and
traditional sampling techniques.
Standard high volume samplers, for example, can provide some information
about the average particulate matter concentration at a sampling site over a
long sampling period, but do not provide samples large enough for other than
total mass determinations. Cascade impactors can provide particle size
distribution information for a relatively small sample and cyclone separators
can collect a fair-sized particulate matter sample in a few size ranges to
provide essentially the same information. Grab sampling of gases or vapors
for subsequent gas-chromatographic analysis can provide data on the chemical
composition and approximate or relative concentration of these emissions, but
is subject to the influence of interaction between emissions or aging of the
samples. No single sampler exists that can collect a particulate matter sample
large enough or a vapor sample stable enough to provide information in all
areas.
This paper describes the progress made to date in the development of a
Fugitive Assessment Sampling Train (FAST) designed to fill the requirements for
a sampler capable of providing a large sample of particulate matter emissions
from the atmosphere in a relatively short sampling period. A smaller organic
vapor sample is obtained from the sampled stream.
The development effort was conducted by TRC-THE RESEARCH CORPORATION of
New England under a contract (68-02-2133) with the Process Measurements Branch
of the United States Environmental Protection Agency's Industrial Environmental
Research Laboratory at Research Triangle Park, North Carolina. Discussions
between TRC personnel and the EPA Project Officers, Dr. Robert M. Statnick and
D. Bruce Harris, resulted in a target design specification for an ideal
sampling train as the development starting point. This ideal sampler was
described as being able to obtain, from the ambient air in the vicinity of an
industrial fugitive emissions source, a 500 milligram sample of suspended
particulate matter and a similar-sized sample of organic vapors in an eight-
hour sampling period. The particulate matter sample would be separated into
respirable (smaller than 3 micrometer) and non-respirable (larger than 3
micrometer) fractions. The sample sizes were selected to correspond to the
then-considered minimum for complete analysis including bio-assay. The sampler
was also to be self-contained and portable; it would require minimum power and,
using commercially available components wherever possible, cost less than
$10,000 to fabricate in the prototype version.
An extensive computerized literature search and review was conducted in
the hope of obtaining sufficient information on ambient concentrations of
industrial fugitive emissions as particulate matter and organic vapors to
prepare a realistic system design specification for the FAST. While this
search and review revealed almost no data on ambient concentrations, it did
provide a wealth of information on emission rates from a wide variety of
-445-
-------
industrial processes. A series of calculations based on the atmospheric
diffusion equations in Turner's Workbook (1) for a range of atmospheric,
topological and wind conditions was then performed to relate the published
emission rates to ambient concentrations. These calculations indicated that
an ambient concentration of 200 micrograms per cubic meter can be found within
100 to 500 meters of sources emitting between 0.6 and 23 kilograms per hour—
a range covering about 90 percent of the industrial sources for which data is
available.
This 200 microgram per cubic meter concentration was used to determine the
sampling rate required to obtain a 500 milligram sample in an eight-hour period
of 5.2 cubic meters per minute (184 CFM) as the initial system design parameter.
A Roots lobe-type vacuum blower, capable of moving the required volume of air
against a pressure drop of about 10 cm Hg, was selected as the particulate
sampling prime mover. A system of drive belts and pulleys was utilized to
operate the blower at the required 3800 RPM from a three horsepower drive
motor. The drive system also provides enough flexibility to adjust the speed
and the sampling rate up to about 20% if required.
To provide for the separation of the particulate matter sample into
respirable and non-respirable fractions, an Air Correction Design 6UP Sanitary
Cyclone Separator was selected. Its design capacity of 6.3 cubic meters per
minute (222 CFM) provides a DSQ at about 2 micrometers at a pressure drop of
about 0.6 cm Hg. The cyclone was selected as preferable to filter- or
impactor-type collectors since the sample is removed from the sampling stream
and minimizes the degradation in sampling rate or effectiveness caused by the
deposition of particulate matter on flow-through filters or impactor plates.
Consultations with Mr. Kenneth Cushing of the Southern Research Institute,
under contract to the Process Measurements Branch in the area of particulate
matter sampling, indicated that Reeves-Angel 934AH glass fiber filter material
would be about 99.95% effective in collecting the fraction of the particulate
matter sample down to about 0.3 micrometers passed through the cyclone. A
circular format was selected for the filter material to provide the most even
distribution of the sample on the filter surface and minimize the pressure drop
buildup. A circular filter holder was designed to accommodate a 929 square
centimeter (1 square foot) filter, limiting the pressure drop across the
unloaded filter to 3.7 cm Hg. A louvered inlet section was also designed to
reject particles larger than 100 micrometers to complete the particulate matter
sampling section of the train.
To provide stable samples of airborne organic vapor emissions, it was
decided to utilize an adsorbent resin in a removable canister that could be
easily transported from the sampling site to a laboratory for extraction and
analysis of the sample. Dr. Philip Levins of Arthur D. Little, Inc., under
contract to the Process Measurements Branch in the organics sampling area,
provided consultation to TRC on the resin. The best available resin, XAD-2
which is almost 100% effective in retaining organic vapors C6 and higher,
was determined to require a canister containing about 75 kilograms to provide
a 500 milligram sample. This was prohibitive from the standpoints of size and
cost, and the design criterion was revised to obtain the minimum sample
required for a Level 1 assessment of 14 milligrams. This sample size requires
only 2.1 kilograms of resin and a sampling rate of only 0.14 cubic meters per
minute (5 CFM). A canister was designed and an oil-less Cast vacuum pump
-446-
-------
selected to draw the sampling stream from the main stream after the
particulate matter is removed.
The system design was reviewed and approved and procurement and
fabrication efforts started. At this time, the EPA's Health Effects Research
Laboratory suggested that an additional size fraction of the particulate
matter sample be included to help in the assessment of the inhalable (less
than 15 micrometer) portion of the emissions. It was decided to add a battery
of six single stage Sierra Instrument impactors to the system to effect this
additional fractionation between the inlet and the cyclone. These impactors
were designed to provide a 095 for 15 micrometer particles at the system
sampling rate with a pressure drop of only 0.05 cm Hg, and could therefore be
added without affecting the system design.
The final system design is shown schematically in Figure 1. Design flow
rates and pressure drops for each system element are shown enclosed in
brackets. Samples retained by each element are shown in parentheses.
[184 CFM]
UOOX>100/j>
[0.05 CM HI]
FILTER
AP
VACUUM
GAGE
EXHAUST
FILTER
noox>o.3//i
[3.7 CM H,]
CYCLONE
!50X>2/i)
[0.8 CM H,]
MAIN VACUUM
BLOWER
XAO-2 MODULE
HOOX>C4)
[179 CFM]
TRAP
OUTLET
T.C.
FILTER
VACUUM
PUMP [5 CFM]
FIGURE 1 - FUGITIVE ASSESSMENT SAMPLING TRAIN
DESIGN OPERATING CONDITIONS
The procured and fabricated elements of the prototype system were then
loosely packaged onto a space frame about 75 cm (2.5 feet) square by 183 cm
(6 feet) high to allow easy access to the elements during development testing.
The main sampling blower and the organic vapor sampling pump'were separately
mounted to improve the system's portability and permit the location of the
-447-
-------
blower and pump exhausts away from the sampling Inlet. The packages are shown
in Figure 2 as they appeared for the initial operational tests.
FIGURE 2
PROTOTYPE FAST SYSTEM INITIAL ASSEMBLY
After a successful operational test and a few modifications to the system
were completed at TRC, the FAST was shipped to the Southern Research Insti-
tute's laboratory for calibration testing of the particulate sampling section.
Tests were run by Southern using monodisperse ammonium fluorescein aerosols
provided by their vibrating orifice aerosol generator at 3, 10 and 15 micro-
meters. The test results for the cyclone, shown in Figure 3 as points plotted
on the manufacturer's design curve, are in very good agreement. The test re-
sults for the impactor, also shown in Figure 3, indicate good agreement with
the design curve for smaller particles but are considerably lower than expected
for the 15 micrometer particles of major concern.
-448-
-------
X
>^
IOO
«0
•0
7O
o
UJ
a
CJ
so
40
3O
JO
10
I I I I I I I
I I
GLASS FIBER
SUBSTRATES
IMPACTOR ---
CYCLONE ----
CYCLONE DESIGN CURVE
. IMPACTOR DESI6R
f ' CURVE
' » I M II
X
»•— r* i i MIM
.4 .J .« .1 .« .VIA
4 S 6 7 • * 10
2O
PARTICLE DIAMETER MICROMETERS
FIGURE 3: COLLECTION EFFICIENCY VERSUS PARTICLE DIAMETER FOR THE FUGITIVE
ASSESSMENT SAMPLING TRAIN PRE-SEPERATOR IMPACTOR AND CYCLONE
Since it was felt that this discrepancy was caused by bouncing of the
larger particles off the glass fiber substrate used in the impactors, a test
was run using a grease substrate in an attempt to reduce the bounce. This
resulted in a slight improvement in performance, but not to a level considered
satisfactory for further development. A joint effort by TRC and Southern
Research has been initiated to design and fabricate an elutriator to replace
the impactor as the 15 micrometer fractionator. It is expected that the
elutriator, envisioned as a battery of parallel settling chambers, will replace
and perform the functions of both the inlet louvers and the impactors.
A field test of the FAST has been planned for the near future at a coke
oven battery, where the system will be tested simultaneously with two Battelle
mega-vol samplers and a standard hi-vol sampler in the measurement of emissions
from coke pushing operations. This test, to be run prior to the fabrication
and installation of the elutriator, should provide sufficient data for contin-
ued development of the cyclone and filter sections as well as an initial
indication of the effectiveness of the adsorbent canister train design.
It is expected that the FAST will require some modifications and a repeat
of the calibration and field testing cycle after the addition of the elutriator.
Verification tests of the total system will be performed after the modifica-
tions and additions have been completed and an operating procedures manual will
be prepared and published for the final version.
-449-
-------
The efforts to date in the development of this Fugitive Assessment
Sampling Train have been quite successful and encouraging. The completion of
the planned effort will provide a useful tool for rapid, reasonable assess-
ments of fugitive particulate matter and organic vapors from a wide variety of
industrial sources.
Reference
(1) Turner, D. Bruce. Workbook of Atmospheric Dispersion Estimates. Public
Health Service Publication No. 999-AP-26, U.S. Department of Health, Edu-
cation and Welfare, Cincinnati, Ohio. Revised 1969. 84 pp.
-450-
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. REPORT NO. ,
EPA-600/7-79-182
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Third Symposium on Fugitive Emissions Measurement
and Control (October 1978, San Francisco, CA)
6. REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
7. AUTMOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
J. King, Compiler
9. PERFORMING ORGANIZATION NAME AND ADDRESS
1O. PROGRAM ELEMENT NO.
TRC--The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
INE623
11. CONTRACT/GRANT NO.
68-02-2615, Task 201
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings; 10/23-25/78
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES JJERL-RTP prOJCCt OfflCCr Ifi D.
541-2557.
Bruce Harris, Mail Drop 62, 919/
16. ABSTRACT
The proceedings are a compilation of technical papers prepared for presentation
at the Third Symposium on Fugitive Emissions, October 23-25, 1978, at San Fran-
cisco, CA. The papers discuss the scope and impact of fugitive emissions (non-
point sources) and present techniques which have been used to measure the emis-
sions. Fugitive emission control technologies are also discussed.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
cos AT I Field/Group
Pollution
Measurement
Processing
Leakage
Pollution Control
Stationary Sources
Fugitive Emissions
Non-point Sources
13B
14 B
13H
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. Of PAGES
456
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (»-73)
451
------- |