vxEPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
EPA-600/9-80-041
December 1980
Research and Development
Proceedings
Fourth Symposium on
Fugitive Emissions
Measurement and
Control
New Orleans, LA
May 1980
-------
EPA-600/9-80-041
December 1980
PROCEEDINGS
FOURTH SYMPOSIUM ON FUGITIVE EMISSIONS
Measurement and Control
(New Orleans, LA - May 1980)
Compiler
Christine Wibberley
TRC—Environmental Consultants, Inc.
125 Silas Dean Highway
Wethersfield, CT 06109
Contract No. 68-02-3115
Project Officer
D. Bruce Harris
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
INDUSTRIAL ENVIRONMENTAL RESARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
FOREWORD
The technical papers included in this volume were prepared for presentation
at the "Fourth Symposium on Fugitive Emissions: Measurement and Control,"
held in New Orleans, Louisiana on May 28-30, 1980.
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 the General Chairman of the Symposium.
Christine Wibberley of TRC-Environmental Consultants, Inc. was the Symposium
Coordinator and Compiler of these Proceedings.
EPA REVIEW NOTICE
This document has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the views
and policies of the Agency, nor does mention of trade names
or commercial products constitute endorsement or recommen-
dation for use.
ii
-------
TABLE OF CONTENTS
(Speakers ' names appear in italics)
Page
OVERVIEW OF FUGITIVE EMISSIONS 1
Thompson G. Pace, EPA/MDAD-RTP
FUGITIVE EMISSIONS CONTROL AFTER ALABAMA POWER 18
David I. Brandwein, TRC-Environmental Consultants, Inc.
PARTICLE PRODUCTION FROM SURFACE MINING - PART 1: VERTICAL
MEASUREMENTS 37
James A. Armstrong and Philip A. Russell, Denver Research
Institute and Dennis C. Drehmel, EPA/IERL-RTP
PARTICLE PRODUCTION FROM SURFACE MINING - PART 2: SURFACE
PARTICULATE AND METEOROLOGICAL MEASUREMENTS 64
David L. Dietrich and William E. Marlatt, Colorado State
University and Douglas G. Fox, Rocky Mountain Forest and
Range Experiment Station
FUGITIVE EMISSIONS CONCERNS FOR COAL STORAGE AND HANDLING AT
UTILITY GENERATING STATIONS 81
Peter W. Kalika and Pietro Catizone, TRC-Environmental
Consultants, Inc.
DESIGN, PERFORMANCE TESTING. AND FIELD OPERATION OF AN ISOKINETIC
ELECTROSTATIC PARTICLE SAMPLER 101
Bengt Steen, Swedish Water and Air Pollution Research Institute
ASSESSING HAZARDOUS WASTE TREATMENT FACILITY FUGITIVE ATMOSPHERIC
EMISSIONS 119
Tim S. Sekulia and B. T. Delaney, Fred C. Hart Associates, Inc.
RESULTS OF FUGITIVE EMISSION MEASUREMENTS AT REFINERIES AND CURRENT
ACTIVITIES IN PETROCHEMICAL UNITS 136
Donald D. Rosebrook, Radian Corporation
EVALUATION OF FUGITIVE EMISSIONS AT A LARGE WOOD-PRODUCTS PLANT . . 155
Peter D. Spawn, GCA Corporation
A METHOD FOR MEASURING FUGITIVE EMISSIONS FROM CAST HOUSE
OPERATIONS 168
James H. Geiger, Betz-Converse-Murdoch, Inc.
STEEL MILL PARTICULATE CHARACTERIZATION AND SOURCE/RECEPTOR
ANALYSES 179
Philip A. Russell, Denver Research Institute
ill
-------
TABLE OF CONTENTS (Continued)
DEVELOPMENT OF HORIZONTAL ELUTRIATORS FOR SAMPLING INHALABLE
PARTICULATE FUGITIVE EMISSIONS ................... 208
Kenneth M. Gushing, Southern Research Institute
TECHNIQUES FOR EVALUATING SURFACE AND GROUND WATER EFFECTS OF
DRY ASH DISPOSAL .......................... 243
James F. Villaume, Pennsylvania Power and Light Company and
Dennis F. Unites, TRC-Environmental Consultants, Inc.
MEASUREMENT OF FUGITIVE EMISSIONS FROM INCO'S COPPER CLIFF SMELTER
REVERBERATORY FURNACES ....................... 279
Alan D. Church, W. J. Middleton, and P. Gatha, Inco Metals
Company
CONTROL OF FUGITIVE EMISSIONS FROM COAL STORAGE PILES ....... 300
Aivo E. Veel and C. H. Carr, The Steel Company of Canada, Ltd.
USE OF ROOF MOUNTED TYPE ESP'S IN IRON AND STEEL INDUSTRIES
IN JAPAN .............................. 306
Seniohi Masuda, University of Tokyo
FUGITIVE HYDROCARBON EMISSIONS FROM AN IN-SITU OIL SHALE RETORT . . 322
Gerald M. Rinaldi and David R. Tierney, Monsanto Research
Corporation
A WIND TUNNEL STUDY OF FUGITIVE DUST FROM TACONITE STORAGE PILES . . 334
Robert B. Jaako and George M. Palmer, Purdue University
COMPUTING DESIGN CHARACTERISTICS FOR COAL PILE RUNOFF TREATMENT . . 357
Pamela B. Katz and John A. Ripp, TRC-Environmental Consultants,
Inc.
EMISSIONS AND EFFLUENTS FROM RAIL AND TRUCK TANKCAR CLEANING .... 375
Thomas R. BlaakDood* Monsanto Company
A NEW CONCEPT FOR THE CONTROL OF URBAN INHALABLE PARTICULATES BY THE
USE OF ELECTROSTATICALLY CHARGED FOG ................ 388
John S. Kinsey and Carol E. Lyons, AeroVironment, Inc., Stuart
A. Hoenig, University of Arizona, and Dennis Drehmel , EPA/IERL -
RTP
CONTROL METHODS FOR FUGITIVE AREA SOURCES ............. 402
Dennis J. Martin, TRC-Environmental Consultants, Inc. and
Dennis C. Drehmel, EPA/IERL - RTP
Iv
-------
TABLE OF CONTENTS (Continued)
Page
CIVIL ENGINEERING FABRICS APPLIED TO FUGITIVE DUST CONTROL
PROBLEMS 416
Dennis C. Drehmel3 EPA/IERL - RTP and Barry Levene, EPA,
Region VIII
SPRAY CHARGING AND TRAPPING SCRUBBER FOR FUGITIVE PARTICLE
EMISSION CONTROL 426
Seymour Calvert, Shui-Chow Yung, Richard-Parker, and Julie
Curran, Air Pollution Technology, Inc., Dennis C. Drehmel,
EPA/IERL - RTP
-------
OVERVIEW OF FUGITIVE EMISSIONS
Thompson G. Pace
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
May 1980
INTRODUCTION
Fugitive emissions comprise a major portion of the total national emis-
sions for both Particulate Matter (PM) and Volatile Organic Compounds (VOC).
There are 395 counties or parts of counties in nonattainment with the TSP
National Ambient Air Quality Standard (NAAQS) and 538 in nonattainment of the
ozone standard for which VOC is a major precursor. The EPA is currently
reviewing the particulate matter standard; one possible outcome may be the
regulation of particles smaller than 15 ym aerodynamic diameter, referred to
as Inhalable Particulate (IP).
Several regulatory programs are concerned with fugitive emissions. State
Implementation Plans require the control of PM and VOC to effect compliance
with the NAAQS for Total Suspended Particulate and ozone. Regulatory programs,
such as the Prevention of Significant Deterioration and regulatory policies
such as the rural fugitive dust policy and the bubble policy are cognizant of
fugitive emissions. The New Source Performance Standards are beginning to
regulate fugitive emissions and many pollutants regulated under the National
Emissions Standards for Hazardous Air Pollutants have fugitive emissions.
Each of these will be discussed.
-------
NATIONAL EMISSIONS ESTIMATES
Tables 1 and 2 give national emissions estimates for 1978 for participate
matter and Volatile Organic Compound (VOC) emissions. The total emissions
are taken from background information for the National Air Pollutant Emissions
Estimates - 1970-78.l Area source fugitive dust emissions from unpaved roads,
entrainment of dust from paved roads, and unpaved airstrips based on 1977
National Emissions Data System (NEDS) data are also included.
The distinction between stack and fugitive emissions is based on a
combination of published reports and engineering judgment.2,3 Thus, the
emissions should be considered as estimates only. The general criteria for
defining stack or fugitive is as follows:
Stack emissions include
- Process and fuel combustion sources vented directly through a
stack.
- Processes vented through cyclones or bag filters considered to be
an integral part of the process.
- Processes vented through stacks serving exhaust or ventilation
system normally associated with a process for safety and health
purposes.
- Process vents with defined airflow sufficient to resemble stack
exhaust.
- Exhaust emissions from mobile combustion sources.
Fugitive emissions include
- Open sources with no stack or associated ventilation system.
- Storage vessels.
- Material handling and transfer losses not normally collected by
cyclones or bag filters as an integral part of the process.
- Leaks in process equipment.
- Evaporative emissions not collected and vented through an exhaust
hood and ventilation system.
- Dust entrainment by motor vehicles.
- Evaporative emissions from motor vehicle fuel tanks/crankcases.
- Wind erosion of open storage piles.
Fugitives are shown in Table 1 to account for 90 percent of total
particulate emissions. Major categories include resuspended dust from highway
vehicles, burning of trash and agricultural land, forest and structural fires,
and industrial processes. Industrial fugitives account for 70 percent of all
industrial emissions with the major sources being iron and steel, primary
copper, cement, crushed stone, brick manufacturing, coal strip mining, and
grain elevators.
-------
Table 2 gives a national estimate of VOC emissions. Total VOC emissions
are about 40 percent from fugitives, with the major source categories being
transportation, burning, solvent use, and industrial processes from petro-
chemicals, crude and gasoline storage and transfer, degreasing, dry cleaning
and surface coating.
Controls of fugitive sources are usually tailored to the specific situation.
Typical control techniques include containment and control, surface treatment,
improved seals and equipment design, modified operating practices and in some
cases changing the process itself. Table 3 summarizes typical control tech-
niques for various fugitive sources. Several novel techniques being explored
are road carpet as an underlayment for unpaved roads and a charged droplet
fog spray for treating either confined or unconfined plumes.
AIR QUALITY IMPACT OF MAJOR FUGITIVE SOURCES
Volatile organic compounds affect the amount of ozone in the air. These
VOC provide a means of converting ambient nitric oxide (NO) to nitrogen dioxide
(N02) which does not consume ambient ozone (03). Subsequently, the resulting
N02 reacts with sunlight to form 03. The net result of these chemical reactions
is a buildup in ambient levels of ozone. The resulting high concentrations of
ozone tend to occur on an urban or regional scale.
Many studies which interpret particulate air quality and emissions data
to estimate source contributions can give some indication of the source con-
tributions as stack versus fugitive. An example of this would be studies using
optical microscopic analysis of hi-volume sampler filters. Table 4 summarizes
the results of the analysis of 300 filters from 14 U.S. cities.1* Using the
component categories defined by the microscopist, it is possible to make a
rough estimate of the fugitive source contributions to ambient air. First,
assume that mineral matter comes mostly from windblown and mechanically
induced fugitive dust sources and that rubber and paper come primarily from
vehicle resuspension of tire particles and litter. Also, assume that fugitive
windblown dust from coal storage piles,, iron and steel fugitive sources and
starch from grain handling contributes in part to particulate ascribed to
these categories. It can then be seen that fugitive dust and emissions con-
tribute perhaps from 50-90+% of ambient! TSP concentrations.
Fugitive particulate emissions differ from most stack emissions in three
respects: 11 most of the emissions are larger particles and tend to settle
out relatively nerr the source; 21 the emissions are usually emitted at or
near ground level; and 3) emissions generally are at or near ambient tempera-
ture and undergo little thermal plume rise. These differences result in a
large gradient in ambient concentrations decreasing with distance from the
source. This has been demonstrated both with modeling studies and with ambient
monitoring studies.5,6 The following is a brief discussion of the ambient
impact of selected sources.
Unpaved Roads. The Lincoln-Lancaster County (Nebraska) Health Department
and the EPA performed a study of short term ambient concentrations around an
unpaved road in Lincoln, Nebraska.7,8 For a one hour period, the traffic on
a road in a rural area southwest of Lincoln was increased to over 100 vehicles
per hour. Hi-volume samplers and dichotomous samplers were located at distances
of 10, 30 and 50 meters from the roadside.
-------
Typical concentrations were very high because short term measurements
were taken under peak traffic conditions. Hi-vol readings were several thousand
yg/m3, and typical dichotomous sampler readings were in the low hundreds
(yg/m3) range. The concentrations at the 30 and 50 meter distances are plotted
in Figure 1 as a fraction of the 10 meter hi-vol concentration. It can be seen
that the hi-vol concentration is reduced substantially at 30 and 50 meters,
but little decrease is seen in the dichotomous data which measures only those
particles smaller than 15 ym. This suggests that many of the larger particles
are settling to the ground very close to the source.
Paved Road Sources. Data have been collected at many locations which
suggest a high partlculate matter concentration around paved roads. Studies
in Boston,9 Kansas City10 and Hamilton, Ontario11 all show these high concen-
trations which decrease rapidly within 75 meters of the road. An empirical
study of data at 79 sites found a relationship between average daily traffic
(ADT), height of the monitor (HGT) and distance from the road (DIS):
concentration * ln
/ HG'F + DI52
Typically, a paved road might contribute an average of 15-25 yg/m3 at sites
located only 4 meters from a typical heavily travelled paved road.5 Current
EPA monitor siting guidance suggests that monitors be located outside of this
zone of direct influence. Typically, a monitor might be placed at least
25 meters away from a road and the contribution of that road would probably
be less than 5 yg/m3 at that point. This contribution can be much higher if
the road were particularly dusty or dirty.
Industrial Sources. Many industrial sources have been sampled by upwind-
downwind ambient techniques to estimate the emission factors for processes.
One recent study was conducted by the National Crushed Stone Association to
estimate emissions from construction aggregate crushing operations. Table 5
summarizes the results of instantaneous ambient measurements of particles
smaller than 10 ym at a distance of 30 feet downwind of uncontrolled crushing
operations. The table shows that significant dust levels are found just down-
wind of these operations. One would not expect fencellne concentrations to be
this great.12 However, one would expect much higher concentrations if particles
larger than 10 ym were included 1n the study.
EMISSION MEASUREMENTS
Two new measurement methods are being considered for proposal along with
the New Source Performance Standards (NSPS) for the Synthetic Organic Chemical
Manufacturing Industries (SOCMI) and Metallic Minerals Industries.
Method 21 applies to the determination of volatile organic compound leaks
from organic process equipment. Specifications and performance criteria for
a portable instrument will be specified under this method. Method 22 is
concerned with documenting the presence and duration of visible plumes from
various partlculate fugitive emission sources. It will specify such parameter*
and how the emissions are to be observed and recorded.13
-------
Several of the more established methods for measuring fugitive particulate
emissions are being documented in a sampling protocol.11* These methods include
roof monitor, quasi stack, upwind downwind and exposure profiling. These
methods are being employed in a major research effort to determine emission
factors for particulate matter. The program, being coordinated by the EPA's
Industrial Environmental Research Laboratory, is a major program measuring
particulate emissions by particle size from the major stack fugitive source
categories. New instruments being used in this program include a size selective
inlet for the high volume sampler for use in upwind/downwind sampling and a
horizontal elutriator which will be used in conjunction with the roof monitor
method. A new sampling train for IP has been developed which uses cyclones to
provide data for 15 pro particles.
REGULATORY PROGRAMS
New Source Performance Standards (NSPS) and National Emissions Standards
for Hazardous Air Pollutants (NESHAPS) are beginning to address fugitive organic
and particulate sources. Fugitive emissions are of concern in both TSP and
ozone State Implementation Plans (SIP's). Also, programs such as the bubble
policy and PSD are affected by fugitive emissions.
NSPS for Particulate Matter — Basic industrial process fugitive emission
sources have been addressed in prior NSPS. These include such operations as
material transfer, crushing, grinding, process equipment leaks, etc. Examples
of this include the NSPS for lime plants, asphalt concrete plants, copper,
lead and zinc smelters, coke ovens and grain elevators. Only recently, however,
have the emissions from windblown and vehicle entrained particulate sources
been mentioned in connection with NSPS. The upcoming NSPS for the metallic
minerals industries and gypsum plants may include provisions for storage piles,
tailing piles, and haul road emissions within and near the plant boundary.
A new concept is being considered in developing a NSPS for metallic
minerals. This is the proposal of Method 22 for visible emissions, as
mentioned earlier. This method will provide for documenting the duration of
a visible plume from specific process points or operations. NSPS regulations
in the future may specify the percentage of time that specific processes must
be void of visible emissions.13
NSPS for Volatile Organic Compounds CVOC) — New Source Performance
Standards for VOC are beginning to consider fugitive emissions. Two main
categories of sources, petroleum refining and Synthetic Organic Chemical
Manufacturing Plants (SOCMI) are being considered for control of leakage from
pumps, valves, flanges, compressors and cooling towers. Regulation may require
specific hardware such as double mechanical seals for pumps, or intermittent
monitoring with portable hydrocarbon monitoring devices for such sources as
valves and flanges. This could be a significant consideration since the SOCMI
produces 350 to 400 basic chemicals at facilities which contain almost 50,000
pumps, 800,000 values and 1.8 million flanges. Other types of sources such as
storage tanks for volatile organic liquids may follow the recently promulgated
Petroleum Storage Tank NSPS and require floating roofs.15,16
A third major category of VOC fugitive emissions is the coatings
industry. Eight new NSPS are being developed now with seven of these scheduled
for proposal this year. Regulation will probably be based on the emissions
-------
allowable per volume of solids applied. Controls may be through increasing the
solids/solvent ratio, collection by solvent recovery systems, such as carbon
adsorption, or by incineration.17
NESHAPS — Section 112 of the Clean Air Act, as amended in 1977, requires
the establishment of emission standards for hazardous air pollutants. Six
pollutants have already been listed and NESHAPS have been, or are currently
being established. These pollutants are: mercury, beryllium, asbestos, vinyl
chloride, benzene, and radionuclides. Nine other compounds or elements are
currently being considered for listing as hazardous pollutants; these are
listed in Table 6. Of these, arsenic, polycyclic organic matter, perchloroety-
lene, tricholoethylene, acrylonitrile, and cadmium seem to have a higher prob-
ability of being listed.18
Fugitive emissions may be a concern with all of these compounds.
Certainly, arsenic sources have been identified as major contributors of fugi-
tive emissions and ROM's are associated with coking operations. Acrylonitrile
storage tanks are a source of fugitive emissions.19 Cadmium has not yet been
assessed by the EPA for fugitive emissions.
TSP SIP — Generally, States in nonattainment areas are expected to
regulate industrial process fugitive paniculate emissions. However, the policy
for urban fugitive dust emissions is less clear. The EPA and the States/locals
are planning a series of demonstration studies to evaluate the effectiveness of
controls on a small scale. Some of these studies are listed in Table 7.
Originally, it was thought that many other areas would be proposing demonstra-
tion studies at this time. However, many of these are delayed pending a more
complete understanding of the ambient contribution of urban fugitive dust
sources in various particle size ranges of interest.20
Ozone SIP — The 1979 ozone SIP's generally called for the States to
adopt regulations for a certain group of hydrocarbon sources (Group I). These
Included the coil, can and paper coating Industries, degreasing and gasoline
marketing. These regulations must be equivalent in stringency to the levels
specified in the control techniques documents for these sources. The States
must also commit to adopt regulations for a second set (Group II) of sources.
Such sources Include petroleum refinery fugitive emissions, miscellaneous
metals coatings, flatwood paneling, graphic arts, synthesized pharmaceutical
products, tire manufacture, perchoroethylene dry cleaners, gasoline tanker
trucks, vegetable oil manufacture, and petroleum liquid storage in external
floating roof tanks.20,21 All of these sources have a significant part of
their emissions from fugitive sources.
Bubble Policy — The EPA adopted a new "bubble" concept to consider the
control of certain 1n-plant fugitive sources 1n Heu of placing additional
control on certain stack emissions within a given facility. Where a source
wishes to trade open source dust controls for the more significant sources of
process emissions* the bubble policy allows sources to demonstrate the equiva-
lency of such trades by Installing the open source controls and then monitoring
the results.22
-------
Prevention of Significant Deterioration — In the past, fugitive emissions,
as well as point or stack emissions, were used In determining if a source were
large enough to be covered by PSD regulations. As a result of a recent court
suit brought by Alabama Power and Light, (AP&L) and other plaintiffs, the EPA
is undertaking special rulemaking to define the manner in which fugitive emis-
sions from certain source categories are considered. Fugitive emissions will
still be used in calculating the increment consumed by a source, since this
was not affected by the AP&L case. However, until the rulemaking is complete,
potential emissions are being calculated from stack emissions only for purposes
of determing if a source is subject to PSD.22
-------
REFERENCES
1. National Air Pollutant Emission Estimates, 1970-78, EPA-450/4-80-002, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina,
in preparation.
2. Assessment of Fugitive Particulate Emissions Factors for Industrial
Processes, EPA-450/3-78-107, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, September 1978.
3. Compilation of Air Pollutant Emission Factors, AP-42, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, August 1977.
4. F. A. Record, National Assessment of the Urban Particulate Problem.
Volume I, National Assessment, EPA-450/3-76-024, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, June 1976.
5. T. G. Pace, An Empirical Approach for Relating TSP Annual Concentrations
to Particulate We™ inventory Emissions Data and Monitor Siting
Characteristics, EPA-450/4-79-012, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, June 1979, page 14.
6. Personal communication between T. G. Pace and E. Burt, Monitoring and
Data Analysis Division, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina, February 1980.
7. D. W. Safriet and T. G. Pace, "Preliminary Analysis of Sampling Around
Unpaved Roads in Lincoln, Nebraska," U. S. Environmental Protection
Agency, Research Triangle Park, North Carolina, October 1978, unpublished,
8. Personal communication between T. G. Pace and Gary Walsh, Lincoln-
Lancaster County Health Department, Lincoln, Nebraska, October 1978.
9. R. M. Bradway and F. A. Record, "Program to Measure, Analyze and Evaluate
Suspended Particulates in Massachusetts," GCA Technology Division, Bedford,
Massachusetts, Monthly Progress Reports, prepared for Massachusetts
Technology Development Corporation, Wakefield, Massachusetts, 1978.
10. Lead Analysis for Kansas City and Cincinnati, Final report from PEDCo
Environmental to U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina, under Contract No. 68-02-2515, Task Order No. 5,
June 1977, pages 22-23.
11. Personal communication between T. G. Pace and David Kane, United
Technology and Science, Toronto Ontario, Canada, January 1980.
12. An Investigation of Particulate Emissions from Construction Aggregate
Crushing Operations and Related New Source Performance Standards,
National Crushed Stone Association, National Industrial Sand Association,
Associated General Contractors of America, National Sand and Gravel
Association, Washington, D.C., December 1979.
8
-------
13. Personal communication between T. G. Pace and Gilbert Wood, Emission
Standards and Engineering Division, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, April 9, 1980.
14. Protocol for the Measurement of Inhalable Participate Fugitive Emissions
from Stationary Industrial Sources, TRC Environmental Consultants,
Wethersfield, Connecticut, draft being prepared for U.S. Environmental
Protection Agency under Contract No. 68-02-3115, Task No. 114, March 1980.
15, D. G. Erikson and V. Kalcevic, "Emissions Control Options for the
Synthetic Organic Chemicals Manufacturing Industry," draft report
prepared for the Emission Standards and Engineering Division, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina,
February 1979, page 2-14.
16. Personal communication between T. G. Pace and Susan Wyatt, Emission
Standards and Engineering Division, U.S. Environmental Protection Agency,
Reseach Triangle Park, North Carolina, April 11, 1980.
17. Personal communication between T. G. Pace and William Tippett, Emission
Standards and Engineering Division, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, April 11, 1980.
18. Personal communication between T. G. Pace and John Fink, Standards and
Air Strategies Division, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina, April 9, 1980.
19. "Source Assessment: Acrylonitrile Manufacture," Monsanta Research
Corporation, Dayton, Ohio, NTIS No. PB-271-969, September 1977, page 7.
20. Personal communication between T. G. Pace and Brock Nicholson, Control
Programs Development Division, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, April 14, 1980.
21. Stephen V. Capone and Malcolm Petrorcia, Summary of Group II Control
Techniques Guideline Documents for Control of Volatile Organic Emissions
from Existing Stationary Sources. EPA-450/2-80-001, December 1979.
22. Personal communication between T. G. Pace and Kirt Cox, Control Programs
Development Division, U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina, April 14, 1980.
-------
TABLE 1. Estimate of Particulate Matter Emissions (103 tons/yr)
Source Category
Transportation
Highway Vehicles
LDV
LOT
HDG
MC
HDD
Railroads
Aircraft
Vessels
Misc. off-highway-gasoline
Misc. off-highway-diesel
Stationary Fuel Combustion
Industrial Processes
Iron and steel
Primary aluminum
Primary copper
Primary lead
Primary zinc
Ferroalloys
Gray iron foundries
Secondary non-ferrous meteJls
Chemicals
Cement
Sand and gravel
Crushed stone
Phosphate rock
Brick and clays
Lime
Glass
Asphalt batching
Concrete batching
Coal cleaning
Coal strip mining
Gypsum
Asphalt Roofing
Pulp mills
PIywood
Cotton ginning
Grain elevators
Food and grain mills
Cattle feed lots
Petroleum refining
Solid Waste
Incinerators
Open burning
Miscellaneous
Forest Fires
Agricultural Burning
Structural fires
Unpaved road vehicle travel
Stack Emissions
884
537
308
57
75
4
93
69
82
28
16
152
4150
2097
225
67
35
4
1
39
30
68
189
80
0
0
16
185
116
29
127
18
15
0
66
8
263
26
13
395
1
0
81
219
219
0
0
0
0
0
0
TOTAL 7350
Fugitive Emissions
395 (5354)
395 (4950)1
257
48
25
0
65
0
<9)2
0
0
0
4666
6903
39
3223
373
203
19
123
12
19
78i
47
1478
30
296
48
0
36
92
0
195
1
0
0
0
0
349
7
25
0
256
0
256
735 (38,735)
574
92
69
(38,000)"
6052 (49,011)
1 NEDS area source total shown in parentheses represents entrainment of
dust from paved roads. Tranportation fugitive total also includes tire wear.
2 NFDS area source total for unpaved airstrips.
3 Includes ore mining emissions all assumed to be fugitive.
u NEDS area source total for unpaved roads. Other totals shown in parentheses
include this total.
10
-------
TABLE 2. Lstimate of National VOC Emissions (10' tons/yr)
Source Category Stack Emissions Fugitive Frissions
Transportation
Hiqhwdy Vehicles'
LDV
LDT
HOG
HC
HDD
Railroads
Aircraft
Vessels
Misc. off-hiqhwdy-gasoline1,2
Misc. off-hiqhway-diesel3
Stationary fuel combustion
Industrial Processes
Petroleum refining
Petrochemicals1 ]
8459
7035
4257
1038
1356
174
210
188
227
476
361
172
326
8640
617
758
3266
3167
2327
577
227
36
0
0
0
0
99
0
6322
343"
Crude oil and natural gas
production 0 2566
Crude oil and gasoline
storage tanks7 0 627
Gasoline bulk terminals/plants8 0 615
Gasoline service stations9 0 1082
Gasoline/crude oil transfer1^ o 572
Plastics 374 0
Other Chemicals11 583 0
Adhesives 254 28
Coke 0 49
Fermentation 0 90
Vegetable oil 21 0
Bakeries 46 0
GTass 2 64
Dry cleaning 100 258
Degreasing 0 571
Graphic arts12 348 39
Surface coating13 2304 256
Misc. organic solvent11* 3060 340
Otner processes1" 173 19
Solid Waste Disposal 380 480
Incineration 380 0
^pen burning 0 480
Miscellaneous 0 2815
Forest fires 0 642
Agricultural burning 0 85
Structural fires 0 47
Architectural coating 0 331
Cutback asphalt paving 0 686
Miscellaneous solvent use 0 1024
TOTAL 17,805 12,883
1 Based on classification of exhaust emissions as stack emissions and evaporatlve/crankcase emissions as fugitive emissions
2 Includes farm, construction. Industrial engines, snowmobiles, lawn and garden tractors, and off-highway motorcycles.
3 Includes farm, construction, industrial engines.
** Includes emissions from process drains, cooling towers, asphalt blowing, valves and seal leaks, sampling, etc.
5 Assume estimated emissions from storage and handling losses, and other fugitive sources as fugitive emissions.
6 Assumed all fugitive stack emissions from fuel combustion equipment Included under stationary fuel combustion.
7 Includes tanks at oil fields and refineries only.
8 Storage tanks only.
9 Stage I and Stage II.
10 Includes ship and barge transfer and filling, rail and truck tankers.
11 Petrochemicals include all products 1n SIC 2865, 2869. Other chemicals Include other organic and inorganic chemicals.
12 Assumes that solvent vapors are collected by an exhaust system and vented through a stack. It is assumed that lot of
emissions are fugitive, escaping the exhaust system or carried out with products and evaporated later (allowance).
13 Assumes emissions occur from spray booths and baking ovens vented to stacks. Evaporative emissions from flash-off of
solvent are assumed to be ICO! of total emissions (allowance).
i* Assumed to be predominantly stack emissions, allowance IDS of total emissions reported as fugitive.
11
-------
TABLE 3
TYPICAL CONTROL TECHNIQUES FOR VARIOUS FUGITIVE SOURCES
OPEN SOURCES
STORAGE VESSELS
MATERIAL HANDLING AND TRANSFER
LEAKS IN PROCESS EQUIPMENT
PROCESS EVAPORATIVES
VEHICLE DUST ENTRAINMENT
VEHICLE EVAPORATIVES
WIND EROSION
VARIOUS (FOGGERS. WATERING. OPERATING PRACTICE)
FLOATING ROOFS/SEALS, VAPOR COLLECTION/RECOVERY
CONTAINMENT, WATER SPRAYS, FOGGERS
IMPROVED SEALS/EQUIPMENT DESIGN, ROUTINE INSPECTIONS
IMPROVED MAINTENANCE/OPERATING PRACTICES
VAPOR COLLECTION/RECOVERY, SOLVENT SUBSTITUTION,
SOLIDS - SOLVENT RATIO
PAVED - STREET CLEANING, DEPOSITION CONTROLS
UNPAVED - PAVING, CHEMICAL STABILIZATION, ROAD CARPET
PCV, CONTAINMENT
PILE CONTOURING, WIND BREAKS, CONTAINMENT,
WATERING, CHEMICAL STABILIZATION
-------
TABLE 4
U.S. COMPOSITE SUMMARY OF FILTER ANALYSES
COMPONENTS
MINERALS
* QUARTZ
* CALCITE
* FELDSPARS
* HEMATITE
* MICA
OTHER
COMBUSTION PRODUCTS
SOOT:
OIL
* COAL
MISC. SOOT
GLASSY FLY ASH
INCINERATORY FLY ASH
OTHER
BIOLOGICAL MATERIAL
POLLEN
SPORES
* PAPER
* STARCH
MISC. PLANT TISSUE
LEAFTRICHOMER
MISCELLANEOUS
* IRON OR STEEL
* RUBBER
OTHER
QUANTITY, percent
AVERAGE
(65)
29
21
5
10
1
1
(25)
7
5
5
6
2
1
( 3)
1
1
1
1
1
1
( 7)
1
7
1
RANGE
3-99
1-84
1-93
0-35
0-65
0-15
0-46
1-89
0-86
0-52
0-88
0-30
045
0-19
0-90
0-45
0-2
0-3
0-10
0-8
0-18
0-50
0-25
0-50
0-15
13
-------
TABLE 5
MEAN DUST CONCENTRATIONS FOR PARTICLES
SMALLER THAN 10 jum MEASURED 30 FEET DOWNWIND
PLANT
A
B
C
D
E
F
G
STONE TYPE
GRANITE
SAND & GRAVEL
SAND & GRAVEL
TRAPROCK
LIMESTONE
LIMESTONE
LIMESTONE
PRIMARY
CRUSHER
200 ± 853
130 ± 170
170± 140
640 ± 590
SECONDARY
CRUSHER
350 ± 210
440 ± 390
570 ± 200
930 ± 640
68 ± 43b
170±640
2200± 1000
TERTIARY
CRUSHER
11401740
210 ± 290
810 i 450
FINES
CRUSHER
1600 1 290
410 1 280
a(STANDARD DEVIATION)
&WET PROCESS
-------
TABLE 6. Compounds Currently Being Considered for Listing as
Hazardous Pollutants
Arsenic
Polycyclic Organic Matter
Perch!orethylene
Tri chloroethy1ene
Acrylonitrile
Cadmium
Methyl Chloroform
Methylene Chloride
Toluene
a Compounds listed first have a higher probability of being listed.
15
-------
TABLE 7
DEMONSTRATION STUDIES PLANNED BY STATES AND THE EPA
TO EVAULATE CONTROLS FOR FUGITIVE DUST SOURCES
STUDY
CONTROLS EVALUATED
PORTLAND, OREGON
DENVER, COLORADO
MINNEAPOLIS, MINNESOTA
LAS VEGAS, NEVADA
TUCSON, ARIZONA
VACUUM SWEEPING OF PAVED STREETS IN
INDUSTRIAL AREA
REDUCE FUGITIVE DUST DUE TO SANDING
FOR SNOW CONTROL
CONSTRUCTION GENERATED DUST- ESPECIALLY
MUD CARRYOUT
PAVING OF GRAVEL ROADS AND MODIFY
SUBDIVISION CONSTRUCTION PRACTICES
CONTROL OF TAILINGS PILES AND
UNPAVED ROADS
-------
FIGURE 1
RELATIVE SHORT TERM CONCENTRATIONS MEASURED
BY DICHOTOMOUS SAMPLERS AND HI-VOLS AT DISTANCES
OF 10, 30 AND 50 METERS FROM AN UNPAVED ROAD
1.0
£
Z 0.8
E
o
T—
O
0.6
>
H
HI
cc 0.4
O
0.2
o
z
o
o
.SAMPLER
I
s
I I I
10 30
DISTANCE FROM ROAD, meters
50
17
-------
FUGITIVE EMISSIONS CONTROL AFTER ALABAMA POWER
DAVID I. BRANDWEIN
ASSOCIATE ENVIRONMENTAL COUNSEL
TRC ENVIRONMENTAL CONSULTANTS
WETHERSFIELD, CONNECTICUT
MAY, 1980
ABSTRACT
Recent studies indicate that traditional stack-oriented control
strategies 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
emission sources will also be required in order to comply with EPA's
regulations on Prevention of Significant Deterioration (PSD). This paper
attempts to summarize the federal regulations and policies applicable to
fugitive particulate matter sources under the Clean Air Act. Pertinent
provisions of EPA's regulations and guidelines for nonattairunent and PSD
areas are discussed. Particular emphasis is given to discussing the impact
of the recent Alabama Power decision and EPA's proposed regulations of
September 5, 1979 dealing with PSD and nonattainment new source review.
18
-------
INTRODUCTION
On June 18, 1979, the United States Court of Appeals for the District
of Columbia issued its initial ruling in the case of Alabama Power Company
v. Costle.1 This decision made substantial changes to EPA's regulations
for preventing significant deterioration (PSD) of clean air. Due to the
close relationship between the PSD program and EPA's regulations for
nonattainment areas, the Alabama Power decision will affect both sets of
regulations.
The issues addressed in Alabama Power cover the .full spectrum of the
PSD regulations promulgated by EPA in mid-1978. The decision directly dealt
with how fugitive emissions may be regulated under the Clean Air Act and
with EPA's fugitive dust exemption. Perhaps more importantly, through its
broad discussion of what sources and modifications are subject to review,
which pollutants must be considered, and how geographic location affects the
type of review, the Court indirectly made fugitive emissions the subject for
increased regulation in the future.
OVERVIEW OF DEVELOPMENTS SINCE THE CLEAR AIR ACT AMENDMENTS OF 1977
The 1977 Amendments to the Clean Air Act were the foundation for
today's PSD and nonattainment new source review program. The 1977 Amend-
ments gave rise to EPA's 1978 PSD regulation which, in turn, directly lead
to the court challenge resulting in Alabama Power. Alabama Power, both
directly and indirectly, will force major substantive changes in EPA's
regulation of fugitive emissions and to EPA's PSD and nonattainment new
source review program. Thus, it is useful to understand these post-1977
developments.
Nonattainment
The widespread failure of many of the nation's air quality control
regions (AQCR's) to attain the National Ambient Air Act Standards (NAAQS) by
the original July 1, 1975 deadline created a dilemma concerning the legal
approvability of new sources which would add to levels of pollution already
in violation of the law.
EPA responded to this problem with promulgation of its national Offset
Policy on December 21, 1976.2 The essence of the Offset Policy was that
major new growth should be allowed in nonattainment areas only if air
quality is improved as a result of the growth. This improvement comes by
way of contemporaneous emission reductions from existing sources so as to
more than "offset" the emissions which will be added by the proposed new
source.
19
-------
The Offset Policy imposed 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 in compliance or on a schedule of compliance with the
State Implementation 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 (SIP) 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
were to have submitted plan revisions to EPA which conform to new Part 0 of
the Act. These revised plans are designed to replace EPA's Offset Policy
and were to have been approved by no later than June 30, 1979. Failure to
adopt and receive EPA approval of the revised plan by this deadline triggers
a statutory ban on major new sources (of the nonattaining pollutant) in the
nonattainment area. Many states have incorporated elements of the Offset
Policy into their SIP.
Table I summarizes the development of the nonattainment provisions and
the Offset Policy.
Prevention of Significant Deterioration
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 unimportant phrase
in the original law 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 did not prevent
significant deterioration of clean air.
This challenge was ultimately successful, and on December 5, 1974, EPA
promulgated regulations to prevent emissions of sulfur dioxide and particu-
late matter from significantly deteriorating air quality in areas where
concentrations of those pollutants were lower than the applicable national
ambient standards.3 EPA incorporated its PSD regulations into the
implementation plan of each state pursuant to Section 110(c) of the Act and
established a procedure by which EPA could delegate its PSD responsibility
to states.
10
-------
TABLE I
CHRONOLOGICAL DEVELOPMENT
OP EPA'S OFFSET POLICY
AND THE ACT'S NONATTAINMENT PROVISIONS
Original Deadline for Attaining the Primary NAAQS
7/1/75
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
Revised SIPs Due to EPA
1/1/79
EPA Promulgates Major Revisions to Its Offset Ruling
1/16/79
Initial Ruling in Alabama Power vs. Costle
6/18/79
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
EPA Proposes Revisions to PSD and Nonattainment Rules
9/5/79
EPA Publishes Final Revisions to PSD and Nonattainment Rules
6/80
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
21
-------
The regulations prohibited construction of stationary sources in any of
19 specified categories unless EPA (or a delegated state) had issued a
permit evidencing that the source would apply "best available control
technology" (BACT) for sulfur dioxide and 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 that system,
clean air areas could be classified as Class I, II, or III. In Class I
areas, small increases of sulfur dioxide and particulate matter would be
significant. In Class II areas, moderate increases were allowed, and in
Class III areas, increases were permissible up to an ambient standard. The
regulations initially classified all PSD areas as Class II, but gave states,
Indian governing bodies, and federal land managers the opportunity to
reclassify their lands under specified procedures.
The 1977 Amendments affirmed EPA's development of the PSD concept.3a
The new statutory scheme follows the outline of the pre-existing regula-
tions,4 but is generally more comprehensive and restrictive. Some of the
more significant changes introduced by the 1977 Amendments include:
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 indeter-
minate status) must be formually listed by EPA. EPA
issued this list on March 3, 1978.5 Several
revisions have been made since March 3, 1978.6
o More Restrictive "Increments" - The new law reduces
the allowable short-term increments for SC-2 and
particulates in Class II and III areas. Class I
increments remain the same.7
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
Amendments were enacted (8/7/77) are now designated
as mandatory Class I areas. These areas may not be
redesignated.8
o More and Different Sources Subject to PSD Review -
The new Act increases the number of source categories
subject to PSD preconstruction review from 19 to 28.
New and modified sources within one of these 28
categories are subject to PSD if they have the
potential 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 pollutant.9 Major sources of fugitive
dust were covered by the 250 ton criteria.
22
-------
o More Restrictive Definition of BACT - Under the
original PSD regulations, Best Available Control
Technology (BACT) could not be more restrictive than
the New Source Performance Standards (NSPS) appli-
cable 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, technology, 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 SO 2 and
particulate matter as under the prior regulations.1°
o Substantially Increased Monitoring and Modeling
Requirements - Under the original PSD Regulations,
applicants merely had to demonstrate, through
atmospheric diffusion modeling, that the proposed
emissions would not violate the allowable incre-
ments. The new law 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 - EPA's original PSD regulations applied only
to SO2 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, photo-
chemical oxidants, and nitrogen oxides) by no later
than August 7, 1979. These regulations will take
effect one year after promulgation and must utimately
be incorporated into all State implementation plans.
PSD regulations for lead must be promulgated in
1980.12
On November 3, 1977, EPA took four regulatory actions toward implement-
ing the Act's new PSD requirements.13 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 proposal of regulations giving guidance to the
preparation of SIP revisions called for by the new PSD review 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 preconstruc-
tion review 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.
23
-------
EPA formally promulgated its new PSD regulations and SIP revision
guidelines on June 19, 1978, three and one-half months after its original
self-imposed deadline of March 1, 1978.I4 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 pollution permits prior to March 1, 1978 had to commence
construction on or before March 19, 1979 in order to be exempt from the new
requirements. This was the deadline for the submission of revised SIPs to
EPA.
Immediately upon promulgation, the 1978 PSD regulations were challenged
by a number of industry and environmental groups, as well as state regula-
tory agencies. These individual lawsuits were consolidated into a single
action before the U.S. Court of Appeals for the District of Columbia. On
June 18, 1979, the court issued its initial ruling in Alabama Power Company
v. Costie.
The Alabama Power opinion was issued by the court "per curium" which
means that the specific opinions of individual judges of the court were not
included. While the opinion reflected the court's final rulings on the
questions presented by the petitions for review, more detailed opinions of
the court were promised in the late summer or early fall of 1979. EPA's
response to the court's remand order was expected to follow the issuance of
these detailed opinions. However, when it became apparent that the Court
would not issue its supplemental ruling until late in the year, EPA decided
to go ahead and propose its new regulations.
On September 5, 1979, EPA proposed comprehensive changes to its PSD
regulations in conformance with the Alabama Power summary opinion.15 At
the same time, the Agency proposed an overhaul of its nonattainment new
source review guidelines to make them consistent with the proposed PSD
provisions.
The D.C. Circuit Court of Appeals issued its supplemental opinion in
Alabama Power on December 14, 1979.16 With two exceptions, the final
opinion made no major changes to the summary opinion. ^ EPA is now
finali2ing its regulations for implementing the final Alabama Power ruling.
These regulations are expected to be promulgated in June 1980.
Table II summarizes the development of the PSD program.
NEW SOURCE REVIEW AND FUGITIVE EMISSIONS
New source review is the regulatory program under which major new
sources and major modifications to existing sources undergo a preconstruc-
tion review process, either under PSD (generally implemented by an EPA
Regional Office) or under nonattainment (generally implemented by a State
through its SIP).
24
-------
TABLE II
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
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
6/19/78
Major Sources with Allowable Emissions in Excess
of 50 Tons Per Year May 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
Initial Ruling in Alabama Power Co. vs. Costle
6/18/79
Deadline For EPA Approval or Disapproval of the Revised SIPs
7/19/79
EPA Proposes Revisions to PSD and Nonattainment Rules
9/5/79
EPA Publishes Final Revisions to PSD
and Nonattainment Rules
6/80
25
-------
There are two aspects to new source review:
1) an applicability determination to decide what source
or modification is subject to review, what pollutants
must be reviewed/ and whether PSD or nonattainment
(or in some cases, both) rules apply; and
2) a substantive review, after the applicable rules are
identified, to determine what technology-based
standards apply (BACT, LAER), what air quality
analyses must be performed/ and so on.
Alabama Power focused on the applicability issues and so will this
paper. The substantive issues can be briefly summarized as follows:
For PSD;
1) The application of Best Available Control
Technology (BACT);
2) Ambient air quality analyses (modeling, monitor-
ing, and other impact analyses); and
3) Consumption of the PSD increments.
For Nonattainment;
1) The application of Lowest Achievable Emission
Rate (LAER);
2) Certification of statewide compliance for all
major sources by the owner;
3) Obtaining of offsets (emissions reductions) from
existing sources in the area of the proposed
source; and
4) A net air quality benefit and reasonable progress
toward attainment of the NAAQS.
It is important to note that PSD rules apply in any area where at least
one NAAQS is attained.18 Because almost every area in the U.S. is
attainment for at least one criteria pollutant, the PSD review will be
required for major new and modified sources, regardless of where they
locate, if the source is "major" for the attainment pollutant. Accordingly,
most discussions of new source review, including this paper, focus on PSD.
The remainder of this paper will focus on four specific issues and how
each was addressed in a) EPA's 1978 PSD regulations, b) Alabama Power, and
c) EPA's proposed PSD and nonattainment rules of September 5, 1979:
1) What sources or modifications are subject to new
source review?
26
-------
2) What pollutants are to be included in the review?
3) How does the source's location in a "PSD area" (an
attainment or unclassified area) or nonattainment
area affect the review? and
4) What is the status of the fugitive dust exemption?
Before beginning the discussion of these issues, it is useful to
clarify what is meant by "fugitive emissions" and "fugitive dust".
The Clean Air Act does not provide defintions of these terms. Fugitive
emissions are generally understood to refer to emissions from a stationary
source that emanate from other than a stack or chimney. As discussed later
EPA has proposed to define the term "fugitive emissions" as those emissions
which do not pass through an opening which the owner or operator uses for
ventilation, such as a stack, chimney, roof vent or roof monitor.19
Fugitive emissions may be particulate matter, volatile organic compounds
(VOC's) or any other pollutant.
On the other hand, fugitive dust is uncontaminated soil particles which
become airborne as a result of human activities.20 The most common
sources are haul roads, exposed soil surfaces, and soil storage piles.
Fugitive dust is particulate matter only.
Source Applicability
Section 165 of the Act imposes the new PSD preconstruction review
requirements on "major emitting facilities" seeking to locate in any PSD
area. The term major emitting facility is then defined as any source in one
of 28 categories having a potential to emit 100 tons per year or more of any
pollutant regulated under the Act. The term also embraces any source in a
category other than one of the 28 listed categories with a potential
emissions' increase of 250 tons per year or more of any regulated pollutant.
The term "potential emissions" (or "potential to emit") is not defined
in the Act and was one of the issues under Alabama Power.
Under EPA's pre-Alabama Power regulations (hereafter, the "old"
regulations), EPA defined "potential to emit" for purposes of both PSD and
nonattainment new source review, as a source's maximum uncontrolled
emissions. Sources with "potential" emissions in excess of the PSD
thresholds (100 or 250 tons, depending on the source category) were deemed
to be "major stationary sources," and, therefore, were subject to PSD.
Sources could escape most of the PSD requirements on a pollutant-by-pollu-
tant basis, however, if that pollutant's "allowable" emissions could be cut
to below 50 tons per year. The basic goal of this approach was to encourage
sources to reduce emissions below the 50 tpy level in order to qualify for
an expedited and less expensive review.
EPA included fugitive emissions in calculating a plant's potential
emissions as a matter of policy. Surface mines were considered as major
sources even if their potential emissions were solely fugitive.
27 -
-------
The court in Alabama Power reversed EPA and held that a source's
potential to emit must reflect emissions after controls are incorporated.
Major new stationary sources must be defined on the basis of their actual
allowable emissions.
On the inclusion of fugitive emissions towards a source's potential to
emit, the Court was quite specific. The court held that, although EPA does
have the authority to regulate fugitive emissions and fugitive dust,
generally, and mining operations specifically, EPA must address the
inclusion of fugitive emissions through rulemaking, as required under
Section 302(j) of the Act. The court, therefore, remanded the PSD regula-
tions as they applied to fugitives, and directed EPA to accomplish its
original objectives through formal rulemaking.
To the surface mining industry's arguement that a surface mine is not a
source because a mine is not within the scope of the definition of source
("a building, structure, facility, or installation"), the Court stated that
EPA had the discretion to define the term "facility" by the concept of
"common sense industrial groupings" so as to encompass a source of fugitive
emissions such as a major mining operation. Again, however, EPA would have
to do so through rulemaking.
In response to the court's decision EPA proposed to amend the defini-
tion of "potential to emit" in the existing regulations. As proposed, the
new definition would provide that the term means "the capability at maximum
capacity to emit a pollutant after the application of air pollution
control equipment." However, the agency has incorporated an irrebuttable
presumption into the definition that a source will operate at maximum
capacity around the clock, 365 days per year.
In response to the court's holdings, EPA also proposed that certain
industrial fugitive emissions, to the extent reasonably quantifiable, may be
included in determining whether a source or modification is "major". This
proposal affects the definition of the term "potential to emit" and this
applies to both PSD and nonattainment new source review.
First, EPA proposed to define "fugitive emissions" as those emissions
which do not pass through an opening which the owner or operator uses for
ventilation, such as a stack, chimney, roof vent or roof monitor.21 (EPA
would also delete the existing definition of fugitive dust at 40 CFR
51.24(b)(6) and 52.21(b)(6) (1978).)
Second, EPA proposed to incorporate into the existing regulations the
principle that rulemaking must precede the inclusion of "fugitive emissions"
in an applicability determination.2^
Finally, the Agency proposed to list the stationary sources in
Table III whose fugitive emissions are to be taken into account. Fugitive
emissions will also be counted if they are from any other stationary source
category which, at the time of the applicability determination, is being
regulated under section 111 or 112 the Act.
EPA will be considering during 1980 the need for additional source
types to be added to the list, including surface mines.
28
-------
TABLE III
1) Coal Cleaning Plants
2) Kraft Pulp Mills
3) Portland Cement Plants
4) Primary Zinc Smelters
5) Iron and Steel Mill Plants
6) Primary Aluminum Ore Reduction Plants
7) Primary Copper Smelters
8) Municipal Incinerators
9) Hydrofluoric, Sulfuric, or Nitric Acid Plants
10) Petroleum Refineries
11) Lime Plants
12) Phosphate Rock Processing Plants
13) Coke Oven Batteries
14) Sulfur Recovery Plants
15) Carbon Black Plants
16) Primary Lead Smelters
17) Fuel Conversion Plants
18) Sintering Plants
19) Secondary Metal Production Plants
20) Chemical Process Plants
21) Fossil Fuel-Fired Boilers
22) Petroleum Storage and Transfer Units
23) Taconite Ore Processing Plants
24) Glass Fiber Processing Plants
25) Charcoal Production Plants
26) Fossil Fuel-Fired Steam Electric Plants
29
-------
Before leaving the discussion of source applicability and fugitive
emissions it is important to note three activities dealing with coal
conversion that are exempted from new source review. These types of coal
conversion are not subject to new source review because, either by statute
or by regulation, they are not considered "modifications": Conversion to
coal (or other alternative fuel):
1) Pursuant to an order under the Energy Supply and
Environmental Coordination Act of 1974 (ESECA) or any
superceding legislation (such as the Power Plant and
Industrial Fuel Use Act of 1978);23
2) If, prior to January 6, 1975, the source was capable
of accommodating such fuel24. or
3) Pursuant to an order or rule under section 125 of the
Act ("Measures to Prevent Economic Disruption or
Unemployment").
These provisions were not under review in Alabama Power and remain
unchanged.
Pollutant Applicability
Once one knows whether a major source or major modification is being
proposed, one must determine which pollutants must be reviewed.
Under the "old" regulations, a source was subject to new source review
only for those pollutants for which it was "major" (i.e., emitted in amounts
greater than or equal to 100/250 tons per year). If a modification was less
than 100/250 tons per year for any pollutant, the modification was not major
and there was no new source review.
Alabama Power sent EPA back to the drawing board on its definition of
modification, holding that EPA's interpretation was inconsistent with the
language of the Act. Specifically, the court ruled that any (non-exempt)
modification of a stationary source is subject to the Act's PSD provisions
if there is any significant net increase in emissions. Thus, the 100/250
ton threshold was rejected.
The court did hold that an insignificant, or "de minimus", net increase
in emissions may be exempted by EPA. EPA has proposed a list of de minimus
levels for each pollutant regulated under the Act. (See Table IV.)
This change will have far greater effect for PSD review than for
nonattainment review because after Alabama Power many more pollutants have
been swept into new source review. Where previously review was required for
only "major pollutants", now review is triggered for all pollutants emitted
in greater than de minimus amounts once one pollutant becomes major.
A BACT review must be performed for each pollutant emitted in greater
than de minimus amounts. However, some pollutants may be exempted from the
30
-------
TABLE IV
EPA'S PROPOSED DE MINIMIS LEVELS
FOR DETERMINING "SIGNIFICANT NET INCREASE"*
Carbon monoxide—100 tons Per year.
Nitrogen dioxide—10 tons per year.
Total Suspended particulates—10 tons per
year.
Sulfur dioxide—10 tons per year.
Ozone—10 tons per year of volatile organic
compounds.
Lead—1 ton per year.
Mercury—0.2 tons per year.
Beryllium—0.004 tons per year.
Asbestos—1 ton per year.
Fluorides—0.02 ton per year.
Sulfuric acid mist—1 ton per year.
Vinyl chloride—1 ton per year
Total Reduced Sulfur:
Hydrogen sulfide—1 ton per year.
Methyl mercaptan—1 ton per year.
Dimethyl sulfide—1 ton per year.
Dimethyl disulfide—1 ton per year.
Reduced Sulfur Compounds:
Hydrogen sulfide (see above).
Carbon disulfide—10 tons per year.
Carbonyl sulfide—10 tons per year.
* Note - EPA has added a second set of ambient impact "de minimis" levels to
limit the air quality review requirements only for certain pollutants
which the source would have the potential to emit, in significant amounts
but which have an insignificant ambient impact.
31
-------
air quality analysis part of new source review if, as a result of screening
modeling, the pollutant is below a second de minimis level which is based on
air quality impacts.
For nonattainment review, pollutant applicability is far narrower. To
trigger new source review, a source or modification must be major for a
pollutant for which the area is nonattainment. And once nonattainment
review is triggered, other pollutants are not swept in as under PSD rules.
An example will illustrate this change. Under the "Old" rules, the
addition of a coal pile with fugitive emissions of 15 tons per year would
not result in new source review because the modification was not maior
( 100/250 tons per year). After Alabama Power and under EPA's proposed de
minimus levels, the addition of the coal pile is a major modification (being
greater than the 10 tpy de minimus level for particulates) , and a BACT
review for the coal pile would be required. However, detailed air quality
analyses may be avoided if the facility can demonstrate that the impact is
less than 5 ug/m3 (24 hour average).
Geographic Applicability
The next stop in the new source review applicability determination is
deciding which set of regulations apply for the pollutants subject to the
review - PSD, nonattainment of both.
Under the 1977 Clean Air Act Amendments, all areas of the nation have
been designated as subject either to PSD or nonattainment requirements.
With respect to new source review, however, the "old" EPA regulations
focused on the area impacted by the source, irrespective of the attainment
status of the area in which the source sought to physically locate.
Specifically, EPA's PSD regulations extended the permit requirements of
section 165 to all sources, wherever located, if the emissions from the
source had a signification impact on any clean air area.
The court held in Alabama Power that the Act's PSD requirements are not
triggered by sources located in nonattainment areas simply because they have
an adverse impact on a clean air area within the same State. Only where the
source is located in a nonattainment area of a neighboring state will its
impact on adjacent clean areas trigger the PSD review. The court's
reasoning is based upon a number of provisions in the Act which seek to
prevent interstate PSD conflicts. With respect to intrastate air quality
management planning, however, the court noted a "federalist policy" of
deferring to the States on matters relating to their internal growth
management. The final court ruling indicates that PSD permits cannot be
required for sources in nonattainment areas even if interstate impacts are
involved. SIP's, however, must address such impacts.
Although EPA has petitioned the DC Circuit for reconsideration on this
aspect of the decision, it appears doubtful that the agency will succeed in
changing the court's mind. In its September 5, 1979 proposed regulations,
EPA has, therefore, developed a change in prior policy designed to overcome
a potential gap in the New Source Review program which could materialize.
32
-------
Specificallyf under the "old" regulations, proposed major sources in
designated nonattainment areas with no significant impact on the violations
can be exempted from nonattainment requirements because they are made
subject to PSD preconstruction review. But, under Alabama Power and the
regulations now being proposed, such major sources would be able to avoid
both PSD and nonattainment NSR requirements.
EPA has determined that such a result is contrary to the intent of the
Clean Air Act and has, therefore, proposed to extend the nonattainment NSR
requirements of both the Offset Policy and Section 173 of the Act to cover
major sources everywhere in the designated nonattainment area.25 This
rule would apply regardless of whether the source could demonstrate that it
would not significantly impact the specific point(s) of violation.
The Fugitive Dust Exemption
Following the 1977 Amendments to the Clean Air Act, EPA created through
rulemaking a partial exemption from new source review requirements for major
emitting facilities of fugitive dust. The regulation maintains the
requirement that such sources apply best available control technology
(BACT), but exempts them from the otherwise-required showing that particu-
late emissions from the facility will not exceed either the applicable
national ambient air quality standards (NAAQS) or the allowable
increments.26
In Alabama Power, the court vacated (i.e., eliminated) the exemption
and sent it back to EPA for further consideration. As in the case of
fugitive emissions, the court held that EPA had the legal authority to
accomplish its stated objective, but had failed to comply with the rule-
making requirement imposed by the Act.
In remanding the fugitive dust exemption for agency reconsideration,
the court suggested a regulatory approach under which fugitive dust could
become a "regulated" pollutant (by controlling it under Section 111 of the
Act) but would not be a "criteria" pollutant within the meaning of the
national ambient standards or the PSD increments. Under this approach,
major new and modified sources of fugitive dust would have to control
emissions with BACT, but would not have to model its air quality impact with
respect to the ambient particulate increments or standards, thereby
achieving the same result as under the exemption.
However, EPA has proposed to eliminate the exemption entirely, both
from the PSD regulations and from a parallel provision in the offset
ruling.27 The agency gave no reason for its decision to eliminate the
exemption. It is important for surface mining operators to note, however,
that under the Surface Mine Control and Reclamation Act (SMCRA), the Office
of Surface Mining (OSM) within the Department of the Interior has begun to
regulate the fugitive dust aspects of mining in a manner substantially more
stringent than EPA. OSM regulations will require, among other things, the
preparation of a fugutive dust plan and the imposition of fugitive dust
control technology.28
33
-------
CONCLUSION
The Alabama Power decision will have a significant impact on the
regulation of fugitive emissions for the forseeable future. But interest-
ingly, the impact will not be due to the sections of the decision directly
addressing fugitive emissions and fugitive dust. The impact will come from
the greater number of sources and modifications subject to new source review
and EPA's implementation of the de minimus concept.
The court's holding that EPA may include fugitive emissions toward
determining a source's "potential to emit" in order to determine whether
there is a major source or modification only placed a temporary road block
in EPA's path. By the time this paper is published, EPA will have amended
the definition of "potential to emit" to include fugitive emissions to the
extent reasonably quantifiable for most major source categories (although
surface mines have not yet been included).
Of more concern is the whole change in the concept of "major modifica-
tion". Because this term now means, for PSD purposes, any significant net
increase of any pollutant once the source is major for any (including a
different) pollutant, there will potentially be many more "major modifica-
tions" and many more new source reviews required. (For nonattainment
purposes, the "significant net increase" must still be for the pollutant for
which the source is major).
This increase in PSD new source reviews will be substantial unless the
bubble policy can be easily used by a source to achieve, through an
increasingly complex "numbers game", no net increase in a source's emission
rate. As fugitive emissions become more quantificable, sources will have to
obtain offsets for such activities as material storage piles and haul
roads. Depending on the final de minimus level for particulate matter, the
addition of coal piles alone could trigger new source review. Any facility
contemplating conversion to coal should, however, carefully review the
available exemptions from new source review for certain conversions.
Once new source review is begun/ either partially or entirely due to
sources of fugitive emissions, the substantive aspects of the review come
into play. What is BACT or LAER for sources of fugitive emissions - for
coal piles, for haul roads, for surface mines? What is involved in
performing the air quality analysis - the modeling and the monitoring - of
these emissions? These two questions, together with how to "reasonably
quantify" these emissions, will be the fugitive emissions issues of the
'80's.
34
-------
FOOTNOTES
1. 13 ERC 1225 (C.A.D.C. 1979).
2. 41 FR 55524.
3. 39 FR 42510] 40 CFR 52.21.
3a. For a discussion of the development of the PSD program, see Raffle,
Bradley I., "Prevention of Significant Deterioration and Nonattainment
Under the Clean Air Act - A Comprehensive Review" Environment Reporter,
Monograph No. 27 (May 4, 1979).
4. 39 FR 42510 (December 5, 1974).
5. 43 FR 8962.
6. 40 CFR 81.
7. Clean Air Act (CAA) Section 162.
8. CAA Section 162.
9. CAA Sections 165 and 169(1) . The Amendments defined a major source in
terms of its "potential to emit" but do not define the term
"potential." EPA's 1978 PSD regulations define "Potential to emit" as
the capability at maximum capacity to emit a pollutant in the absence
of air pollution control equipment. This was at issue in Alabama Power.
10. CAA Sections 165(a) (4) and 169(3).
11. CAA Section 165
-------
18. CAA Section 161.
19. See Proposed Sections 51.24(b)(20) and 52.21(b)(20) (1979).
20. 40 CPR 51.24(b)(6) and 52.2Kb) (6) (1978).
21. Supra, n. 19.
22. See 44 FR 51930 and Proposed Sections 51.24(b)(3) and 52.21(b)(3)
(1979).
23. 40 CPR 52.21{b)(2)(c) (1978) and 40 CFR 51, Appendix S, Paragraph II(A)
(5) (ii) (d).
24. 40 CFR 52.2(b) (2) (d) (1978) and 40 CFR 51, Appendix S, Paragraph II(A)
(5) (ii)(c).
25. See Proposed Sections 51.24(1)(2)(i) and 52.21(1)(8)(i) (1979).
26. 40 CFR 51.24(k)(5) and 52.21(k)(5) (1978).
27. 44 FR 3274.
28.' 30 CFR Parts 782-784. See especially Sections 783.18 and 784.26.
36
-------
PARTICLE PRODUCTION FROM SURFACE MINING:
PART 1--VERTICAL MEASUREMENTS
James A. Armstrong and Philip A. Russell
Denver Research Institute
Denver, Colorado
Dennis C. Drehmel
U.S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina
ABSTRACT
An investigation of the vertical extent and downwind transport of
fugitive dust plumes resulting from various operations at a surface coal
mine was conducted using a tethered balloon sampling system. This EPA-
sponsored field study took place concurrently with a Bureau of Land Manage-
ment-sponsored field program conducted by the USDA's Forest Service and
Colorado State University to assess the overall air quality impact of
strip mine activities on the surrounding environment.
A balloon was used to carry aloft three lightweight wind directional
particle samplers. One sampler was attached directly to the balloon; the
other two were attached to the tetherline at selected distances from the
balloon. This arrangement allowed for the simultaneous collection of
dust samples at three heights above the ground at various downwind
distances from the mine operations. A fourth sampler was located near
the balloon launch site to monitor ground level concentrations. Dust
samples were collected on Nuclepore and Millipore filter substrates
which were analyzed by optical and scanning electron microscopy.
The size distribution, concentration and composition versus height
of the collected dust samples are presented. Time lapse movies of mining
activities at the test site during tethered balloon sampling are also
discussed.
37
-------
INTRODUCTION
For the past 3 years, the Denver Research Institute (DRl) has been
involved in developing and utilizing balloon-borne particle sampling
systems which possess both vertical and horizontal mobility and are
capable of sampling at specified altitudes for extended periods of time.
Two of the systems developed employ remote controlled samplers and have
been used to monitor emissions from the plumes of coal-fired power
plants1 and a portland cement plant.2 The third sampling system, which
is described in this paper, was developed to investigate the vertical
extent and downwind transport of fugitive particle emissions from point
and nonpoint sources. The system was recently employed to monitor dust
emissions from various operations at a surface coal mine and to evaluate
the feasibility of using the tethered balloon sampling method to ultimately
determine emission factors using an exposure profiling technique.
The DRl field study at the surface coal mine in Colorado was sponsored
by the Industrial Environmental Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina. The study
took place concurrently with a Bureau of Land Management-sponsored field
program performed by the U.S. Forest Service and Colorado State University
to assess the overall air quality impact of strip mine activities on the
env i ronmen t.
A preliminary field test at the mine was conducted from June 4 to
June 8, 1979« to check out and evaluate sampling techniques. A more
extensive field investigation was conducted at the mine from July 30
through August 8, 1979. Results of the DRl study for the latter test
period are presented, including the evaluation of selected airborne dust
samples from a dragline operation versus dust from hauling, dragline,
dressing and scraping operations, as well as dust samples from an active
haul road. The mineral content of selected soil and road bed samples is
also presented. Documentation of balloon sampling and mine site activities
using time lapse photography is discussed.
BACKGROUND
It is now widely recognized that fugitive particle emissions contribute
significantly to the measured levels of total suspended particulates
(TSP) found in both urban and rural environments. These emissions are
believed to be responsible for the failure of many areas to attain the
national primary air quality standards for TSP. Major open sources of
fugitive emissions include unpaved roads, paved roads, surface mines,
storage piles, construction activities, agricultural tilling, wind
erosion of harvested cropland and forest and brush fires, including both
wildfires and prescribed burns.3 Other sources include iron and steel
foundries, smelters, and refineries.1"5'6 Considerable effort is being
devoted by numerous researchers in establishing the particle emission
levels and rates from fugitive sources so that cost effective control
strategies can be developed for such emissions.
38
-------
Of the fugitive sources listed above, open sources (except forest
and brush fires) emit particles primarily in a coarse size range of
greater than approximately 2.5 ym in equivalent aerodynamic diameter.
These particles, classified as mechanically generated, are believed to
be the most significant sources of fugitive emissions contributing to measured
ambient TSP levels. Fugitive particle emissions produced by combustion
processes, such as from forest and brush fires and from furnace operations
at foundries and smelters, are typically in a fine particle size range
less than 2.5 3-im. The chemistry of the coarse versus fine particles
also differs: coarse particles are usually more basic; and fine particles,
more acidic.7
The traditional method of monitoring airborne particles from fugitive
sources has been to use ground-based high-volume samplers. The samplers
are usually at considerable distances both upwind and downwind from the
source being sampled. Dispersion modeling is normally used to interpolate
the source strength from the HiVol data. Considerable error can be
incurred when using this technique because of the often limited number
and location of sampling sites and due to the current simplicity of the
atmospheric dispersion models. Complications are further increased when
the source is not wel1 defined.
Samplers mounted on towers have been used successfully to characterize
fugitive dust emissions from paved and unpaved roads employing an "exposure
profiling" technique which is based on the isokinetic profiling concept
used for conventional source testing.8 Particulate emissions are measured
directly downwind of the source being sampled by placing a grid of
samplers in the effective cross-section of the plume. Emission factors
are subsequently determined by a mass-balance calculation. The exposure
profiling technique eliminates the need to use dispersion models to
estimate emission rates from fugitive sources. A two dimensional vertical
array of samplers must be used when sampling a point or area source with
this technique while line sources (e.g., haul roads) require a one
dimensional vertical grid of samplers.
Major limitations of the exposure profiling technique are the
practical height of the sampling towers and their lack of mobility. Results
of the study presented here indicate that these limitations can be
overcome using balloon systems so that the exposure profiling technique
can be extended to a larger number of fugitive source sampling applications.
TETHERED BALLOON SAMPLING SYSTEM
The sampling system basically consists of a balloon system, wind
vane samplers and a ground-based sampler.
Balloon System
A commercially available blimp-shaped balloon having a buoyant gas
capacity of 3.25 m is used as the airborne platform for the three wind
vane samplers discussed below. Tethered balloons up to this size do not
require Federal Aviation Administration waivers to Federal Aviation
Regulations, Part 101, in order to be operated. When inflated with
39
-------
helium, this balloon has a static lift of 1.9 kg at sea level. The
balloon is constructed of polyethylene and, for ease in observation and
safety, is red. The balloon is attached via a tetherline to a portable
winch which contains a battery operated forward/reverse, variable speed
motor for tetherline control, a 12-volt sealed lead/acid battery
pack and charger system have been incorporated into the winch housing so
that the winch is totally self-contained. The winch weighs 27 kg including
the battery pack modification. The extremely lightweight tetherline
has a breaking strength of 535 newtons and a mass to length ratio of 0.4
kg/km.
This balloon system is capable of lifting a package weighing 1.2 kg
to maximum altitudes of 800 to 1,000 m in winds of 6 to 8 m/s. Flight
performance of the balloon is still reasonable in winds of 10 to 12 m/s
but the maximum altitude attainable is somewhat reduced. The balloon is
designed to survive in winds up to 20 m/s.
Wind Vane Samplers
Three wind directional particle samplers, each weighing ^00 g, were
employed in this study. A schematic of this sampler type is shown in
Figure 1. The vertical tube of the sampler attaches to the horizontal tube
at a point where the wind vane balances the forward tube extension such
that the extension is always horizontal. The tubes and wind vane are fiber-
glass. A 37 mm filter cassette ,is attached to the front end of the
horizontal tube. An assembly housing the pump, battery pack and voltage
regulator is attached to the bottom of the vertical tube. The weight of
these components ensures that the sampler maintains the orientation
shown in Figure 1. The samplers are attached to the balloon or tetherline
by nylon lines. Swivels are tied to the ends of the nylon lines that
attach to the samplers while quick disconnect clamps are tied to the
ends that attach to the balloon or tetherline. The wind vane configuration
ensures that the samplers point Into the wind during sampling. Observations
have shown this to be the case even in very light winds (<1 m/s).
Two of the wind vane samplers are equipped with a diaphragm pump;
the third uses a piston pump. All pumps are powered with battery packs
consisting of three lithium cells wired in series which provide 8.1
volts. This voltage is regulated down to 5 volts to ensure constant
pumping rates. Flexible tubing is used to connect the pumps to the 37
mm filters held in plastic cassettes. The sampling probe shown attached
to the filter cassette in Figure 1 was not used during this field study:
all airborne dust samples were collected on open-faced filters. Each
sampling pump is activated at the time that the associated sampler is
launched. For this study, 0.2 urn Nuclepore filter substrates were employed;
the corresponding air sampling rates of the diaphragm pumps and the piston
pump were 2.75 i/m and 1.2 l/m, respectively. Millipore filters having
an effective pore size of 0.8 urn were also used, resulting Fn air flow
rates of 5-1 A/m and 3.k H/m for the diaphragm and piston pump samplers,
respectively.
40
-------
FILTER
CASSETTE
LITHIUM
BATTERIES
SWITCH
VOLTAGE
REGULATOR
Figure 1. Wind vane particle sampler.
-------
-ts
1-0
OVERBURDEN
PILE
ACTIVE PIT
r
MAIN GROUND-BASED
SAMPLING NETWORK
-TOPSOIL
SCRAPING
AREA
08/l/79
°8/5/79
B
NOTES:
O BALLOON LAUNCH LOCATIONS
AND DATES
M.T. METEOROLOGICAL TOWER
A TIME LAPSE CAMERA LOCATED
~3.5 km FROM M.T. ON 7/30/79
B TIME LAPSE CAMERA LOCATED
-2.5 km FROM M.T. ON
7/31 to 8/5/79
08/2/79
08/3/79
100 m
Figure 2. Strip mine sampling site.
-------
Ground-Based Sampler
In addition to the three balloon-borne samplers, a fourth sampler
equipped with a piston pump was always located near the balloon launch
site to monitor dust concentration at ground level. This sampler was
not wind directional. It was positioned at a height just above the
local vegetation with the open-faced filter cassette pointing upward.
FIELD OPERATIONS
A description of the field site, the operation of the balloon sampling
system, the collection of soil samples, and the documentation of sampling/
mining operations are discussed.
Field Site
The mine site, in arid western Colorado, is on the north face and
near the bottom of a gently sloping mountain range. Its elevation is 1,950
m. In terms of dispersion modeling, the terrain would be classified as
relatively complex. Dry conditions existed during field sampling. Morning
temperatures varied between 8°C and 15°C; afternoon temperatures varied
between 31°C and itO°C. Surface radiational inversions existed most
mornings. Winds were predominantly from the west. During the mornings,
the winds were light and sometimes variable; by early afternoon, they
would become strong and often gusty. The site in Figure 2 shows balloon
launch locations and dates of balloon sampling. Main elements of the site
consisted of an active coal pit, a coal storage pile, haul roads and a
topsoil scraping area. A dragline removed overburden from the pit, coal
was loaded in the pit and, on most days, the pit coal seam was blasted.
During several of the sampling days, topsoil was removed from an area
east of the pit and coal from the storage pile was loaded onto haul
trucks. The storage pile and the area adjacent to the pit were dressed
periodically.
Balloon Sampling Operations
Balloon sampling consisted of first locating the balloon system at
a selected distance downwind from the mine operation(s) to be sampled.
The first sampler would then be attached to the balloon an.d the pump of
this sampler would be activated by connecting the battery leads shown in
Figure 1. The balloon would be launched until a selected length of
tetherline was reeled off the winch. The winch spool would be stopped,
the second sampler would be attached to the tetherline and its pump
would be activated. Again the winch spool would be actuated until the
appropriate length of tetherline was reeled out. The winch would be
stopped, the third sampler would be attached to the line and its pump
would be activated. Finally, the winch would be started and additional
tetherline would be run out until the balloon reached the desired uppermost
sampling height. Appropriate distances were previously marked on the
tetherline to assist in placing the samplers at the desired vertical
separation distances. The ground-based sampler would also be activated
at this time.
43
-------
Flight operations were conducted at downwind distances varying from
150 to 1,000 m from selected mine operations with sampling heights up to
J20 m. Sampling times were typically 4 to 6 hours.
Three balloon sampling operations, in which the collected dust
samples have been examined in detail, are discussed below. A log of
these operations is presented in Appendix A.
Dragline Versus Additional Mining Operations
From 0700 hours to 11 A3 hours MDT on August 1 and again from 0830
hours to 151& hours on August 5, balloon sampling was conducted at the
same launch site shown in Figure 2. During both sampling periods, the
wind at an altitude of 75 m was predominantly from the west. The surface
wind velocity measured by a ground based meteorological station varied
from 1.8 m/s to 3.1 m/s and from 1.3 m/s to 2.7 m/s during the sampling
perfods on August 1 and August 5, respectively. The launch site was
approximately ^00 m east of the dragline operation at the active coal
pit. Mine activities during the sampling period on August 1 included
coal hauling, overburden removal by the dragline, dressing operations
and topsoil scraping, while only the dragline was in operation on August
5. Dust samples were collected at elevations of 0.5 m, 10 m, 20 m and
75 m.
Haul ing Operations
A balloon sampling operation was conducted from 1155 hours to IkkQ
hours on August 6 to investigate the extent of dust emissions from an
active haul road. During this sampling period, the surface wind velocity
was 2.2 m/s to 3-1 m/s. The balloon was located approximately 150 m
directly downwind from a curve in the haul road and dust samples were
collected at an elevation of 30 m.
Soil Sample Collection
Loose soil samples were collected from the middle of the main east-
west haul road, from the spoils of an overburden drill hole and from the
overburden pile adjacent to the dragline. These samples were used to
identify the typical mineral composition of the mine site. The samples
were collected by hand and were stored in sealed plastic bags for later
laboratory analysis.
Documentation of Sampling/Mining Operations
A 16 mm zoom lens Bo lex movie camera equipped with an MR I intervalometer
was used to take time lapse movies of all tethered balloon sampling operations,
The movies were taken to document the downwind sampling distances, the
activities at the mine site and the genera] atmospheric conditions during
sampling. The camera locations are shown in Figure 2. The camera framing
rate was typically one frame every 3 seconds.
44
-------
RESULTS
Analyses of airborne particles, bulk soil samples, and time lapse
movies are discussed.
Particle Analysis
Both optical and scanning electron microscopy were utilized to
determine the composition, size distribution and concentration of the
collected particles. In general, optical microscopy was limited to
larger particles (diameters >_ 5 Um) which permitted rapid determination
of coal and mineral components. Scanning electron microscopy was employed
to perform detailed size and compositional analyses of particles ranging
from 0.5 ym to 100 ym in diameter.
All Millipore filters exposed in the field were examined using
optical microscopy. In the laboratory, mid-sections of the Millipore
filters were carefully excised and permanently mounted on microscope
slides. Examination during the mounting procedure assumed that particle
displacement was minimal. A mixture of k percent flexible collodian in
amyl acetate was used to clear the Millipore filter material and adhere
particles to its surface. After the collodian/amyl acetate mixture
dried, a cover slip was added using Permount to make a permanent slide
for future observations.
Upon returning to the laboratory, it was observed that some of the
particles collected on the Nuclepore filters had become dislodged from
the filter substrate and had redeposited on the walls of the capped
cassettes. This was especially noticeable for the fraction of the
larger particles. Subsequently, the particles on these filters and
filter cassettes were resuspended and transferred to glass vials using
filtered (0.2 ym filter material) ethyl alcohol. Resuspension was
carried out in a positive pressure glove box. Air to the glove box was
filtered through dual absolute filters. Samples for subsequent optical
and electron microscopy were prepared by refiltering the resuspended
particles through 0.2 ym Millipore media. Filter mid-sections to be
used for electron microscopy were mounted on SEM stubs using double
sided adhesive tape. A gold/palladium alloy was then vacuum sputtered
onto the samples to make them electrically conductive for high resolution
electron beam observations. The thickness of the gold/palladium coating
was approximately 10 nm.
The optical analysis was performed using a Zeiss Photomicroscope II
at magnifications of 36x, I60x and 250x. Polarized light with the
polarizing elements slightly out of phase was used to permit simultaneous
recording of mineral and coal particle images. Highly opaque images
were assumed to be coal. Inhomogeneous particles containing both coal
and mineral were commonly observed. For the purpose of the optical
analysis, any particle containing coal was classified as a coal particle.
Particle counting was performed from photomicrographs, examples of which
are shown in Figure 3-
-------
August 1, 1979
August 5, 1979
75 Meter Elevation
75 Meter Elevation
-
-
0.5 Meter Elevation
0.5 Meter Elevation
Figure 3. Optical micrographs of sampled particles (mine operations: August 1
hauling, dragline, dressing and scraping; August 5 - dragline only).
-------
More detailed information was obtained using a Philips 501B scanning
electron microscope equipped with an EDAX 9100/60 energy dispersive x-
ray analyzer. A number of randomly selected fields of 80x and 640x were
examined for each sample. The EDAX analyzer was used to identify the
relative elemental composition of the particles. An example of the x--
ray analysis of three types of particles collected during this field
program is illustrated in Figure k. The gold peak recorded in the
sample spectrum is produced by the gold/palladium alloy coating. It is
notable that the coal particle contains significant amounts of elements
normally associated with minerals (Al, Si, K, Ca, Fe) whereas the carbon
black particle, probably diesel, produces no observable x-ray spectra.
Carbon black particle aggloremates were only rarely observed in the
samples examined. In general, particles having a gold peak the same as
or less in magnitude than Si (or other major elemental peaks) were
classified as mineral in this analysis. As with the optical microscopy,
particle counting was performed from photomicrographs. The composition
of the particles was noted directly on the micrographs prior to counting.
Examples of the SEM micrographs are shown in Figure 5.
Equivalent spherical diameters of the basically irregular mineral
and coal particles were estimated by averaging the maximum length of
each particle and its maximum width taken at right angles to the length.
All particles observed on a micrograph were classified by size and composition
and tallied into a size class based upon a geometric progression of the
/2.9 For example, 13 size intervals, ranging from 0.6-1.0 ym to 60.0-
85.0 ym, were employed for mineral and coal particles observed on the
SEM micrographs. After classifying the mineral and coal particles
separately, the particles counted per interval at 640x were extrapolated
to the number which would be present at a magnification of 80x. For
intervals which overlapped the two magnifications, weighted averages of
the number of observed particles were employed. The total number of
particles in each interval across the entire active filter was then
calculated and, in turn, divided by the total volume of air sampled
through the f i 1 ter.
The mid-size diameter of each size interval was used to calculate a
spherical particle volume. These volumes were then multiplied by the
appropriate number counts per volume of air sampled and by assumed par-
ticle densities so that cumulative and differential mass distributions
as functions of particle size could be plotted. The assumed density of ,
the mineral particles was 2.6 g/cm ; that for the coal particles, 1.4 g/cm .
Cumulative mass distributions for particles sampled on August 1,
5 and 6 are presented in Figures 6, 7 and 8, respectively. The straight
lines through the data points were fitted using a least squares routine
applied in log-normal space. If a distribution is log-normal, it will
occur as a straight line when plotted on a log-probability graph. In
general, the data is approximately log-normal. The mass median diameters
d , are the 50 percent values of the distributions while the geometric
standard deviations, O , are the 84.1 percent values divided by the 50 percent
values. Table 1 summarizes d , a , and the total number of mineral and
coal particles counted from the micrographs for each sample analyzed.
47
-------
IO.O K>V
Figure k. SEM micrograph showing particle types sampled at
the strip mine and associated x-ray spectra.
-------
-•
c
75 Meter Elevation
0.5 Meter Elevation
Figure 5.
SEM micrographs of particles sampled on August 1, 1979 (mine
operations: hauling, dragline, dressing and scraping).
-------
SflMPLE 5N MINERflL
SRMPLE 5N CORL
99.9
39.0
590.0
§50.0
10.0
1.0
.1
T
4-
1
10
D (micron*)
100
10
0 (microns)
100
75 Meter Elevation
SflMPLE 8N MINERflL
SflMPLE 8N COflL
89.9
18
D (»1oron«>
10
0 (micron*)
0.5 Meter Elevation
Figure 6. Cumulative mass distributions for mineral and coal particles
sampled at two elevations on August 1, 1979.
50
-------
SRMPLE 17N MZNERflL
SflMPLE 17N COflL
99.9
99.8 • -
18
D
-------
SRMPLE 41M MINERRL
99.9
99.0
?90.0
J 90.0
O
£ 10.0 j-
i.0|-
.1
r
4-
1 10
0 (mforona)
100
99.a
99.0
590.0
$80.0
O
1.0
.1
SRMPLE 41M CORL
4-
10
0 (micron*)
Cumulative Mass Distributions
SRMPLE 4iM MZNERRL
Q
01
O
10
10
0 Cnlorona)
Q
9
O
10
SRMPLE 41M CORL
T
10
0 (mtoron*)
Differential Mass Distributions
Figure 8. Cumulative and differential mass distributions
of mineral and coal particles sampled at an
elevation of 30 meters on August 6, 1979.
52
-------
Table 1. Sampling Results
Date
8/1 /79a
8/5/79b
8/6/79°
Filter
Number
5N
6N
7N
8N
17N
20N
41M
Sampl ing
Elevation
(m)
75
20
10
0.5
75
0.5
30
Total Parti
Counted
Mineral
438
414
327
262
523
382
138
cles
Coal
221
165
234
137
283
130
523
d
gm
Mineral
15-4
17-7
18.1
24.4
10.6
17.1
32.2
(ym)
Coal
13.8
17.3
21.1
28.2
13.2
15.5
34.5
a
9
Mineral
2.42
2.27
2.40
2.59
2.33
2.33
2.51
Coal
2.21
2.34
2.60
2.61
2.60
2.30
2.64
.Mine operation:
Mine operation:
Mine operation:
hauling, dragline, dressing, scraping,
dragline only.
haul ing only.
-------
For samples 5N through 8N collected on August 1, in which the mine
operations consisted of coal hauling, dragline overburden removal,
dressing and topsoil scraping, d values for both mineral and coal
particles are reported in the taBTe to be highest at ground level (0.5 m
elevation) and successively decrease, as would be expected, for samples
taken at elevations of 10, 20 and 75 meters. The vertical profile of
d for mineral and coal particles sampled on August 1 appears to follow
an exponential type relationship. For samples 17N and 20N collected on
August 5 (at the same balloon launch location as that of August 1),
during which time the only mine operation was overburden removal by the
dragline, d values are again highest for the ground level sample
versus the 9§ meter sample. The ground level d values for mineral and
coal on August 5 are considerably less than those of August 1 but this
is reasonable when one considers, besides the reduced mining activities,
that the topsoil scraping operation of August 1 occurred approximately
100 m closer to the balloon launch site. The fact that coal particles
were present in concentrations comparable to the mineral particles on
the August 5th samples is interesting. The source of this coal may
have been the pit itself, assuming that the dragline was removing over-
burden near the coal seam, or from the storage coal pile.
Dust from a haul road was monitored only at an elevation of 30 m on
August 6 (sample 4lM). The sampling occurred at a considerably shorter
downwind distance from the source than on either August 1 or August 5
(see Figure 2). Here d values for mineral and coal are significantly
greater at the 30 m elevation than those measured at ground level on the
previously reported sampling days. In addition, the total number of
coal particles observed on the 41M micrographs was almost four times the
number of mineral particles. This is probably because some coal often
spilled from the haul trucks at the haul road curve. This coal would
become crushed and entrained into the air by trucks passing by later. In
all instances, d values of mineral versus coal are comparable at the
same elevations.
The average of all geometric standard deviations reported in Table 1
is 2.A3 with a corresponding "standard deviation" of 0.14. This shows
good agreement in the data such that all of the samples are believed to
be from the same general population of particles.
All of the cumulative mass distribution plots of Figures 6, 7 and 8
show a characteristic inflection point in the vicinity of particle
diameters from 10 to 15 ym. This may be attributable to the fact that
the 80x and 6*»0x magnifications were sufficiently far apart that minimal
overlap occurred in the area where the two magnifications were least
accurate.
Differential mass distributions for the sampled mineral and coal
particles are presented in Figures 8, 9 and 10. The graphs are log-log
plots of the mass per interval divided by the difference of the log D
interval boundaries (ug/m ) versus particle diameter (ym). The data
points have been joined by smooth curves using a spline fit routine. It
is apparent that the mineral plots are more well behaved than the coal
54
-------
SRMPLE 5N MJNERRL
SRMPLE 5N CORL
1980
Q
en
o
\
£
"O
100
ie
10030
1000
Q
at
o
\
•o
100
10
10
D (microns)
100
IS Meter Elevation
10
D (microns)
100
SRMPLE 8N MINERRL
SRMPLE 8N CORL
1000
a
a)
o
•o
£
-o
100
10
1000
Q
O)
o
£
TJ
100
10
10
0 (mlorons)
100
10
D (microns)
100
0.5 Meter Elevation
Figure 9- Differential mass distributions for mineral and coal
particles sampled at two elevations on_August 1, 1979.
(The unit of the ordinate axis is yg/m ).
55
-------
SRMPLE i?N MINERRL
10000
1080
Q
OB
O
•o
r
100
10
J_
SRMPLE 17N CORL
10000 i ' i'
1000 -
Q
0»
O
13
\
X,
100 -
10
D (micron*)
100
75 Meter Elevation
10
D (microns)
100
SRMPLE 28N MINERRL
10000
1000
Q
09
o
•a
r
•o
100
10
10000
SflMPLE 20N CORL
1080 -
Q
09
O
100 -
1
10
D (micron*)
100 *
0.5 Meter Elevation
10
D (microns)
100
Figure 10. Differential mass distributions for mineral and coal
particles sampled at two elevations on,August 5, 1979,
(The unit of the ordinate axis is yg/m ).
56
-------
plots. This can be attributed to the fact that in general, except for
sample **1M, 50 percent fewer coal particles were observed on the micrographs
compared to mineral particles. The distributions show that the majority
of mass for both mineral and coal sampled at any elevation is associated
with the larger particles. Note that tethered balloon sampling was conducted
with open faced filters such that nonisokinetic sampling took place in
which the filter face velocity was normally considerably less than the
mean wind speed. This has undoubtedly biased the samples to some extent
with large particles. Still, as expected, the vast majority of the
captured dust particles from the mining activities were in a coarse size
range of greater than 2 ym in actual diameter for all samples.
Soil Sample Analysis
To determine the major crystalline mineral composition of the
selected soil samples, each sample was first pulverized, by means of a
ball mill, to fine material which would pass through a 200 mesh screen.
The powder samples were then semiquantitatively analyzed by a Philips X-
ray Diffractometer. Diffraction spectra were obtained using copper ka
x-rays from 2° to 6^° 26. Minerals were identified by peak position and
semiquantified by comparing peak intensities to those of other known
mineral samples. The results of this analysis are presented in Table 2.
Time Lapse Movie Analysis
Time lapse movies have been analyzed to confirm the downwind distances
of the tethered balloon sampling system from mining operations and to
establish the types and levels of activities occurring at the mine site
during sampling. An example of the information obtained concerning mine
operations for one day is presented in Table 3. The time lapse photography
was also useful in documenting the relative amount and vertical extent
of dust from the various mine operations and in qualitatively assessing the
effectiveness of watering to suppress dust from active haul roads. During
hot dry afternoons, haul road watering markedly reduced the amount of
visible dust from the roads for periods of 30 to kO minutes.
CONCLUSIONS
The tethered balloon sampling method, coupled with microanalysis,
can provide useful definitive insights into the nature of multispecied,
size discriminate particulate dispersions from fugitive dust sources.
The sampling method permits rapid horizontal and vertical mobility and
simultaneous sample collection at selected elevations and distances
downwind from dust sources with relatively little expense. Optical and
electron microscopy provides useful estimates of particle size distributions,
concentrations and composition of lightly loaded samples. This investigation
concludes that the tethered balloon sampling method can extend the exposure
profiling technique to more fugitive source sampling applications.
Dust samples collected from the plumes of various strip mine operations
follow a log-normal distribution. Sampling at various elevations downwind
57
-------
Table 3. Mine Site Activities on July 31, 1979
Time
of
day
(hrs)
0730
0800
florin
U jUU
1000
1100
1 200
t 300
Win
Oir
from
W
- E
- W
— t
_ W
N
W
- E
- E
_ W
- E
- W
- E
1
J
Eu
E
)ragl in
Iperat i
0836
up
-0
d
*36
3wn
i
_I010
walk
-1030
up
-1123
down
1
1
1232
up
.
Total Number of
Passes
Topsoil Scrapin
Operation
1 umber o
Scraper
Passes
t
fc
4-
t
13
J
t
1
30
i
J
t
1
28
1
1
—
16
1
T
1
1
16
i
1
t
28
i
1
159
Time o
Water in
Trucks
- '0855
- 0?42
_ 1029
-1114
-1348
-1426
-1506
-1553
- 1648
9
E-W Haul Road L
Truck Traffic
4umL
\«r
Haul
Truck
Passes
17
30
T
t
!
3&
!
I
29
1
-
2
t
11
4
15
I
23
i
1
I
1
26
1
1
f
33
i
222
Time o
Waterin
Trucks
1311
1323
1352
1356
1409
1413
1434
I44j
1451
14^4
1515
1518
1531
1633
1652
16
Other
Activities
0836-0842
Cat dressed crushed
coal pile
0836-1140
;oal loading south end
of pit A
OS34-1006
:at dressed dragline
>ase area
Charge dri 1 1 rig in
operation all morning
;at dressed south end
of O.B. pile (W of
).L.) all morning
0«36-1045
1136-1216
Cat dressed O.B. dri 1 1
area (E of D.L.)
232-154"
Cat dressed O.B. pile
^W of D.L.)
1331
Coal blast in Pit A
1232-165?
Cat dressed O.B. drM
area (E of D.L.)
Charge dri 1 1 rig in
operation all after-
noon
1405 to end coal
loading south end of
Pit A
1232-1300
1405-1558
i £oc i tnt\
1O25-17''1'
Cat dressed D.L. base
area
1658
O.B. blast
Remarks
_0745
Start of balloon
sampling operation
0836
Start of time
lapse movie-Rol 1 2
Winds light all
day steady from W
until 1155 then
extremely variable
Balloon samples
taken near CSU
stakes 17 & 1«
1216
End of Roll 2
Start of Rol 1 3
-1321
Clock blown over-
event times calc.
from frame f
1412
End of bal loon
sampl ing operation
65"
nd of Roll 3
58
-------
Table 2. Soil Sample Analysis
Mi neral
Quartz
Feldspars
Do 1 om i te
Calci te
Kaol ini te
Plagioclase Feldspar
Montmori 1 loni te
Sampl
Haul Road Drill
Major
Minor
Trace
Trace
--
Trace
--
e Composition
Hole Spoi Is
Major
Major
Major
Mi nor
Trace
Trace
—
Overburden Pile
Major
Minor
Minor
Minor
Trace
Trace
--
aMajor - >_ 25% of sample.
Minor - distinct portion of sample (5~25%).
Trace - slight peak present (~ < 5%).
- absent from sample (~ < 1%).
59
-------
of mining activities has demonstrated that the mass media diameters
of mineral and coal particles decrease with increasing sampling height.
Still, considerable dust concentrations are present at a sampling height
of 75 m and a significant fraction of this dust consists of particles
having diameters greater than 10 ym. The large particles are responsible
for the major mass fraction of all samples.
Time lapse photography has been shown to be a useful technique to
document balloon sampling operations and surface mining activities.
This technique can be applied to visually assess the effectiveness of
dust control measures and strategies.
PLANNED ACTIVITIES
Tethered balloon sampling at higher elevations and over a wider
range of downwind distances will be conducted at surface mines to more
accurately determine the vertical extent and downwind transport of
fugitive dust plumes.
A series of balloon systems will be utilized in conjunction with
the exposure profiling technique to establish improved emission factors
from mining and other fugitive emission sources. Typical sampling
schemes employing four balloons are presented in Figures 11 and 12.
When sampling a line source, such as the haul road shown in Figure 11,
it is envisioned that the three downwind balloons will be aligned parallel
to the prevailing wind. This network will allow the downwind dust
transport to be evaluated as well as the validity of extending the
exposure profile technique to greater sampling distances. The characterization
of emissions from point and area sources will require a two dimensional
sampling network, such as that shown in Figure 12. As with the line
source sampling scheme, an upwind balloon will be operated to determine
background dust concentrations.
Basic improvements to the wind vane particle samplers are also
planned. Elutriators will be added to the samplers to allow for the
sampling of inhalable particles having an aerodynamic diameter of 15 ym
or less.10 Inhalable particles can further be subdivided into groups of
coarse and fine fractions with a cut point being approximately 2.5 ym.
The fine particles are primarily responsible for air quality visibility
degradation and are primarily produced by combustion processes. The
coarse fraction of inhalable particles are those between 2.5 and 15 ym
and are normally classified as mechanically generated. Most wind generated
dust, for example, would consist of coarse size particles. Depending on
the type of source being sampled, it may sometimes be desirable to be
able to separate inhalable particles into coarse and fine fractions.
For example, fugitive particulate emissions from an iron foundry are
expected to be a mix of fine and coarse particles. By being able to
quantify the fractions in terms of amount and chemical composition, the
actual emission sources should be able to be identified. When desired,
the balloon samplers will employ sequential filtration to separate
inhalable particles into coarse and fine fractions.11
60
-------
LAYOUT OF TETHERED BALLOON SAMPLING SYSTEM FOR PROFILING THE PLUME
FROM A LINE SOURCE AND FOR DETERMINING THE DOWNWIND TRANSPORT OF PARTICLES.
Figure 11
LAYOUT OF TETHERED BALLOON SAMPLING SYSTEM FOR PROFILING THE PLUME
FROM AN AREA SOURCE.
Figure 12
61
-------
REFERENCES
1. J. A. Armstrong and P. A. Russell, "The Monitoring of Participates
Using a Balloon-Borne Sampler," in Symposium on the Transfer and
Utilization of Particulate Control Technology, Volume 4, EPA-600/7-
79-044d, (NT1S PB 295229), February 1979, pp. 357-376.
2. J. A. Armstrong, P. A. Russell, M. N. Plooster and L. E. Sparks,
"Tethered Balloon Plume Sampling of a Portland Cement Plant." To
be submitted for publication.
3. J. S. Evans, D. W. Cooper, M. Quinn and M. Schneider, "Setting Prior-
ities for the Control of Particulate Emissions from Open Sources," in
Symposium on the Transfer and Utilization of Particulate Control
Technology, Volume 4, EPA-600/7-79-044d (NTIS PB 295229), February
1979, PP 85-103.
4. W. D. Scott and C. E. Bates, "Measurement of Iron Foundry Fugitive
Emissions," in Symposium on Fugitive Emissions: Measurement and
Control, EPA-600/2-76-246 (NTIS PB 261955),September 1976, pp 211-237.
5. R. N. Lindstrom and S. E. Sundberg, "Fugitive Particulate Emission
Rate and Characterizatfon for Electric Arc Steelmaking Furnaces,"
presented at the 72nd Annual Meeting of the Air Pollution Control
Association, Cincinnati, Ohio, 79-39-3, 1979.
6. P. W. Kalika, R. E. Kenson and P. T. Bartlett, "Development of
Procedures for the Measurement of Fugitive Emissions," EPA-600/2-76-284
(NTIS PB 263992), December 1976.
7. F. J. Miller, D. E. Gardner, J. A. Graham, R. E. Lee, W. E. Williams,
and J. E. Bachmann, "Size Considerations for Establishing a Standard
for Inhalable Particles," APCA Journal, 29_:6, 1979.
8. C. Cowherd, C. M. Maxwell and D. W. Nelson, "Quantification of Dust
Entrapment from Paved Roads," EPA-450/3-77'072 (NTIS PB 269944),
July 1977.
9. J. D. Stockham and E. G. Fochtman, Particle Size Analysis, Ann Arbor
Science, Ann Arbor, 1977, P. 6.
10. W. B. Smith, K. M. Cushing, M. C. Thomas and R. R. Wilson, "Some
Aerodynamic Methods for Sampling Inhalable Particles," in Proceedings:
Advances in Particle Sampling and Measurements, EPA-600/9-80-004
(NTIS PB 80-187487), January 1980, pp. 316-347.
11. T. A. Cahill, L. L. Ashbaugh, J. B. Barone, R. A. Eldred, P. J. Feeney,
R. G. Flocchini, G. Goodart, D. J. Shadoan and G. W. Wolfe, "Analysis
of Respirable Fractions in Atmospheric Particulates via Sequential
Filtration," APCA Journal, 27;7, 1977.
62
-------
Appendix A. Sampling Log of Selected Balloon Operations
Date
8/1/79
8/5/79
8/6/79
Filter
Number
05N
06N
07N
08N
17N
18N
19N
20N
41M
Sampl ing
Elevation
(m)
75
20
10
0.5
75
20
10
0.5
30
Start
Time
(hrs-MDT)
0659
0708
0711
0720
0828
0830
0835
0848
1155
Total
Time
(min)
283
269
264
263
408
?
265
372
165
Air Vol.
Sampled
(&) Remarks
778
740
317
421
1122
? Pump stal led
during sampl ing
318
595
553
Note: Launch sites shown on Figure 2.
63
-------
PARTICLE PRODUCTION FROM SURFACE MINING: PART 2
SURFACE PARTICULATE AND METEOROLOGICAL MEASUREMENTS
David L. Dietrich
Colorado State University
Fort Collins, Colorado
William E. Marlatt
Colorado State University
Fort Collins, Colorado
Douglas G. Fox
Rocky Mountain Forest & Range Experiment Station
Fort Collins, Colorado
Federal land managers are now being challenged to evaluate the scope of
potential environmental effects in the development of western mineral
resources. The influence that surface mining for coal will have on air quality
is one impact that must be considered. This paper describes the first phase
(summer 1979) of a study to investigate the sources, characteristics, and
transport of particulate emissions from surface mining. Air quality and
meteorological measurements were made on a surface mine in western Colorado.
Near source suspended and deposited particulate loads were quantified as they
relate to meteorological conditions and mining operations. The results char-
acterize the variety, variability, and complexity of particulate emissions from
mining operations in the western states. The study provides the basis for a
continued research effort (Phase II) to be conducted in the summer of 1980.
-------
INTRODUCTION
There is an Increasing demand for coal in the United States. A large por-
tion of our coal reserves exist on federally managed land in the semi-arid
regions of the western states. Federal land managers are now being challenged
to evaluate the scope of potential environmental effects in the leasing and
development of these western energy resources. The influence that surface
mining for coal will have on air quality is one impact that must be considered.
Most air quality regulations that apply to surface mining are still unre-
solved. The Environmental Protection Agency (EPA) has proposed regulatory
procedures (44 FR 51924, September 5, 1979) that alter their previous exclusion
of fugitive dust from prevention of significant deterioration (PSD) review.
Fugitive dust is defined as particles of native soil uncontaminated by pollu-
tants resulting from industrial activity. The proposed EPA procedures now
state that fugitive emissions should not be considered in determining air
quality degradation except for 26 named source types. The procedures also
state that the Administrator of the EPA will consider the need for additional
source types, including surface mining, over the next few months. However,
even when fugitive dusts are excluded they cannot be ignored. Proposed regula-
tions point out that: 1) the applicant must determine the amount of total
emissions that are fugitive dust; 2) mines emitting greater than 250 tons/year
(225 metric tons/year) including fugitive dust must apply best available con-
trol technology; and 3) the fugitive dust exclusion will be reviewed periodi-
cally by the EPA. In addition to EPA regulations, the Office of Surface Mines
and most states also consider air quality in their regulations.
Despite these regulatory clarifications and exemptions, the federal land
manager remains challenged with making environmentally sound, resource related
decisions. These decisions may be affected by the lack of knowledge about the
sources, nature of particles, and transport of all emissions from surface min-
ing, including fugitive dust.
The purpose of this study is to conduct field research on the sources,
nature of particles and transport of air pollutants from surface mining. The
study is funded by the Bureau of Land Management and is being performed in com-
bined effort by the U.S.D.A. Forest Service, Rocky Mountain Forest and Range
Experiment Station and Colorado State University. The broad goal of the study
is to provide insights into the nature and magnitude of air quality impacts
that federal land managers must consider. The study is being conducted in two
phases over a two year period. Phase I, performed during the summer of 1979,
emphasized near-source analysis of surface mine emissions from a test site in
western Colorado. Phase II, to be conducted in 1980, will emphasize analysis
of air quality at various upwind and downwind distances from the surface mine.
This paper discusses the preliminary results from the Phase I field program.
65
-------
BACKGROUND
In most of the non-industrialized arid west, air quality and visibility
are dependent on the amount of native or fugitive dust in the air. The dust is
emitted by human activity and wind erosion of undisturbed and disturbed lands.
The more people, agriculture, mining or other activity, the greater the amount
of dust released into the atmosphere. In addition, the increasing number of
industrial sources, such as coal fired power plants and mineral processing fa-
cilities, is influencing air quality. What, if any, contribution surface min-
ing emissions have on local or regional air quality can not be easily deter-
mined. The wide variety of pollutant sources associated with mining areas such
as mine mouth power plants, boom-town growth, and increased traffic make iso-
lating the pollutant contribution of a surface mine even more difficult.
Measurement techniques to quantify and separate each of these sources to evalu-
ate their respective impacts are limited, varied, and complicated.
This study was not designed to address or answer all of the air quality
problems of the west, but was designed with four major objectives:
1. determine the sources and quantity of particulate emissions from
surface mining;
2. determine the physical and chemical nature of the emissions;
3. determine the background particle loading, deposition, and visibility
in semi-arid regions of the west; and
4. develop and validate various particle sampling technologies.
The overall goal in addressing these objectives is to provide information
related to the decisions federal land managers face in leasing and managing our
nation's mineral-rich western land.
To accomplish these objectives, a two phase field program was developed.
Phase I, conducted during the summer of 1979, emphasized the characterization
of particulate emissions near the broad range of sources in a surface mining
operation. Field measurements were taken on an active surface mine during two
periods (June 4-7 and July 30-August 7, 1979). Along with the measurements tak-
en by the Colorado State University/Rocky Mountain Forest and Range Experiment
Station team, the study site was shared with independently supported research
teams from the Denver Research Institute and the University of Minnesota. In
addition to the on-site mine activities, Phase I work also established an
atmospheric deposition station and a visibility monitoring station. The depo-
sition station is operated as part of the National Atmospheric Deposition Net-
work. Visibility measurements will be used in analyses that will attempt to
consider the effects of fugitive emissions on atmospheric visibility.
Phase II, to be conducted during the summer of 1980, will emphasize the
measurement of atmospheric particulate loading at increased upwind and downwind
distances from the surface mine. The measurements from both phases will be
combined and applied to evaluate transport mechanisms and to estimate what, if
any, effect particulate contribution from mining may have on local and regional
air quality.
66
-------
PHASE T FIELD PROGRAM
The Study Site
A fully operational surface coal mine served as the study site. All asso-
ciated mining activities cover an area of approximately four square miles
(10.3 km2). The topography of the region is rolling terrain with an average
elevation of approximately 6,500 feet (1,980 m). The mine was located on a
gentle, north facing slope.
Weather plays a major role in the emission and transport of fugitive dust
in the arid west. The study site is located in a region that receives a year-
ly average of 13.5 inches (34 cm) of precipitation. Precipitation is distri-
buted evenly throughout the year. Winters are cold and windy and light snow
cover is often present. In the summer months, intense solar radiation, per-
sistant winds and low relative humidity contribute to hot and dry conditions,
providing an environment for maximum fugitive dust emissions.
Land use in the region includes dryland agriculture (primarily winter
wheat) and livestock grazing. Native vegetation is sparse and dominated by
sagebrush.
Particulate Sources
The sources of particulate emissions from the mine include:
dragline operation grading
coal loading in pit haul road traffic
drilling coal storage pile dressing
overburden blasting coal crushing (only during June 4-7)
coal seam blasting loading from storage pile
top soil scraping
The emissions generated from each source are dependent on wind speed,
moisture, type of equipment, and scheduled time of the activity. The nature
and strength of the source can be affected by a wide range of circumstances.
For example, a haul truck that is slightly overloaded can spill a portion of
its load while rounding a corner. Subsequent haul trucks will crush and en-
train the spilled coal into the atmosphere. Under these circumstances, a haul
road, normally a source of primarily fugitive dust, will become a source of
coal particulates until a road grader cleans and dresses the haul road. With
these intricate variations in sources and source strength, isolating individual
particulate sources in an operating coal mine is extremely difficult. There-
fore, throughout the analysis of Phase J. data, no attempt was made to calculate
the emission factor of any individual particulate source.
Two short-term field programs were conducted during the summer of 1979
(July 4-7 and July 30-August 7). Hot, dry, and often windy conditions coupled
with all mining operations being in full swing provided an environment where
maximum particulate emissions could be expected. If measurements were taken in
the winter or wet spring^ very different results would be expected.
67
-------
Instrumentation
The need to measure particulate emissions from complicated area sources
such as a surface mine has stimulated the technical imaginations of many a
researcher. The instrumentation used in Phase I employed both standard and
imaginative technology. The primary sampling strategies used on the mine site
included active and passive particulate measurements and meteorological mea-
surements. Off-site measurements were taken of atmospheric deposition and
visibility. Though associated with the study, these off-site measurements are
also committed to other research goals.
Active Particulate Samplers
1. High Volume Samplers: Two high volume samplers (hi-vols) were operat-
ed during both Phase I field programs. The hi-vols were located approximately
100 and 150 meters from a major particulate source area. For the June program
the hi-vols were operated with standard filters. During the July-August pro-
gram, the hi-vols were outfitted with Andersen Head particle size samplers in
order to obtain information about the particle sizes contributing to the near
source air quality. Due to the location of the hi-vols, the particulate load-
ing they monitored was probably dominated by haul road traffic and coal stor-
age pile activities. A schematic showing the location of the hi-vols and other
on-site instrumentation relative to the location of mining activities is pre-
sented as Figure 1.
2. Wind Erosion Threshold Velocity Measurements; As part of the June
sampling program, a portable wind tunnel, operated by Dr. Dale Gillette
(National Center for Atmospheric Research), was used to determine the threshold
wind velocities required to erode a variety of surfaces. Analyses were per-
formed over freshly dressed and weathered coal storage piles, dry and wetted
haul roads, wheat fields, spoil piles, graded reclamation areas, native cover,
and other surfaces. Analysis of the data will highlight the erodability of
various stages of disturbed land and coal storage as compared to undisturbed
a*nd agricultural land.
3. Elemental Analysis; As part of the July-August sampling program, a
spatial and temporal variety of particulate samples were taken by Dr. Thomas
Cahill (University of California, Davis) for elemental analysis. Samples taken
both upwind and downwind from the mine will be analyzed for their chemical and
physical characteristics.
4. Balloon-Borne Particulate Samples; In a separate project, Dr. James
Armstrong and his associates from the Denver Research Institute collected par-
ticulate samples throughout the study area via a balloon-borne sampler. The
results of their research are presented in the preceding paper of this sympos-
ium.
5. University of Minnesota Mobile Laboratory; During the summer of 1979
Dr. Virgil A. Marple and his associates were conducting air quality experiments
throughout the southwest as part of the U. S. Environmental Protection Agency
VISTTA visibility research program. In support of both EPA and Bureau of Mines
funding, the University of Minnesota took advantage of the July-August sample
68
-------
Native Shruh
Agrlcu
SCALE
Blow-Up
I
FIGURE 1
LOCATION OF INSTRUMENTATION AND MINING ACTIVITIES
-------
period to investigate surface mine emissions. The instrumentation in the
mobile laboratory was capable of making fast response measurements of the chem-
ical and physical nature of particles, particle transport and visibility. The
mobile laboratory permitted specific monitoring of individual emission sources.
Passive Particulate Samples
1. Petri Dish Dust Fall Network: Dust fall collectors were fabricated
for the study. Each collector consisted of a standard petri dish equipped
with a dust catchment media (astroturf). Each dish was mounted at 0.5 meters
above the soil. Dust fall was collected over a one day period in June and
over two four-day periods in the July-August program. Collected dust fall was
weighed and converted to dust fall rates (g/m2 day). Dust fall samples were
also qualitatively analyzed for the amount of coal, and, during the June field
program» for presence of seeded tracer material.
A 60 dish network was established for the June field program. The dis-
tance between dishes was approximately 60 meters. The network was expanded to
120 collection sites for the July-August program.
Meteorological Measurements
1. Tethersonde; A tethered helium balloon with an instrument package was
used to measure verticle profiles of wind, temperature, and humidity. These
soundings were generally taken three times per day (early morning, noon and
early evening). On the average, the soundings extended from the surface to 450
meters.
2. MRI Weather Stations; Wind speed, wind direction and temperature
were monitored continually by MRI weather stations. One station was located
near the meteorological tower and within the dust fall network. Measurements
taken at this station are representative of the local meteorology within the
major sampling area. A second MRI station was located at a topographic high
point overlooking the mine and was considered more representative of the
synoptic flow.
3. Atmospheric Turbulence Measurements: A 15 meter tower was erected
near the center of the data collection activities. Three sets of Gill (u,v,w)
anemometers were placed at heights 1, 3, and 15 meters on the tower. The
instrumentation was connected to an automatic data logging system that record-
ed instantaneous signals from the anemometers for 2.5 minutes each hour (at a
rate of three scans per second). The data will be used to evaluate the nature
of atmospheric turbulence at the field site.
Off-Site Instrumentation
1. Atmospheric Visibility Site; Daily visibility measurements started
in April 1979 and continue to date. Two different instruments are used, a
contrast telephotometer and an integrating nephelometer. The visibility site
overlooks the mine and is located on a hill top approximatly ten kilometers
from the mine site. The telephotometer is aimed at four different distance
targets once each day at mid-morning. The mine site is located between the
telephotometer and one of these targets. The nephelometer runs continuously
sampling local air.
70
-------
RESULTS AND DISCUSSION
Preliminary results from the first phase (summer 1979) highlight a number
of interesting observations pertaining to the sources, nature, and transport
of fugitive emissions from surface mining. Emission sources at the mine in-
cluded a complex array of point, line, and area sources. The primary study
area concentrated on emissions generated in the vicinity of the pit and coal
storage pile. A primary haul road in this vicinity appeared to be a major
contributor to the particulate loading of the local atmosphere. The following
observations and results were recorded by the various instrumentation utilized
in the June and July-August sample periods of Phase I.
Active Particulate Samples
High volume sampler results indicate that the total suspended particulate
concentrations, near the mining and coal handling operations were quite high
(200 - 1000 yg/m3) and well above the background (annual average approximately
40 yg/m3). However, indications are that groundlevel air concentrations de-
crease rapidly downwind from the sources. Due to the location of the hi-vols
and frequency of wind direction, the majority of the dust collected by the hi-
vols was probably generated from coal storage pile and haul road activities.
On the average, total suspended particulates measured at the hi-vol located
150 meters from the haul road and edge of the coal storage were 33 percent
less than measured at 100 meters. During the July-August sampling periods,
Andersen Head particle sizing plates were operated on the hi-vols to provide
information on the particle size distribution of the suspended particulate
loads. Results indicate that approximately 35 percent (be weight) of the par-
ticles sampled were above 7 ym and 35 percent (by weight) of the particles
were below 1 ym.* Table 1 presents the hi-vol measurements for the July-
August sample periods. Figure 2 presents a representative sample (August 6-7
measurement) of Andersen Head particle size distributions.
Analyses of the wind erosion threshold velocity measurements and elemental
analysis are continuing. Therefore, the results are not available for this
report.
Passive Particulate Samples
The dust fall network provided the opportunity to observe how deposition
rates vary with distance from the mining operations. Rates as high as
25 g/m2-day were recorded immediately adjacent to haul roads, but usually
within 600 meters of a source the rates had dropped to less than 1 g/m2-day.
Topography, meteorology, and mining activity had an effect on the deposition
pattern. An example plot of dust fall rates for the July 30-August 3 period
is presented as Figure 3. In addition to total dust fall two other qualita-
tive analyses were performed on the dust fall samples. The amount of coal in
the dust fall sample relative to the amount of non coal dust was estimated for
each sample. Results indicated that coal amounts were highest around the coal
storage pile, near a haul road turn where coal was being spilled from haul
trucks, and immediately downwind from the pit (perhaps due to coal seam
*It should be pointed out that size separation was accomplished using the
standard hi-vol sampler. No 15 ym cyclone cut-off device was used; therefore
the sample may be somewhat biased to the smaller particle sizes.
71
-------
TABLE 1
HI-VOL RESULTS
(24 Hour Samples at 1200 MDT)
Dace
7/30-7/31
7/31-8/1
8/1-8/2
8/3-8/4
8/4-8/5
8/5-8/6
8/6-8/7
(meters from
major sources)
150
100
150
100
150
100
150
100
150
100
150
100
150
100
Particulate Concentrations in ug/m3 and Percent
Anderson Head Results
(Plate No. and 50% Cutof
1-1
(7 ura)
-
-
_
176
(4C.7)
_
84
(19.4)
199
(41.2)
-
113
(40.8)
197
(31.6)
77
(37.9)
102
(24.6)
191
(51.1)
264
(36.5)
2-3
(3.3 urn)
-
-
_
5]
(11.8)
_
43
(9.9)
65
(13.5)
-
39
(14.1)
78
(12.5)
£9
(14.3)
43
U0.4)
54
(14.4)
93
(12.9)
3-4
(2.0 urn)
-
-
-
29
(6.7)
_
33
(7.6)
46
(9.5)
-
33
(11.9)
58
(9.3)
25
(12.4)
23
(5.5)
48
(12.8)
78
(10.8)
: Dla. in urn)
4-5
(1.1 um)
-
-
-
23
(5.4)
_
14
(3.3)
28
(5.8)
-
22
(7.9)
44
(7.0)
10
(4.9)
13
(3.1)
25
(6.7)
44
(6.1)
Back-up
(<1 urn)
-
-
-
153
(35.4)
-
259
(59.8)
145
(30.0)
-
70
(25.3)
247
(39.6)
62
(30.5)
234
(56.4)
56
(15.0)
244
(33.7)
Total
Particulates
Equip. Fall.
Equip. Fail.
Equip. Fail.
432
262
433
483
916
277
624
203
415
374
723
72
-------
FIGURE 2
SURFACE MINE EMISSION STUDY
HI-VOL SUSPENDED PARTICULATE CONCENTRATIONS
AND ANDERSON HEAD PAUTICLE SIZE
1000
bl
i
W
K
u
§
1 i
H
, i
i:
11
W
n
900
800
700
600
500
400
300
200
100
D
DATE: 8/6 -
TIME: 1200
- 8/7
- 1200
|//yf 150 m HI-VOL
| 1 100 m HI-VOL
7/
//
/
;-:
XXXXX^I
—
k\\\\XX
ibi fb rb5
K
Total >7.0
Particulates pm
3.3-7.0 2.0-3.3
um iim
1.1-2.0
1 1;
ANDERSON HEAD PARTICULATE SIZE RANGES (\im)
-------
FIGURE 3
DUST FALL DISTRIBUTION
7/30/79 - 8/3/79
Relative
Wind Rose
-. -> - —
,
j
lw>
$££
>6
^
:"--
SCALE
; ,
-------
blasting). A plot of the amount of coal in the dust fall samples for the
July 30-August 3 period is presented as Figure 4.
During the June sampling period coal haul roads were seeded periodically
with a florescent particle tracer. When a truck would pass over the haul road
the particles would become entrained and later fall out on the dust fall net-
work. The amount of tracer material found in the dust fall samples during
analysis correlated well with the wind direction. The tracer provided confi-
dence that the dust fall network was sampling material generated from mining
activities.
Meteorology Measurements
Daily wind roses were plotted throughout the sample periods. Daytime
winds were predominantly from the west and northwest. The orientation of the
topography appeared to channel the synoptic flow into westerly local flow at
the study site. Strong, gusty winds often occurred between 1400 and 1600,
induced by local thermal convective cells. At night, a downslope (southerly
flow) was often established. Inversions occurred on most mornings and were
usually broken by 11:00 a.m. The tethersonde soundings effectively illustrat-
ed a number of these inversions. Figure 5 provides an example of soundings
at 7:00 and 12:00 on August 6. Throughout the study periods, a number of
dust devils were observed. These mini-cyclones would entrain and carry par-
ticulates to altitudes of hundreds of meters.
Air turbulence data taken by the u,v,w anemometers is still being analyzed
and,therefore,will not be presented in this report. When the analysis is com-
plete, the results will contribute to dispersion modeling exercises using data
collected during Phase I and anticipated data from Phase II.
Visibility Measurements
The integrating nephelometer draws an air sample, passes a light through
the air sample and measures the scattering of the light. It is possible to
relate the light scattering to visibility reduction by assuming that no light
is absorbed by the particles present in the air. This assumption is expected
to be quite good for low concentrations of particles of soils. Carbon parti-
cles obviously would be expected to absorb significantly, however, so that as
the percentage of carbon present increases, the actual visibility degradation
might be greater than indicated by the instrument. A major disadvantage of
the nephelometer is the fact that it takes a local, small air sample. Obvi-
ously, if the atmosphere is homogeneous, then this small sample is representa-
tive. For looking at long distances, especially in mountainous topography,
such a sample must be viewed as an indication of local conditions only.
Generally, visibility is reported in terms of visual range, defined as
the longest theoretical distance that one can distinguish a target against a
background. The theoretical limit to visibility is Raleigh scattering, namely
scattering by the air molecules totally free from any suspended aerosol.
Raleigh scattering suggests a visual range on the order of 390 km.
Table 2 and Figure 6 provide summaries of the nephelometer data taken in
western Colorado (10 km from the mine site) over an 11 month period. Visual
ranges vary from a maximum of 377 km to a minimum of 36 km. The interpretation
75
-------
FIGURE 4
QUANTITY OF COAL IN DUST FALL SAMPLE
7/30/79 - 8/3/79
Relative
Wind Rose
;
•
!
i
i
i
i
•
>
•
•
g
KEY:
Light (0-33%)
¥777% Medium (3A-66%)
Heavy (>67%)
76
-------
FIGURE 5
TEMPERATURE, RELATIVE HUMIDITY AND
WIND VELOCITY TETIIERSONDE SOUNDINGS
9VO-
MO
4OO
e
3100
I
100
too
0
DAT
TI;
t
i
1
1
j
1
I
^
\
{
Vs
"{
\
— • . -T-»— 1— — T ^
TEMP.
REL-
103O40
M )• IB M 11
TIMPtMATUflf I°C1
9O «O TO iO 00 WO
HUMIDITY \\\
WIHD SKID
• DIM err OH
PATE: S/b/79
TIME; 120O MDT
• • M> l>
MCLAftVf HUMID) rr
M l« H » J> » !• 71 30 31
WtHO SPtf D
4 OinECtlON
77
-------
TABLE 2
NEPHELOMETER
Visual Range
(km)
Month/Year
4/79
5/79
6/79
7/79
8/79
9/79
10/79
11/79
12/79
1/80
2/80
Averag««
Max
166
307
377
207
280
278
279
311
329
316
327
289
Mln
69
81
81
86
74
58
93
70
46
45
36
67
Std. Dev.
29
45
49
27
43
45
40
56
59
58
71
-
Mean
107
158
151
134
175
141
160
169
124
137
153
146
FIGURE 6
NEPHBLOMETER RESULTS
400
300
100-
78 9 .10 11
MONTH (1979-80)
12
78
-------
of these data, however, is not straight forward. Visibility decreases and in-
creases from day to day for a great many reasons. A longer record will be
needed in order to reach valid conclusions about any effects of mining activ-
ity on visibility. It is interesting to note one comparison with a telephoto-
meter observation; Malm, et al. (1979) presented results from a number of lo-
cations. One location, Dinosaur National Monument, in northwestern Colorado
could be considered to be in a visual environment similar to our study. The
average visual range they measured for summer 1978 was 195 km and for spring
1979 was 169 km. The mean value we measured for spring 1979 was approximately
139 km, while summer 1979 was 150 km. The comparison should not be inferred
to indicate anything other than the nephelometer at our site is indicating
visual ranges the same order of magnitude as at Dinosaur.
Contrast telephotometer data taken over the same 11 month period are
still being reduced, and are not presented in this report.
CONCLUSIONS
During the summer of 1979, Phase I of a two phase study was conducted to
evaluate the sources, physical and chemical characteristics, and transport of
particulate emissions from a surface mine in the semi-arid west. The study
emphasized analysis of particulates near the emission sources within a working
coal mine. A variety of air quality and meteorological instrumentation was
utilized to collect data during two sampling periods. Results highlight the
temporal and spatial variability of emission sources relative to meteorologi-
cal conditions, type of equipment used, and type and scheduling of mining
activities. Though this source variability complicates measurement and anal-
ysis of emissions, valuable information pertaining to the particulates that
mining contributes to air quality can be investigated. Phase I field programs
were conducted during the dry summer months under conditions where maximum
fugitive emissions would be expected. Particulate measurements taken during
the winter or wet spring would yield different results.
The uniqueness of a surface mining site (topography, soils-spoils, local
meteorology, depth of coal seam, surrounding land use, and other factors)
coupled with the engineering decisions employed to extract, transport and
store the mineral resource, emphasize the difficulty in translating informa-
tion gathered at one mine to other mines.
Phase II of this study will be conducted during the summer of 1980.
Phase II will emphasize the measurement of particulates at increased upwind
and downwind distances. Results from both phases will be integrated to
thoroughly address the objectives and goals of the research effort. Increas-
ing land disturbances and increasing numbers of industrial sources in the west
will increase the level of particulates in the atmosphere. What, if any,
contribution surface mining will have on air quality and visibility is a par-
ticularly difficult question to answer, but one that should be pursued. Land
managers must have information available to assist them in making environment-
ally sound decisions relative to the utilization of the nation's valuable
resources in the semi-arid west.
79
-------
REFERENCES
Environmental Research and Technology, A Comparison of Alternative Approaches
for Estimation of Particulate Concentrations Resulting from Coal Strip
Mining Activities in Northeastern Wyoming, Prepared for U.S. Dept. of
Energy, Doc. P-3545-106, October 1979.
Gillette, D. A., "Fine Particulate Emissions Due to Wind Erosion," Transactions
of the ASAE. Vol. 20, No. 5, pp. 890-897, 1977.
Malm, W. C., E. G. Walther, K. O'Dell and M. Kliene, Visibility in the South-
west, Unpublished, 1979.
Marlatt, W. E. and D. L. Dietrich, Progress Report — Surface Mine Emission
Study. U.S.D.A. Forest Service Contract No. 16-864-CA, Colorado State
University, Fort Collins, CO, October 1979.
Zeller, K. F., D. G. Fox and W. E. Marlatt, "Estimating Dust Production From
Surface Mining," Proceedings, Third Symposium on Fugitive Emissions. 1979.
80
-------
FUGITIVE EMISSIONS CONCERNS FOR COAL STORAGE AND HANDLING
AT UTILITY OPERATING STATIONS
Presented at the Fourth Symposium on Fugitive Emissions:
Measurement and Control
May 28-30, 1980
Monteleone Hotel, New Orleans, Louisiana
Peter W. Kalika and Pietro Catizone
TRC Environmental Consultants
ABSTRACT
This paper discusses the potential impact on utility operations of fugi-
tive particulate matter emissions from coal storage and handling. It is
based primarily on a study completed by TRC for a large utility.
Utilities seeking to convert to coal firing or preparing to reactivate
older coal systems will be faced with a number of concerns, including:
o What are the probable magnitudes of the fugitive emission?
o What type of fugitive emission control systems are available?
o What modeling techniques are available to assess the impact on air
quality of fugitive emissions?
o What are the design characteristics of various types of coal hand-
ling components with respect to fugitive emission potential?
o What are the ramifications of fugitive emissions control with re-
spect to BACT and LAER under the Clean Air Act?
In developing a fugitive emissions inventory and applying the findings to
decisions regarding the feasibility of resuming coal firing, the utility and
TRC found it necessary to formulate answers to the foregoing questions. The
paper presents discussions related to each of these concerns, with the ob-
jective of providing a general guideline to the utility managers who must
address these problems.
81
-------
1.0 FUGITIVE EMISSIONS CONCERNS FOR COAL STORAGE AND HANDLING
AT UTILITY GENERATING STATIONS
1.1 Introduction
The U.S. dependence on imported oil for utility power generation has
continued to grow in the years since the 1973 oil embargo. Coupled with the
Three Mile Island incident, this has given major impetus to the early as-
sessment of opportunities for increased power generation by coal combus-
tion. However, the increased utilization of coal has been cited as carrying
an inherently greater air pollution potential than any other form of power
generation. Among the potentially serious sources of air pollution from
coal-fired power generation are fugitive emissions of particulate matter
from coal storage and handling operations.
Utilities considering the switch to coal firing, or preparing to reacti-
vate older coal systems, must be concerned with fugitive emissions from the
delivery, loading/unloading, transport, storage and processing of coal. TRC
has recently completed a study for a large utility which inventoried the
fugitive emissions for a projected resumption of coal firing at selected
generating stations. The major technical problem of the study was the
meaningful quantification of the particulate matter fugitive emissions.
Direct measurement of such emissions from individual sources is an exceed-
ingly expensive process, with costs at least an order of magnitude greater
and accuracy inherently lower than for conventional stack testing. Thus
estimation of emissions by emission inventory/emissions factor techniques
although not nearly as well defined as for point source emissions, was
selected. Similarly, the nature and performance of emissions control sys-
tems for coal operations fugitive emissions are not well established, and
there is a serious lack of information on the degree to which fugitive emis-
sions can be designed out of the available options in coal handling equip-
ment.
Given this framework, a stringent schedule, and a limited budget, TRC
sought to predict fugitive emissions quantities and control performance from
information available in the open literature, and by the application of
judicious engineering judgment. The calculated emissions were then to be
used as input to a mathematical model which would predict the air quality
impact of the emissions and allow individual source contributions to be iso-
lated.
This paper reviews the findings of the study, with the objective of pro-
viding general guidelines to the utility managers who must address coal
fugitive emissions problems. The discussion will be directed to the fol-
lowing questions:
o What are the probable magnitudes of the particulate matter fugitive
emissions and how are they distributed among the coal system com-
ponents?
82
-------
o What types of fugitive emissions control systems are available for
each coal system component, and what are their probable efficien-
cies? What is considered BACT?
o What are the design characteristics of various types of coal hand-
ling components with respect to fugitive emission potential? To
what degree are these quantifiable?
o What are the ramifications of coal handling fugitive emissons con-
trol, with respect to BACT and LAER under the Clean Air Act?
o What modeling techniques are available to assess the impact on air
quality of particulate fugitive emissions, and how well can they
isolate the contributions of the major sources within the coal sys-
tem complex?
83
-------
2.0 EMISSION FACTORS
The answers to the first three questions raised in the Introduction were
sought through a comprehensive literature survey. The objective of this
survey was to determine the range of fugitive particulate matter emission
factors applicable to coal storage and handling, and the range of control
efficiencies to be expected from appropriate control techniques. The liter-
ature review encompassed the recent studies by PEDCO and Midwest Research
Institute, and the information generated by TRC's ongoing studies for EPA.
In all cases, the literature sources were critically reviewed in terms of
the means used to arrive at the emissions factors given, the type of opera-
tion for which they were developed, and their applicability to existing or
visualized coal handling operations. A literature pool of 12 references
resulted from this review, and formed the basis for the emission factors
used in the study. These are given in the reference section of this paper.
The fugitive emission factors from the literature were classified as
follows:
o coal delivery: rail car
barge
o coal transfer: belt conveyors
transfer operations
o coal storage: active storage
inactive storage
Table I summarizes the emission factors which were selected from the
literature for these classifications, and also shows further refinement of
the factors, based on a critical assessment of currently installed and ex-
pected equipment at the generating stations in question. It is this step
which requires the application of engineering judgment based on experience
The "refined" factors which appear in Table I therefore should be viewed as
somewhat site-specific, and would probably differ for other installations.
In applying the refined factors, the use of the ranges is recommended
as opposed to so-called "best estimate" single values. The carrying of a
range of calculated emissions throughout the process will serve to remind
all concerned that the emission values being developed incorporate a consi-
derable degree of uncertainty.
The form of the emission factors for storage and transfer operations
raises questions as to their physical validity. For example, the literature
did not provide a specific factor which reflected the influence of the
length of a conveyor, which seems the logical way of assessing the emissions
from this type of source (emissions per unit length). Thus the factors ex-
pressed on the basis of through-put of coal conveyed or transferred were
used, despite their obvious shortcoming in excluding the length factor.
84
-------
In the case of inactive storage, one would expect the emissions to be
related to windspeed and frequency of rainfall. While there are empirical
emission factor equations available which take these into account, most of
the literature emission factors were given in terms of the quantity placed
in storage (annual basis) and this convention was adopted despite the lack
of accounting for other physical aspects of the situation.
In the case of the active stockpile, it is recognized that there are
several specific emission-generating processes which take place; i.e.
load-in, load-out, vehicular traffic around the storage pile, and wind
pickup. Each process would be expected to have a specific emission factor.
Many literature sources did, in fact, attempt to distinguish among these
individual sources. The typical active storage operations are usually lo-
cated relatively close together, and the processes do not occur occur simul-
taneously. Thus, a single emission factor was judged to be most appro-
priate. Fortunately, such a factor was available in the literature (Ref. 4)
and was adopted for use in the study. The single value given (0.22
kg/metric ton) was assigned a range of 0.11 to 0.33, in recognition of the
likelihood that, even if obtained in a well-run experiment, such a factor
would be determined from data scatter exhibiting at least a +50% range.
85
-------
TABLE 1
SUMMARY OF UNCONTROLLED PARTICULATE MATTER FUGITIVE EMISSION FACTORS
FOR COAL OPERATIONS AT UTILITY
POWER PLANTS
Operations
Uncontrolled Refined Uncontrolled
Emission Factor (kg/metric Emission
From Literature ton) Factorj
Reliability
Ratings Found
In Literature2
00
Coal Delivery;
(1) Railcar
(2) Barge
Coal Storage;
(1) Inactive
Storage Piles
(2) Active
Storage Piles
Coal Transfer;
(1) Belt Conveyor
(2) Transfer
Operations
0.0001 - 0.2
0.0001 - 0.2
0.0553
0.223
0.02 - 0.5
0.0755
0.025 - 0.2
0.01 - 0.1
0.03 - 0.084
o.n - 0.334
0.02 - 0.2
0.04 - 0.124
B, C, D, E
C, E
B
B
D, E
N/A (E)5
-------
3.0 EFFICIENCY OF CONTROLS
Information on availability and expected effectiveness of control sys-
tems for fugitive particulate emissions from coal operations was also
gleaned from the literature. Again, application of engineering judgment led
to the refining of the literature values to provide efficiencies which could
be applied to the existing or anticipated installations under study. In the
case of control efficiency, a single "refined" value was selected from the
literature range. This specificity could seem unwarranted in view of prior
discussion regarding the uncertainties of the whole emission calculation
process. However, it was felt that "engineering judgment" is more consis-
tent with a single efficiency value than a single-valued emission factor,
and there seemed to be insufficient justification for further widening of
the range of uncertainty by assigning a control efficiency range.
Table 2 summarizes the control efficiency values which were selected
from the literature for the same coal operation classifications used in
Table 1. The "refined" efficiency value is given along with the nature of
the control technique which could be expected to provide that efficiency.
The "remarks" column gives information as to the probable maximum control
efficiency available if cost of control were not a factor. The "refined"
efficiency value and associated control type could also be viewed as pro-
viding a "practical" level of control, at least as far as the operations
which were the subject of the TRC study are concerned.
87
-------
TABLE 2
SUMMARY OF CONTROL EFFICIENCIES
FOR PARTICOLATE MATTER FUGITIVE EMISSIONS
FOR COAL OPERATIONS AT UTILITY POWER PUNTS
Operations
Control
Efficiency Range
From Literature
Refined
Control Efficiency
& Nature of Control
Remarks
Coal Delivery
(1) Rail Car
Unloading
(2) Barge
(a) Unloading
(b) Loading
90 - 99Z
70 - 90Z
50 - 90Z
95*
00
00
Coal Storage
(1) Inactive Storage 50 - 100Z
Pilea
2) Active Storage 50 - 75Z
Piles
Ventilation of
Enclosure to
Baghouse
75Z Unventilated
Enclosure
75Z Telescopic
Unloading Chute
& Local Ventilation
and Control
50Z Wetting and Chenical
Stabilization
75Z Wetting & Chenical
Stabilization Plus
Local Ventilation
and Baghouse
Complete Enclosure is Difficult
With Unit Train Dumping Systems
95Z Feasible With Enclosure,
Ventilation and Control, But
Very High CFM
Virtually 100Z if Fully Enclosed,
But Not Usually Practical With 30,
60 or 90 Day Piles at Large
Stations
95Z Feasible With Enclosure,
Ventilation and Control, But Very
High CFM
Coal Transfer
(1) Belt Conveyor
(2) Transfer
Operations
95 - 99Z
90 - 99Z
97Z Conveyors Enclosed
But Not Ventilated
98Z Enclosed; Ventilated
to Baghouse
99Z If Conveyors Properly Enclosed
& Ventilated to Control, But Very
Sensitive to Leaks
Usually Enclosed to Start With
-------
4.0 BEST AVAILABLE CONTROL TECHNOLOGY (BACT)
The requirements of the Clean Air Act Amendments of 1977 specify that
new or modified sources in certain process categories, whose emissions could
impact criteria pollutant air quality attainment areas, be required to in-
stall Best Available Control Technology, (BACT). BACT is determined on a
case by case review basis, and only limited information, as shown on Table
3, is presently available on BACT for coal handling fugitive emissions.
Table 3 shows that the required efficiency of BACT for coal handling and
storage and similar operations could conceivably range from an unspecified
"do the best you can" to complete enclosure with ventilation and baghouse
control and 99.7% efficiency. The control efficiencies obtained from the
literature and "refined" by TRC as shown on Table 2, are generally consis-
tent with the BACT values given. Only the 90% efficiency of the raw coal
storage pile dedusting agent for the Region V Coal Cleaning Plant seems in-
consistent. Lacking more detailed information, an efficiency of 50% would
be more in keeping with the information encountered in the TRC study. At
best, we could visualize no more than 75% control if the dedusting agent
were used in conjunction with carefully placed windscreens/barriers at the
active sites of the storage piles. 90% control is expected to require a
substantial degree of enclosure or covering of the pile, in conjunction with
a dedusting agent.
89
-------
TABLE 3
SUMMARY OF 8EST AVAILABLE CONTROL TECHHOLOGT DETERMINATIONS!
Process /Operation
Petrochemical
Coal Crusher
Coal Bunker
Coal Handling
FOR PARTICULATB MATTER
FOR COAL HANDLING AND
EPA BACT Strategy
Region Description
VI
Baghouse Filters
Baghouse Filters
Shrouded Conveyors;
FUGITIVE EMISSIONS
STORAGE OPERATION
Control
Efficiency
Specified
99.71
99.71
Remarks
Controlled efficiency
to include capture ven-
tilation system
(Totally inclosed)
Wa
Duie Collection at
Transfer Paints
!Tot Given
Coal Cleaning Plant
Truck Sump
Rotary breaker
Raw Coal Transfer
Jeltl
•taw Coal Sampling
Suilding
Preparation Plant
Duct Belt
Capture a Fabric Filter SOX
Fabric Filter 99*
Capture Plus Fabric Filter 30Z
Oaduscin; Agent 90Z
Dust Collector 90S
Dust Collector 991
Coal Cleaning Plant
Truck Dump
Screen'i Transfer
Points
Crusher
Surge Sins
Stock Piles
Loading
17
Enclosed Wet Suppression 90Z
Enclosed 'Jot Suppression 90Z
Enclosed Vee Suppression <>OZ
Not Given 90Z
Enclosed Wet Suppression 90X
'Jet Suppression 90Z
Assumed chat no ventila-
cion used in conjunction
with these
Electrical
Generating Station
Coal Handling
Spray Application
And Dust Collectors
Mot
Given
Detailed Control
Stragaty Info is
Given Ln Penait
Coal Classification
Plant
Aggregate Coal
Storage
Loading Areas
•HI Shale Project
Roads, Parking Areas(
3rill Pads and Shale
Deposit Areas
VTII
Water Sprays
Water Sprays 6
Chemical Controls
Hot Given
"To Greetest
Extent Possible"
NOTES:
I. SOURCE: "Compilation of SACT/LAER De
90
-------
5.0 MAGNITUDE OF EMISSIONS
The significance of particulate fugitive emissions may best be judged by
the application of the foregoing emission factors and control efficiencies
to a typical coal handling and storage situation. The calculations are sum-
marized on Table 4. These are abstracted from the TRC study, with appro-
priate simplification of the system and modification of throughput values.
The "typical" plant is assumed to burn 5000 metric tons of coal per 24
hours, which provides approximately 5500 X 106 BTU/hr heat input and is
about equivalent to a 600 MW plant.
The total fugitive particulate emissions estimated for this typical
plant are significant, even after application of controls which could cer-
tainly be considered BACT. At the high end of the range of annual emis-
sions, over 100 TPY would be emitted. The active stockpile is seen to be
the largest contributor (40% of the total), which is not suprising, since
this operation involves actively disturbing and working the coal in the
open. Significant additional control could only be achieved by completely
enclosing the operation, and by ventilating to a baghouse or other control
device. The maximum contribution of the conveyors (30 metric tons, or 21%)
seems intuitively too high. Although the selected upper value for the un-
controlled emission factor of 0.2 kg/metric ton is well supported by the
literature, its application to each conveyor section may be suspect. The
available information did not clearly state whether the conveyor uncon-
trolled factors should be applied to the entire conveyor system, or to each
section. In the TRC study, we adopted the latter, more conservative ap-
proach, and this is reflected in Table 4.
91
-------
TARLK 4
CALCINATION OF PARTICWUATE FUGITIVE EMISSIONS
FROM COAL HANDLING AND STORAttK OPFAATIONS
FDR A TYPICAL COAL FIRED GENERATING PIANT
10
Emission Source
An
-------
6.0 COAL HANDLING COMPONENTS
The design of the coal handling system components can significantly in-
fluence the magnitude of uncontrolled fugitive emissions. That is, the
degree to which the design minimizes the mechanical agitation of the coal,
will be reflected in reduced emissions. The operations which are most in-
volved with such mechanical manipulation of the coal, and which contribute
significantly to the total fugitive emissions are:
o coal unloading
from trains
from barges
o coal loading
to barges
to trains
o active stockpiles
coal delivery to the piles
coal reclamation from the piles
The emission factor literature rarely distinguishes among various avail-
able equipment arrays for accomplishing these functions. Ideally, one would
assume that the uncontrolled emission factor would range from the high ex-
treme to the low extreme as the equipment's inherent capability to disturb
the coal is decreased. Unfortunately, there is no specific information to
support this speculation and one can only surmise in a general way as to
which method (e.g., of delivering coal to a storage pile), generates the
least fugitive emissions.
Figure 1 illustrates this, schematically showing two ways of accom-
plishing several of the functions in the order of (probably) decreasing
emissions.
For barge unloading, the continuous unloader, which is a shrouded and
articulated bucket conveyor, would be expected to generate substantially
less fugitive emissions than the grapple bucket unloading method. The con-
tinuous unloader also offers the opportunity to apply reasonably effective
local ventilation.
In the case of trains, it is not apparent that a rotary car dumper de-
signed to invert coal rail cars while they remain coupled, could generate
more fugitive emissions than a single car dumper, except to the extent that
a single car is more readily enclosed. The unit train system would operate
more rapidly and thereby generate more emissions in a given time period,
given equal control effectiveness. In either case, the train unloading sys-
tem can be expected to be enclosed, ventilated and controlled, so in this
case, the degree to which one version can generate more fugitive emissions
than the other may be a moot point.
93
-------
Active stockpile coal delivery is shown in two versions. One entails a
free drop to the stockpile, while the second incorporates a telescoping
chute to essentially enclose the coal stream during the drop. it is ap~
parent that the latter method will generate less fugitive emissions.
Coal pile reclamation is shown in two systems which could be construed
as "manual" vs. "automatic". The use of earth moving equipment around the
pile, such as bucketloaders, bulldozers or graders which carry or push the
coal to underground hopper/conveyor systems, can be expected to generate
substantially greater fugitive emissions than an articulated bucket wheel
stacker/reclaimer with associated conveyors. The latter would be shrouded
and can be subjected to local ventilation and control.
In summary, the design features of coal handling component and systems
can be a positive factor in minimizing fugitive emissions from the operar-
tion. Quantification of emission characteristics for various designs is
beyond the present state of knowledge, and only educated estimates can be
used to recognize that one component array may generate less emissions than
another.
94
-------
FIGURE 1: THE EFFECT OF COAL HANDLING
COMPONENTS ON FUGITIVE EMISSIONS
UHLOftSING.BARGES
GRAPPLE
BUCKET '
„,. CONTINUOUS
'OPPE1< BASGE
•CONVEYORS UNLOAOtR
;BUCKET
COKVEVORj
UNLOAD 116
"~ TOWER
',E.'_:JHT;NUOUS
CNCLCSURE
UN3EP
HOW'S
L*i JK1T
(8) S1KCLL CAR OUHPfR
ACT.L CTOCy.'I^'.
C0nu JEUVERV
'OWES
ENCLOSED
iOOH COKVEYOP
RECLtHtTION FROM PILES
ACTIVE
PILE
tUCkET LOADER
—0!! OThtP. C3AL
MOVIN& MACHINERY
(S)-CQ
RECLM1ATIOH F?0>< ACTIVE
STOCl.PiLE Br COAL MOVISu
CHUTE EHCLOSED DW TO STQQ.PlLi
_ 6UCSET WHEEL
STACKER-HECL^MEP
/
(0; RECLAHAT1W* FROM ACTIVE
STOCf.rlLE BY SUO.n il^ii.i
-------
7.0 MODELING OF FUGITIVE PARTICULATE MATTER EMISSIONS
The foregoing sections nave described concerns and methodology related
to the quantificaton of fugitive particulate matter emissions, and the means
available for their control. Determination of the impact of these emissions
on ambient air quality and their specific contribution to ambient concentra-
tions requires the use of sophisticated mathematical modeling techniques.
These techniques also are needed to assess tradeoffs in the application of
controls to individual fugitive emissions sources, and to nearby point
sources.
The immediate need for such refined analytical techniques has been
acknowledged and met by EPA with the recent development of advanced disper-
sion models. The Industrial Source Complex (ISC) Dispersion Model developed
by the H.E. Cramer Co. Inc., and the CDMDEP Model, a modified version of the
Climatological Dispersion Model (COM), developed by TRC Environmenal Consul-
tants are two such models prepared for the EPA.
The ISC Dispersion Model combines and enhances various EPA dispersion
model algorithms into a set of two computer programs that can be used to
assess the air quality impact of emissions from a wide variety of sources
associated with an industrial source complex such as a Power Generating
Station.
The ISC and CDMDEP models have been designed to account for mechanisms
which influence the dispersion and quantification of particulate plumes
These mechanisms include:
o dry deposition
o gravitational settling
o washout
o building wake effects
o vertical wind profile variations
o plume rise
Stack, Area and Volume sources can be simulated by the ISC Model programs.
Long-term concentrations of total suspended particulate matter due to emis-
sions from point and area sources can be estimated by CDMDEP.
Short term or long-term source-receptor relationships can be derived
using the ISC Dispersion Model. The ISC Short Term Model (ISCST) is de-
signed to calculate concentrations or deposition values for time periods of
1, 2, 3, 4, 6, 8, 12 and 24 hours. The ISC Long Term Model (ISCLT) uses
STAR Meteorological Summaries to calculate seasonal and/or annual ground
level concentrations or deposition values.
96
-------
The input data required to run either ISC or CDMDEP consists of four
categories:
o Meteorological Data
o Source Data
o Receptor Data
o Program Control Parameters
The details of the input data are given in references 13 & 14. The output
from these mathematical models provides a quantitative estimate of the
spatial distribution of air quality, averaged over the time period con-
sidered. Contributions from major sources within the industrial complex can
be identified and appropriate controls prescribed. It should be noted that
these advanced, complicated models are intended to be used by Air Quality
Engineers and/or Meteorologists familiar with effluent dispersion charac-
teristics and computer techniques. The models, with proper fundamental
knowledge, can be applied to other types of studies such as:
o Stack design
o PSD analyses
o Source Permit Applications
o Monitoring Network Design
o Regulatory Variance Evaluation
o Control Strategy Evaluations
o Fuel Conversion Studies
o New Source Reviews
o Analyses of Emission Offsets
97
-------
8.0 CONCLUSIONS
This paper has reviewed a methodology for estimating the quantities and
degree of control for fugitive particulate matter emissions from coal hand-
ling operations. Available information in the form of emission factors and
estimates of control efficiencies was found to be relatively sparse and to
vary widely. Information on BACT controls is also limited at present. Con-
siderable judgment must be incorporated in the application of information.
The estimation of well-controlled fugitive particulate emissions from a
typical coal-fired plant results in significant residual emissions, of which
a large proportion is due to actively working the stored coal in the open.
More information on the relationship between coal handling equipment design
characteristics and fugitive emissions is needed.
Mathematical Models are available which will discriminate among various
fugitive emission sources and elevated point sources to isolate their in-
dividual impacts on ambient concentrations of suspended particulate matter.
Application of these models completes the assessment of fugitive emissions
and leads to the formulation of cost-effective control strategies.
There is at present enough information to only generally guide those
managers who must address coal fugitive emission problems at utility sta-
tions. Each site will have its specific problems and constraints. The
manager is urged to approach .the problems utilizing a broad team of ex-
perts. Air pollution control engineers, power generation specialists, dif-
fusion meteorologists, coal systems designers, environmental consultants,
and environmental attorneys can all provide important inputs to the problem
solutions. The success and cost effectiveness of the solutions will depend
on how well the efforts of the team are orchestrated.
98
-------
9.0 REFERENCES
The following references are pertinent to the quantification and control
and modeling of fugitive particulate matter emissions from coal handling and
storage operations.
1. U.S. Environmental Protection Agency. Compilation of Air Pollutant
Emission Factors. 3rd. ed., Including Supplements 1-7. U.S. EPA.,
Research Triangle Park, NC. August 1977. AP.-42. pp. 11. 105; 11.2-1
to 11.2-5; 11.2.3-1 to 11.2.3-2.
2. Currier, Edwin L., and Neal, Barry D., "Fugitive Emission from
Coal-Fired Power Plants", paper 79 - 11.4, Air Pollution Control
Association, 72nd Annual Meeting, Cincinnati, OH., June 24-29, 1979. pp.
15.
3. Jutze, George A.; Zoller, John M; Janszen, Thomas A.; Amick, Robert S.;
Zimmer, Charles E.; and Gerstle, Richard W. Technical Guidance for
Control of Industrial Process Fugitive Particulate Emissions. PEDCO.
Environmental, Inc., Cincinnati, OH.March 1977.EPA-450/3-77-010. pp.
2-5 to 2-43.
4. Cowherd, Chatten, Jr., et al. Development of Emission Factors for
Fugitive Dust Sources. Midwest Research Institute, Kansas City, MO.
June 1974. EPA-450/3-74-037.
5. Blackwood, T.R. and Wachter, R.A. Source Assessment! Coal Storage
Piles. Monsanto Research Corporation, Dayton, OH. Sponsored by U.S.
EPA, Cincinnati, OH. EPA-600/2-78-004k. May 1978. 83 p.
6. Bohn, Russel; Cuscino, Thomas Jr.; and Cowherd, Chatten Jr. Fugitive
Emissions from Integrated Iron and Steel Plants. Midwest Reseach
Institute, Kansas City, MO. Sponsored by U.S. EPA, Research Triangle
Park, NC. March 1978. EPA -600/2-78-050. p. 262.
7. Axetell, Kenneth Jr. Survey of Fugitive Dust from Coal Mines. PEDCO
Environmental, Inc., Cincinnati, OH. February 1978. EPA-908/1-78-003.
8. Ambrose, D.; Brown, D. and Clark, R., "Fugitive Monitoring at a Coal
Clenaing Plant Site." In Second Symposium on Fugitive Emissions;
Measurement and Control. May 1977, Houston, TX. TRC - The Research
Corporation of New England, Wethersfield, CT. December 1977.
EPA-600/7-77148. p. 64-113.
9. Boscak, V., and Tandon, J.S. "Development of Chemical(s) for Supression
of Coal Dust Dispersion from Storage Piles." Paper presented at 4th
Annual Environmental Engineering and Science Conference, University of
Louisville, Louisville, KY. March 4-5, 1974. p. 14.
99
-------
10. Amick, Robert S. and Wisbith, Anthony S. Fugitive Dust Emission
Inventory Wasatch Front, Utah. PEDCO-Environmenta 1 Specialists, Inc/"~
Cincinnati, OH. Sponsored by U.S. EPA, Denver, CO. July 1975.
EPA-908/1-76-001. p. 20.
11. Cross, Frank L., Jr., and Forehand, Gerald David, editors. Air
Pollution Emissions from Bulk Loading Facilities. (EnvironmentaT
Monograph Series, V.6.) Technomic Publishing Co., Inc., Westport GT
1975. p. 22
12. Zoller, John M.; Wood, Gilbert H.; Jansen, Thomas A., "Current Status of
Process Fugitive Particulate Emission Estimating techniques.: In Second
Symposium on Fugitive Emission; Measurement and Control. May 19777"
Houston, TX. TRC - The Research Corporation of New England*
Wethersfield, CT. December 1977. EJPA-600/7-77-148. p. 385-456. *
13. Bowers, J.F.; Bjorklund, J.R.; and Cheney, C.S. Industrial Source
Complex (ISC) Dispersion Model Users Guide, Volume 1. EPA, Contract Nb7
68-02-2507, H.E, Cramer Co. Inc., April 1980.
14. Bowne, N.E.; Wackter D.J.; Lazorick, S.W. TRC; Final Report to EPA
CDMDEP - A Climatological Dispersion Model for Sources of Suspended
Particulate Matter! EPA, Contract No. 68-02-2615, Task No. 4, September"
100
-------
Design, Performance Testing and Field Operation of
an Isokinetic Electrostatic Particle Sampler
Bengt Steen
Swedish Water and Air Pollution Research Institute,
P.O. Box 5207, S-402 24 Gothenburg, Sweden.
Abstract
An isokinetic electrostatic particle sampler is described. Its perfor-
mance is investigated both in laboratory and field operation.
The sample flow is generated from the inertia of the surrounding air
flow. The particles in the sample flow are deposited inside a cylindrical
collector electrode, which is weighed before and after the sampling period.
Afterwards, the sample can be further analyzed for its chemical composi-
tion or physical properties. The particle flux is obtained by dividing
the inlet nozzle area with the sample mass and the sampling time.
The collection efficiency has been studied under various controlled con-
ditions. It is seldom lower than 90%, but often above 97%. Wind tunnel
tests have been performed to study the deviation from isokinetic sampling.
The maximum deviation found was 8% in the 0.5 - 26 m/s range.
The sampler may be used for measurement of fugitive dust emission, its
operation being wind-directed and the sampling being merely in a 45-
degree-sector or similar. Evaluation procedures for fugitive emissions from
ideal point sources, line sources and surface sources are presented.
Experiences from field measurements are reported. Measurements have been
performed in roof ventilators and around an open stock pile area of a
smelter industry. In the latter case 13 samplers placed in 4 masts were
operated during one year.
101
-------
Design, Performance Testing and Field Operation of
an Isokinetic Electrostatic Particle Sampler.
Bengt Steen, Swedish Water and Air Pollution Research Institute
General
Measurement of fugitive dust emissions are more difficult to perform than
other dust emission measurements. Several factors complicate the measure-
ment:
- particle sizes are often large
- changes in both flow velocity and direction may occur
- part of the emission may be occational, for instance
in connection with high wind speed. This requires
long sampling periods to obtain necessary statistics
- the extension of areas with high fugitive emission
source strength is often vaguely known
- areas with fugitive emission are often large, which
calls for many measurement points
Taking these complicating factors into account emphasis should be laid upon
two qualities of a good measurement method for fugitive dust emissions*
- isokinetic sampling
- low cost
Five years ago we developed a sampler at our institute to meet these
quality demands (1).
Design
After some field experience the sampler was modified. The new design is
shown in Figures 1 and 2.
The sampler consists basically of a nozzle, a cylindrical collecting elec-
trode, an emission electrode, a flow regulator, a shaft and a high voltage
generator.
During sampling, the axis of the cylinder is parallel with the flow lines
and the fluid is allowed to pass isokinetically through the nozzle by means
of its own inertia. The particles are charged and collected on the inner
cylinder wall by electrostatic forces. The cylinder is weighed before and
after the sampling to determine the particle mass. The particles may also
be analysed by chemical or physical means.
102
-------
The particle flux is determined by dividing the sample mass with the nozzle
area and sampling time.
The nozzle diameter is such that approximately 25 times reduction of the
air velocity is obtained inside the sampler. The combined dimensions of
the nozzle, the collecting electrode, the outlet holes and the venturi
screen result in an isokinetic flow through the nozzle tip.
Performance tes ti ng
Measurements of the flow velocity in the nozzle and in the main gas stream
showed little or no difference from isokinetic conditions, (table 1).
The deviations are such that to some degree they compensate a decreasing
sampling efficiency at high wind speed.
Flow velocity Flow velocity
in main gas stream in nozzle tip of Note
the sampler
(m/s) (m/s)
0.5
0.8
2.3
3.4
5.5
9.8
12.3
25.8
0.5
0.8
2.3
3.4
5.6
10.0
13.0
28.0
measurements made by hot
wire anemometer at room
temperature
_«_
-»-
„"-
_u_
measurements made with
static and dynamic pressure sensors
at 260 °C.
Table 1 Results of flow velocity measurements in nozzle tip and
in main gas stream
The collection efficiency of the sampler has been tested at various con-
ditions. The influence of particle size, concentration, flow velocity,
amount of collected material, particle material, and battery current
has been studied (1).
The experimental set-up consisted either of: (a) the sampler, a filter,
a dry gas meter and a pump; or (b) the sampler, a Royco 220 light scatter-
ing particle counter and a pump.
Table 2 shows the results obtained by using a Royco particle counter when
determining the concentration of different particle size classes present
103
-------
in a room atmosphere before and after the sampler. The sampler was used
without a nozzle and with a flow velocity of 1 m/s inside the collection
electrode. The particle number concentrations were low, around 1000/1 in
the 0.3-0.4 \im range and 10-21/1 in the 3.4-6.6 ym range.
Latex equivalent
particle diameter
Sampling efficiency (%)
Battery current (mA)
6 8 10 50
196
Table
Table
0.3-0.4 73.7 83.2 93.7 97.6
0.4-0.6 72.9 83.8 94.5 97.5
0.6-1.0 75.6 86.1 96.0 98.0
1.0-1.8 77.6 92.2 96.3 99.3
1.8-3.4 78.7 92.4 99.7 100
3.4-6.6 78.2 92.9 98.7 100
6.6-13 88.3 (80.0) 100 100
2 Collection efficiency of sampler for different parti
size classes at different battery currents
3 Collection efficiency obtained from measurements of
types of particles.
99.0
98.7
99.2
100
100
100
100
cle
different
Gas velocity
at nozzle
(m/s)
0.8
0.8
13.6
0.8
0.8
0.8
Nozzle Battery Type of Amount of Collection
d1a. current particles collected efficiency
(Tim) (mA) | dust (mg) %
25 50 I Redispersed 27.8 99
1 soot from
oil combustion
25 50 | Redispersed 8.7 98
Na^SO. from
kraft pulp mill
recovery furnace
25 50 Talc powder(mass 23.6 98
mean aerodynamic
diameter « 2. 5 urn)
5 -x.150 Emissions from 389 97.5
electric arc
steel furnace
25 50 | Fe2 03 8.4 100
25 50 Particles in 5.3 91
ambient air
25 50 Sulphate in 0.201 87
particles in
ambient air
104
-------
The collection efficiency obtained by the filter method for various types
of particles are shown in Table 3.
Comparative measurements between two samplers at different angles were
made in a wind tunnel. The purpose was to study the influence of deviations
between the direction of the main flow and sample flow on sample flow velo-
city.
Talc powder with a mass mean aerodynamic diameter of 2.5 ym was used at a
wind velocity of 4 m/s. The result is shown in Figure 3.
Selection of sample points and data evaluation
Since the position and extension of a fugitive dust source generally are
not known, many measuring points in an envelope surface around the source
are needed to determine the total emission. By using an infinite amount of
measuring points the location and evaluation are fairly easy to make.
However, economic realities will put a limit to the number of measuring
points. Presumptions or knowledge of the position and character of a certain
source usually enable adequate measurements of the emission even by using
a few samplers. A few such cases will be discussed below.
In all cases it has been assumed that the background is zero and that nothing
of the emitted dust has been deposited between the emission point and the
sampler.
An ideal point source emits dust from one single well-defined point.
Welding, loading- and unloading operations, minor storing places and open
fires usually could be treated as ideal point sources. The size of the
emission from an ideal point source is here supposed to be independent of
the wind direction.
If one puts an imaginary cylinder with vertical axis and with the bottom
surface concentrically with the source the particle flow through the enve-
lope surface of the cylinder will be identically with the emission, provided
that the cylinder wall is high enough and that the background is low.
The particle flow through the envelope surface could be determined by using
samplers located at different heights in the cylinder wall. If one mast
with n samplers is used the emission will be the following
mi • 8 rh
n • dz • f
105
-------
where
m. = quantity of sample in the ith filter
n = number of samplers
r = radius of cylinder
h = height of cylinder
d = dia. of the sampler nozzle
f = relative frequency of wind direction in the sampling sector
(Frequency of wind direction in the sector divided with the
frequency in the sector if the wind was blowing with equal
frequency from all directions)
If the background is high the particle flow coming in through the opposite
side of the cylinder has to be measured and subtracted from the flow of
the leeward side of the source.
An ideal line source is here defined as a source which form looks like a
line and which emission pro length unit is the same along the whole line.
A road, a ditch or belt conveyor usually could be regarded as an ideal
line source.
If we look at a length increment, dL at the line source as a point source,
the following is valid
Zmi,L
8-h-r
= dL- a
n-d2 • f
where
a = the emission per length unit, and ID. , is that part of the sample
which emanates from the length unit dL. If we integrate over the whole
length of the line source (L«) in the measuring sector we get:
' 8>h C-Q HI
' = a . I u _dk_ (i)
n • d • f L=0) r
This integral will here be solved for the case that the line source is
a straight line which crosses both sides of the sampling sector (see fig. 4).
106
-------
The equation of this straight line equation expressed in polar coordinates
is:
r = cosa - k sina where k and p are constants.
Further L = -v/(p - r-cosa)2 + r2sin2a
_)2 , P^ ' sin2a «
/ />****».» If & ^ _ . i ~
. .
cosa - k • sina (cosa - k • sina )
/k2 • sin2a + sin2a = p . sina . f~\
^ (cosa - k • sina)2 ~ cosa -sina V k +
dL _ _.,A.2 , sina • (sina + k • cosa ) + (cosa - k-sina) • cos a
-r- - fjy K. 1- I • - -< - - - - ' -
(cosa - k • sina ) *
= pjk2 + I
v
(cosa - k • sina )
(1) then would be expressed as:
a,
,- 8 • h ( /2
= a- I p-yk + 1 -(cosa - k • sing) da =
I
n • d • f JL (cosa - k • sina ) * • p
a
/ I,i\/l^i1 J.4-,
= a-ln
If a is released we get:
£m. • 8 • h / (k + ^/k2 + l +tg|) (k -y/k2 + 1 ) "
n - d2 • f I " (k-./k2 + l +tg|) (k+Jkz + 1 )
Suppose that the extension of the surface source is limited by four lines
which in polar coordinates are the equations r = r, r = r2,a = 0 and
a = a, (a is given in radians). The measuring point 'is placed in origo.
In thi same way as earlier we will get the flow through a vertical slit
in origo:
107
-------
'-' 4 '• '- r • da • dr • e • f
n • IT • d 1 | 2-ir.r-h
e = the emission per surface unit.
m. - 8 • h
Then we will get: e = ~~
f • n • d • a1(r2 - r, )
Field experience
The field experience with the sampler includes measurements in roof venti-
lators and a one-year measurement of the emission from a stock yard of a
smelter industry in northern Sweden.
An example of the location of samplers in a roof ventilator is shown in
Figure 5.
Figure 6 shows a particle sample collected in a roof ventilator above an
electric arc steel furnace. The particles were collected on an aluminum
foil.
With a conventional technique including flow measurements and isokinetic
sampling the same measurements would require several times the man power
and equipment investments.
£2£lLy§rd_measurement
Fugitive emissions from a stockyard were measured using 13 samplers during
one year.
The samplers were mounted at light masts in the area as shown in Figure 7.
They were directed towards different fugitive dust sources (Figure 8).
The sampling took place with the wind blowing from these sources against
the sampler and in a sector of - 22.5 degrees from the main direction.
One month samples were taken. The particles which had isokinetically passed
through the sampler nozzle were collected on an aluminum foil, which weight
was determined before and after the sampling period. The aluminum foil then
was rinsed in distilled water and the suspension was analysed for Cu, Pb,
Zn, As and mercury by means of ASV and atomic absorption spectrometry.
108
-------
The fugitive emission was calculated in three ways. The results differ
only slightly and is shown in Table 4. The losses of Cu and As is approx-
imately 1 - 2 % of the handled amount of material.
Table 4 Total amount fugitive emission calculated in three ways
I Estimated total emission ton/year
1 0 13 I
Calculation criteria | Cu Pb Zn As Hg Dust
Total emissions from single ,
point sources. I 15 7.5 140 1.7 0.23 2800
Total fugitive dust transportation
from material handling yards, cal- '
culated from the transportation |
above the south bank (the poles A
and B) which was multiplied with I
3.7 (the area is presumed surrounded j
by a square. Through one of this
sides the flow was measured) I 6.9 (25)x 110 1.4 0.055 2200
Mean dust flow in the area without '
consideration of the location of |
the samplers multiplied with the
envelope surface of a cylinder with '
a radius of 400 m and a height of [
25m (will surround the area) , 14 (35) 220 2.5 0.18 2200
the lead storage yard at the
A-pole disrupt the result
109
-------
References
(1) B. Steen, "A new simple isokinetic sampler for the determination of
particle flux". Atmospheric Environment, Vol 11, pp. 623-627 (1977).
no
-------
Emission electrode
Nozzle
Isolating support
Water proof
box
FIGURE 1 Isokinetic electrostatic particle sampler
Principal sketch
111
-------
1) Housing for high
voltage generator
2) Shaft with high
voltage wire
3) Connection with
teflon plug
4) Flow regulator
5) Ball bearings
6) Collecting electrode
^ m.
© il
7) Relay
8) Magnet for 7)
9} Nozzle
10,11) Two position nozzle valv«
12) Magnet
13) Adjustment screw
14) Vanes
FIGURE 2 Wind-directed measurements
112
-------
Sampling efficiency (weight %)
Flow
06 direction
Angular
deviation
(00°)
0 10 20 30 40 50 60 70 80 9O
FIGURE 3 Percentage of main flux of talc dust obtained in a
sampler oriented angular to the main flow direction
at 4 m/s wind velocity
113
-------
Line source
FIGURE 4 Line source
114
-------
FIGURE 5 Measurement of fugitive emissions from roof
ventilators. Location of samplers.
115
-------
FIGURE 6 Unfolded foils with dust collected in a roof
ventilator above an electric arc furnace.
116
-------
Cu- material
unload :2.5 kton/yr
material
unload:
12-14
kton/yr
material
unload:
8-10 kton/yr
Pb-
aterial
unload:10:12 kton/yr
load /^10 12
FIGURE 7 Simplified overview of raw material handling
and measurement sectors
117
-------
FIGURE 8 Samplers mounted on masts.
118
-------
ASSESSING HAZARDOUS WASTE TREATMENT FACILITY
FUGITIVE ATMOSPHERIC EMISSIONS
By
Tim S. Sekulic
and
B.T. Delaney
Fred C. Hart Associates, Inc.
New York, New York
ABSTRACT
A program of sampling and analysis for the identification and quantifi-
cation of atmospheric emissions and the development of emission factors for
all aqueous waste storage, treatment, and disposal processes at a hazardous
waste management facility has been developed. Samples for atmospheric
emission characterization will be obtained via at least two complementary
methodologies, wherever possible, as follows. At the facility's aqueous
waste receiving and physical/chemical treatment lagoons, composite liquid
samples will be obtained for equilibrium vapor analyses, and in situ surface
sampling, using an enclosure technique, will be performed. At the aerated
biological oxidation ponds, these methods will be supplemented by an emis-
sion profiling method.
The organic vapor content of all air samples will be analyzed with a
Century Systems Organic Vapor Analyzer (OVA) with a self-contained gas
chromatographic/flame ionization detector system. Initally, OVA analyses
will be duplicated using a laboratory gas chromatograph/mass spectrometer
for verification and calibration of results. Additionally, impinger absorp-
tion samples will be collected for the determination of ammonia, amine, and
acidic vapor emission rates.
Measured pond and lagoon emission rates will be used to develop emis-
sion factors (mass emission rate per unit of surface area), and mass trans-
fer coefficients will be determined for a theoretical emission rate ex-
pression which has been developed. Further, sampling results will be used
to identify potentially troublesome (e.g., toxic or odorous) emissions from
any process. A meteorological monitoring station is also being installed
for correlation of meteorological conditions to sampling results and for
future analysis and prediction of atmospheric emission problems.
Sampling results are not yet available for discussion; therefore,
concentration is given to the details of the planned sampling and analytical
program and the computation methodology, for characterizing the facility's
waste treatment pond and lagoon emissions.
119
-------
INTRODUCTION
Hazardous waste management facilities may contain significant sources
of both stack or point and fugitive atmospheric emissions, depending upon
the nature of the waste materials and the particular processes and opera-
tions employed. Methods for measuring point or stack emission rates are
well defined, and emissions of many point sources have been fairly exten-
sively quantified. Fugitive emissions, however, have received much less
attention and have not been well quantified.
The sampling and analysis program described here has been developed for
the identification and quantification of fugitive atmospheric emissions from
aqueous hazardous waste treatment processes. The results of this program
will be used to develop emission factors and mass transfer coefficients for
use in a theoretical emission rate expression which has been developed, and
to identify potentially troublesome (e.g., toxic or odorous) emissions.
Although a comprehensive program for emission characterization involving all
of the treatment, storage, and disposal operations of hazardous waste treat-
ment facilities has been developed, the discussion here concentrates pri-
marily on the aqueous waste ponds and lagoons. Sampling results are not yet
available, so this discussion is limited to the details of the planned
sampling and analytical program and the emission rate computation metho-
dology.
BACKGROUND
Hazardous Waste Management Facility -- Aqueous
Waste Treatment Processes/Emissions
The facility for which this atmospheric emission characterization
program has been developed is a relatively advanced, comprehensive pro-
cessing, recovery, and disposal plant for a wide range of liquid and solid
hazardous wastes. The capabilities of this facility include the bulk trans-
fer of wastes; solvent and fuels recovery; physical, chemical, and biologi-
cal treatment of aqueous wastes; and secured scientific landfill of solid
material. Since the emphasis of this paper is the characterization of
atmospheric emissions from the facility's aqueous waste ponds and lagoons,
the aqueous waste treatment system will be reviewed here.
The aqueous waste treatment system consists of four unit operations:
(1) chemical oxidation and reduction; (2) neutralization and precipitation;
(3) carbon adsorption; and (4) biological treatment (Figure 1). Incoming
aqueous wastes are sent to the oxidation-reduction system, consisting of
120
-------
FIGURE 1
SLUDGE TO
-^•SCIENTIFIC LANDFILL (SLF)
HOLDING
LAGOON
REDOX
SYSTEM
ACID
NEUTRALIZATION
SYSTEM
AIR
. COLD WEATHER ONLY
FACULTATIVE
POND
(COLD WEATHER)
±
pH ADJUSTMENT
TANK
HIGH RATE
BIOAEROBIC
TREATMENT
CARBON
ADSORPTION
SYSTEM
AIR
*. DISCHARGE
FACULTATIVE
POND
SYSTEM
AQUEOUS WASTE
TREATMENT SYSTEM
-------
three lagoons in series, in which the wastes are chemically oxidized and
reduced. Wastes containing reducing agents are sent to one lagoon; wastes
high in organic matter or containing oxidizing agents are sent to a second
lagoon; and wastes from these two are mixed together into a third lagoon,
which is always maintained in a slightly reducing state to promote heavy
metal precipitation.
Following oxidation-reduction, the waste enters the neutralization-
precipitation system. Here the pH of the waste is increased to approxi-
mately 11.5 by the addition of lime in an agitated vessel. The resulting
alkaline slurry is then conducted to a gravity separation area, the salts
lagoons, where the heavy metals settle out as metal hydroxides.
The supernatant from the salts lagoons is transferred to a holding
pond, after which its pH is adjusted to approximately neutral, and it is
sent through sand filters prior to treatment in the carbon adsorption sys-
tem. In the carbon adsorption system, the organic content of the waste,
particularly the high molecular weight organics, is reduced substantially.
Following carbon adsorption, the waste enters the biological treatment
system. The first stage of biological treatment is a high rate aeration
system consisting of a 500,000 gallon tank with a 40 horsepower surface
aerator. This is followed by the facultative pond system, consisting of a
series of large, aerated ponds, where biological reduction of the waste's
organic matter continues. The final step is the storage of the treated
wastewater in aerated ponds prior to discharge into natural water bodies.
The many open ponds and lagoons of this aqueous waste treatment system
expose large surface areas from which dissolved contaminants may be emitted
to the atmosphere. The raw waste contains a wide range of volatile orga-
nics, as well as dissolved inorganics, acids, and bases. Oxidation-
reduction reactions and changes in pH in the redox and neutralization sys-
tems result in the formation and potential volatilization of partially
oxidized organic species as well as inorganic gases such as ammonia, amines,
and H2S. Further, in the biological treatment system, the emission poten-
tial is enhanced by the aeration of the liquid. Here, volatile organic
matter remaining dissolved in the water as well as products of biological
degradation of these organics (e.g., amines, ammonia, aldehydes, ketones,
alcohols, and H^S), may be released to the atmosphere.
Due to the widely varying sources of waste handled by this facility,
the nature and composition of the waste varies extensively. Wastes received
at the aqueous waste treatment system include such materials as the fol-
lowing: acids and alkaline solutions, some of which contain heavy metals;
aqueous solutions containing various organic contaminants; and water conta-
minated with any of a number of potentially hazardous materials such as
phenol, PCB, or heavy metals.
Because of this vast variation in waste composition, atmospheric emis-
sions from this sytem cannot be predicted based on the concentrations of
specific components of the waste, and a sampling and analysis program for
characterizing the emissions cannot be designed for specific chemicals in
122
-------
the entering waste streams. Rather, the prediction of emissions and the
design of a sampling program must be comprehensive and generic in nature.
Emissions considered must include those categories of classes of materials
most likely to be emitted based on (1) their presence in substantial concen-
trations, either because they are major constituents of the types of waste
materials typically treated, or because they are typically formed during
biological degradation of these wastes; and (2) their volatility and diffu-
sivity.
Emission Estimating Methodology
A methodology has been developed for estimating atmospheric emissions
from aqueous waste ponds and lagoons based on the concentrations, volatili-
ties (as determined by vapor pressure and activity coefficient, or Henry's
law constant), and diffusiyities of the wastewater contaminants. The metho-
dology is founded upon basic two-phase mass transfer theory, and preliminary
values for empirical constants of the methodology's mass transfer expression
have been developed from data reported in the literature. Work dealing
specifically with the evaporation of volatile components of aqueous waste
provided limited data for the computation of wastewater atmospheric emission
rates (Bondurant 1973, Englebrecht 1961, Fruman, Gaudy 1961, Ledbetter 1978,
Thibodeaux 1971 and 1978, and Tischler). However, a great deal of data,
useful for determining mass transfer rates, has been reported in association
with related subjects such as evaporation from spills (Wu), evaporation of
water from exposed surfaces (Liss 1973, Harbeck 1962), evaporation of compo-
nents of dilute solutions (Dilling 1977), dissolution of gases into the
ocean (Broeker 1974, Liss 1974, National Academy of Sciences 1976, Schooley
1969, SI inn 1978), dissolution of oxygen into water (Owens 1964), and gas
scavenging by rain (Hales 1972).
The basic premise of the methodology is that the rate of emission of
dissolved materials from the surface of ponds and lagoons is a function of
the driving force for mass transfer across the water-air interface. The
main bodies of the water and air phases are assumed to be well mixed, so
that the water-air interface, considered as a two-layer film, is assumed to
offer the primary resistance to the transfer of mass from the liquid to the
gas phase, resulting in the concentration profile depicted in Figure 2.
The transfer of mass across the interface may be expressed by applying
Pick's first law of molecular diffusion to the two-layer situation as fol-
lows:
F=KAC=kL(CLB-CLI)=kG(CGI-CGB);
where F is the interfacial flux of mass per unit of surface area, k, and kr
are the liquid- and gas-phase mass transfer coefficients, and C represents
the solute concentration in the bulk liquid (LB), liquid interface (LI), gas
interface (GI), and bulk gas (GB). Further, the liquid- and gas-phase mass
transfer coefficients may be expressed in terms of molecular diffusivity
(D), and the thickness of the interfacial film (Z) as follows:
123
-------
FIGURE 2
CONCENTRATION GRADIENT FOR MASS TRANSFER ACROSS
THE WATER-AIR INTERFACE"
BULK
LIQUID
LIQUID
FILM
GAS
FILM
BULK
GAS
'GB
Figure Notation
C -- Solute Concentration
LB — Bulk Liquid
LI -- Liquid Interface
GI -- Gas Interface
GB — Bulk Gas
Z -- Film Thickness
124
-------
k =
The gas film interface concentration may be expressed in terms of the liquid
film interface concentration, using an appropriate partition coefficient, M,
assumed to be the Henry's law constant or the product of the solute's acti-
vity coefficient and vapor pressure, as follows:
Substitution into the expression for flux, assuming the bulk air phase
concentration equals zero, and simplifying, yields the following expression
for transfer of mass across the water-air interface:
kl
F = k,C1Q (1 - L *
1WLB Vi kGM + kL }
The liquid film thickness is not highly variable (National Academy of Sci-
ences 1976), and the liquid diffusivity, which is a function of the system
temperature and the solute molecular weight and inter-molecular attraction,
may be assumed to be fairly constant for the range of conditions and solutes
of concern. Thus, based on published data (National Aceademy of Sciences
1976, Liss 1973 and 1974, Broeker 1974, Thibodeaux 1978, Schooley 1969,
SI inn 1978, Owens 1964, Hagbeck 1962), a value may be derived for k. of
approximately 3 to 5 X 10 cm/sec, the variation in the gas phase mass
transfer coefficient is much greater, due to the greater variability of gas
phase diffusivities and because of the sensitivity of the gas film thickness
to environmental conditions, particularly wind speed. Based on limited
published data (Liss 1974, Schooley 1969), a first approximation of a mean
value may be derived for the gas film thickness of 0.26 cm.
Emissions from aerated surfaces may be treated in a manner similar to
the above treatment for quiescent surfaces, by accounting for the increased
surface area created by the aerators and the enhancement of the gas film
mass transfer coefficient. Data are available for estimating the increased
surface area created by aerators (Bondurant 1973), and estimates of mass
transfer coefficients for such systems have been developed (Thibodeaux
1978).
As present, however, emission rates may not be predicted with confi-
dence for either quiescent or aerated surfaces of wastewater ponds and
lagoons because of the paucity of data for determining mass transfer coef-
ficients for these systems. One purpose of the proposed atmospheric emis-
sion sampling program is to furnish these data.
125
-------
PROCEDURE
Sampling
A number of sampling methodologies were considered for use in the
emission characterization program. These included upwind-downwind sampling
(Kolnsberg 1976a), a quasi-stack technique (Kolnsberg 1976b), emission
profiling (Cowherd 1977), source enclosure methods (Zimmerman 1978, Adams
1979), and vapor space analysis techniques. The program which has been
developed involves the use of a vapor space analysis technique in conjunc-
tion with an enclosure sampling method for the non-aerated aqueous waste
lagoons, and emission profile sampling for the aerated ponds.
Vapor Space Analysis
The vapor space analysis is used to identify and determine the concen-
trations of species in the vapor phase at equilibrium with the aqueous
solutions of concern. This analysis provides for the identification of
solutes which are present in solution at sufficient concentration and which
are sufficiently volatile to be released from the pond or lagoon into the
air above it. Further, since the vapor phase equilibrium concentrations of
the solutes are determined, this analysis provides a direct indication of
the driving force for mass transfer because it identifies and quantifies the
evaporative potential of the volatile species which are of interest and is
not confounded by interferences of other materials present.
The first step in performing a vapor space analysis is collecting a
composite liquid sample from the pond or lagoon of interest. The sample
collection container is filled completely and sealed tightly to prevent the
evaporation of volatile components during handling. In the laboratory, the
sample is thoroughly mixed and an aliquot of approximately 800 ml is trans-
fered to a 1 liter flask with a side arm. (The flask should be at least
three-quarters full). The flask is sealed tightly with a septum, the con-
tents are agitated, and the flask is placed in a 70°F constant temperature
bath. After allowing sufficient time for equilibrium between the liquid and
gas phase at 70°F to be reached, a micro-liter syringe sample of the vapor
is withdrawn.
The gas sample is syringe injected into the carrier gas stream of the
(Organic Vapor Analyzer) gas chromatograph (GC). After the chromatogram is
obtained, the GC column is backflushed directly to the flame ionization
detector chamber, providing a quantitative measurement of any residual
compounds in the gas sample. In this way, the total concentration of orga-
nic material as well as the concentrations of individual species may be
determined. To ensure that the vapor samples analyzed represent equilibrium
conditions, gas chromatograms will be run at 15 minute intervals, after an
initial equilibrization period of one hour, until constant results are
obtained for successive samples.
126
-------
Surface Enclosure Sampling
Surface enclosure sampling provides data for the determination of the
dynamic rate of transfer of material from the liquid surface of a pond or
lagoon into the air above it. To obtain a dynamic sample of this air, an
air-tight surface enclosure, with a sample outlet and fresh air inlet at
opposite ends, is floated on the surface of the water. An air-tight enclo-
sure is created by covering one end of the opening of an inflated truck
innertube with translucent plastic, sealed to the tube's perimeter. As the
air sample is withdrawn at a constant rate through the sample outlet, puri-
fied sweep air enters the opposite end of the enclosure through an inlet
tube which projects down to just above the liquid surface. The sweep air is
purified prior to entering the enclosure by passing through a dual adsorp-
tion column of activated carbon and XAD2, when sampling for organic vapors,
and through appropriate absorption solutions, when sampling for ammonia,
amines, and H2S.
For organic vapor analyses, the sample air is drawn directly into the
OVA. The gas sampling rate to the OVA is adjusted, while monitoring the
total organic vapor concentration, until a concentration of 10-20 ppm is
obtained. This concentration is high enough for accurate measurement and
low enough so as not to cause vapor phase resistance to mass transfer (the
equilibrium partial pressure of volatile species will correspond to a con-
centration much greater than 20 ppm).
Once the desired flow rate is determined, sampling at this rate is
continued until a steady state condition is reached, as indicated by a
stable concentration reading. At this point, the flow rate and concentra-
tion of total organic compounds are noted, and a portion of the sample being
drawn to the OVA is injected into the GC column for identification and
quantification of the organic species present. A surface-based emission
rate is determined from the data obtained by dividing the product of the gas
sample rate and the measured concentration by the area of the liquid surface
exposed to the inside of the enclosure. That is:
Emission Rate Per Unit of Surface Area = (Gas Sampling Flow Rate) X
(Sample Concentration)/(Surface Area Covered by Enclosure).
The dynamic emission rate can be correlated to equilibrium vapor phase
concentration data, determined via vapor space analysis, for a determination
of the resistance to mass transfer posed by the liquid and gas interfacial
films, or, conversely, the interfacial mass transfer coefficients.
Profile Sampling
Because the surface enclosure method cannot be used where the surface
is disrupted by the action of aerators, an alternative sampling approach
employing a perimeter or downwind air sampling technique must be used at the
aerated biological oxidation ponds. For this program a perimeter profile
sampling technique was selected.
127
-------
In profile sampling, the concentrations of contaminants are measured at
an array of points in the source emission plume as it passes some downwind
location; in this case, the source perimeter. From these measurements, a
vertical plume concentration profile can be developed. Initially, screening
is performed, using the total organic vapor analysis capability of the OVA
to determine the necessary extent of the vertical sample point array and the
sample point spacing needed to develop a meaningful profile. Then, at each
sampling point, samples are drawn into the OVA for total organic vapor
analysis and for GC determination of the organic components. Additionally,
samples are collected in appropriate absorption solutions for analyses of
inorganic vapor emissions. An upwind sample is taken in each case so that
background contaminants levels can be accounted for.
The vertical plume profile is plotted and integrated over the vertical
extent of the plume in order to determine the average plume concentration.
The emission rate per unit of surface area can then be determined by multi-
plying the average plume concentration by the amount of diluting wind which
crossed a one unit wide vertical cross section of the plume (i.e., wind
speed times the vertical extent of the plume) and dividing by the pond
surface area over which the one unit width column of wind traveled (i.e.,
pond width). That is:
Emission Rate Per Unit of Area = Average Plume Concentration X
Wind Speed X Vertical Extent of the Plume r Pond Width.
Sample Analysis
Organic Vapor Analysis
A portable Organic Vapor Analyzer (OVA) produced by Century Systems
Corporation will be used for determining the total concentration of organic
material and the concentrations of specific organic compounds in gaseous
emission samples. The portable OVA is designed to measure small concentra-
tions (1-10,000 ppm) or organic material in air. The OVA uses a hydrogen
flame ionization detector (FID) which has been used for many years as .a
detector for organic compounds, and it has a built-in gas chromatograph for
separation of the components of sampled gases. The portability of the OVA
coupled with its intrinsic safety make it an ideal instrument for the pro-
posed field work. (The OVA is certified intrinsically safe by Factory
Research Corporation for Plot One, Division One, Groups A, B, C, and D
hazardous environments).
The FID system used in the OVA consists of a diffusion flame of pure
hydrogen which is free of ions and is thus nonconducting. However, when a
sample of organic material is introduced into the flame, ions are formed and
the flame becomes conductive. The conductivity of the flame can then be
measured. The amount of conductivity that the flame exhibits with the
12c
-------
introduction of organic compounds is related to many parameters, but is
thought to be primarily based on the number of carbon atoms present and the
efficiency of combustion.
In the typical laboratory FID, the sample to be analyzed is entirely
mixed with the hydrogen prior to introduction to the burner. This premixing
of the sample with the hydrogen does not occur in the OVA. In the OVA, the
sample is brought in with the combustion air. The sample contained in the
combustion air is introduced into the combustion/detection chamber through a
porous bronze filter which disperses the air and organic compounds around
the hydrogen flame.
This method of introducing the sample to the hydrogen flame modifies
the typical ion formation process and consequently changes the OVA's FID
response to various compounds from that of a conventional FID. In the usual
laboratory FID, a normal C3 would have a response of three times the methane
response (for the same concentration)and a normal Cfi would have a response
of six times that of methane. The OVA FID, however, has a response for
nearly all commonly encountered hydrocarbons of between 50% and 150% the
response for methane.
The precise reason for the difference in response between the OVA and
other FID's has not been completely determined. One hypothesis is that the
size of the reaction envelope and the energy available from the hydrogen
flame are altered by intoducing the sample at the periphery of the flame
reaction zone. This allows for less contact time, which reduces the number
of ions formed, and consequently changes the response mechanism.
The Century Systems OVA, which will respond to almost all organic
compounds, may be calibrated with hexane. However, since the response of
the OVA differs somewhat for the various compounds, it is best to calibrate
the instrument with the organic compounds that are expected to be encoun-
tered in the field. The basic method for developing a calibration curve
consists of recording the response of the OVA to a range of concentrations
of the expected organic gases. The OVA responses to the known hydrocarbon
concentrations are then plotted to obtain a calibration curve for the test
mixture.
The compositions of the test mixtures used for initial calibrations
will be determined from the presence of volatile organics in the wastes
received and from GC/mass spectrometer analyses of equilibrium vapor samples
of pond and lagoon wastewater. During the sampling program, any additional
peaks which occur, which were not present in the original test mixtures,
will be evaluated based on their relation to known peak locations and
heights, to the extent possible. Recurring unknown peaks will be evaluated
by performing additional GC/mass spectrometer analyses, as needed.
Inorganic Vapor Analysis
Inorganic gases do not contain carbon atoms and, therefore, they are
not detected by the OVA FID. Further, because the anticipated inorganic
129
-------
constituents of waste treatment pond and lagoon emissions are highly reac-
tive,- they are not well suited to analyses, such as GC/mass spec, which
require the collection of neat samples in glass bombs or sample bags.
Therefore, absorption sampling will be used for collecting air samples for
the analysis of inorganic vapor constituents by wet chemical methods.
Although many types of wash bottles are available for air sampling, impin-
gers are recommended for collecting readily absorbable gases which rapidly
dissolve in the washing liquid. For the proposed program, a series of
Greenburg-Smith impingers are to be used. A constant flow rate pump is used
to draw air through these impingers at a rate of up to 1 cubic foot per
minute (28 liters per minute) for the desired sampling duration (typically
30 to 60 minutes). A 50 milliter aliquot of washing solution is typically
employed in each impinger. Effective washing of the air by the absorption
solution is enhanced by the small diameter of the narrow end of the impinger
inlet tube (2.3 mm) and the proximity of the inlet to the base (5 mm).
Examples of analytical methods for the inorganic vapors of concern,
selected based on sensitivity and specificity (Leithe 19/1 and Ruch 1970),
are as follows:
Ammonia - Indophenol Reaction (Leithe 1971)
Determination Method
Absorption Solution
Sensitivity
Interference
spectrophotometri c
0.01N H2SO.
ppb range
monoalkylamines
Nessler's Reaction (Elkins)
Determination Method
Absorption Solution
Sensitivity
Interference
spectrophotometri c
sulfuric acid
3 ppm
any material producing
ammonia
As Trichloramine (Zitomer 1962)
Determination Method
Absorption Solution
Sensitivity
Interferences
spectrophotometri c
distilled water
10 - 300
- amines
Phenol - Hypochlorite Reaction (Weatherburn
1967)
Determination Method
Absorption Solution
Sensitivity
Interferences
spectrophotometri c
distilled water
suitable for air pol-
lution work
amines
130
-------
Aliphatic Amines - Acylation with Cinnamic Anydride
(Hong 1968)
Determination Method - spectrophotometric
Absorption Solution - acetonitrile solution
Sensitivity - ppm level
Interferences - none of concern
Aromatic Amines - With 9-Chloroacridine (Stewart 1969)
Determination Method - spectrophotometric
Absorption Solution - ethyl alcohol
Sensitivity - suitable for air pollu-
tion work
Interferences - primary amines
Sulfide - N,N - dimethyl-p-phenylenediamine (Jacobs
1967)
Determination Method - spectrophotometric
Absorption Solution - alkaline cadmium
sulfate
Sensitivity - ppm range
Interferences - none of concern
Hydrogen Chloride - Silver Nitrate Titration (Hanson
1965)
Determination Method - titration
Absorption Solution - distilled water
Sensitivity - 2.5 ppm
Interferences - halides
DISCUSSION AND CONCLUSIONS
The proposed sampling program will provide a comprehensive characteri-
zation of the organic and inorganic vapor emissions from aqueous hazardous
waste physical/chemical treatment lagoons and aerated biological oxidation
ponds. Further, the organic vapor emission data will be used to develop
organic vapor emission factors for wastewater ponds and lagoons.
The use of complementary sampling and analysis methods will provide for
the verification of results. Further, complementary sampling will furnish
data for determining both the driving force for mass transfer across the
water-air interface and the dynamic mass transfer rate (emission rate), so
that empirical mass transfer coefficients can be evaluated.
The driving force for mass transfer can be determined from the equili-
brium vapor concentration data from the vapor space analyses. A further
131
-------
value of this analysis technique is that it provides a relatively concen-
trated gas sample so that organic components of concern can be detected and
measured accurately.
The dynamic emission rate may.be assessed via surface enclosure, expo-
sure profiling, or upwind-downwind sampling. Surface enclosure sampling
provides certain advantages over sampling performed at the perimeter or
downwind of an aqueous pond or lagoon. The concentrations of the components
of interest in the gas sampled can be varied by varying the rate of enclo-
sure sweep gas, and a concentration suitably high for accurate instrumental
analysis can be obtained. In contrast, concentrations at perimeter or
downwind sampling points will be much lower. Further, the concentrations of
downwind samples are affected by the wind speed and direction and by disper-
sion, which is a function of atmospheric stability. These meteorological
parameters introduce sources of error in relating measured downwind concen-
trations to surface emission rates.
On the other hand, by enclosing the liquid surface, the experimenter is
creating an artificial situation and must guard against the introduction of
errors due to consequent effects. The enclosure will stifle the natural
escape of heat from the surface, so the sampling period must not be of so
long a duration nor the sweep gas of so low a flow rate as to allow heat to
build up within the enclosure. Further, the sweep gas rate must be great
enough that equilibrium between the components in the liquid and vapor phase
is not reached, as this would result in the computation of artificially low
emission rates. However, the rate of sweep gas must not be so high as to
disturb the liquid surface, artificially inducing volatilization, nor so
high that the concentrations of gaseous contaminants are too low to be
measured accurately.
Where the surface enclosure technique cannot be used because of the
disturbance of the liquid surface from the action of aerators, an alter-
native method must be employed. For this purpose, profile sampling at the
source perimeter offers certain advantages over upwind-downwind sampling.
Since samples are obtained closer to the source (i.e., at the source peri-
meter rather than at some downwind distance), the concentrations of conta-
minants from the surface will be higher and, thus, can be measured more
accurately. Further, although the results are sensitive to wind speed and
direction and atmospheric dispersion, the effects of these factors are. not
as great as they are for the upwind-downwind method.
132
-------
REFERENCES
Adams, D.F., 1979, "Sulfur Gas Emissions from Flue Gas Desulfurization
Sludge Ponds," JAPCA, Vol. 29, No. 9, pp. 963-968.
Bondurant, J., S. Luce, and J. Townsend, 1973, Chemical Engineering Depart-
ment Report, Univ. Arkansas, Fayettville, Arkansas.
Broeker, W.S., and T.H. Peng, 1974, "Gas Exchange Rates Between Air and
Sea," Tell us, Vol. 26, pp. 21-35.
Cowherd, C. Jr., 1977, "Measurement of Fugitive Particulate," Second Sympo-
sium on Fugitive Emissions: Measurement and Control, (EPA-600/
7-77-148), Midwest Research Institute, Kansas City, Missouri.
Oil ling, W.L., 1977, "Interphase Transfer Process II. Evaporation Rates of
Chloro Methanes, Ethanes, Ethylenes, Propanes, and Propylenes from
Dilute Aqueous Solutions. Comparisons with Theoretical Predic-
tions," Environmental Science and Technology, Vol. 11, No. 4, pp.
405-409.
Elkins, H.B., 1952, The Chemistry of Industrial Toxicology, John Wiley and
Sons, Inc., New York, New York, p. 292.
Englebrecht, R.S., A.F. Gaudy, and J.M. Cederstrand, 1961, "Diffused Air
Stripping of Volatile Waste Components of Petrochemical Wastes,"
Journal Water Pollution Control Federation, Vol. 33, No. 2, pp.
127-135.
Fruman, R.A., "Stripping of Hazardous Chemicals from Surface Aerated Waste
Treatment Basins," Monsanto Company, St. Louis, Missouri.
Gaudy, A.F., et al, 1961, "Stripping Kinetics of Volatile Components of
Petrochemical Wastes," Journal Water Pollution Control Federation,
Vol. 33, No. 4, pp. 382-292.
Hales, J.M., 1972, "Fundamentals of the Theory of Gas Scavenging by Rain,"
Atmospheric Environment, Vol. 6, pp. 635-659.
Hansen, N.W., D.A. Reilley, and H.E. Stagg, (Eds.), 1965, The Determination
of Toxic Substances in Air, W. Heffer and Sons, Ltd., Cambridge,
England.
Harbeck, G.E., 1962, Geol. Survey Professional Paper 272-E, Washington, D.C.
Hong, W.H. and K.A. Conners, 1968, "Spetrophotometric Determination of
Aliphatic Amines by Acylation with Cinnamic Anhydride," Anal.
Chem., Vol. 40, No. 1273.
133
-------
Jacobs, M.B., M.M. Braverman, and S. Hochheiser, 1957, "Ultramicrodetection
of Sulfide in Air," Anal. Chem. , Vol. 29, No. 1349.
Kolnsberg, H.J. , 1976 a, Technical Manual for the Measurement of Fugitive
Emissions: Upwind/Downwind Sampling Method for Industrial Emis-
sions (EPA-600/2-76-089a), TRC, Wethersfield, Connecticut:
Kolnsberg, H.J., P.W. Kalika, R.E. Kinson, and W.A. Marrone, 1976b, Techni-
cal Manual for the Measurement of Fugitive Emissions: Quasi' -Stack
Sampling Method for Industrial Fugitive Emissions (EPA-600/2-76-
089c), TRC, Wethersfield, Connecticut.
Ledbetter, J.O. and C.P. Richardson, 1978, "Hydrocarbon Emissions from
Refinery Wastewater Aeration," Industrial Wastes, July/August pp
26-28. '
Leithe, W. , 1971, The Analysis of Air Pollutants, Ann Arbor Science Pub-
lisher s,~~
Liss, P.S., 1973, Deep Sea Res., Vol. 22, No. 221.
Liss, P.S. and P.G. Slater, 1974, "Flux of Gases Across the Air-Sea Inter-
face," Nature, Vol. 247, pp. 181-184.
National Academy of Sciences, 1976, Halocarbons: Effect on Stratospheric
Ozone, National Research Council, Washington, D.C.
Owens, M. , R.W. Edwards, and J.W. Gibbs, 1964, "Some Reaeration Studies in
Streams", International Journal of Air and Water Pollution, Vol
8, pp. 469.
Ruch, W. E. , 1970, Quantitative Analysis of Gaseous Pollutants, Ann Arbor-
Humphrey ScTeTicTTubTFIhirTriifnTn^^
Schooley, A.H. , 1969, J. Mar. Res., Vol. 27, pp. 335.
Slinn, W.G.N. , et al, 1978, "Some Aspects of the Transfer of Atmospheric
Trace Constituents Past the Air-Sea Interface," Atmospheric Envi-
ronment, Vol. 12, pp. 2055-2087.
Stewart, J.T. , T.D. Shaw, and A.B. Ray, 1969, "Spectrophotometric Determi-
nation of Primary Aromatic Amines with 9-Chloracridine, Anal.
Chem. , Vol. 41, No. 360.
Thibodeaux, L.J., R.B. Estridge, and B.G. Turner, 1971, "Measurement of the
Relative Volatilization of the Water-Mi scible Fractions in an
Aqueous Effluent," AIChE Symposium Series No. 124, Vol. 68, pp
169-179.
Thibodeaux, L.J. and D.G. Parker, 1978, AIChE Symposium Series No. 156, Vol
72, pp. 424-434.
134
-------
Tischler, L.F. and J.F. Malina, Updated, "Evaluation of the Potential for
Air Stripping of Hydrocarbons During Activated Sludge Wastewater
Treatment," University of Texas/Engineering-Science, Austin,
Texas.
Weatherburn, M.W., 1967, "Phenol-Hypochlorite Reaction for the Determination
of Ammonia," Anal. Chem., Vol. 39, No. 971.
Wu, J.M. and J.M. Schroy, "Emissions from Spills," Monsanto Company, St.
Louis, Missouri.
Zimmerman, P., 1978, "Procedures for Conducting Hydrocarbon Emission Inven-
tories of Biogenic Sources and Some Results of Recent Investiga-
tions," in: Emission Inventory/Factor Workshop, Volume 2 (EPA-
450/3-78-042b), Office of Air Quality Planning and Standards,
Research Triangle Park, N.C., May 1978, pp. 25-1 through 25-32.
Zitomer, F. and J.L. Lambert, 1962, "Spectrophotometric Determination of
Ammonia as Trichloramine," Anal. Chem., Vol. 34, No. 1738.
135
-------
RESULTS OF FUGITIVE EMISSION MEASUREMENTS AT
REFINERIES AND CURRENT ACTIVITIES IN
PETROCHEMICAL UNITS
by
Donald D. Rosebrook
Radian Corporation, Austin, Texas
ABSTRACT
Final results of fugitive emissions monitoring in petroleum
refineries are presented as a series of emission factors for valves, flanges
pump seals, compressor seals, relief valves and drains. The emission factors
appear to be dependent on the type of service in the line, consequently the
emission factors have been developed for gas/vapor, light liquid and heavy
liquid service. No dependency on temperature, pressure or line size was
noted. Extensive quality control data allowed the calculation of confidence
intervals for all emission factors. Correlations between a hydrocarbon
"sniffer" value and measured mass emissions rates in pounds per hour are
given.
Currently EPA is sponsoring studies in the synthetic organic chemi-
cals manufacturing industry. The. current program includes both screening and
sampling selected valves and pumps and screening all "baggable" sources in a
variety of chemical manufacturing units. The screening studies are being
conducted by four separate EPA contractors, Radian, PEDCo, TRW and Acurex
with Radian doing all data reduction. Screening data for target chemicals are
being obtained from at least two different sites if possible. Sampling data
is being obtained for three chemicals at two sites each. Correlations between
the hydrocarbon "sniffer" value and the mass emission rate are being developed.
Studies of the effect of maintenance on leak rate for valves are also being
conducted. Hydrocarbon "sniffer" response factors are being developed for
over 100 chemicals with respect to known concentrations of methane.
136
-------
The intent of this paper is twofold: first, to discuss the final
results of the recently completed fugitive hydrocarbon emission portion of an
EPA sponsored Petroleum Refinery Assessment Program. Secondly, to discuss
a program which is just beginning to study fugitive emissions in various
processes of the synthetic organic chemical manufacturing industry (SOCMI).
Thus, in Part 1 of this paper we will be discussing final results, while in
Part 2 we will be discussing philosophy and design.
Part 1; Results of Fugitive Hydrocarbon Emission Measurements in
the Refining Industry
Study Design
The study was designed to include a representative portion of the
United States petroleum refining industry for an examination of fugitive
emissions from point sources including valves, pump and compressor seals,
drains, relief valves and flanges. In the design of the study, the variables
of refinery age, refinery size, and refinery location were carefully con-
sidered. Thirteen refineries across the United States were selected for the
study. They ranged from less than ten years old to nearly 100 years old. The
U. S. was divided into four primary geographical areas, that is, the Northeast,
Southwest, Midwest, and the West coast, where the bulk of the petroleum
refining activity is conducted. At least three refineries were chosen in each
of these geographical areas and within each area at least one new and one old
refinery and one large and one small refinery were selected. For each fugitive
source to be sampled, several variables were considered to be important in
selecting those fittings to be tested. Table 1 shows the type and range of
variables which were used in developing the sampling matrix for each type of
fugitive emission source. Those fittings selected for inclusion in the study
were chosen from P&ID diagrams, sight unseen, by Radian engineers, then lo-
cated in the refinery, screened with a portable hydrocarbon detector, and
eventually bagged or tented for actual mass emission rate measurements.
Results
Table 2 presents the results of the fugitive emission measurements
by source type. As indicated in the table, the stream service for the various
sources proved to be the major source of variation. Types of service were
divided into gas vapor streams, light liquid or two-phased streams and heavy
liquid streams. The dividing line between light liquid and two-phased streams
and heavy liquid streams proved to be at the approximate volatility of kero-
sene. Valves in hydrogen service were streams which were at least 50 percent
hydrogen with the remainder being hydrocarbons. Open ended valves were also
singled out. These are valves primarily used for quality control sampling
purposes within a refinery unit. The values shown in the table for pump and
compressor seals are expressed per seal rather than per pump or per compressor.
The emission factor for compressor seals was the largest of any of the
emission factors and the percent of compressor seals leaking was the largest
of the percentages for any of the sources. Conversely, the percent of flanges
leaking was the smallest of any of the sources and its emission factor was
also the lowest. Expanding this data to an entire refinery would show that
valves are the most significant of all of the leaking sources because of their
number, while flanges are the least significant of all leaking sources.
137
-------
TABLE 1. RANGE OF CHOICE VARIABLES FOR
SCREENED BAGGABLE SOURCES
Baggable Source
Choice Variable
Variable Ranges
for Screened Sources
Valves
Flanges
Pump Seals
Compressor Seals
Drains
Relief Valves
Pressure
Temperature
Fluid State
Service
Function
Size
Pressure
Temperature
Fluid State
Service
Size
Pressure
Temperature
Capacity
Shaft Motion
Seal Type
Liquid RVP
Pressure
Temperature
Shaft Motion
Seal Type
Lubrication Method
Capacity
Service
Pressure
Temperature
Fluid
-10 - 3000 psig
-190 - 925°F
Gas, Liquid, 2-phase
In-line, Open-ended
Block, Throttling, Control
0.5 - 36 inches
-14 - 3000 psig
-30 - 950°F
Gas, Liquid, 2-phase
Pipe, Exhanger, Vessel,
Orifice
1-54 inches
0 - 3090 psig
0 - 800°F
0 - 100,000 gpm
Centrifugal, Reciprocating
Mechanical Seal, packed seal
Complete range
0 - 3000 psig
40 - 300°F
Centrifugal, reciprocating
Packed, labyrinth,
mechanical
Hydrocarbon lubricant
0.06 - 66.0 MMSCFD
Active, Wash-up
0 - 1350 psig
40 - 1100°F
Gas, Liquid
138
-------
TABLE 2. SUMMARY STATISTICS AND ESTIMATED VAPOR EMISSION FACTORS
FOR NONMETHANE HYDROCARBONS FROM BAGGABLE SOURCES
Source Type
Volvos
One Vapor Stream
Light Llquld/Two-Fliaae
Heavy Liquid
Hydrogen
O|ien-Ended Volveu
Light Liquid Streams
Heavy Liquid Streams
lira Ins
Flangea
Relief Valvca
Cowprcaaora
Hydrocarbon Service
Hydrogen Service
Niwlx-r
Screened
563
913
485
135
129
470
292
257
2094
148
142
83
Nimher
Leaking
154
330
32
59
30
296
66
49
62
58
102
69
I'l-ICl'Ut
Leaking
27
36
6
43
23
63
22
19
3
39
71
63
.4
.1
.6
.7
.3
.0
.6
.1
.0
.2
.8
.1
95Z Confidence
Interval for
Fcrccnt l.rnklng
(2*.
(33.
( 4,
(35.
(16,
(59,
(18.
(14,
( 2.
(31.
(64,
(75,
3D
39)
9)
52)
3D
67)
27)
24)
4)
47)
79)
91)
Knlaeloii
Factor
F.vtlautte
(lb/hr/Bource)
0
0
0
0
0
0
0
0
0
0
1
0
.059
.024
.0005
.018
.005
.25
.046
.070
.00056
.19
.4
.11
95Z Confidence
Interval fur
Ealaalon Factor
(Ib/lir/aource)
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
030.
017.
0002
007.
0016
16,
019.
023.
0002
070.
66,
05,
0.110)
0.036)
. 0.0015)
0.045)
. 0.016)
0.37)
0.11)
0.20)
, 0.0025)
0.49)
2.9)
0.23)
-------
The distribution of leak rates was another important result of the
study. The study results indicate that the majority of sources leak at a very
low level and that the significant contributions to the emission factor come
from those few sources leaking at high rates. This is illustrated in Table 3
for valves in gas/vapor service, where it is shown that of all the valves in
the study, seven of them leaked at greater than one pound an hour and that
these seven valves contributed 70 percent of the total mass of the leak.
Figures 1 and 2 provide the same information in graphical form relating screen-
ing value to percent of sources and percent of emissions, respectively.
Figure 3 shows a nomograph for predicting total nonmethane leaks
from screening values for valves in gas/vapor service. This demonstrates that
within reasonable confidence limits, the screening value obtained with a hydro-
carbon detector can be correlated to mass emission rates for a group of valves
This correlation should prove to be a valuable tool for both maintenance and
regulatory activities. The nomograph in Figure 4 demonstrates the same type
of relationship for valves in light liquid service. Similar nomographs have
been developed for all of the source types investigated during this program.
It should be stressed that these nomographs are not meant for direct point by
point correlation of single valves, but rather apply to the mean of a number of
sources.
Correlations
In Table 4 the correlation coefficients between operating/design
variables and leak rate are presented. As you can see, the correlation
coefficients presented in the table are by and large rather small. They are
however, statistically significant in those cases where they have been denoted
by an asterisk. Even though some of the correlations were statistically sig-
nificant, no discernible pattern emerges relating leak rate to process/design
variables for all sources.
Figures 5-7 are schematic plots demonstrating the lack of varia-
tion in emission rates in relation to some of the parameters examined in the
study. Figure 5 shows the data for pumps with single and double seals, and
indicates little difference in the mass emission rate on the basis of seal
type. However, there may be some service severity factor which causes the
apparent lack of correlation. A service severity factor was not anticipated
during the study, thus no data are currently available for ready review. in
Figure 6 the same type of information for single seals and double seals is
presented for compressors. Once again the cells overlap and the means are
extremely close together. Figure 7 examines the block and control function of
valves, and again, it is demonstrated that there is very little difference
between the mass emission rates in any type of stream between block and control
valves. Figure 8 shows the relationship between leak rate and diameter for
compressor seals, and again it should be noted that the correlation is poor.
Figure 9 shows leak rate versus pressure for compressor seals and again the
relationship is poor.
Finally, in Figure 10 the variation in a given leak rate as the
function of time is presented. This figure demonstrates the reason for wide
confidence intervals in many of the sources and illustrates the difficulty
which will be encountered in both maintenance and regulatory activities. As
one can see on the right hand portion of the figure, the leak rate in pounds
140
-------
TABLE 3. DISTRIBUTION OF NONMETHANE
HYDROCARBON LEAK RATES
Values, Gas/Vapor Streams
(563 Screened)
Leak Range
(Ib/hr)
1.0
0.1 - 1.0
0.01 - .1
0.001 - 0.01
0.00001 - 0.001
No.
7
18
43
49
37
154
Leaking Sources
% of
Leaking
Sources
4.6
11.7
27.9
31.8
24.0
100%
Within Range
% of Total
Sources
Screened
1.2
3.2
7.6
8.7
6.6
20.3%
Total
Total
Leakage
(Ib/hr)
17.7654
5.9187
1.4867
0.2052
0.0133
25.3893
Leakage Within Range
% of Total
Source of Leakage
70.0
23.3
5.9
0.8
0.1
100%
-------
o
90
80
70
60
S 50
«^
c
t>
i; 40
u
a.
30
20
10
Upper Limit of 951 Confidence Interval
Estimated Percent of Sources
x
Limit of 95*
Confidence Interval
I 111 t
I I
1 2 345 10 50 100 1,000 10.000 100.000 1,000.000
Screening Value (ppmv) (Logi0 Scale)
Percent of Sources - indicates the percent of sources with screening
values greater than the selected value.
Figure 1. Cumulative Distribution of Sources and Total Emissions
By Screening Values for Valves - Gas/Vapor Streams
-------
u>
100
90
80
«•
a
- 70
«•
Ul
M 60
3 50
2
o 40
*J
C
i 3°
c.
20
10
Estimated Percent of
Total Mass Emissions
I i i
j i
Upper Limit of 901
Confidence Interval
Lower Limit of the
901 Confidence Interval
1 2 345 10 50 100 1,000 10.000 100.000
Screening Value (ppav) (Logjo Scale)
1,000.000
Percent of Total Mass Emissions - indicates the percent of total emissions
attributable to sources with screening
values greater than the selected value.
Figure 2. Cumulative Distribution of Source and Total Emissions By
Screening Values for Valves - Gas/Vapor Stream
-------
0.07
£ 0.06
0.05
e
e
0.04
•«
c
£ 0.03
a
2 0.02
o
0.01
m Lt»k Ritt) • -7.0 * 1.23 Logt, (fUx Scrtenlnj Vtlue)
CorttUrion Coefficient • 0.76
Number of Dm ?«1rj • 0.79
Standard Error of E»t1mzt* • 0.78 Log>t (KM Leak Rats)
lias Correction ftaar • 4.31
/
/
/
/
/
/
/
/
/
/
/
/
Upper Licit of SOI Confidence
/ Interval for Msan
/
/
/
/
/
/
/
/
/
Mean
Lower Lir.it of SOS Confidence
. Interval for Mean
1,000
s.ooo
10,000
Maximum Screening Value (ppmv as Hexane)
Using J.W. Sacharach TLV Sniffer at the Source.
Figure 3. Nomograph for Predicting Total Nonmethane
Hydrocarbon Leak Rates from Maximum Screening
Values - Valves, Gas/Vapor Streams
144
-------
0.07
w
JS
^»
«•
^0.06
•
«^
£
2 0.05
9 0.04
•5 0.03
-0.02
0.01
Upper Limit of 901 Confidence
Interval for Mean
Mean
Lower Limit of 901 Confidence
Interval for Mean
I t I I I I I I I
1.000
5,000
10.000
Maximum Screening Value (ppmv as Hexane)
Using J.U. Bacharach TLV Sniffer at the Source.
Figure 4. Nomograph for Predicting Total Nonmethane
Hydrocarbon Leak Rates from Maximum Screening
Values - Valves, Light Liquid/Two-phase Streams
-------
TABLE 4. CORRELATIONS BETWEEN CONTINUOUS VARIABLES AND LOG LEAK RATE
LINE STROKE
PRESSURE TEMPERATURE AGE SIZE DIAMETER AREA RPM CAPACITY LOAD LENGTH
VALVES
Gas/Vapor Streams
Light Liquid Streams
Heavy Liquid Streams
Hydrogen Service
Open-Ended
PUMP SEALS
Light Liquid Service
Heavy Liquid Service
FLANGES
COMPRESSOR SEALS
Hydrocarbon Service
Hydrogen Service
DRAINS
RELIEF VALVES
.230*
.103'
-.351'
-.088
.236
.088
,097
.072
.346*
.398*
-
.045
.077
.051
.144
.129
.212
-.012
-.098
.021
.218*
.312'
-.108'
.096
.263' .150' - - - -
.096 .143* - - -
.220 .046 - - - -
-.531* .288* - - - -
.230 -.078 - - - -
.062 - .021 - -.064 - -
,237 - .128 - -.182
-.180 .336" - - -
.105 - .278* - -.143° -.138 -.087 -.012
.052 - .343* - -.034 .218 -.099 -.074
- -.039 -.191 -
- -.075 -
•Correlation Coefficient statistically different from zero (P> .90).
3Log10 RPM was correlated with Iog10 leak rate.
-------
2.00
0.667
C£
nr
CQ
-0,667
-2,00
UJ
£ -3.33
o
CD
S -4,67
-6,00
N = 33
N = 231
LIGHT LIQUIDS
DOUBLE SINGLE
SEAL SEAL
= 9
N = 54
HEAVY LIQUIDS
DOUBLE SINGLE
SEAL SEAL
KEY
: Detached Value (1 In 200)
~*
Upper Quortlle
Heon
Median
Lower Quortlle
0 Outside Value (1 In 20)
0
Figure 5. Schematic Plot for Pumps by Seal Variables
-------
2.00
0,833
-0.333
-1,50
-2,67
CD
o
-3.83
-5.00
• t
t
N = 3 N = 76
HYDROCARBON STREAMS
DOUBLE SINGLE
= 21 N = 22
HYDROGEN STREAMS
DOUBLE SINGLE
KEY
* Detached Value (1 in 200)
Upper Quart lie
Mean
Median
Lower Quart lie
0 Outside Value (1 In 20)
0
Figure 6. Schematic Plot for Compressor Seals by Single/Double Seal Variable
-------
l.UU
-0,333
i
> -1 R7
cu j. i \ji
— I
UJ
1 —
_
- -3,00
9
UJ
i -4,33
I
fc
o
V5.67
3
7 nn
I*
0 0
p
(
t
t
t
i
i
*
t
*--.* t
•v*»* *•«
f f "-- *
It * <
/ < «
* t t
1 t »-•
»..-.*
N=108
c
N=50 N=;
BLOCK CONTROL
1
t
t
t <
f «.-.
-* > * •- — +
* •**.* 4.*-* ft «
* * * If •-*-* .4- f «
< * ».--4 »_.-*
*._-* «
N=8
Q M*^ O O
N=91
0
268 N=29 N=39
0
BLOCK CONTROL
KEY
j Detached Value (1 In 200)
p-i— - Upoer Quart Me
•f- MCOD
Median
I-|-J Lower Ouartlle
1
1
0 Outside Voiue (1 in 20)
0
™*°««
Figure 7. Schematic Plot for Valves by Block/Control Variable
-------
Ui
o
UJ
s
1.6
0,8
0,0
-0,3
-1.1
CD
LD
O
t A
ci> _2ti
-3,2*
-4.0
LEGEND: A = 1 OBS, B = 2 OSS, ETC,
Correlation Coefficient (r) = 0.278
Number of Data Pairs = 88
1,5
2,0 2,5 3,0 3,5 4.0
DIAMETER (INCHES)
,5 5,0 5,5 6.0
Figure 8. Leak Rate vs. Diameter -
Compressor Seals, Hydrocarbon Service
-------
o:
1.6
0.8;
0.0
-o.s;
-1.6;
2-2.4
-3.2
LEGEND: A = 1 DBS, B = 2 DBS, ETC.
Correlation Coefficient (r) = 0.346
Number of Data Pairs = 102
-4.0
-4.8 ;
0 40 80 120 160 200 240 280 320 360 400 440 480 520 560
PRESSURE (PSIG)
Figure 9. Leak Rate vs. Pressure - Compressor
Seals, Hydrocarbon Service
-------
0.028 •
I 0.024'
aa
^ 0.020-
* 0.015-
bU
= 0.012 •
2 O.OC8-
LU
£
0.004-
0
*•-..
1/18
ID - 15VA231
* :
#
1/27
DATE (Ofl. fiM£j
L31 2/22/3 I/ =1 ' ' *
Figure 10. Short-Term Variation in
Leak Rate - Valves
152
-------
per hour varied by an order of magnitude over a period of approximately two
hours. Such occurrences were not uncommon.
Part 2; Fugitive Emission Testing in SOCMI Process Units
The second part of this paper deals with current work in various
SOCMI process units. This work is being supported again by EPA, 1ERL both in
RTF and in Cincinnati and is being closely coordinated with OAQPS. There are
essentially two programs being conducted simultaneously and in some cases in
the same production unit. One of the programs is a fugitive emission screening
survey which will be further described here.
The second is a program to determine the long and short term effects
of maintenance on valves and the occurrence and reoccurrence of fugitive leaks
from valves and pumps. This program will be more fully described in a paper to
be presented by Weber and Smith at the APCA Meeting in Montreal later this
year. This program will also develop correlations between screening values
and leak rates for valves and pump seals in chemical process units.
The screening program being supported by IERL-RTP will screen all
potential fugitive emission sources in some 20 process units except that only
5 to 15 percent of the flanges will be included. The work is being conducted
by four companies: Radian, PEDCo, Acurex, and TRW. The last three contractors
are forwarding their data to Radian for processing and interpretation, and
Radian has the responsibility for the final report. In this study we are
examining a wide variety of chemicals and attempting to study at least two
process units for each type of chemical. Currently, screening has been con-
ducted at ethylene, ethylene dichloride, vinyl acetate, vinyl chloride monomer,
methylethyl ketone, phenol/acetone, and formaldehyde units. These data are not
yet sufficiently reduced to give any estimate of the frequency or of variations
between process units producing the same type of chemical.
In addition, this program is developing response factors for two
types of hydrocarbon sniffers for a large number of industrial synthetic or-
ganic chemicals. This effort consists of a laboratory study where the response
of specific gases is examined using two Century OVA and two Bacharach TLV hy-
drocarbon detectors. The response versus concentration data are being
developed at several points between 100 and 10,000 ppm. It is the intent of
this study to determine the utility of various types of hydrocarbon sniffers/
portable hydrocarbon monitors in screening for maintenance or regulatory pur-
poses for the wide variety of synthetic/organic chemical manufacturing pro-
cesses. Once again the response factor data are not complete. It is, however,
safe to say from a preliminary look at the results that response factor does
vary with concentration and may vary between chemicals by one or more orders
of magnitude.
This study was scheduled for completion in the summer of 1980 but
has, however, suffered severe delays due to difficulties in accessing the vari-
ous process units. This was caused by the recent chemical workers strike and
by some reluctance on the part of plant managers to have the screening teams in
their units. Currently, 12 units have been screened and the remainder of the
screening visits have been scheduled. We anticipate that the data from this
program will be ready in the Fall of 1980.
153
-------
Acknowledgment s
The work described in the refinery assessment portion of this paper
was supported by EPA, IERL, RTF under Contracts 68-02-2147, Exhibit B and
68-02-2665. The cooperation of the participating refineries and of the Ameri-
can Petroleum Institute is also to be acknowledged. Finally the support of
Radian's field crews and statistical department is gratefully acknowledged.
The work in Part 2 of this paper is supported under the contract from IERL
RTP 68-02-3171, Task 1. The authors would also like to acknowledge the support
of those companies who have made available their facilities and of the assis-
tance and advice of EPA, OAQPS.
154
-------
EVALUATION OF FUGITIVE EMISSIONS AT A
LARGE WOOD-PRODUCTS PLANT
by
Peter D. Spawn
Environmental Engineer
GCA CORPORATION
GCA/TECHNOLOGY DIVISION
Bedford, Massachusetts
ABSTRACT
Under sponsorship of the U.S. Environmental Protection Agency (EPA), this
program assessed fugitive particulate emissions at a large particleboard plant.
Major and minor sources of fugitive particulate emissions were identified in
a pretest survey. Emissions from major fugitive sources — open doorways in
large material handling buildings — were directly sampled during a 2-week
field program. Concurrently, ambient TSP data were collected daily at four
sites outside plant boundaries using 24-hour continuous hi-vols and also wind-
actuated units in an upwind/downwind arrangement. This paper presents the
results of the sampling program and identifies control options applicable to
the fugitive emissions.
ACKNOWLEDGMENT
This study was funded under EPA Contract No. 68-01-4143, Technical
Service Area 1, Task No. 34. John R. Busik served as Project Officer for
the overall contract while Norman Edmisten of EPA Region X served as Task
Manager for this particular task. The author also wishes to thank Tim Curtin
and Barbara Myatt of GCA for their invaluable assistance in conducting the
field program.
DISCLAIMER
This paper was prepared for the Fourth Symposium on Fugitive Emissions in
New Orleans on 28 May 1980. The paper is based on the Final Report furnished
to EPA by GCA Corporation, GCA/Technology Division, Bedford, Massachusetts
01730, in fulfillment of Contract No. 68-01-4143, Technical Service Area 1,
Task No. 34. The opinions, findings, and conclusions expressed are those of
the author and not necessarily those of EPA or of cooperating agencies.
Mention of company or product names is not to be considered as an endorsement
by EPA.
155
-------
PROJECT BACKGROUND AND SCOPE
This project assessed the magnitude of fugitive particulate emissions at
a large Northwest particleboard plant. Although the plant had taken a number
of steps towards reducing fugitive emissions, violations of ambient air stan-
dards were still recorded at the local TSP monitoring station. Previous work
indicated the plant was the primary industrial source responsible for these
violations, based on correlations between wind patterns and measured TSP
levels. Consequently, the Oregon Operations Office of EPA Region X requested
GCA to assess the magnitude of fugitive emissions originating from the plant.
A 2-week sampling program directly measured particulate emissions from
major fugitive sources. These data provided a range of emission rates for a
variety of plant operational and ambient wind conditions. Ambient air quality
was monitored for TSP at four sites outside plant boundaries by upwind/downwind
sampling with source-oriented hi-vols. This arrangement was designed to pro-
vide an indication of the plant's contribution to ambient TSP levels.
Observation of daily plant operations enabled the identification of
minor sources of fugitive emissions such as conveyor spillage, leaky ductwork
and wind entrainment of spilled material. These observations coupled with a
review of control technology generated a preliminary assessment of methods for
reducing fugitive emissions.
DESCRIPTION OF FUGITIVE EMISSION SOURCES
Particleboard is manufactured by mixing prepared raw wood fibers with
organic resin binders and heating the mixture under pressure to form hard
solid sheets which are then trimmed and sanded to size and shape for sale and
use. The plant facilities consist of a major building complex housing the
three production lines and the storage and loading areas, and a number of
smaller buildings housing raw material storage and preparation operations.
Fugitive particulate emissions originate primarily from handling,
transport and storage of raw materials, which consist of green and dry planer
shavings and plywood trim. Sources of fugitive emissions which leave plant
boundaries can be summarized as:
• Major Sources — Material handling within four large buildings — the
truck dump, dry shavings storage and two plywood trim storage build-
ings - and one open storage pile (which will soon be enclosed).
• Minor Sources - Small leaks in the air transport system, an occasional
leak in a cyclone, conveyor belt losses, an occasional process equip-
ment upset in the mill and flake area.
• Wind entrainment of particulate matter which originates from the
above sources and accumulates over the entire plant.
156
-------
MAJOR FUGITIVE PARTICULATE SOURCES - OPEN DOORWAYS
IN MATERAL HANDLING BUILDINGS
Figure 1 shows the major fugitive emission sources which were sampled in
this study. Each source is described below.
Raw materials — dry and green planer shavings and plywood trim material —
are received from local sawmills in tractor-trailer trucks which are emptied
into two large hoppers within the truck dumper building. A large hydraulic
lift raises the entire truck at an angle, emptying it through a rear (truck)
door. When prevailing winds are from the west, south or southeast, dust
generated from the dumping operation is carried out the northeast facing
doorway. A vibrator located on one of the dumpers to loosen compacted shavings
somewhat increases dust generation. Green shavings (40 to 60 weight percent
moisture) generate substantially less dust than dry shavings (10 to 11 percent
moisture) or plywood trim. Under normal operation, about 9 to 10 trucks per
hour are dumped.
From truck dump hoppers, material travels to storage silos by two belt
conveyor systems, also shown in Figure 1. When silos are full, incoming
material is diverted to storage. Green shavings enter a large outdoor stor-
age pile. Dry shavings enter conveyors located underneath the dry storage
building roof and fall to storage piles within the building. Plywood trim
is similarly deposited into one of the two ply trim storage buildings by
overhead conveyors.
When incoming trucks cannot maintain necessary levels in the storage
silos, silos are refilled with material previously placed in storage.
Bucket loaders "scoop" raw material from storage areas to reentry hoppers
and material travels to appropriate silos by the main belt conveyor system
shown in Figure 1.
In summary, fugitive emissions from the three material storage/handling
buildings exit to the atmosphere through large open doorways. The doorway
emissions are generated by the following processes:
• Discharge of raw material to storage piles from overhead conveyors.
• Bucket-loader scooping of stored material and dumping to conveyor
hoppers.
• Wind reentrainment of particulate during periods of no internal
building activity, which only occurs with high winds.
MINOR FUGITIVE PARTICULATE EMISSION SOURCES
In addition to the major fugitive sources, i.e., the open doorways, a
number of minor sources were also identified. External conveyors, although
covered on three sides, lose some material. Air systems also used to transfer
raw material occasionally leak to some degree. An occasional upset in the
mill and flake area generates some fugitive emissions. All of these minor
sources discharge some particulate directly to the ambient air, and also
create accumulations of dust which are then reentrained by winds. The upwind/
downwind sampling program addressed the ambient air impacts of these (and all
other) plant emissions.
157
-------
W
NOTE-PNUEMATIC TRANSFER TO
MILL AND FLAKE FROM
DRY STORAGE SILOS AND
PLY-TRIM SILO
GREEN STORAGE
SILO
Ui
a-
PLY-TRIM
REENTRY
HOPPERS
(TO SILOS)
AAF
SCRUBBER
PLY-TRIM
STORAGE
[DRYER ios|
SANDERDUST
BOILERS
AND
DRYER
NO. I PRODUCTION LINE
GREEN
SHAVINGS
STORAGE
NO.2 PRODUCTION LINE
SILOS FOR
PREPARED
II REENTRY
li HOPPERS
RAW MATERIAL
GREEN
REENTRY
HOPPER
NO.3 PRODUCTION LINE
METALLIC
REJECT BELT
KEY:
BELT CONVEYORS-OUTSIDE BUILDING
BELT CONVEYORS-INSIDE BUILDING
OPEN DOORWAY
NOTE
TRUCK DUMP
BUILDING
ARROWS REPRESENT FUGITIVE SOURCE
EMISSION POINTS ADDRESSED IN THIS STUDY.
Figure 1. Major Fugitive Emission Sources.
-------
HISTORY OF FUGITIVE EMISSION CONTROL AT THE PLANT
Handling of raw materials by bucket loaders, conveyors and active storage
piles is inherently a very dusty operation. When the plant was designed and
constructed in 1960, little attention was given to control of these fugitive
dusts. All raw material was stored outdoors in piles or open-sided sheds,
belt conveyors were exposed to prevailing winds, plant grounds were unpaved
and the truck dumpers were unenclosed. Since this time, the plant has made
a number of operational improvements which have served to significantly reduce
the potential for fugitive emissions. These are summarized in Table 1.
TABLE 1. SUMMARY OF MAJOR CONSTRUCTION ACTIVITY AT DURAFLAKE
WHICH REDUCED FUGITIVE EMISSIONS
Year Activity
1969 Constructed truck dump building over the two dumpers.
Purchased used street sweeper.
Paved parking lot and eastern plant grounds.
Closed off east end of dry storage building with partition.
1970 Covered roof area between dry hemlock and fir storage sheds.
Installed truck dump negative air exhaust system and found only
partially effective.
1972 Enclosed conveyor belts on dry shavings roof.
Enclosed flight conveyors to dry storage silos.
1973 Paved western plant areas.
Installed roof over plywood trim storage buildings.
1974 Purchased vacuum truck for cleaning silos and spilled material.
1976 Installed canvas curtain below green shuttle in conveyor discharge to
reduce wind entrainment.
Corrected belt: conveyor spillage problems in mill and flake areas.
1977 Installed self-closing personnel doors on dry storage and truck dump
buildings.
Installed door to close off eastern opening of the dry storage
building.
1978 Installed fire hose connections to allow hosing of dusts accumulated
on roofs of dry and ply trim storage buildings, and mill and flake
areas.
Began construction of green shavings pile enclosure building.
Ordered new street sweeper.
DESIGN AND OPERATION OF SOURCE SAMPLING TRAINS
Fugitive emissions from the open doorways were directly measured by
traversing the plume exiting each doorway with a specially designed sampling
train consisting of a calibrated hi-vol motor with filter, and a cyclone pre-
separator to provide a 10 ym particle size cut. Each doorway was uniformly
traversed in a rectangular grid pattern to obtain an average concentration
of exiting particulates during the hour-long sampling run. The exit velocity
159
-------
of the particulate plume was monitored and recorded at each traverse point.
With these data, mass emission rates could be calculated based on the average
particulate concentration, the average exit velocity and the doorway cross-
sectional area.
Each sampling train consisted of an inlet nozzle, a cyclone preseparator
a standard 8 x 10 inch hi-vol filter, a hi-vol blower motor, and auxiliaries
as shown in Figure 2. Sampling flow rate was controlled by a variac (rheostat)
and monitored by a dial gauge on the blower exhaust and also a manometer across
the cyclone. The 4-inch diameter inlet nozzle was designed to provide sampling
at a rate approximately equal to typical doorway exit velocities so sampling
would approach isokinetic conditions. The train was calibrated by techniques
similar to those used for hi-vol ambient air samplers, both before and after
the sampling program. Design of the train allowed for easy component disas-
sembly to facilitate sample recovery from cyclones and connecting piping by
acetone washing.
For each source sampling run, the train was suspended in front of a
particular doorway from a trolley assembly similar to monorails used for
stack sampling. A pulley and rope arrangement allowed manual positioning of
the train at each traverse (sampling) point. Monorails were installed in a
semipermanent fashion above two doorways (the truck dump and western door of
dry storage) while remaining doorways were served by a portable design which
could be leaned against the doorway. The inlet nozzle was always located
about 12 inches just outside each doorway, to ensure measurement of actual
atmospheric emissions, not internal building concentrations.
On a few occasions when winds were light and/or variable, the doorway
exit flow reversed for brief periods such that no emission occurred. In
these instances, sampling was temporarily halted by stopping the blower, and
resumed when measurable flow was observed to exit the doorway. This pro-
cedure was necessary so sampling data would be compatible with the technique
used to estimate annual emissions.
FUGITIVE SOURCE SAMPLING RESULTS AND EMISSION ESTIMATES
A total of twenty-seven 1-hour sampling runs characterized emissions from
the four major sources under a variety of plant operational and wind conditions
as shown in Table 2. From an assessment of annual wind patterns and plant
operation data, annual emission estimates were developed for each source as
shown in Table 3. Since doorway emissions are nonexistent when winds blow
into the doorway, the several years of wind data available for this site were
used to estimate the amount of time emissions would blow out of the doorways.
On an annual basis, fugitive particulate emissions from the four material
handling buildings are estimated at about 50 tons or 15 percent of process
stack emissions. However, GCA feels the sampling data are biased low due to
the wet weather and light winds encountered during the study. The source
sampling found that peak emission rates from these buildings could easily
amount to 40 Ib/hr or more on weekends or nights with moderately southwesterly
and westerly winds, compared to process stack emissions of 100 Ib/hr.
160
-------
MONORAIL INSTALLED
o rr
ABOVE DOORWAY
8xlG inch HI-VOL
FILTER and HOLDER(S"
FLOW (\
GAUGE
\
/TROLLEY/ROPE ARRANGEMENT TO ALLOW
COMPLETE HORIZONTAL AND VERTICAL
POSITIONING FOR TRAVERSE
PRESSURE TAPS TO
MANOMETER
' > INLET NOZZLE
CYCLONE DROP-OUT
BOTTLE
Figure 2. Front view of source sampler constructed for Duraflake sampling program.
-------
TABLE 2, SUMMARY 07 SOURCE SAMPLING DATA FOR MAJOR FUGITIVE SOURCES
N.
Source
Truck dump
Dry storage
building
west door
Dry storage
building
east door
Ply trim
building*
Ply trim
hopper
Roof monitors
production
line 3
Number /
of
samples
5
1
2
2
1
5
1
4
1
1
4
13
Average
<10V
0.3
0.9
0.1
0.1
0.8
0.5
1.0
0.1
0.3
0.2
0.1
emissions Ib/hr
>10u
3.7
12.0
1.9
4.7
13.0
12.0
37.0
7.0
25,0
8.1
8,7
0.1
Internal activity
Normal operation, light winds
One truck, very dusty material, light wind
No activity; front door closed, light wind
Light activity; front door open, light wind
Continuous conveyor discharge, front door open, light wind
Normal weekend payloader operation, light to moderate wind
Heavy payloader operation near door; moderate winds
Peak emissions during conveyor discharge, light winds
One dusty conveyor discharge, moderate winds
Average emission during payloader removal of stored
material
Dust from payloader discharge to reentry hopper, time
averaged emission
Survey measurements of roof monitor emissions during
normal operation
*Reported values for one of two doors in building emissions normally escaped only through one door.
Data for particles less than about 25 ym.
-------
TABLE 3. ANNUAL EMISSION RATES FOR MAJOR FUGITIVE SOURCES
Estimated Percent °f A T0 T
^ year with Average TSP Net annual
Source . dust emission emission
operating . ,,. „ , , . N
r, ° generating (lb/hr) (tons/yr)
wind
Truck dump 2,850 51 3.7 2.7
Dry storage-west door
normal daily activity 6,260 54 1.9 3.2
payloader operation 4,000 54 12.0 13.0
Dry storage-east door
normal daily activity 313 46 1.9 0.14
payloader operation 4,000 46 12.0 11.0
Ply trim buildings
conveyor discharge 830 53 7.1 1.6
payloader activity 4,000 75 8.1 12.0
Reentry hopper at ply trim 4,000 53 8.7 9.3
TOTAL ANNUAL DOORWAY
EMISSIONS
Inventoried process
(stack emissions)
53
8,100 100 96.2 391
Particle size data collected during all sample runs showed that less than
5 weight percent of emissions were smaller than 10 vim aerodynamic diameter.
However, this does not necessarily indicate that most particles are heavy and
soon settle to the ground. Observations during sampling found most emitted
particles were entrained in prevailing winds and significant fallout on plant
grounds did not occur.
UPWIND/DOWNWIND AMBIENT SAMPLING APPROACH
To evaluate the impact of particulate emissions from the plant on ambient
air quality, upwind/downwind sampling was conducted concurrently with the
source sampling programs. Based on an analysis of historical wind data, four
ambient sampling sites were selected, two upwind and two downwind of the plant.
To the north, one hi-vol operated continuously for 24-hour periods while
a colocated, wind-actuated hi-vol only operated when winds were from the plant
sector. An additional single hi-vol, also located to the north, operated
continuously for 24-hour periods.
A similar arrangement of hi-vols was operated to the south of the plant.
Since winds tended to blow north-south, or south-north, these hi-vol stations
were designed to cope with anticipated wind patterns.
Hi-vols operated daily for 12 days with filter changes conducted at mid-
night of each day. An electric clock wired into the wind controls recorded the
actual time the wind-actuated hi-vols operated. Wind direction and wind speed
were continuously monitored and recorded on a strip chart unit.
163
-------
AMBIENT TSP MONITORING RESULTS
The ambient monitoring program collected a total of forty 24-hour TSP
samples at six hi-vols to the north and south of the plant for 12 continuous
days. The ambient data for each site is summarized in Table 4.
Upwind/downwind sampling showed substantial increase in ambient TSP when
winds originated from a sector including the plant boundaries, but excluding
adjacent industrial facilities. Based on data from this monitoring program,
particulate emissions from the plant apparently increased ambient TSP roughly
two to four times above measured background levels, as shown in Table 5.
Weather conditions during the sampling program did not appear representa-
tive of those which would produce maximum fugitive emissions from the plant.
Steady rainfall was experienced over roughly one-half of the period and winds
were generally calm or light. Further, TSP measured by GCA at the permanent
site to the north was significantly lower than that recorded by the state
agency over past years. For these reasons, TSP levels measured at ambient
sites and also from fugitive (doorway) sources at the plant are considered to
represent a lower limit to emission potential, not a maximum.
CONTROL TECHNOLOGY ASSESSMENT
In general, control options for fugitive particulate emissions fell into
two categories:
• Complete control by capital-intensive techniques; i.e., evacuating
all buildings to baghouses, and redesigning raw material handling
processes.
• A lesser degree of control, achieved by minor structural modifica-
tions, operational changes, and a dramatic increase in attention to
plant housekeeping and cleanliness.
Options for control of each type of fugitive emission are summarized in Table 6
and discussed briefly below.
Complete control of the open doorways can only be accomplished by evacua-
tion and diversion of the emission flow to a baghouse. The buildings cannot be
sealed because of the potential for explosions and also worker safety/health
concerns. Although technically feasible (the metallurgical industry often uses
complete building evacuation to control plants containing a number of diffuse
emission sources), the capital costs per building would be on the order of one-
half million dollars. This cost for complete control must be weighed against
the potential emission reduction. Fugitive emissions from these buildings
account for an estimated 15 percent of process stack emissions on an annual
basis and a maximum of 40 percent of stack emissions on a peak daily basis.
As an alternative to the baghouse option, emissions from these buildings
could be reduced, possibly by as much as 50 percent, by implementing a number
of structural and operational modifications listed in Table 6; i.e., reducing
doorway size, constructing enclosure type sheds over reentry hoppers, install-
ing conveyor discharge downspouts, and retraining plant operators.
164
-------
TABLE 4. SUMMARY OF AMBIENT TSP SAMPLING CONDUCTED IN SEPTEMBER 1978, yg/m3
U'
Date satellite
9/7
9/8
9/9
9/10
9/11
9/12
9/13
9/14
9/15
9/16
9/17
9/18
Mean
- THU
- FRI
- SAT
- SUN
- MON
- TUE
- WED
- THU
- FRI
- SAT
- SUN
- MON
Maximum
28
77
70
137
179
99
73
65
122
69
58
120
91
179
Mortuary Mortuary
24-hour W. D.
-
99
95
162
91
41
58
32
43
31
48
73
70
162
-
216
233
269
207
-
100
60
180
(1-
hour)
-
175
302
194
269
South
Main
24-hour
38
46
25
45
61
86
57
44
137
143
91
66
70
143
South
Main South
W. D. satellite Wind and weather
-
-
-
160
-
152
73
63
161
192
80
116
192
125
- Light south wind,
clear until noon,
rain until 6 a.m.
Light south wind,
rain all day
70 Light south wind,
rain
- Light south-se wind,
rain in mornings,
clear thereafter
37 Light south wind,
dry by evening
- Light northerly
winds, variable
40 Light variable wind,
no rain
Light variable wind,
no rain
78 Light north wind, dry
Light north wind, dry
- Moderate south wind,
dry - a.m., rain - p.m.
57 Light south wind, dry
56
78
-------
TABLE 5. AVERAGE MEASURED TSP BY WIND DIRECTION, ug/m3
North North South South
North Main wind Main wind South
satellite 24-hour directional 24-hour directional satellite
Average of 7 95
days with
south wind
Average of 3 97
days with
north wind
96
38
200
53
122
55
168
Note: Winds not always constant in direction.
TABLE 6. SUMMARY OF CONTROL OPTIONS APPLICABLE TO DURAFLAKE
Emission
source
Control
option
Comments
Major sources -
open doorways
Minor sources
General dust
accumulation
Evacuation to baghouse
Structural modification
- doorway size reduction to
reduce emission flow rate
- enclosure sheds over reentry
hoppers to minimize dust
entrainment
- conveyor downspouts to mini-
mize free fall distance
Operational changes
- retraining of bucket loader
operators to operate equipment
more carefully
- more diligent closure of
existing doors
Complete enclosure of belt conveyors
Installing more effective belt wipers
More effective control of leaky
ductwork, cyclones, etc.
Vacuum sweeper - daily use
Hosing of roofs and ductwork - daily
More careful operation of
vacuum truck
General reduction in dust
generating activity
Capital intensive
Very efficient
Relatively inexpensive
Less effective
than baghouse
Requires cleaning
New sweeper ordered
Hose connections
currently installed
166
-------
Improved control of minor point-type sources of fugitive emissions can be
achieved primarily by more effective conveyor belt enclosures to eliminate
spillage, use of more effective belt wipers, and more attention to repair of
leaky duct work, cyclones, and other minor sources.
A major problem at the plant is the substantial accumulation of dust over
plant surfaces, which contributes to windblown fugitive emissions. The solu-
tion here is simply more attention to housekeeping details. A three-wheel,
brush-type street sweeper is reportedly on order and should be used daily on
all plant grounds. Hosing of rooftops and other plant surfaces that collect
dust should be conducted daily since GCA's observations indicated enough
material can accumulate in 1 day to constitute a windblown dust problem. The
plant currently operates a vacuum truck for cleaning major spills and/or
storage silos and hoppers, and more frequent use of this equipment should be
encouraged.
CONCLUSIONS AND RECOMMENDATIONS
Based on GCA's evaluation of fugitive emissions during the 2-week sampling
program, the following comments and recommendations are offered.
• Major fugitive emission sources contribute about 50 tons per year
of TSP annually, or about 15 percent of inventoried stack TSP.
• Violations of ambient TSP standards were recorded outside of
plant property.
• Upwind/downwind hi-vol sampling showed increased ambient TSP when
winds were from the plant sector.
• Based on the 2 weeks of hi-vol data, all plant emissions increase
ambient TSP by a factor of 2 to 4 above background levels.
• All sampling data are considered to represent minimum emission
levels since the sampling period experienced low wind speeds and
wet weather.
• State-of-the-art control of fugitives involves capital-intensive
measures — complete building evacuation and/or process redesign.
• A moderate level of control can be achieved by minor structural
modification and minor operational change.
• A dramatic increase in overall housekeeping is essential to
reduce windblow fugitive dusts.
• Although only 5 percent of the fugitive emission particulate was
measured at less than 10 ym, visual observations indicated that
much of the TSP became entrained in ambient wind currents and
did not descend on plant grounds.
167
-------
A Method for Measuring Fugitive Emissions
From Cast House Operations
James H. Geiger
Betz«Converse»Murdoch»Inc.
Abstract
An attempt to measure fugitive emissions from blast furnace cast
house operations presented a variety of problems. A method was
developed to determine a mass emission rate. The method utilizes
high volume samplers placed in a multipoint array where temperature
and velocity measurements also are made. The reasoning behind the
method is discussed, and the method is compared to other "state of
the art" procedures for measuring fugitive emissions. Certain
factors which affect the emission measurements are discussed.
Results indicate that the method is reasonably accurate for
determining mass emission rates from cast house operations.
This paper is a synopsis of several measurement programs and,
although the basic methodology was the same, the type of equipment
used varied for each program. The method, currently being used in
an ongoing program to quantify emissions from cast houses, is
continually undergoing refinement.
160
-------
The emissions generated by a blast furnace during casting
originate from relatively small point, line and area sources (tap
hole, runners, hot metal cars); they expand quickly to the
dimensions of the cast house and exit by a number of routes.
Emissions from the tap hole are caused by the cooling of the hot
metal as it exits the furnace and encounters the atmosphere. More
emissions are created by a violent cast with a lot of hot metal
turbulence than are generated by a smooth, controlled cast. Oxygen
lancing of the tap hole at the beginning of a cast and the
enlargement of the tap hole by erosion during casting also
contribute to increased emissions.
The generation of emissions from the iron runners, iron trough
and slag runners is apparently a function of the surface area of
metal exposed to the atmosphere and the metal temperature. As the
metal cools, carbon emerges from the hot metal solution as "kish", a
form of graphitic carbon that is light, flaky and readily
air-borne. The trough and runners are lined with clay and coke
breeze which must be thoroughly dry prior to casting. Any residual
moisture reacts violently with the molten metal, resulting in
greater emissions.
The hot metal cast from a blast furnace is transported to
further processing facilities via hot metal cars (i.e., ladle cars,
torpedo cars). The hot metal cars are filled by the iron runner
spouts. If the hot metal cars are too cool, if a "skull" has formed
over the car opening, or if the hot metal splashes onto the inside
of the car opening, the hot metal cools very rapidly, and the
emissions from the cars can become extremely dense.
Because fugitive emissions are not confined, they are inherently
difficult to quantify. Three basic strategiesl»2,3 nave been
developed for the measurement of fugitive emissions: the roof
monitor, the quasi-stack, and the upwind-downwind methods. All
three methods, or variations thereof, have been used in an attempt
to quantify blast furnace cast house emissions^.
The roof monitor measurement technique assesses emissions
exiting through monitors, doors, windows, ventilators or other
openings. Usually high volume samplers^ are utilized to measure
the concentration of particulate passing through the opening. A
velocity is determined and a mass emission rate is calculated for
each opening. The numerous openings present in cast houses, and the
physical problems associated with sampling at those openings,
precludes the use of the roof monitor sampling method.
169
-------
The quasi-stack sampling method employs a temporary hood which
is exhausted by a fan to a temporary stack. The hood and exhaust
system are designed to capture most (theoretically all) of the
emissions from the process being investigated. Then, the mass
emission rate is measured by standard stack sampling techniques.
The quasi-stack sampling method is impractical for cast house
testing because it interferes with production equipment and
personnel. Also, the design for the temporary hoods would have to
meet strict criteria, including the ability to withstand high
temperatures.
The third method, upwind-downwind sampling, as well as a fourth
method, measurement in the plume, seemed the most promising methods
to use. The plume measurement method was developed by Jacko? for
measuring coke oven pushing operations. The method used a high
volume sampling device suspended from a boom into the plume. The
boom also held velocity and temperature measuring instruments. The
push was filmed with a 16 mm movie camera. The resulting film was
then projected onto a screen and measurements were made to determine
the cross sectional area of the plume. The film was also used to
determine when the samplers actually intercepted the plume so that
an accurate sampling time could be determined.
Observations of a number of casts indicated that neither the
plume measurement technique nor the upwind-down wind method would
work. The numerous points of exit for emissions from a cast house
result in a plume whose boundaries are not as well defined as those
from a coke oven, making it difficult to determine the plume
cross-sectional area. Also, interferences from surrounding sources
would have had a substantial impact on either sampling method.
The sampling methodology that eventually evolved was one that
had been dismissed earlier as being too difficult to implement. The
method could be described as an in-situ stack method because the
entire cast house was considered to be the stack. The
cross-sectional area of the cast house was divided into sub-areas
and each sub-area was sampled simultaneously. Samples were taken
with standard high volume samplers with constant flow controllers.
The samplers were placed on a scaffolding which was erected on the
roof girders just above the crane.
The philosophy behind such a sampling arrangement was simple.
Observations indicated that emissions exited the cast house via five
routes: hot metal car (ground) level; cast house floor doors;
windows; roof monitor; and around the furnace shell at roof level.
The emissions from each route were present in varying degrees during
each cast, but the bulk of the emissions appeared to be exiting via
the higher exit points (windows, monitor, roof top). By sampling at
the roof girder level, (approximately window height), the bulk of
the emissions would be intercepted.
170
-------
One objective of the first test program was to determine
emission factors for each source within the cast house (i.e., tap
hole, runners, hot metal cars). Consequently, the scaffolding and
samplers were placed to maximize plume interception from the
individual sources while maintaining sufficient coverage of the
entire cast house. The spatial limitations of the cast house girder
system determined the exact placement of the scaffolding.
To develop emission rates, plume velocity and cross-sectional
area must be determined. Conduction and convection transfer
significant quantities of heat to the surrounding air creating a
thermal draft which causes a rising air current. As the heated air
stream rises from a hot surface, it mixes turbulently with the
surrounding air. The higher the air column rises the larger and
more diluted with ambient air it becomes.
Equations have been developed that describe the diameter of a
hot rising jet at any height above a hypothetical point source
located a distance, z, below the actual hot surf ace. 8 The rising
air column expands approximately according to equations 1 and 2 as
indicated in Figure 1. A cross-sectional area at the height of the
samplers was calculated from equations 1 and 2 for each of the hot
sources in question (iron trough, runners, hot metal cars). The
resulting areas are indicated in Figure 2 by the shaded areas.
To complete the flow calculations, a velocity must be determined
and applied to the previously computed areas. The velocities of the
emissions were primarily determined by photogrammetric methods. A
movie camera with a precisely controlled shutter speed was used to
photograph various portions of each cast. The developed film was
projected onto a grid. As the film was advanced a single frame at a
time, an identifiable portion of a particular plume was followed on
the grid. By using simple geometry, the grid was calibrated to the
known size of objects projected onto the grid and distances between
reference points were established. The velocity of the plume was
calculated by counting the number of frames it took the plume to
travel the distance between reference points and multiplying by the
framing speed of the camera. This procedure was followed a number
of times for each plume to arrive at an average plume velocity. For
emission points, such as over the iron runners, which did not have
readily identifiable plume attributes, a different procedure was
used. Neutrally bouyant helium filled balloons were released and
tracked as they were projected onto the grid as described above.
The vertical transport of the balloons was timed for comparison with
the calculated velocities. In addition, a hot wire anemometer was
used to determine velocities outside the areas of the plumes.
171
-------
where:
z
Or
2D.1.138
0.5
(1)
(2)
the diameter of the hot column of air
at the level in question, ft
the diameter of the hot source, ft
the distance from the hot source to the
plane of Dc, ft
the distance from the hot source to the
hypothetical point source, ft.
HYPOTHETICAL
POINT SOURCE
Figure 1'
THEORETICAL EXPANSION OF AIR COLUMN
DUE TO HOT SOURCE
172
-------
SLAG SPOUT W.T
.
--
o
5%H IRON TROUGH PROJECTED AREA
Z~M RUNNER PROJECTED AREA
MM TREADWELL PROJECTED AREA
• AIR SAMPLING LOCATIONS
FIGURE NO. 2
SAMPLING NETWORK SCHEMATIC
-------
Once velocities and areas were determined, the flows at a given
sample location were calculated. The gas flow at a given sampler
was calculated as a time weighted average of each of the velocities
and sub-areas as calculated above. For example (referring to Figure
2), the determinations of flow within the"domain" of sample point
number 5 is composed of the flow from the non-plume area, the flow
from a portion of the runner when filled with hot metal, and the
flow from a portion of the plume from hot metal car number 3 during
the time that hot metal is going into that car. This relationship
can be expressed mathematically as follows:
n
£ V, Af t,
Qa = i'°l (3)
where:
A-j = the itn area, ft2
T - total time of test, min.
t-j = the time that the ith area was affected by the i'th
velocity, min.
Qa = the average flow at the sample point (ventilating
rate), ft^/min
y.j = the i™ velocity, ft/mi n
Note: n is a variable depending upon the specific point.
Ai = f(Dc)
Twelve standard high volume samplers were placed on the
scaffolding for the parti cul ate measurement as indicated in Figure
2. All twelve samplers were activated simultaneously by a common
switch at floor level. The samplers were started when tap hole
drilling was started. Since there are significant emissions after
the furnace is plugged, the samplers were not turned off until five
minutes after the mud gun was placed in the tap hole. Then, the
filters were changed and the samplers were readied for the next test.
During the first test program, the emissions were observed to
have a pronounced horizontal movement toward the furnace which acted
as a natural draft chimney. This unanticipated movement of the
emissions prevented the determination of emission factors for each
individual source and caused a reassessment of the sampler
placement. In subsequent test programs, normal stack testing
methodology has been adopted; i.e., an attempt has been made to
sample equal areas with the samplers placed at the center of each
area and to include the areas at the front of the cast house on
either side of the furnace. An example of such an arrangement is
shown in Figure 3. Final placement of the samplers is, however,
174
-------
--
_-
I||M|| '
n
T
CO
L,^ —
12
•
11
_«;
T
A
L
U
A
___— ^_ —
10
9
•
- L
.
!
I \
i
I
^
7
i
^^*
1 1
^•M
I
-20-
• AIR SAMPLING LOCATIONS
A LOCATION OF NOT METAL POURING STATIONS
Figure 3
SAMPLING NETWORK SCHEMATIC
-------
still dependent upon the spatial limitations of the cast house
girder system.
One other major revision, the elimination of velocity
determination by photogrammetry, has been made to the basic test
method used in subsequent programs. Instead, hot wire anemometers
and thermocouples were placed at each sampler location. The
thermocouple signals were fed to a multipoint recorder. In an
attempt to maximize data acquisition yet minimize cost, the hot wire
anemometer outputs were fed to a multiplexer and then to single
channel strip chart recorders. Velocities at four different sample
locations were recorded simultaneously.
Standard high volume samplers were chosen primarily for their
ease of operation. Additionally, the upright design prevented loss
of sample, and the cover prevented sample contamination from
fallout. The design also provided some size selectivity which
excluded particles that would probably settle out within the cast
house. The samplers were equipped with constant flow controllers
which simplified data reduction and enhanced the data reliability.
Tests during each cast also included particle size data
collection using a five stage cascade collection device attached to
one of the high volume samplers. The cascade device was moved from
sampler to sampler throughout the program. Because only one cascade
device was used, no simultaneous particle size data was collected.
Since it is unknown whether the particle size data variability was
point-to-point variability or test-to-test variability, the data was
treated as if it were from the same population. Averages indicate
that approximately 85% of the particles are less than the 7.2 micron
cut-off size of the first stage. Because the high volume sampler is
fairly effective up to about 100 microns, the samples were probably
indicative of the emissions actually leaving the cast house.
Testing was also conducted between casts so that background data
could be obtained. The data collected for the periods between casts
was considered as "upwind" data and was subtracted from the cast or
"downwind" data in subsequent analyses. This data treatment was
meant to account for the material remaining in the cast house after
the cast was completed and the driving force (heat) was removed.
The residence time of this suspended material could not be accounted
for and the extent to which it contributed to the cast test results
was unknown. Therefore, it was subtracted from them.
In one of the test programs, observed interference of
particulate matter from an adjacent process necessitated the
acquisition of the background data. Because the generation of the
interfering particulate matter was not constant, the impact on the
results of the individual cast tests is unknown, but it is believed
that the average values were reliable estimates.
176
-------
During normal casting operations, the slag which floats on top
of the hot metal is skinmed off. It is shunted via runners to a
slag pit adjacent to the cast house or into slag pots which then are
transported to a remote disposal area. In some cast house
operations, the cinder notch is opened and slag is flushed from the
furnace without casting any hot metal. This flushing operation is
performed just prior to a cast. Since this operation results in
noticeable emissions, testing was performed during the flushes and
the results were added to those obtained from cast tests.
One area of concern not addressed during any of the sampling
programs was any effect on the measurements, caused by local
meteorology. Since cast houses are fairly open, wind direction and
velocity could modify pollutant trajectories. Also, the difference
between summer and winter temperatures can cause variations in
emissions. During one test program, what appeared to be a
convection cell was established within the cast house itself. To
what extent this cell would impact on the measurements is unknown as
is its cause and relationship to the local meteorology.
Although some areas need further investigation and a few
questions need to be answered, the prelimi nary results of the
in-situ stack method are encouraging. To date, the method has been
used to generate mass emission rate data for a process for which
little reliable data previously existed. The method uses reliable,
readily available equipment and is suitable in a wide variety of
applications where conventional sampling methods would be difficult
to implement. With further refinements, the method is capable of
generating more detailed results in a cost-effective manor.
177
-------
References
1) Technical Manual for the Measurement of Fugitive Emissions:
Roof Monitor Sampling Method for Industrial Fugitive Emissions.
Industrial Environmental Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina. May
1976. Publication No. EPA-600/2-76-089b.
2) Technical Manual for the Measurement of Fugitive Emissions:
Quasi-Stack Sampling Method for Industrial Fugitive Emissions.
Industrial Environmental Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina. May
1976. Publication No. EPA-600/2-76-089c.
3) Technical Manual for the Measurement of Fugitive Emissions:
Upwind-Downwind Sampling Method for Industrial Fugitive
Emissions!Industrial Environmental Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, North
Carolina. March 1977. Publication No. EPA-600/2-76-089a.
4) Blast Furnace Cast House Emission Control Technology
Assessment!Industrial Environmental Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, North
Carolina. November 1977. Publication No. EPA-600/2-77-231
5) Private Cormunications
6) T.E. Kreichelt and T.G. Keller, "Roof Monitor Emissions: Test
Methodology," J. Air Pollution Control Association. 22:8. 641.
(Aug. 72)
7) R.B. Jacko, D.W. Neuendorf, and J.R. Blandford, "The By-Product
Coke Oven Pushing Operation: Total and Trace Metal Particulate
Emissions," Presented at APCA Annual Meeting, Portland, Oregon,
1976.
8) Air Pollution Engineering Manual. U.S. Department of Health,
Education and Welfare. Public Health Service Publication No.
999-AP-40.
173 •
-------
STEEL MILL
PARTICULATE CHARACTERIZATION AND SOURCE/RECEPTOR ANALYSIS
U.S. Steel Corporation Geneva Plant
Philip A. Russell
Materials Analysis Laboratory
Metallurgy and Materials Science Division
Denver Research Institute
University of Denver
Denver, Colorado 80208
ABSTRACT
This study (1) characterized particulate emissions from fugitive sources
within a steel mill to determine their potential to be utilized as relatively
unique environmental indicators and (2) determined their relative presence
and quantity in the ambient environment within a region near the plant. Analy-
ses were made using scanning electron microscopy/energy dispersive X-ray
spectroscopy (SEM/EDS) and transmission electron microscopy/selected area elec-
tron diffraction (TEM/SAED). Distinct species of particulates were identified
which could be useful in source/receptor analyses. These include carbon
particles produced during coking, open hearth and blast furnace operations;
iron spheres produced during blast furnace and open hearth operations; angu-
lar Fe and Fe-rich particulates produced during sintering plant operation;
potassium-sulfur rich particle type produced only in the slag flush and slag
tap sections of the blast furnace; and encapsulated iron particles produced by
door leaks after slag charging. Examination of ambient samples further demon-
strated that the carbonaceous particles and iron spheres produced by the steel
mill were the best environmental indicators of the steel mill impact. Angular
Fe and Fe-rich particulates were also useful indicators in the vicinity of the
steel mill itself.
INTRODUCTION
It has been hypothesized that observed light scattering problems and high
TSP levels in the Prove, Utah area are significantly affected by particulate
emissions from the U.S. Steel Corporation (USSC), Geneva Plant, located north
of Provo, Utah. The Materials Analysis Laboratory, Denver Research Institute,
performed a particulate characterization of (1) source emissions from the USSC
Geneva Plant, located north of Provo, Utah, and (2) ambient samples collected
within the region to determine the feasibility of obtaining fugitive emission
source/receptor information using microinventory analyses. Samples submitted
for analysis were collected on 0.2 and 0.4 pm Nuclepore polycarbonate thin
membrane filters in ^ 7-10 LPM low volume air filter systems. Analyses were
designed to provide definitive information on the composition and concentration
of particulate species distributed on nuclepore substrates by using scanning
(SEM) and transmission (TEM) electron microscopy, energy dispersive X-ray
spectroscopy and selected area electron diffraction. The study was conducted
179
-------
in two phases: (1) the characterization of fugitive particulate emissions from
sources within the steel plant to determine their potential to be utilized as
relatively unique environmental indicators specific to the Geneva steel plant-
and (2) determine the presence and quantity of such particulates in the ambient
environment in the vicinity of the plant. Because the project was terminated
for unspecified reasons prior to its completion, much of the material presented
in this paper was derived independently by Denver Research Institute. At this
date the exact locations of the ambient sampling sites have not been made
available to the author.
METHODS AND MATERIALS
All samples were collected by the Environmental Protection Agency, Nation-
al Enforcement Investigation Center, Denver, Colorado, during the fall of 1977
Particulates on thin membrane polycarbonate filters were examined "in situ" by*
coating excised portions of each filter with 20.0 nm of carbon by vacuum
evaporation to make them conductive. Each sample was quickly examined for
even particle distribution and then random portions of the filter were photo-
graphed at magnifications adequate to examine particles of interest. These
particulates were simultaneously subjected to individual qualitative X-ray
analysis; comments were recorded directly on photographs or separate notation
sheets. The number of particulates examined on any photograph ranged from
approximately 10 to 100; numerous photos were analyzed where practical to pro~
vide adequate sample sizes. X-ray emission spectra of typical and unusual
particulate species were recorded for future reference.
Estimates of mass loadings were derived from measured particle size and
the area observed. It was assumed for calculation purposes that all particles
were spheres with a diameter estimated by averaging a particle's longest axis
length and its most representative width at right angles to this axis. Specif-
ic gravity was assumed to be 2.0 for carbon, 2.4 for "CaS", 5.2 for Fe or Fe-
rich, and 2.6 for "KS" particulates. Thus by using (1) the assumed specific
gravity, (2) the number of particles located in a specific area(s) and (3) the
estimated individual average particle volume it was possible to estimate mass
loading in gm cm"2.
Only high magnification photographs (5,OOOX and 10,OOOX) were used to
analyze ambient samples, which restricted analyses to relatively small parti-
cles 0\> < 5.0 urn) for the purposes of this study. For particles less than l.Q
Vim the procedure was to measure all of the particles in 20mm x 20mm grid placed
on a series of randomly selected areas of a photograph. Average particle size
was calculated and the total number of particles of a group in the area are
entered into the equation (# of particles/cm^) (avg. particle vol.-cm )
(density-gm/cnr) = gm cm. For larger or less prevalent particles, a whole
photograph or number of photographs were used instead of a portion of a filter
Samples were also examined by transmission electron microscopy to provide
higher resolution of very fine structural detail, interior particulate morph-
ology and crystallinity (by electron diffraction). The process of membrane
filter dissolution and subsequent examination are detailed in Ruud et al. (1976)
180
-------
RESULTS
A. Source Samples
1. Coking Process
a) Door Emissions
Particulates observed on three samples were very sparse.. The par-
ticulates noted were spheres of carbon ranging from 0.5-2.5 ym in
diameter. In one sample mineral particles approximately 2-5 ym in
effective diameter were also observed.
b) Pushing
Emission observed on two samples contained virtually no particles.
c) Process Charging
No large particles on this sample. Some small carbon spheres 0.27
± 0.11 ym in average diameter in low concentration 8.64 x 10" cm"2
were present. This is approximately 1.78 x 10~7 gm cm"2 assuming
a specific gravity of 2.0.
d) Ambient Conditions Topside
Two samples were examined.
i) The first sample contained two morphologically different spe-
cies of carbon particles (Figs. 1 and 2). The first were
larger spheres 1.18 ± 0.41 ym in diameter in a concentration
of 9.43 x 105 cm~2 which estimated to be 1.62 x 10~6 gm cm"2.
The second group of particles were smaller but more numerous
with an average diameter of 0.15 ± 0.04 ym in a concentration
of 1.5 x 10^ cm2 with an estimated mass concentration of 5.30
x 10~6 gm cm"2. These smaller particles were often observed
to be agglomerated. Both species contained trace amounts of
Al, Si, S and Cr. The larger spheres were observed to usually
be hollow when examined by the TEM (Fig. 3).
ii) Sample C-9-18 contained larger particles similar to those
observed in C-13-21; however, the morphology of some of these
particles was distorted in shape such as to suggest that they
had been fluid at one time (Fig. 4).
e) Cap Leaks
Few particalates were observed on the sample examined.
2. Open Hearth
a) Door Leaks After Slag Charge
Mineral particulates (l-10ym) were observed in moderate
181
-------
Fig. 1. SEM micrograph of coke particles,
10.000X.
CO
to
0.1pm
TEM micrograph of two
species of coke particles
TEM micrograph of coke
particles, lO.OOOX.
Fig. 4. TEM micrograph of "hollow"
particles of coke, 2.000X.
-------
concentrations, they contained various amounts of Si, Fe and Ca.
The smaller particles (Fig. 5) contained Fe and carbon particles.
The Fe particles have an "encapsulated" appearance with a size of
0.40 ± 0.17 ym, concentration of 1.26 x 10 cm and mass concen-
tration of 2.20 x 10~ cm assuming a specific gravity of 5.2
The carbon particles were somewhat spherical or plate-like and had
a size of 0.28 ± 10 ym, concentrations of 1.29 x 10 cm and mass
of 2.97 x 10~6 gm cnT2.
b) Ambient
The moderately loaded filters contained large (> 5.0 ym) Fe-rich
mineral and Fe spherical particles. The smaller particles were
similar in composition to the preceding sample.
c) Hot Metal Charge
This sample was heavily loaded with iron and carbon particles
(Fig. 6). The larger Fe spheres (> 3 ym) were only distinguished
from smaller Fe particles by the presence of more surface morph-
ology: average size 4.07 ± 1.27 ym, concentration = 1.10 x 10^
cm~2, mass concentration = 2.02 x 10~^ gm cm~ . The smaller Fe
particulates (Fig. 7), which were somewhat spherical, had a mean
size of 0.86 ± 43, concentration of 1.10 x 10 cm and approximate
mass of 1.90 x 10~-^ gm cnf . The small carbon particles (e.g.
those not indicated in Fig. 7) were highly concentrated in agglom-
erates; size was difficult to assess but was approximately 0.18 ±
0.09 ym, which gives a concentration of 9.50 x 10 cm and a mass
concentration of 5.80 x 10~° gm cm" . TEM examination shows a
relatively stable but relatively nonelectron dense film coating on
all of these particles (Fig. 8).
d) Door Leaks
The samples contained only relatively light concentration of parti-
cles similar in composition to those produced in the hot metal
charge. Some Fe particulates < 0.3 ym in diameter were observed.
e) Door Leaks During Slag Flush
Some large (1.0-5.0 ym) mineral particles were observed. These
were mainly Ca with some Mg. Smaller carbon particles were ob-
served in similar sizes and concentrations as the hot metal charge.
f) Scrubber
Particulates observed in association with the scrubber were again
similar to the carbon produced by the hot metal charge but in
lower concentrations. Some particulate agglomerates (1-5 ym) con-
tained Fe, K, S. Many of the smaller particles (< 1.0 ym) had
trace amounts of Fe or were composed of Fe.
183
-------
Fig. 5. SEM micrograph of "encapsulated
iron particles produced by door
leaks in the open hearth, 10,OOOX.
03
I
Fig. 7. SEM micrograph of particles pro-
duced during the hot metal charge
process in the open hearth, 10,OOOX.
Fig. 6. SEM micrograph of particles pro-
duced during the hot metal charge
process in the open hearth,1000X.
1.0pm
Fig. 8.
TEM micrograph of open
hearth particles, 62,OOOX.
-------
g) Tap
This was a relatively lightly loaded sample. Some large particles
(5-2 pm) of Ca mineral containing Mg, S and Si were observed.
Smaller particles (< 1.0 pm) were mostly carbon but a few mineral
types were also observed.
h) Limestone Charge
This sample contained particulates similar to the hot metal charge,
but in reduced concentrations.
i) Slag Flush
Almost identical to those produced by hot metal charge, except
less concentrated. Some spheres less than 1.0 pm were not rich in
iron, however.
j) Bank Cleaning
No particles observed.
k) Fettle
No particles observed.
} • Blast Furnace
a) Slag Flush
Large spheres of Fe (^ 90%) and Ca (^ 10%); avg. size = 2.06 ±
1.01 pm, concentration = 1.41 x 10 cm mass concentration = 3.36
x 10 gm cm were observed. But numerous small particles rich
in K and S (Fig. $) dominated the sample. Morphology can be more
easily observed in high resolution transmission electron micro-
scope (Fig. 10). The average size was 0.24 ± 0.15 pm, concentra-
tion = 1.49 x 10 cm~^ and mass concentration - 2.80 x 10"-3 gm
cm~2 assuming a specific gravity of 2.6. The material produces an
electron diffraction pattern and is therefore crystalline. The
particles were also observed to partially disintegrate in an in-
tense focused electron beam.
b) Hot Metal Tap
Lightly loaded - a few large Fe spherical particles similar to
those found on the slag flush sample. Carbon associated with
this sample is composed of agglomerates of carbon made up of par-
ticulates < 0.1 pm in diameter.
c) Hot Metal Runner
Filter virtually without particles.
185
-------
Fig. 9. SEM micrographs (1,OOOX and 10.000X) and X-ray spectra of particles produced during blast furnace
slag flush operations.
-------
SC
•
Fig. 10. TEM micrograph of particles
produced during blast furnace
slag flush operations, 23,OOOX.
Fig. 11. SEM micrograph of particles produced
during sintering plant operations.
1000X.
Fig. 12.
£W$'tf$&&
-•* A-^%' '^ •
SFH micrograph of particles produced
during crushing operatins at the
Heckett Engineering site, 1000X.
-------
d) Slag Tap
Sample composition almost identical to Slag Flush but contains more
large Fe spheres. Surface of these spheres morphologically smooth-
er than those being produced in the open hearth. Average size »
1.74 ± 1.29, concentration B 1.26 x 10^ cm" , mass concentration •
1.81 x lO"-* gins cm~2.
Sintering Plant
All sintering plant samples were characterized by the almost total lack
of carbon particles, spherical particles and particles less than 1.0
ym. The samples contained large angular particles (Fig. 11) rich in
Ca and/or Fe. If we assume an average density of 4.0 the estimated
avg. size = 4.03 ± 2.35 vim and concentration - 7.77 x 10^
mass concentration is estimated at 1.07 x 10"^ gm cm~2
particulates of carbon or coal were also observed.
Heckett Engineering^
cm
-2
the
Some large
Overall, all samples collected in Heckett Engineering location con-
tained almost exclusively large particles of a mineral composition
(i.e. calcites and silicates); small (< 1.0 ym) carbon or Fe particles
were not observed on any of these samples (Fig. 12). The predominant
species was a calcium silicate (Figs. 13 and 14), similar in composi-
tion to the minerals epidote and dolomite.
Fig, 13. Typical x-ray spectra of an Fig. 14.
individual particle collected
during crushing operations.
Typical x-ray spectra of an
individual particle collect-
ed during crushing opera-
tions.
188
-------
B. Ambient Samples
Specific information on ambient sample sites has not been made available
at this time.
1. Station. 1
All samples for this sampling site were heavily loaded. Overall X-ray
emission spectra for November 7. 13, 16 and 18 are presented in Figure
15. From this information it is readily observable that elemental sul-
fur became most abundant on November 16 and elemental iron on November
18 (one of the highest levels observed on ambient samples). In both
cases this is the result of relatively large particles containing cal-
cium and sulfur or, respectively, Fe-rich minerals.
2. Station _2
The November 8 sample was heavily loaded with large mineral particles
which were mainly calcium silicates and silicon (quartz). One very
large Fe sphere (9 ym) was observed. Iron and iron-rich particles
were observed in sizes greater than and less than 1.0 ym in diameter.
Particles rich in sulfur were also observed, sometimes with trace
amounts of elemental lead. Some micron size spheres of carbon were
also observed. Sulfur was associated with carbon particles and was
very high in concentration producing an overall high bulk sample sul-
fur concentration (Fig. 16). The November 11 sample was much less
Fig. 16. X-ray bulk spectra from particulates
collected November 11, at site 2.
139
-------
November 7
November 13
November 16
November 18
Fig. 15. X-ray spectra of bulk samples collected at ambient site 1, November
7, 13, 16 and 18, 1977
190
-------
heavily loaded but contained relatively more large (> 1.0 ym) iron
particles than the November 8th sample.
The November 14 sample was virtually void of particulates and no Fe-
rich particles were observed.
The November 17 sample contained a moderate particulate load. Most of
the larger particles were calcium or calcium silicates. A few Fe
spheres about 2.0 ym were observed. Only a few small (< 1.0 ym) Fe
particles were noted.
3. Station 3
The November 9 sample was heavily loaded. The larger particulates were
almost entirely mineral. Two species of iron particles: individual
spheres approximately 0.5 um in diameter and agglomerates of particles
< 0.2 ym in diameter were observed.
The November 12 sample was less heavily loaded, contained a large Fe
sphere 2 ym in diameter and contained less small (< 1.0 ym) Fe particu-
lates than the November 9 sample. Relatively high concentrations of
sulfur were observed with the small carbon particulates.
The November 15 sample was similar in composition to November 12 except
for reduced amounts of material.
The November 18 sample exhibited increased amounts of material. There
was a significant increase in the observed number of large Fe mineral
particles. Small Fe particles were also present as 0.2 ym spheres.
Sulfur-rich particles were observed; the morphology was consistent with
carbon particles observed earlier.
4. Station 4
The November 8 sample was heavily loaded. Large angular particles of
Fe and Fe silicates were observed as well as a few Fe spheres. Small
particulates included Fe and Fe silicates; a few particles rich in Cu
and Zn were also observed.
The November 11 sample contained only a few large particles (none over
3-4 ym in length). The small particles were predominantly carbon
spheres < 2.0 ym and carbon plates/spheres; one agglomerate of Fe
particles was observed.
The November 14 sample was very clean. No small Fe particles were
observed, and the number of carbon particulates was greatly reduced.
An increase in particulate loading on the November 17 sample was ob-
vious. The large angular particulate fraction contained a high propor-
tion of Fe or Fe-rich particulates. An increase in small carbon and Fe
spheres (^ 0.2 ym) was also evident at higher magnifications.
191
-------
5. Station 5
Samples collected on November 7 and 10 were heavily loaded with large
particles (Figs. 17 and 18). The sample collected November 13 (Fig.
19) was less loaded and the sample collected November 16 had virtually
no particulates of any kind.
On November 7 some of the larger angular particulates contained Fe or
Fe silicates which were 4.97 ± 1.75 um in average effective diameter.
Carbon particulates of two distinct types were observed at higher mag-
nification (Fig. 18); iron-rich particles were present in relatively
low concentrations.
No iron silicates were observed in the sample collected November 1O
however, Fe particulates were observed, most were spherical (^ 2.0 um)
in shape.
The November 13th sample was lightly loaded but did contain a number of
angular particulates ^1.0 ym which were mostly mineral silicates al-
though some were carbon and iron (pollen was also present). Some car-
bon spheres approximately 1 ym in diameter were observed. Particulates
< 1.0 um were present in very low concentration, and only one submicron
Fe particle was observed.
6. Station _6
The November 12th sample was unique in that it contained a number of
large flyash and carbon particles with high concentrations of potassium
Some spheres containing Fe were noted. The smaller size particle frac-
tion contained Fe (0.2 um) and sulfur-rich carbon particulates.
The November 18th sample was void of particles > 2 um and contained
only a few carbon and Fe particles about 0.2 ym.
7. Station 8
Samples were collected at this location consecutively from November 7th
through I Bib. Particle loading was heavy the 7th and increased the 8th
Some reduction was noted the 9th and 10th; the llth and 12th exhibited
even heavier loads. Some reduction was noted the 13th; and by the 14th
the level of particulates had been reduced most significantly. From
the 15th through the 18th the amounts of material continued to rise to
levels experienced November 7.
Most of the large particulates noted on the 7th were mineral; none were
Fe and only a small percent were Fe rich. A significant number of
spheres > 1.0 um were observed which were composed of carbon but also
contained a significant amount of K; others were siliceous or PbBrCl.
The smaller particulates observed were carbon (0.1 um); some spheres
(^ 0.6 um) of carbon, and Fe (0.3 um) were also observed. Material
rich in sulfur was also present.
The sample collected the 8th was very heavily loaded but had a compo-
sition similar to that of the previous day (Fig. 20). More carbon
192
-------
Fig. 17. SEM micrograph, Ambient Site 5,
November 7, 200X.
-
Fig. 19. SEM micrograph, Ambient Site 5,
November 13, 10,OOOX.
* I.Ojim
SEM micrograph, Ambient Site 5,
November 7, 5,OOOX.
Fig. 20. SEM micrograph, Ambient Site 8,
November 8th, 2,OOOX.
-------
particulates (0.5 ym) were evident than previously observed; these, and
the finer carbon particulates, contained trace amounts of sulfur. This
proportion of particulates remained the same through the 13th. On the
llth no large Fe or Fe-rich particulates were observed, though the
sample was heavily loaded; the small Fe particulates, e.g. Fig. 21, had
a mass concentration of 5.95 x 10~^ gm cm~2. The amount of sulfur
associated with the numerous carbon particles was observed to have in-
creased since November 8th.
The 13th, one large (3.6 urn) Fe sphere in a 1000X field was observed.
The November 14th sample was the cleanest of Station 8. However, iron
spheres (1-2.5 ym) were observed, as well as some small (< 1.0 ym) Fe
particles. Particulates observed on the 15th were similar to those of
the 14th; some particles (^ 0.5 ym) containing calcium and sulfur were
also observed.
The large particle fractions of sample 16 were predominantly mineral
and carbon (automotive); a few Fe spheres and Ca-S particles were ob-
served. A few submicron Fe and Ca-S particles were also observed.
Figure 22 illustrates some of the typical small particles observed on
this sample.
Samples collected on the 17th and 18th are very similar to those col-
lected the 16th with a slight increase in concentration.
8. Station 9
Samples were collected consecutively from November 8th through the 17th.
The most distinctive feature of these samples is that no large Fe
particles are observed (only one large angular Fe-rich particle was
observed in all the samples). A few micron size Fe particles were ob-
served in the sample collected November 17th. Particulate loading was
heaviest November 8-11, reducing in amount through the 12th and 13th to
a low level the 14th which then remained relatively continuous through
the 17th. Submicron size particles rich in copper and lead-arsenic
were observed in a number of samples.
On November 9th, large carbon particles 2.92 ± 1.31 ym and smaller par-
ticles > 1.0 ym were observed. Figure 23 illustrates some of these
particle types. Sulfur was observed to be associated with both carbon
types.
9. Station 10
November 8 and 9 - very heavy loading. Though large clay (Al,Si,K,Ca,
Fe) and quartz particles were present, the filters were observed to
also contain high concentrations of small particles. Most of the sub-
micron size carbon particles were rich in sulfur, producing a high sul-
fur concentration for the entire sample (Fig. 24).
November 10 and 11 - heavily loaded but reduced from the 8th and 9th.
Large particles still predominant, but some large carbon particles
apparent. More larger (.4-1.0 ym) submicron size particles observed;
194
-------
Fig. 21. SEM micrograph, Ambient Site 8,
November llth, 10,OOOX.
Fig. 22. SEM micrograph, Ambient Site 8,
November 16th, 10,OOOX.
c
Fig. 23. SEM. micrograph, Ambient Site 9,
November llth, 10.000X.
Fig. 24. X-ray spectra from a bulk sample collected
from ambient sampling site 10, November 9.
The sulfur peak is produced by sulfur adsorbed
onto carbon particles.
-------
they are probably carbon with adsorbed sulfur; some iron agglomerates
also present (Fig. 25). Overall sulfur concentration was the highest
of any ambient or source sample on sample 10 and was almost entirely
associated with particles less than one micrometer in average diameter.
By November llth, sulfur concentration had been reduced to 1/3 of that
of the 10th.
November 12-13 - sample loading reduced from the previous samples; com-
position essentially the same. Individual small particles (> 0.5 urn)
of Fe and Cu observed.
November 14-15 - Samples lightly loaded. On the 14th, only small (0.1
ym) and large (0.5 ym) carbon particles were observed. By the 15th,
particles had increased in concentration; small (< 0.5 ym) Fe particles
were also observed. Carbon particles contained only traces of sulfur.
A few submicron size Cu particles were observed on the 15th.
November 16-18 - Sample loading became progressively heavier after the
15th. Particulates observed in the larger size fraction continued to
be mainly mineral, some Fe spheres and Fe-rich angular particulates.
In the smaller size fraction, carbon particulates were observed to in-
crease. Fe and Ca-S particulates were also observed.
Estimated filter concentration values for large and small carbon par-
ticulates and iron-rich particulates spherical in morphology are pre-
sented in Appendix Table A.
DISCUSSION
The following particulates produced by the U.S. Steel Corporation, Geneva
Plant could be potentially useful in assessing fugitive source emissions from
specific areas within the plant:
1. Carbon particulates small spheres approximately 0.1 ym in diameter, often
found in agglomerates, are produced during the coking, open hearth and
blast furnace processes. Particulates of this type were observed in all
ambient samples. However, there are other sources for these particulates
notably the automobile. Careful analysis would be required to use these
as source emissions indicator.
2. Carbon particulates approximately 1.0 ym in diameter having a definite
spherical shape are produced during the coking process. These particulates
were observed at most ambient stations, usually in low concentrations.
They are apparently unique to the Geneva Steel Works.
3. Carbon particulates - spheres or plate-like particles ^ 0.30 ym in diame-
ter produced by door leaks after slag charge. Similar particulates have
been observed at a number of ambient sites. They are apparently somewhat
unique, although they might be indistinguishable from some particulates
produced during coking operation.
4. Iron spheres - spheres containing mainly Fe produced in a size range from
> 5 ym to < 0.3 ym. The most productive source is probably the hot metal
-------
Fig. 25. SEM micrograph, Ambient Site 10,
November 10, 10.000X.
Fig. 26. TEM micrograph, Ambient Site 10,
sulfur-rich carbon particle, 20,OOOX.
19 /
-------
charge section of the blast furnace, although these particles are produced
in other open hearth and blast furnace sections. They are unique to the
plant. The larger (> 1.0 ym) spheres produced in the open hearth and blast
furnace areas may be morphologically different. Such iron spheres both
large and small are found in low concentrations in almost all ambient sam-
ples.
5. Potassium-sulfur particulates 0.2 ym in diameter are produced only in the
slag flush and slag tap sections of the blast furnace. They are unique to
the steel plant. However, similar particles were found rarely only in a
few ambient samples.
6. Large Fe and Fe-rich angular particles > 2.0 ym in effective diameter are
produced during sintering pJant operations. These particles could be used
as a source indicator, but cannot be considered unique. A carefully
applied statistical program coupled with meteorological information would
be required to assess their potential as a spot source indicator.
7. "Encapsulated" iron particles (0.4 ym) produced during door leaks after
slag charge. These particulates have a non-spherical structure, and mor-
phological "halo" surround a central core. Similar particulates have been
found in some ambient samples.
Carbon particles are usually identified by their conspicuous lack of ele-
ments with an atomic number greater than sodium. An observation made at all
ambient sites was that the levels of sulfur associated with these types of
particles were dependent upon the length of time that the local air mass had
remained stagnant. In Figure 26, a TEM micrograph of one of the sulfur-rich
carbon samples is displayed; these particles did not produce an electron dif-
fraction pattern and were stable in a focused beam. No morphological differ-
ences were noted between particles with and without a sulfur association.1
A separate analysis for large angular iron-rich mineral particles was
conducted to determine the feasibility of using the sintering plant fugitive
emissions as a local specific indicator. A table of these data is provided in
Appendix Table B and a summary is presented in Fig. 27. Sites 5 and 9 are
respectively close to the steel plant.
On about November 13 a storm front moved through the region. Using mass
estimated (Appendix Table A) it is possible to examine changes in particulate
concentration probably associated with the Geneva Steel Plant operation as the
air mass in the Provo area is cleared and the rate at which these particulates
return to previous levels (Fig. 28). It should be noted that at this stage of
refinement these values are not corrected for flow rates and sample times and
are therefore comparative in nature.
I/
This topic is covered in more detail in Russell (1979)
19J
-------
100%-
iFe >80
n
Ft 80 - 20
IF. <20
50%-
Fig. 27. Particle analysis for iron rich species,
199
-------
5 x 10 ' -
CAtlON <0 Jp.
O
o
1 x 10 —
.0-*-
1 t. - I
Fig. 28. Ambient particle analysis summation.
-------
CONCLUSION
A microscopic inventory of particulates can provide invaluable assistance in
assessing the nature and inhomogeneities of fugitive source emitters. Particles
that are, in turn, unique can be identified some distance from the emitter. With
suitable sampling parameters, estimates can be made for number and mass per unit
of air.
201
-------
REFERENCES CITED
Ruud, C. 0., C. S. Barrett, P. A. Russell and R. L. Clark. 1976. Selected
area electron diffraction and energy dispersive x-ray spectrometry, a compari-
son. Micron. 7:115-132.
Russell, P. A. 1979. Carbonaceous particulates in the atmosphere: illumina-
tion by electron microscopy. Pages 133-140 ^n T. Novakov (ed.), Proceedings:
carbonaceous particles in the atmosphere. LBL-9037. NTIS, Springfield, Va.
202
-------
APPENDIX A
MASS ANALYSIS OF PARTICLES COLLECTED ON
AMBIENT THIN MEMBRANE SUBSTRATES
203
-------
Appendix Table A
Estimated mass collected on ambient
site thin membrane filters
n
Estimated Mass (gin cm )
Sample
1-7
1-13
1-16
1-18
2-8
2-11
2-14
2-17
3-9
3-12
3-15
3-18
4-8
4-11
4-14
4-17
5-7
5-10
5-13
5-16
6-12
6-18
8-7
8-8
8-9
8-10
8-11
8-12
8-13
Carbon (< 0.3 ym)
3.02 x 10-6
3.18 x ID"6
2.39 x 10~6
1.75 x 10~6
2.71 ± 0.91 x ID"6
1.05 ± 0.66 x 10~6
1.13 x 10" 7
1.01 ± 0.41 x ID"6
2.19 x 10-6
1.33 x 10-6
5.62 x 10"7
6.22 x 10"7
^ 5.0 x 10~7
1.21 x ID'6
5.47 x 10"7
1.18 x 10"6
4.70 ± 0.73 x 10"6
7.56 ± 0.27 x ID"7
2.12 ± 0.71 x ID"7
2.25 x ID"8
5.32 x KT7
9.22 x 10~8
1.67 x 10~6
2.60 x 10~6
1.76 x 10~6
9.35 x 10" 7
8.17 x 10- 7
3.03 x 10"7
Carbon (> 0.3 ym)
1.11 x 10-6
1.23 x 10-6
2.09 x 10~6
j.,84 x 10~6
^^
_ _
__
—
8.03 x 10"7
1.01 x 10-6
6.26 x 10~7
1.37 x 10~7
1.50 x ID"6
2.07 x ID-6
1.50 x. 10~7
2.37 x 10-7
2.34 x ID'6
2.99 x 10~7
__
—
2.44 x 10~7
—
6.07 x 10~7
2.84 x 10-6
1.57 x 10- 7
9.25 x ID" 7
3.01 x 10~6
1.93 x 10"6
8.07 x 10~ 7
Fe
1.37 x 1C-5
1.50 x 10-6
1.17 x 10"6
1.34 x 10~6
3.26 x 10-6
1.39 x 10~7
—_
2.26 x 10~7
4.69 x 10"6
1.25 x ID"6
8.25 x 10~7
3.19 x 10"7
1.82 x 10~5
(Fe-rich mineral)
1.78 x ID'7
8.67 x 10-8
7.80 x 10~7
3.74 x 10-7
«••
—
1.37 x 10~7
1.71 x 10~8
4.46 x 10~8
2.06 x 10~7
3.14 x 10- 7
7.64 x 10~8
1.12 x 10~6
(1 Irg. part.)
3.25 x 10"8
2.31 x 10~7
Other
8.84xlO~7
(PbBrCl)
4.69xlO~7
(flyash)
—
—
—
—
3.19x10-6
(mineral)
~
(1 part.)
204
-------
Appendix Table A (Con't)
o
Estimated Mass (gm cm
Sample Carbon_(< 0.3 ym) Carbon (> 0.3 ym) Fe
6.19 x 10"7
8-14
8-15
8-16
8-17
8-18
9-8
9-9
9-10
9-11
9-12
9.
1.
1.
2.
1.
1.
1.
1.
3.
7.
28
06
82
98
99
40
66
19
35
79
x
x
X
X
X
X
X
X
X
X
io-7
10~6
10~6
10~6
io-6
io-6
10- 6
10" 6
10- 6
io-7
6
6
6
6
5
5
9
8
.57
.80
.12
.44
.30
.67
.30
.42
x
-
X
X
X
X
X
X
X
—
io-7
-
io-7
io-7
io-7
io-7
10~6
io-7
io-7
3.61 x 10~7
("CaS")
1.44 x 1(T6 1.07 x 10~6
("CaS")
2.77 x 1(T7 2.43 x 10~6
("CaS")
9.17 x l
-------
APPENDIX B
ANALYSIS OF IRON-RICH MINERAL
PARTICULATES
206
-------
Appendix Table B
Analyses of particles >^ 5.0ym on thin membrane samples
S-10-6, 5-7 and ~9-3 from Geneva Steel Works
o
-J
Sample
S-10-6
5-7
9-13
Fe >
80%
35.9
4.8
0
Si
11.5
2.4
4.8
JSr.5.
Si &
Ca
6.5
2.4
0
su*
Ca
1.3
0
0
Total
19.3
4.8
4.8
—
Si
19.2
25.0
39.3
Si &
Ca
7.7
35.7
13.1
Fe __<
Ca
6.4
22.6
40.5
^_
Other
2.8
3.6
1.2
Total
37.1
86.9
94.1
Carbon
7.7
3.6
1.2
-------
DEVELOPMENT OF HORIZONTAL ELUTRIATORS FOR SAMPLING
INHALABLE PARTICULATE FUGITIVE EMISSIONS
by
Kenneth M. Gushing
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
ABSTRACT
The U.S. Environmental Protection Agency (EPA) is required, under the
amended Clean Air Act of 1977, to review the scientific basis for the total
suspended particulate (TSP) ambient air quality standard and determine whether
a revised particulate standard can be promulgated by December, 1980. It has
been recommended that research to develop information for a size-specific
standard should focus on inhalable particulate (IP) matter as defined as air-
borne particles £15 ym aerodynamic diameter. This particle size range relates
to that fraction of particulate matter which can primarily depos.it in the con-
ducting airways and the gas exchange areas of the human respiratory system
during mouth breathing.
This paper addresses the efforts of Southern Research Institute under con-
tract to IERL/EPA/RTP to experimentally apply horizontal elutriation to specif-
ic methods for sampling instack, ambient, and fugitive inhalable particulate
emissions. Theoretical and experimental data are shown. Results of the appli-
cation of horizontal elutriation for the initial collection stage of the
Fugitive Assessment Sampling Train (FAST) developed by TRC Environmental Con-
sultants are presented.
208
-------
INTRODUCTION
The U.S. Environmental Protection Agency (EPA) is required under the
amended Clean Air Act of 1977, to review the scientific basis for the total
suspended particulate (TSP) ambient air quality standard and determine whether
a revised particulate standard can be promulgated by December 1980. It has
been recommended that research to develop information for a size-specific
standard should focus on inhalable particulate (IP) matter as defined as air-
borne particles <15 urn aerodynamic diameteri This particle size range relates
to that fraction of particulate matter which can primarily deposit in the con-
ducting airways and t.;as exchange areas of the human respiratory system during
mouth breathing. Figures 1 and 2 present some of the data which led to the
specification of the 15 urn IP size standard. While 100% deposition of particles
during nose breathing occurs above 10 urn aerodynamic diameter, the 100% depo-
sition size for mouth breathing is approximately 15 ym aerodynamic diameter.1
An ideal IP sampler would then separate those fractions of particulate matter
larger and smaller than 15 ^m aerodynamic diameter.
Figure 3 presents the performance criteria for IP samplers. It was de-
veloped after extensive discussions between EPA staff members and scientists
knowledgeable in particle sampling methods.8 All samples will be required to
fall within the performance specification indicated by the region shown in
Figure 3. It is constructed by plotting two lognormal curves, with geometric
standard deviations of 1.0 and 1.7 through the 15 urn ± 2 ym 50% points. After
allowing 10% wall loss for small particles and 10% penetration of large par-
ticles, the area is completely defined. The tolerance in the Dso is 15 ym ±
2 ym.
Theoretical Considerations
A variety of aerosol sampling systems are being developed for use in mea-
suring the concentration of inhalable particulate (IP). The primary perfor-
mance requirement is that each system must separate particles larger and
smaller than 15 ym aerodynamic diameter according to the specified efficiency
curve.
Cyclones, impactors, virtual impactors, and horizontal elutriators are
devices that possibly can be used to collect the larger particles. All of
these, however, have features that limit their applicability. Impactors may
not retain particles efficiently that are much larger than the stage cut point.
The particles bounce and are reentrained into the gas stream, resulting in
contamination of the small particle fraction.9'10'11 Cyclones are not subject
to that problem but they are larger and their performance is difficult to pre-
dict.12 Virtual impactors may have significant losses of small particles.13
Horizontal elutriators are bulky and their performance is sensitive to orienta-
tion. However, for reasonable velocities reentrainment is not a problem in
horizontal elutriators and the theory describing their performance is straight-
forward and accurate.11*
Horizontal elutriators have been used extensively as devices for measur-
ing and collecting particles and as apparatae for a variety of aerosol
209 •
-------
experiments.15"1' The British Medical Research Council adopted the penetra-
tion curve of a particular horizontal elutriator as the definition of respir-
able particles. Mercer has reviewed many of the practical applications of
elutriators for sampling respirable particles1.* Hamilton and Walton18 have es-
tablished the criterion that reentrainment of deposited particles may occur if
the ratio of the mean air velocity to duct height (rectangular geometry)
reaches a critical value which lies between 240 and 650 sec'1.
This paper describes the results of laboratory studies that were carried
out to develop horizontal elutriators for measuring the IP concentration in
fugitive, ambient, and stack aerosols. Specific designs were developed for
use in conjunction with commercial, instack, cascade impactors, the EPA's Source
Assessment Sampling System (SASS),20 and the EPA's Fugitive Assessment Samplina
Train (PAST)." 9
Background Discussion
The initial concept of the horizontal elutriator was the result of studies
to measure quantitatively the deposition of aerosol particles on surfaces ad-
jacent to aerosols flowing through laboratory apparatus. Natanson22 and
Thomas23 independently solved this problem for particle settling in a circular
horizontal tube. Although the derivations are somewhat complex, the results
can be summarized and simplified in terms of the collection efficiency (Eff) by
the following equation:
-1/3 ./I _ e-Lt 3 J. -,,-„„,•„ ~l/3
Eff = - 2c /I - e"3 e1/3 /I - e"4 + arcsin e1" (1)
where
3V gL
e = — and
8 aV
V is the settling velocity,
s
L is the length of tube,
a is the radius of tube, and
V is the average velocity of the gas.
The theory describing the performance of a rectangular, flat-plate elu-
triator is considerably less complex and is re-derived here to give the reader
a better understanding of the particle motion in the elutriator.
The settling velocity of spheres in a gas stream is given by the equation
V
Cpqd2
s 18n
where
g is the acceleration due to gravity,
C is the slip correction factor,
p is the particle density,
d is the particle diameter, and
n is the viscosity of gas.
2IC
-------
In health-related studies, the aerodynamic behavior of the particles is of
interest. It is therefore convenient to relate the settling velocity of all
particles, whatever their shape or density, to that of spheres having unit den-
sity. The aerodynamic diameter is defined as the diameter of a unit density
sphere having the same settling velocity as the particle of interest. The
aerodynamic diameter is given from Equation 2 by settling p = 1 g/cm3 (the
units must be retained). Also, for particles larger than about 2 ym, the slip
correction factor is approximately equal to unity. A graph of settling velo-
city vs. diameter is shown in Figure 4.
For fully developed flow between parallel plates, the velocity profile of
a gas is parabolic with the maximum velocity equal to 3/2 the average and with
zero velocity at the plates:
where
V is the velocity parallel to the plates, at a point y,
X
y is the displacement above the bottom plate,
h is the spacing between plates, and
V is the average velocity of the gas.
The velocity profile and a typical particle trajectory are illustrated in
Figure 5.
The efficiency with which particles are collected is determined by the
dimensions of the channel, the velocity of the gas, the settling velocity of
the gas, the setting velocity of the particles, and the height at which the
particles.-enter the channel.
Consider a particle which enters the channel at position y as shown in
Figure 6 and which has a trajectory of horizontal length L (the°length of the
channel). Neglecting diffusion, which is negligible for the particles of in-
terest, all particles of this diameter or larger entering at or below position
y will be captured. All particles of this diameter entering the channel at
positions higher than y will penetrate the channel.
Thus
dx = /6y. _ 6£\ -
dt { h h2 / V
but dy is given by dy « V dt, therefore
s
-*)
dy; and
V , (4)
from which the length of the channel needed to capture particles of a particu-
lar diameter may be determined.
211
-------
If the particles are uniformly distributed within the gas, the fractional
collection efficiency of the particle illustrated in Figure 5 is equal to the
ratio of the volume of gas passing below position y , to the total volume; thus
Eff.
V dy
x •*
14
W / V
X
o
where W is the width of the channel.
Now, from Equations 2 and 4,
r,ff
Eff. * -3- » — *-^ — . (5)
hV 18r|hV
The theoretical efficiency curves for a horizontal elutriators with a
circular cross-section (from Equation 1) and for a horizontal elutriator with a
rectangular cross-section (from Equation 5) , both designed to have a cut point
of 15 urn aerodynamic diameter, are shown for comparison in Figure 7.
The efficiency is found to be independent of the vertical velocity profile
and, as Fuchs1** observed, Equation 5 can also be derived assuming plug flow.
Experimental Procedures and Results
Experiments were conducted to verify Equation 5 and to set design para-
meters for the IP samplers.
Figure 8 is a schematic diagram of the experimental apparatus used to in-
vestigate the performance of a horizontal elutriator designed to have a cut
point (D50) of 15 \an aerodynamic diameter. The settling chamber consists of
28 channels, each 7 mm high, 17.8 cm wide, and 38.1 or 20 cm long. A high vol-
ume blower (Model 305, Sierra Instruments) connected in series with a variable"
voltage transformer was used to supply the desired air flow rate through the
chamber. Monodisperse particles of methylene blue were generated using a vi-
brating orifice aerosol generator. During each test, the particles were sam-
pled and checked frequently by optical microscopy to insure constant monodis-
persity. All particles entering the horizontal elutriator were collected
either by settling on the plates or on the 8 in. x 10 in., (20.3 cm x 25.4 cm)
filter placed downstream from the plates. ' '
Upon completion of each test, the plates and filter were washed separatel
with tap water. Samples from each wash were centrifuged to remove debris an*^
the masses of methylene blue collected on the plates and on the filter were d
termined by absorption spectroscopy.
A study of the velocity distribution through the settling chamber was mad
to determine the configuration necessary to obtain uniform flow. The use of
212
-------
two filters and an extended, flared inlet was necessary to provide the desired
velocity distribution. Figure 9 is a velocity profile measured using a thermal
anemometer immediately upstream from the blower before the plates were in posi-
tion. From these data, it was evident that the blower was pulling uniformly
across the rectangular opening where the two 8 x 10 in. filters were positioned
Figure 10 shows the velocity profile measured upstream of the collector plates
with the extended, flared inlet in position.
The velocity profiles shown in Figures 9 and 10 were considered satisfac-
tory, and the overall average velocity was used to calculate the theoretical
performance curves for the system. The average velocity through the chamber
measured using the thermal anemometer was in agreement with previous calibra-
tion of the blower.
Two sets of experimental data were obtained from the testing of the hori-
zontal elutriator. The first set was acquired while operating the settling
chamber at an average gas velocity of 70 cm/sec and a plate length of 38.1 cm.
The second data set was obtained by operating the system at 40 cm/sec after
shortening the plate length to 20 cm. The theoretical curves of the collection
efficiency versus aerodynamic particle diameter shown in Figures 11 and 12 were
developed from Equation 5 using the following parameter values:
Figure 11 Figure 12
p = 1.35 g/cm3 p = 1.35 g/cm3
g = 981 cm/secz g = 981 cm/sec2
L=38.1cm L=20cm
n - 181 x 106 poise n = 181 x 106 poise
h = 0.701 cm h = 0.701 cm
V = 70 cm/sec V = 40 cm/sec
Reynolds No. - 315 Reynolds No. = 180
V/h = 100 V/h = 57
The calculated values from theory for the collection efficiency were
found to be in excellent agreement with the measured values. No corrections
were made for end effects.
The experiments described above indicated that the theoretical equations
can be used to predict particle collection by horizontal elutriators to a high
degree of accuracy. Consequently, the equations were used to create design
nomographs for IP precollectors to be used in conjunction with the three sys-
tems of interest. Design parameters were calculated for horizontal elutria-
tors to be used with: (1) cascade impactors operated at 14.2 i/min and 149°C,
(2) the Source Assessment Sampling System operated at 125 Jl/min and 204°C, and
(3) the Fugitive Assessment Sampling Train operated at 5,282 Jl/min and 23°C.
Figures 13, 14, and 15 are the design nomographs for the three sampling
devices. On the vertical axis is the open area required, neglecting the thick-
ness of the tube walls. The horizontal axis is the length of the elutriator
required to yield the IP performance at the specified flow rate and temperature.
In constructing the graphs, it is assumed that the rectangular channels have
widths much greater than their heights, so that the vertical walls will not
have a significant effect on the gas flow.
213
-------
Fugitive Assessment Sampling Train
The Fugitive Assessment Sampling Train (FAST), developed by TRC Environ-
mental Consultants for EPA, was designed to obtain 500 milligram samples in an
8 hour period downwind of typical industrial sources. The initial prototype
shown in Figures 16 and 17 consisted of a louvered inlet transform, a 15 ym
D-,0 impactor precollector, a 2.5 ym D$o cyclone, and a back-up filter. The
operating flowrate is 5.28 m3/min. A 5 CFM side stream to an XAD-2 solid sor-
bent canister is used to measure the organic vapor content of the sampled gas
stream.
Subsequent laboratory calibration of the precollection impactor by
Southern Research Institute indicated that it was not operating properly; a
large number of particles were bouncing and reentraining. The growing interest
in the inhalable particulate fraction indicated that some type of 15 ym D50
precollector would still be necessary. The success that Southern had had in
the laboratory evaluation of horizontal elutriators led to the decision to de-
sign a 15 ym D5a horizontal elutriator inlet collector for the FAST. Knowing
the flowrate of the FAST and the size and weight limitations required by TRC
it was a straightforward calculation to determine the specifications for the
inlet elutriators. '
The final design was for a horizontal elutriator consisting of 67 slots
of height 7 mm, and a slot width of 45 cm and length of 21.5 cm. in order to
facilitate removal and cleaning of the elutriator section after each test, the
plates are mounted in two removal containers as shown in Figure 18. Alignment
pins in each container hold the 67 plates in position. Slide type covers can
be inserted to close off the elutriator section before and after sampling to
prevent contamination. See Figures 19, 20, and 21. The elutriator is mounted
in front of the FAST as shown in Figures 22 and 23.
After the elutriator was built, it was shipped to Southern Research In-
stitute for calibration tests. The laboratory set up was as shown in Figure
24. Monodisperse 15 ym and 16 ym aerodynamic diameter particles of ammonium
fluorescein were sampled. The particle generator outlet was moved across the
louvered inlet section to simulate a uniform inlet aerosol concentration during
the sampling period. The particle collection efficiency for these two
particle sizes was 47.8% and 58.8%, respectively as shown in Table 1. Figure
25 shows these data points along with the theoretical efficiency curve. The
agreement between theory and experiment is very good.
The new Fugitive Assessment Sampling Train is now awaiting its initial
field sampling test which should occur during the summer of 1980.
Summary
Horizontal elutriation has been shown to be an effective and accurate
method for obtaining inhalable particulate samples. Application of the hori-
zontal elutriator method to the Fugitive Assessment Sampling Train has proven
the design extrapolation to a full scale device. Future field tests wi^l de-
termine the usefulness of horizontal elutriation under actual industrial opera-
ting environments. ""
214
-------
TABLE 1
FAST HORIZONTAL ELUTRIATOR INLET TEST DATA
Test NO.
Aerodynamic Particle Diameter (ym)
Filter Pressure Drop ("Hg)
Sampling Duration (hrs.)
Inlet Louvre Section
Screen
Screen to Elutriator
Transform
Elutriator Plates
Plate Holder Section
Filter Transform and
Filter
1
16.0
2.0
4
2
15.0
2.2
2.5
% of
Total Mass
0.75*
0.0009*
3.63
56.2
0.5
% Coll.
Eff.
0.75
0.0009
3.65
58.8
1.27
% Of
Total Mass
28.2
5.34
1.56
31.2
0.14
% Coll.
Eff.
28.1
8.4
2.4
47.8
0.4
38.9
100.0
33.7
100.0
*Questionable results due to inaccurate washing
technique.
215
-------
References
1. F. J. Miller, D. E. Gardner, J. A. Graham, R. E. Lee, Jr., "Size Con-
siderations for Establishing a Standard for Inhalable Particles, " J. Air
Pollut. Contr. Assoc.f 29(6) :610 (1979).
2. M. Lippmann, "Deposition and cleanance of inhaled particles in the human
nose," Ann. Otol. 70:519 (1970).
3. R. F. Hounam, A. Black, and M. Walsh, "Deposition of aerosol particles in
the nasopharyngeal region of the human respiratory tract," Nature 221:1254
(1969) .
4. G. Giacomelli-Maltoni, C. Melandri, V. Prodi, and G. Tarroni, "Deposition
efficiency of monodisperse particles in human respiratory tract," Amer.
Ind. Hyg. Assoc. J. 33:603 (1972). "
5. A. Martens and W. Jacobi, "Die in-vivo-Bestimmung der Aerosol teil-chendep-
osition in Atemtrakt bei Mund-bzq. Nazentmung," in Aerosole in Physik.
Medizin und Technik, Kongress-bericht der Aerosol tagung im Taunus-Sana-
torium am 17. und 18., Oktober 1973 der LVA Wttbg., V. Bohlau, Ed. Bad
Soden, West Germany, Gesellschaft fur Aerosolforschung, 1973, pp. 117-121.
6. G. Rudulf and J. Heyder , "Deposition of aerosol particles in the human
nose," in Aerosole in Naturwissenschaft, Medizin und Technik, V. Bohlau,
Ed. Proceedings of a conference held in Bad Soden, October 16-19, 1974,'
Gesselschaft fur Aerosolforschung 1974.
7. Airborne Particles, National Academy of Sciences, National Research Coun-
cil, Subcommittee on Airborne Particles, Committee on Medical and Biolo-
gical Effects of Environmental Pollutants, 1977, p. 155.
8. W. B. Smith, "Sampling Techniques For Inhalable Particulate Matter",
Special Report under EPA Contract No. 68-02-2610, January 16, 1979,
Southern Research Institute Report No. SORI-EAS-79-031.
9. M. Corn and F. Stein, "Re-entrainment of Particles from a Plume Surface,"
Am. Ind. Hyg. Assoc. J. , 26:325 (1965).
10. A. K. Rao and K. T. Whitby, "Non-ideal Collection Characteristics of
Single Stage Cascade Impactors," Am. Ind. Hyg. Assoc. J., 38:174 (1977).
11. K. M. Gushing, J. D. McCain, and W. B. Smith, "Experimental Determination
of Sizing Parameters and Wall Losses of Five Source-Test Cascade Impac-
tors, " Ejiyj£pj]u_ScJ^__Technol_. , 13:726 (1979).
12. W. B. Smith, R. R. Wilson, Jr., and D. B. Harris, "A Five-Stage Cyclone
System for In-Situ Sampling," Environ. Sci. Technol., 13:1387 (1979).
13. T. J. Yule and c. G. Bror.iarck, "An Experimental Study of Virtual Impac-
tors,'1 presented at Symposium on Advances in Particle Sampling and Mea-
surement, Daytona, Florida (October. 1979).
14. N. A. Fuchs, The Mechanics of Aerosols, Pergamon Press, New York (1964),
P. 110.
1 r-
-------
15. F. Stein, W. A. Esmen, and M. Corn, "The Shape of Atmospheric Particles in
Pittsburgh Air," Atmos. Environ., 3:443 (1969).
16. D. Rimberg and J. W. Thomas, "Comparison of Particle Size of Latex Aero-
sols by Optical and Gravity Settling Methods," J. Colloid Interface Sci.,
32:101 (1970).
17. M. Corn, F. Stein, Y. Hammad, S. Manekshaw, W. Bell, S. J. Penkola, and
R. Freedman, "Physical and Chemical Characteristics of 'Respirable' Coal
Mine Dust," Ann. N.Y. Acad. Sci., 200:17 (1972).
18. Medical Research Council Panels, cited by R. J. Hamilton and W. H. Walton,
"The Selective Sampling of Respirable Dust," in C. N. Davies, (Ed.),
Inhaled Particles and Vapors, Pergamon Press, Oxford (1961).
19. T. T. Mercer, Aerosol Technology in Hazard Evaluation, Academic Press,
New York, Ch. 8 (1973).
20. D. E. Blake, Source Assessment Sampling System; Design and Development,
EPA Report No. EPA-600/7-78-018, 221 pp. (1978).
21. H. Kolnsberg, Ed., Third Symposium on Fugitive Emission Measurement and
Control (October 1978, San Francisco, California), EPA-600/7-79-182,
August 1979.
22. G. Natanson, cited in N. A. Fuchs, The Mechanics of Aerosols, Pergamon
Press, New York (1964), p. 112.
23. J. W. Thomas, "Gravity Settling of Particles in a Horizontal Tube,"
J. Air Pollut. Contr. Assoc., 8(1):32 (1958).
217
-------
AERODYNAMIC DIAMETER AT 300 l/min,
ui
DC
CO
UI
V)
O
100
1
6 8 10
20
80
lounam et al. 1969
Q Lippmann 1970
A Giacomelli-Maltoni et al. 1972
• Martens & Jaeobi 1973
• Rudolf & Heyder 1974
70
60
- 50
O
rTJ»Uunamet,,.T1969 ' ' ' « H/'
100
1000
10,000
700-«2
Figure 1. Deposition of monodisperse aerosols in the head during inhalation via the nose
versus D^F, where D is the aerodynamic equivalent diameter (pm) and F is the
average inspiratory flow (l/min)2-6 The inspiratory flows in the individual
studies of this composite range from 5 to 60 l/min. The heavy solid line is
the International Commission of Radiological Protection Task Group?
deposition model.
218
-------
100
80
z H
g&
g u 60
UJ Z
05
Q t
UJ
40
20
AERODYNAMIC DIAMETER AT
30 l/min INSPIRATORY FLOW, urn
6 8 10
20 30
1
1
1 1 til
t I i t I mi
100
1000
10,000
700-63
Figure 2. Deposition of monodisperse ferric oxide aerosol in the heads of nonsmoking
healthy males during mouthpiece inhalations as a function of D?F where D is
the aerodynamic equivalent diameter (nm) and F is the average inspiratory
flow in liters/min. An eye-fit line describes the median deposition between
10 and 80%.7 The fitted fine has been extrapolated to 15 jum.
219
-------
to
o
4 6 8 10 20
PARTICLE DIAMETER, /jm aerodynamic
60 80 100
4181-30
Figure 3. Recommended specifications for collection efficiency of samples
of inhalable paniculate matter.
-------
0.001
0.6 1 2 4 8 10 20 40
PARTICLE DIAMETER, Mm AERODYNAMIC 4181-86
Figure 4. Sett/ing velocity in air for unit density spheres.
221
-------
V
Vc
Vs x -y
4181-106
5. Velocity profile and particle trajectory between parallel plates.
122
-------
IT
h v
yo
+ t
4181-73
Figure 6. Zone of 100% particle collection.
223
-------
u
z
01
0
u
0
u
RECTANGULAR
CYLINDRICAL
2.0 4.0 6.0 8.0 10 20 40 60 80 100
GEOMETRIC MEAN DIAMETER, micrometers
4181-233
Figure 7. Theoretical collection efficiency by particle settling in rectangular
and cylindrical tubes.
224
-------
COMPRESSED AIR LINE
REGULATOR
AND TRAP
REGULATOR
DRYER
I
ABSOLUTE FILTER
REGULATOR
VALVE
ROTAMETERS
I I
VIBRATING
ORIFICE
AEROSOL
GENERATOR
FLARED INLET
A|R
DISPERSION AIR
FILTER
n
SETTLING CHAMBER
VARIABLE
VOLTAGE
TRANSFORMER
HIGH VOLUME
AIR SAMPLER
I
SYRINGE PUMP WATER MANOMETER
FUNCTION GENERATOR
4181-74
Figure 8. Apparatus used to measure the collection efficiency of the settling chamber.
-------
V (m/sec)
** 0.30
*• 0.20
^ 0.10
STATISTICAL DATA:
MEAN, x , = 0.24 m/sec
STANDARD DEVIATION,
Sx = 0.02 m/sec
COEFFICIENT OF VARIATION,
8.2%
4181-107
Figure 9. Profile of the air velocity immediately upstream from the blower
of the settling chamber before the plates were positioned.
226
-------
STATISTICAL DATA:
MEAN, x , - 0.23 m/sec
STANDARD DEVIATION.
Sx = 0.02 m/sec
COEFFICIENT OF VARIATION.
•• 8.8%
4181-108
Figure 10. Profile of the air velocity upstream from the plates
of the settling chamber.
227
-------
>
o
ui
5
H
u.
UJ
z
g
K
O
O
O
2 34 6 8 10 20 30 40
AERODYNAMIC PARTICLE DIAMETER, /urn
60 80 100
4181-109
Figure 11. Collection efficiency vs. aerodynamic particle diameter for the
horizontal elutriator. Length = 38.1 cm.
228
-------
u
OJ
LU
0
o
0
U
THEORY (RECTANGULAR)
EXPERIMENT
2 34 6 8 10
AERODYNAMIC PARTICLE
20 40 60 80 100
DIAMETER, ^m im-232
Figure 12. Collection efficiency vs. aerodynamic diameter for the horizontal
elutriator. Length - 20 cm.
229
-------
KJ
U3
O
u
LLJ
Of
ta
O
DC
C
RECTANGULAR ELUTRIATOR
2mm 3mm 456 PLATE SEPARATION
NS^Cs^
CYLINDRICAL ELUTRIATOR
RECTANGULAR ELUTRIATOR
TUBE DIAMETER 2mm
CYLINDRICAL ELUTRIATOR
2.0
3.0 4.0 5.0 6 7 8 9 10
LENGTH, cm
20
30
40 50 60 70
4181-124
Figure 13. Relationship of design parameters for horizontal elutriators with
DSQ outpoints of 15 yjm aerodynamic diameter used as precollectors
for instack cascade impactors.
-------
CROSS-SECTIONAL AREA VS. LENGTH FOR 15 /urn AERODYNAMIC CUT POINT
RECTANGULAR ELUTRIATOR
8 PLATE SEPARATION
CYLINDRICAL ELUTRIATOR
RECTANGULAR ELUTRIATOR
i 3mm ; 4mm p56;
CYLINDRICAL ELUTRIATOR
LENGTH, cm
40 50 60 80 100
4181-123
Figure 14. Relationship of design parameters for horizontal elutriators with
outpoints of 15 nm aerodynamic • SASS.
-------
RECTANGULAR ELUTRIATOR
2mm 3mm 4mm 5 6 7 8 - PLATE SEPARATION
<
_
DC
u
LJ
tfl
S
£
U
600
500
400
300
200
CYLINDRICAL ELUTRIATOR
RECTANGULAR ELUTRIATOR
100
50
40
10
80 100
LENGTH, cm
200
300 400 500
700 1000
4181-125
Figure 15. Relationship of design parameters for horizontal elutriators with
outpoints of 15 ysn aerodynamic diameter - FAST.
-------
Figure 16. Fugitive assessment sampling train with cascade impactor precollector.
233
-------
CYCLONE
INLET
CYCLONE
Figure 17. Fugitive Assessment Sampling Train components.
-------
Figure 18. Horizontal elutriator container for FAST.
235
-------
Figure 19. Horizontal elutriator positioned downstream of inlet transform.
236
-------
Figure 20. Slide covers protecting FAST horizontal elutriator.
237
-------
Figure 21. Slide covers of FAST horizontal elutriator partially removed.
233
-------
Figure 22. Horizontal elutriator positioned on inlet of FAST.
230
-------
Figure 23. Louvered inlet positioned on front of FAST.
240
-------
Figure 24, Laboratory set-up for calibration of FAST horizontal elutriator.
241
-------
u
z
L.I
u
u
LV
LU
Z
u
:
t
Pffl
RECTANGULAR
- CYLINDRICAL
1.0 2.0 4.0 6.0 8.0 10 20 40 60 80 100
GEOMETRIC MEAN DIAMETER, micrometers
4181-234
Figure 25. Laboratory calibration data for the FAST horizontal elutriator inlet
collector. Also shown are theoretical collection efficiency curves for
rectangular and cylindrical tubes.
242
-------
TECHNIQUES FOR EVALUATING SURFACE AND
GROUND WATER EFFECTS OF DRY ASH DISPOSAL
by
James F. Villaume, Project Scientist
Pennsylvania Power & Light Company
Two North Ninth Street
Allentown, PA
Dennis F. Unites, Principal Scientist
TRC Environmental Consultants, Inc.
125 Silas Deane Highway
Wethersfield, CT
ABSTRACT
Utilities are finding it necessary to switch to dry fly ash handling to
minimize water quality impacts. At the same time, regulations regarding
solid waste disposal are becoming increasingly more stringent. In
switching from a wet sluiced system to dry ash disposal at the Montour
Steam Electric Station, Pennsylvania Power and Light Company required
data as to the surface and ground water effects of ash disposal for both
design and permit purposes. To provide these a study program of labora-
tory testing and computer modeling was conducted in conjunction with a
detailed site investigation.
Three types of laboratory testing were involved. They included: an
extraction leachate test to assess ash variability; a serial batchwise
extraction test to determine ash leachate quality and potential ground
water effects; and a runoff simulation. To determine runoff quality the
simulation results were then used as input into a modified version of
EPA's SSWMM computer model.
The paper will discuss the testing techniques and how the program results
were incorporated into the disposal site design and plan of operations.
-------
INTRODUCTION
Pennsylvania Power & Light Company (PP&L) operates a 1500-megawatt coal-
fired (eastern bituminous) steam electric generating station (the Montour
Station) near Washingtonville, Montour County, Pennsylvania. In 1976
PP&L was cited by the state environmental agency for discharging arsenic
at levels which caused the water quality criteria of the station's
receiving stream to be exceeded. An investigation by the company revealed
that most of the arsenic was being released from the fly ash as a result
of the wet sluicing and ponding process. Attempts to identify the
mechanisms controlling the arsenic levels in the ash pond were largely
unsuccessful, and so after evaluating various other alternatives (waste
recycle, discharge relocation, and stream augmentation) the company
decided to convert the existing wet ash handling system to a dry system.
This meant that any ash which could not be immediately marketed for
reuse would have to be landfilled. After considering twelve possible
sites, a 170-acre company-owned tract adjacent, to the station was finally
selected as having the least potential environmental impact after develop-
ment (Figure 1). To determine just what this impact would be and how
best to minimize it, a study of the ash and site characteristics was
undertaken. Of greatest concern were the effects the ash leachate and
runoff might have on local ground water supplies and area streams. What
was being sought ultimately was a basic landfill design concept and plan
of operations.
244-
-------
FIGURE 1
PLOT PLAN
245
-------
STUDY PROGRAM
Ash Variability
The variation in ash chemistry was investigated for two reasons. First,
the full range of chemical variability had to be established so that
representative samples of the ash could be collected for the later
leaching and runoff simulations. It was also necessary to know just how
this range fit into the possibly larger range of ash types likely to be
produced by the station over time, so that the results of the present
study would have some applicability beyond just the immediate situation.
Several likely sources of variation were considered.
Generally speaking, fly ash represents the non-combustible fraction of
the coal being burned. Thus, the chemical characteristics of the fly
ash are largely determined by the composition of this non-combustible
fraction, which can vary from coal to coal, even for coals from the same
mine (1). The Montour Station receives the majority of its coal from
two supplies in western Pennsylvania, one a captive deep mine working a
single seam ("Greenwich" coal) and the other a stripping operation
working up to five separate seams ("Benjamin" coal). The proportion of
coal coming from each is about 2:1 Greenwich:Benjamin. To the extent
possible the fly ash samples collected for the study were keyed to these
two supplies in an attempt to control this one important source of
chemical variation.
Once in the boiler, where the fly ash is actually formed, the coal is
subjected to various operating conditions which could also contribute to
ash variability. They include combustion temperature, excess oxygen,
and station load. Rather than try to isolate each source of variation,
the extent to which these operating conditions might be expected to
change was quickly checked by reviewing past station operating records.
In no case was the change found to be significant enough to warrant
further detailed investigation.
After its production in the boiler, the fly ash is exhausted through the
station's electrostatic precipitators, where it is removed from the
combustion gases and then transferred to a series of hoppers for collec-
tion (32 for each generating unit arranged in two banks of 16, as shown
in Figure 2). From work done earlier at the station it was known that
the smaller ash particles tended to be concentrated in the outlet hoppers
(those furthest from the boiler) and the larger size fraction in the
inlet hoppers (those closest to the boiler). Since the partitioning of
chemical elements by ash particle size is well documented in the litera-
ture (2, 3, 4, 5), most of the effort to quantify the chemical variability
of the ash was concentrated in this area.
The principal technique used to evaluate the effects of size sorting on
ash chemistry was to take a sample of ash from an inlet, outlet, or mid-
range hopper, add it to distilled water at a low solid:liquid ratio (50
246
-------
OUT
IN
STflCK
BOILER
OUT
B-4
IN
FIGURE 2
FLY ASH HOPPER LflYOUT
-------
g:800 ml), mix for 10 minutes, let the mixture stand for an hour, and
then decant and filter. The parameters were standardized for each test
run so as not to introduce any new variables into the system. The
mixing time selected was determined by separate test and represents a
stable condition characterized by constant pH and conductivity. The low
solid:liquid ratio was designed to prevent suppression of one dissolved
species by another more soluble species ("first ion effect"). The
results of the analyses performed on the resulting solutions are presented
in Tables 1 and 2. As a gross check on these results samples of the ash
itself were also analyzed by atomic absorption and emission spectrographic
methods. The results of these analyses are presented in Table 3.
This data clearly shows that the Greenwich and Benjamin ashes produce
very different aqueous solutions. In particular, the Greenwich ashes
are generally strongly alkaline (pH 10-11), have a high calcium:iron
ratio (characteristic of the strong alkalinity (6)), and contain no
detectable manganese or nickel. The Benjamin ashes, on the other hand,
are strongly acidic (pH 3-4) and have a- low calcium:iron ratio. This
large difference in pH is significant because it indicates that any
future changes in ash chemistry will probably fall within the range of
characteristics already studied.
The data also shows a marked variation in ash chemistry with particle
size sorting, as evidenced particularly by the increasing concentrations
of arsenic, selenium, and chromium in the extracts from the hopper
samples collected at increasing distances from the boiler. Thus, ash
samples from the outlet hoppers could be expected to produce "worst
case" solutions having the highest dissolved metals concentrations.
This was important to know as far as planning future sampling activities.
The data from the first phase of the study also served to identify which
chemical elements would most likely be leached out of the ash in large
amounts (the potential "bad actors") and which in insignificant amounts.
For example, cadmium and mercury were present in such low concentrations
in all of the extracts that they were judged not to be worth looking at
in the following leaching and runoff simulations. Antimony and lead,
meanwhile, were never found in any of the extracts, even though the
levels of lead in all of the whole ash samples were quite high (70-100
mg/kg).
An interesting result of the variability data is the comparison of the
recent ash samples with those collected in 1976, before the flue gas
conditioning systems (S0_ injection) were installed. The largest dif-
ferences appear to be in the Benjamin ash. The older samples produce
much lower concentrations of extractable arsenic and selenium and much
higher concentrations of manganese, iron, nickel, and zinc and they have
a much lower calcium:iron ratio. The results from the whole ash analyses
248
-------
TABLE 1
RESUITS OF ASH VARIABILITY TESTS
LEACHABLE CONSTITUENTS (ug/g)
Hopper
A-l
A-4
A- 8
A-9
A-12
A- 16
A-l
A-4
A-8
A-9
A-12
A-16
A-l
A-4
A-8
A-9
A-12
A-16
2-B Inlet
2-B Inlet
2-B Inlet
2-B Inlet
2-B Inlet
2-B Inlet
B-l
Coal Type
Unit 112 Greenwich
"
11
11
"
Unit #2 Benjamin
"
"
"
11
ti
Unit #2 Greenwich
it
ii
"
"
Greenwich
Benjamin
Benjamin
Greenwich
Banjamin
Benjamin
Unit *2 Greenwich
Date
4/20/79
ft
It
II
It
II
4/19/79
11
ii
"
M
ii
6/08/79
It
"
lr
II
(1
5/14/76
5/06/76
5/05/76
5/13/76
12/07/76
12/15/76
7/06/79
As
4.0
2.4
2.9
2.7
3.5
2.7
7.4
4.3
1.1
4.2
3.7
7.8
3.4
3.0
3.8
3.7
5.1
3.5
2.5
0.8
0.85
2.0
0.03
0.11
2.2
Se
4.0
1.8
1.12
—
—
—
0.32
4.6
2.2
—
—
—
3.7
4.2
3.5
—
—
—
1.12
nd
nd
1.6
nd
nd
2.24
Sb
—
—
—
—
—
nd
—
—
—
—
—
__
—
—
—
—
— •
nd
nd
nd
nd
nd
nd
nd
Mo
—
—
—
—
—
4.3
—
—
—
—
—
__
—
—
—
—
—
11.2
nd
nd
9.6
nd
nd
31.2
B F
—
—
—
—
— —
14.9 14.7
—
—
—
— —
— —
_.
—
—
—
—
— —
3.2 38.7
3.2 81.6
7.2 35.2
9.6 26.4
nd 3.0
8.8 51.2
7.2 5.3
Al
333.
151.
102.
250.
192.
78.6
176.
261.
282.
73.1
12.2
230.
370.
293.
317.
782.
198.
140.
349.
414.
285.
126.
261.
576.
237.
Mn
nd
nd
nd
nd
nd
nd
4.5
nd
nd
0.5
0.8
nd
nd
nd
nd
nd
nd
nd
nd
22.2
25.1
nd
52.0
30.6
nd
Fe
nd
i.n
nd
1.6
nd
] .0
8.6
1 .1
1.0
nd
2.6
nd
1.8
nd
nd
2.1
nd
nd
nd
86.4
88.8
nd
138.
59.2
nd
Ca
4,336.
4,360.
4.8SO.
2.6S8.
2,304.
2,6'2.
2,656.
2,1«4.
3,008.
2,192.
2,000.
1,632.
2,720.
3,008.
2,800.
2,240.
2,096.
1,920.
1,512.
717.
1,096.
1,808.
653.
2,080.
1,664.
Total Cr
3.7
2.4
nd
—
—
—
2.2
nd
nd
—
—
—
1.9
nd
nd
—
—
—
nd
3.8
nd
nd
nd
nd
1 .6
Zn
1.7
4.5
0.32
—
—
—
10.4
0.2
0.3
—
—
—
__
0.3
0.2
—
—
—
0.64
99.2
191 .
0.48
501.
87.2
0.48
Ni
nd
nd
nd
—
—
—
7.7
—
—
—
—
—
nd
nd
nd
—
—
—
nd
31.8
35.4
nd
302.
62.4
nd
Cd
nd
nd
nd
nd
nd
nd
0.5
—
—
—
—
—
nd
nd
nd
nd
nd
nd
—
—
—
—
—
~~~
Pb
nd
nd
nd
—
—
—
nd
nd
nd
—
—
—
nd
nd
nd
—
—
—
—
—
—
—
—
Hg
nd
nd
0.006
—
—
—
—
—
—
—
—
nd
nd
nd
—
—
—
—
—
—
—
—
™
NOTES :
"nd" indicates none detected.
-------
TABLE 2
SELECTED pH, CONDUCTIVITY, AND TDS RESULTS
FROM ASH VARIABILITY TESTS
to
ui
o
GREENWICH
Unit 2
Hopper it
A-l
A-4
A-8
A- 9
A-12
A-16
B-l
B-8
B-9
B-12
B-16
A-l
A-4
A-8
A-9
A-12
A-16
B-l
B-4
B-8
B-9
B-12
B-16
Date
4/20/79
It
tt
ft
If
ft
ft
IT
II
tt
6/08/79
tl
II
If
tt
II
tt
tt
M
It
It
If
10.2
11.3
11.4
11.0
11.2
11.6
10.3
11.2
11.0
11.1
10.9
10.3
10.1
10.4
10.2
10.6
10.7
10.3
11.2
10.2
10.5
10.6
10.4
Conductivity
(uMHO)
1496
2095
2645
1277
1127
2095
1496
1195
1417
1187
1602
1101
1251
1216
1031
887
953
1197
1100
1050
964
914
844
TDS
(mg/L)
1482
854
1199
1021
809
795
1360
1014
858
783
724
982
1207
1056
933
682
568
1163
935
920
823
795
651
Date
4/19/80
IT
II
IT
It
II
II
It
||
6/21/79
6/21/79
—
6/21/79
6/21/79
—
BENJAMIN
Conductivity
pH (uMHO)
4.1
9.9
8.8
8.6
7.4
10.4
4.0
5.2
4.7
4.3
8.1
3.9
3.7
—
3.7
3.9
—
1376
886
1367
917
877
707
1338
1308
988
958
808
1100
1400
—
1150
1100
—
TDS
(mg/L)
1320
748
1292
775
727
567
1272
1195
844
1267
648
936
1446
—
1126
948
—
-------
TABLE 3
RESULTS OF WHOLE ASH ATOMIC ABSORPTION AND
EMISSION SPECTROGRAPHIC ANALYSIS
A1203
Fe203
CaO
MnO
As
Cd
Pb
Se
Fe203
CaO
MnO
As
Cd
Pb
Se
Benj amin
Inlet
Composite
6-21-79
25.6%
13.1%
2.08%
0.033
440ppm
1.4
75
28
Benjamin
Inlet
5-6-76
26.6%
16.0
1.62
0.039
-290ppm
0.9
100
50
Greenwich
Composite
4-20-79
25.9
9.59
3.98
0.039
230
1.5
70
14
Greenwich
Inlet
5-13-76
25.1
8.22
2.67
0.046
130
1.2
70
21
Greenwich
Composite
6-8-79
26.0
9.60
3.15
0.036
190
1.6
80
16
Greenwich
Inlet
5-14-76
24.9
8.56
2.63
0.047
100
1.1
75
21
Benj amin
Composite
4-19-79
26.0
12.9
2.35
0.032
210
1.2
80
23
Benj amin
Inlet
12-2-76
26.2
15.2
1.38
0.050
150
1.2
80
70
Benj amin
Composite
5-5-76
26.2
17.1
1.72
0.036
280
1.1
80
45
Benj amin
Inlet
12-15-76
26.4
15.4
1.60
0.045
300
1.0
90
44
251
-------
confirm these differences. While this change in ash chemistry cannot be
attributed to an exact cause, it is worth noting because it shows the
magnitude of the changes which are possible from factors other than
simple size sorting and chemical conditioning.
252
-------
Leaching Simulations
Leaching is the process whereby certain constituents of the ash become
dissolved in water as it percolates downward through the ash pile. This
process can be simulated in the laboratory—at a great expense of
time—by packing columns with the ash and then allowing water to pass
through them. A quicker way, and one which has been shown to produce
comparable results (7), is to mix the ash with water at a high solid:
liquid ratio. To find out what is happening at a given level in the ash
pile over time, "batches" of water representing a certain pore volume
can be added to the ash and then extracted. By knowing the flow velocity
through the ash under field conditions, the equivalent time of penetration
for each batch of water can be calculated. A portion of each extract
can then also be used to challenge another type (layer) of ash or some
other material, such as a site soil. This new extract would show how
the leachate changes as it moves through the pile. The overall simulation
method just described is referred to as serial batchwise extraction.
The specific technique followed in the present study and the associated
equation for calculating equivalent time of penetration (leaching) are
discussed in detail by Houle and Long (7).
Using this method three scenarios representing three different field
conditions were simulated in the laboratory. The first involved a
combination of Greenwich and Benjamin ashes from the outlet hoppers
("worst case" combination) overlying a typical site soil. The second
involved the same combination of ashes overlying that combination again,
but after partial leaching, overlying bottom ash (coarse boiler slag).
The partially leached material was intended to represent ash which had
been exposed for a time on the pile (aged ash). The bottom ash was
being evaluated as a possible drainage blanket material. The third
scenario involved a combination of Benjamin ashes from across the hopper
series ("average" combination) overlying the aged Greenwich-Benjamin
ash combination overlying bottom ash again. These three scenarios are
shown schematically in Figure 3. The general extraction sequence is
shown in Figure 4. The results of the analyses performed on the extracts
are presented in Tables 4 through 6.
The site soil was also subjected to the extraction procedure to determine
what constituents it might contribute to the leachate. The extraction
sequence which was followed is the same as that shown in the first
column of Figure 4. The analytical results are presented in Table 7.
Figure 5 illustrates how the data from the scenarios can be interpreted.
Similar graphs can be prepared for all of the analysis parameters. What
this one example from Scenario 1 shows is that very little arsenic is
dissolved from the ash initially, apparently due to the greater solubility
of other constituents, such as sulfate, which actually control the
solubility of the arsenic. Selenium and chromium follow the same general
253
-------
SCENARIO I
'HORST-CASE'
COMBINATION OF
FRESH GREENNICH
A BENJAMIN ASHES
3 SUCCESSIVE
BATCHES OF
'SITE SOIL'
SCENARIO II
'HORST-CASE'
COMBINATION OF
FRESH GREENNICH
A BENJAMIN ASHES
'HORST-CASE'
COMBINATION OF
AGED GREENWICH A
BENJAMIN ASHES
BOTTOM ASH
DRAINAGE
BLANKET
SCENARIO III
'AVERAGE' COMBINATION
OF FRESH
BENJAMIN ASHES
BOTTOM ASH
DRAINAGE
BLANKET
I
FIGURE 3
SCENARIO SCHEMATIC
-------
FILTRATE
N5
U>
Ol
HATER
HATER
HATER
ASH
i
FILTRATE
SOIL
BATCH I
1
S-A1-1
FII TRATF
FILTER
CAKE
FILTRATE
FILTRATE
SOIL
BATCH II
1
S-B1-1
FILTRATF '
FILTER
CAKE
1
S-A1-2
FTI TRATF
FILTER
CAKE
FILTRATE
1
FILTRATE
»•
FILTRATE
SOIL
3ATCH III
1
S-C1-1
FII TRATF
FILTER
CAKE
\
S-B1-2
FILTRATF
FILTER
CAKE
FILTRATE
1
FILTRATE
FILTER
CAKE
1
S-C1-2
FII TRATE
FILTER
CAKE
FILTRATE
i
FILTER
CAKE
FILTRATE
1
S-01-1
FILTRATE
1
S-01-2
FILTRATE
1
S-A1-3
S-A1-4
S-A1-5
S-A1-6
S-A1-7
S-A1-8
S-B1-3
ETC.
S-C1-3
ETC.
S-01-3
ETC.
FIGURE 4
GENERAL EXTRACTION SEQUENCE
-------
to
Ln
TABLE 4
SCENARIO 1
EXTRACTABLE CONSTITUENTS (ug/g)
Sample Nuuber
New Fly Ash
S1-A1
S1-A2
SI -A3
S1-A4
S1-A5
S1-A6
S1-A7
S1-A8
Sice Soil Batch 1
S1-B1
S1-B2
S1-B3
S1-B4
S1-B5
S1-B6
S1-B7
S1-B8
Site Soil Batch 11
01 JM
3X^^*1.
S1-C2
S1-C3
S1-C4
S1-C5
S1-C6
S1-C7
S1-C8
Site Soil Batch III
Sl-01
S1-D2
S1-D3
S1-D4
S1-D5
S1-D6
S1-D7
S1-D8
As
0.048
0.67
1.29
3.5
22.5
24.0
40.1
22.7
nd
nd
0.072
0.17
2.04
29.8
35.5
24.0
On&
. UH
nd
0.03
0.04
0.38
9.8
25.3
24.2
nd
—
nd
nd
0.17
—
—
~
Se
0.68
0.5
0.70
2.58
1.92
nd
0.29
nd
nd
0.054
0.06
0.12
nd
1.44
nd
(7.7)
n 09?
U • \j£.£.
0.03
nd
nd
—
—
—
nd
—
nd
nd
—
—
—
—
Sb
nd
nd
nd
nd
—
—
—
—
nd
nd
—
nd
—
—
—
—
nd
—
nd
—
—
—
—
nd
—
—
nd
—
—
—
—
Mo
46.6
12.0
3.42
1.8
nd
nd
nd
nd
6.84
9.3
9.0
7.2
20.2
(67.2)
nd
(67.2)
4.8
17.1
33.6
—
—
—
—
nd
—
15.0
52.8
12.0
—
—
—
B
5.0
1.35
nd
4.8
nd
nd
nd
(18. 4)
4.2
1.05
2.7
3.6
nd
—
—
nd
1.86
nd
5.4
nd
—
—
—
1.16
—
2.4
nd
nd
—
—
—
F
1.62
0.6
1.62
10.9
13.0
5.76
4.8
5.76
1.5
1.23
4.08
17.5
—
—
—
—
1.86
3.78
12.1
31.4
—
__
—
1.4
—
4.08
9.36
31.9
—
—
—
AI
48.6
30.0
49.0
250.0
—
—
—
nd
nd
—
nd
—
—
—
—
nd
nd
—
—
—
—
—
nd
—
—
nd
—
—
—
~
Hn
nd
nd
nd
nd
nd
nd
nd
nd
0.68
0.57
0.54
nd
—
—
—
—
OCA
. 7*4
0.78
nd
nd
nd
—
--
—
0.54
—
nd
nd
nd
—
—
~
Fe
nd
0.42
0.72
1.32
__
—
—
—
nd
nd
—
nd
—
—
—
—
A T
u . /
nd
nd
—
—
—
—
—
nd
—
—
nd
—
—
—
~
Ca
500
372
431
598
955
979
1085
(8448)
504
348
597
521
1764
1334
(6144)
(15360)
1296
756
1019
—
—
—
—
328
—
1512
2052
2098
—
—
—
Total
Cr
0.5
nd
nd
nd
nd
nd
nd
nd
1.48
0.84
nd
nd
nd
nd
nd
nd
\ 7ft
J . £o
1.74
nd
nd
—
—
—
—
2.82
—
2.46
nd
nd
nd
—
—
Zn
0.14
0.24
0.18
0.24
0.48
nd
nd
nd
0.14
nd
nd
nd
—
—
—
—
n 7R
t > . <1O
0.21
nd
1.8
—
—
—
—
0.18
—
0.3
nd
nd
—
—
—
Ni
nd
nd
nd
nd
—
—
nd
—
nd
—
—
nd
—
—
—
—
r»H
nQ
nd
nd
nd
—
—
—
—
nd
—
—
nd
—
—
—
—
S°4
3550
4530
5160
1596
864
96
96
192
3150
5100
5400
1956
—
—
—
—
4305
6300
—
—
—
—
—
3150
—
5550
—
—
—
—
—
pH
10.3
10.2
10.0
10.5
9.8
9.2
8.9
8.1
7.5
7.5
7.8
8.2
8.4
8.1
7.8
7.5
7 J
1 * £.
7.2
7.2
7.3
7.4
7.5
7.5
7.4
7.3
7.2
7.2
7.1
7.3
7.3
7.3
7.4
IDS
5620
7350
8916
4788
3038
2702
3302
3821
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
4.6
—
—
—
—
—
—
Equiva lent
Days of
Penetration
1.8
4.6
10.1
21.2
43.2
87.4
175.7
352.4
1.22 Yr
3.04
6.7
13.9
28.6
57.8
116.2
233
1 7"? YT
-L i Z Z if
3.04
6.7
13.9
28.6
57.8
116.2
233
1.22 Yr
3.04
6.7
13.9
28.6
57.8
116.2
233
NOTES:
Scenario 1 involves a combination of fresh Greenwich and Benjs
"nd" Indicates none detected.
"O" indicate a possible "outlier" value.
tin fly ashes from hoppers A-l and B-l In direct contact with site .soil.
-------
TABLE 5
SCENARIO 2
EXTRACTABLE CONSTITUENTS (ug/g)
N>
Sample
Number
New Fly Ash
S2-A1
S2-A2
S2-A3
S2-A4
S2-A5
S2-A6
S2-A7
S2-A8
Aged Fly Ash
S2-B1
S2-B2
S2-B3
S2-B4
S2-B5
S2-B6
S2-B7
S2-B8
Bottom Ash
S2-C1
S2-C2
S2-C3
S2-C4
S2-C5
S2-C6
S2-C7
S2-C8
As
0.072
0.24
0.70
3.34
8.62
23.7
39.8
30.3
0.68 '
0.73
1.07
3.66
11.0
36.6
73.2
77.6
nd
nd
0.23
0.17
7.27
33.4
73.2
86.0
Se
0.58
0.54
0.60
nd
1.68
1.92
0.96
nd
0.14
0.03
0.36
1.56
2.88
2.88
3.84
nd
nd
nd
nd
nd
0.65
1.44
3.84
nd
Sb
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
Mo
39.
9.54
3.78
1.8
2.64
nd
nd
15.36
39.
8.34
3.3
2.76
(28.8)
3.84
7.68
15.36
0.76
0.36
6.27
11.4
3.6
3.84
nd
nd
B
4.32
nd
nd
nd
nd
nd
nd
nd
3.42
nd
—
nd
nd
nd
nd
nd
4.24
4.53
2.64
nd
nd
(21.12)
nd
nd
F
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
nd
0.93
1.20
11.88
13.68
17.28
15.36
11.52
Al
43.6
72.
95.6
99.7
—
—
—
—
9.8
24.0
72.
168.
—
—
—
—
nd
9.8
12.8
60.2
—
—
—
—
Mn
nd
nd
nd
nd
—
—
—
—
nd
nd
nd
nd
—
—
—
—
14.
1.
0.
nd
—
—
—
—
Ca
308
732
1278
594
725
266
(4512)
1674
600
426
1734
525
1397
2150
3936
(10330)
0 36
11 1116
48 1908
802
1488
3034
(6192)
(12000)
SO^
1720
5370
5040
606
—
—
—
—
3280
4440
5760
1392
—
—
—
—
5024
5160
7680
2076
—
—
—
—
pH
9.9
10.2
10.0
9.8
9.6
9.3
7.9
7.1
9.2
9.3
9.2
10.0
9.8
8.9
8.6
8.4
3.0
4.6
7.8
8.7
9.0
8.6
8.5
7.3
TDS
5260
7350
8220
1740
2388
2981
2707
(10330)
—
—
8880
3492
3936
3989
5741
5779
—
—
8340
3420
3888
3744
5453
6278
Equivalent
Days of
Penetration
1.8
4.6
10.1
21.2
43.2
87.4
175.7
352.4
1.8
4.6
10.1
21.2
43.2
87.4
175.7
352.4
0.005
0.013
0.03
0.06
0.12
0.25
0.5
1.0
NOTES :
Scenario 2 involves a combination of fresh Greenwich and Benjamin fly ashes from hoppers A-l and B-l challenging a combination
of "aged" Greenwich and Benjamin fly .ashes and bottom ash.
"nd" indicates none detected.
"()" indicate a possible "outlier" value.
-------
TABLE 6
SCENARIO 3
EXTRACTABLE CONSTITUENTS (ug/g)
to
m
00
Sample
Number
New Fly Ash
S3-A1
S3-A2
S3-A3
S3-A4
S3-A5
S3-A6
S3-A7
S3-A8
Aged Fly Ash
S3-B1
S3-B2
S3-B3
S3-B4
S3-B5
S3-B6
S3-B7
S3-B8
Bottom Ash
S3-C1
S3-C2
S3-C3
S3-C4
S3-C5
S3-C6
S3-C7
S3-C8
As
0.26
0.48
0.78
5.52
18.0
26.9
127.
21.1
0.86
1.26
1.56
3.84
16.8
49
57.6
3.84
0.06
0.01
0.06
0.24
3.6
39.8
70.0
57.6
Se
0.58
0.48
0.69
1.72
0.34
0.16
0.10
nd
0.15
0.08
0.91
0.95
2.42
0.10
0.23
0.33
nd
nd
0.02
0.01
0.89
0.72
1.44
1.11
Mo
28.0
4.8
1.14
nd
—
—
—
—
20.6
7.7
2.1
nd
—
—
—
—
1.54
nd
3.5
8.5
—
—
—
—
B
5.84
nd
—
—
—
—
—
—
4.22
3.6
rd
nd
--
--
—
—
4.32
3.18
nd
nd
—
—
—
—
F
7.50
3.45
5.46
6.72
4.80
3.36
4.80
5.76
0.68
0.90
4.50
10.56
—
—
—
—
4.0
1.02
2.88
16.20
—
—
—
—
Mn
nd
—
—
—
—
nd
(3.84)
nd
nd
—
—
—
—
—
—
—
15.62
1.08
nd
nd
—
—
—
—
Total
Cr
0.18
nd
nd
nd
nd
nd
nd
nd
0.72
0.39
nd
nd
nd
nd
nd
nd
1.14
nd
nd
nd
nd
nd
nd
nd
Ni
nd
nd
nd
nd
nd
nd
(13.4)
nd
nd
—
—
—
—
—
—
—
162.5
22.6
0.66
nd
—
—
—
—
S°4
3820
4620
1170
415
214
86.4
355
384
—
—
2340
900
797
749
643
288
—
—
3870
1158
1104
739
739
384
PH
9.7
10.1
10.1
9.9
9.2
8.8
7.8
7.2
9.6
9.1
9.4
10.2
9.8
9.3
90
S.9
3.1
3.7
8.2
8.7
9.0
8.9
8.5
7.6
TDS
5798
6519
2550
1632
1226
1094
3168
6624
—
—
3798
2808
3120
3811
4848
6624
—
—
5172
2532
2784
3758
3514
6451
Equivalent
Days of
Penetration
1.8
4.6
10.1
21.2
43.2
87.4
175.7
352.4
1.8
4.6
10.1
21.2
43.2
87.4
175.7
352.4
0.005
0.013
0.03
0.06
0.12
0.25
0.50
1.0
NOTES:
Scenario 3 involves a combination of fresh Benjamin fly ashes from hoppers A-4, A-8, A-9, A-12, A-16, B-8, B-9, B-12, and B-16
challenging a combination of "aged" Greenwich and Benjamin fly ashes and bottom ash (same as in Scenario 2).
"nd" indicates non detected.
"()" indicates a possible "outlier" value.
-------
TABLE 7
SITE SOIL EXTRACTIONS
EXTRACTABLE CONSTITUENTS (ug/g)
N5
Ul
VO
Sample Total Total
Number As Se Sb Mo B F Al Mn Fe Ca Cr Zn Ni C SO^ pH TDS
SX-1 0.012 0.008 nd 0.44 1.26 1.6 nd 0.2 nd 156. 1.48 (1.12 nd 30 670 7.1 1402
SX-2 0.018 nd nd 1.32 nd 3.75 nd nd nd 52.8 1.23 nd nd 39 153 7.4 486
SX-3 0.072 nd nd 3.48 nd 7.32 10.2 nd 1.62 232. 0.78 0.36 nd 42 60 7.6 2292
SX-4 0.108 .nd nd 8.4 5.4 11.3 nd nd nd 30.6 nd nd nd 72 12 7.7 2200
NOTES :
Equivalent
Days of
Penetration
1.22
3.04
6.7
13.9
SX-1 indicates the first site soil extraction, etc.
"nd" indicates none detected.
-------
ON
O
ASH
5^120
no
100
EQUIVALENT DAYS PENETRATION
THROUGH ASH
FIGURE 5
INTERPRETATION OF SCENARIO 1
ARSENIC DATA
-------
release pattern. After about three simulated months of leaching the
concentrations of the metal in solution increase dramatically as the
more soluble constituents are removed from the system. This increase
continues for a while and then the concentrations begin to slowly decline.
In the soil the arsenic is retained at first, but then it passes through
without appreciable attenuation.
Another interesting aspect of the results is presented by the pH data.
In the variability tests the pH values ranged between 4 and 11 for the
Benjamin and Greenwich ashes, respectively, while in the batchwise
extractions they never got below 7 after one simulated year of leaching.
This shows the value of not relying exclusively on a simple "shaker"
type of test to determine the leaching characteristics of a solid waste.
Since the serial batchwise extractions were carried out under essentially
equilibrium conditions (the reaction vessel was stirred until the time
rate of change of pH and conductivity reached zero) and at a constant
solid:liquid ratio, a question remained as to what the very first leachate
generated at the moving front of the wetting mass would look like. To
answer this question a separate series of extractions (wetting front
tests) was performed on the worst-case Greenwich and Benjamin ash combin-
ation and on the average Benjamin combination. Basically, what was done
was to add fresh ash to each successive extract to determine maximum
solubilities of the various constituents of interest; however, mixing
times did not exceed 30 minutes, compared to two hours for the batchwise
extractions. The extraction sequence is shown by the top row of Figure
4 and the analytical results are given in Table 8. Generally, this data
served to verify the concentration ranges already established in the
variability tests and scenarios, except that the pH values tended to be
very low (in the range of about 3 to 4). This was probably due to the
short mixing times involved, which apparently favor the release of early
acid-producing constituents from the ash.
Using these ranges, which are summarized in Table 9, it was then possible
to determine the concentrations of certain pollutants likely to appear
in the leachate at the bottom of the landfill. This was done by assuming
that 2-5% of all the precipitation which falls on the ash will infiltrate
it (these are laboratory values, although they have also been verified
by actual field experience). By knowing the volume of water which
infiltrates the ash pile (about 20,000 gal/yr per 5-acre segment based
on an annual precipitation for the site of about 43 inches and a 2%
infiltration rate) and the pounds of extractable constituents per ton of
ash, the concentrations of those constituents in the leachate can then
be calculated for different stages of development of the landfill. The
results of these calculations are given in Table 10 for arsenic, selenium,
and sulfate, which when compared to the EPA's primary (health based)
261
-------
TABLE 8
WETTING FRONT TEST //I
EXTRACTABLE CONSTITUENTS (ug/g)
Sample
Number
Wl-1
Wl-2
Wl-
3
As
0.74
1.22
2.44
Se
nd
nd
nd
*
Mo
0.16
0.34
0.7
B F
13.2 1.22
25.2 0.2
39.2 0.18
Al
39.4
252.
398.
Mn
5
11
18
.08
.9
.4
Ca
408
564
298
TDS
6646
9348
13372
NOTES:
This test involved a combination of fresh Greenwich and Benjamin fly ashes
from hoppers A-l and B-l.
"nd" indicates none detected.
WETTING FRONT TEST #2
EXTRACTABLE CONSTITUENTS (ug/g)
Sample Number
W2-1
W2-2
W2-3
W2-4
As
0.18
0.8
0.76
0.5
Mo
9.64
0.34
0.86
2.9
B
7.82
14.96
25.0
27.2
F
1.26
0.2
9.1
0.72
Mn
0.74
4.34
5.86
5.96
TDS
5766
6398
7408
7946
NOTE:
This
test involved a combination of fresh Benjamin fly ash from hoppers A-4
V-8, A-9, A-12, A-16, B-8, B-9, B-12 and B-16.
262
-------
TABLE 9
AVERAGE AND MAXIMUM CONCENTRATIONS OF CONSTITUENTS APPEARING AT
THE BOTTOM OF INDIVIDUAL LAYERS IN SCENARIOS 1, 2, AND 3 (ug/g)
Scenario 1
Scenario 2
Scenario 3
Parameter
As
Se
Sb
Mo
B
F
Al
Mn
Fe
Ca
Cr (Total)
Zn
Ni
SO,
TDS
Concentration
Avg.
Max .
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Avg.
Max.
Ash
14.4
40.1
1.11
2.58
nd
16.0
46.6
3.72
4.8
5.51
13.0
94.4
250.
nd
0.82
1.32
703.
1085.
0.5
0.5
0.26
0.48
nd
2011.
5160.
4942.
8916.
Soil
0.04
0.17
nd
nd
19.9
52.8
0.89
2.4
11.7
31.9
nd
0.13
0.54
nd
1498.
2098.
1.06
2.82
0.12
0.3
nd
4350.
5550.
na
Fly Ash
25.6
77.6
1.67
3.84
nd
11.5
15.4
3.42
3.42
nd
68.5
168.
nd
na
1538.
3936.
na
na
na
3718.
5760.
5303.
8880.
Bottom Ash
25.0
86.0
0.74
3.84
nd
3.28
11.4
1.63
4.53
10.3
17.28
20.7
60.24
3.9
14.0
na
1397.
3034.
na
na
na
4984.
7680.
5187.
8340.
Fly Ash
16.8
57.6
0.05
2.42
na
10.1
20.6
3.91
4.22
4.16
10.56
na
nd
na
na
0.56
0.72
na
nd
853.
2340.
4168.
6624.
Bottom Ash
21.4
70.0
0.52
1.44
na
3.4
8.5
1.9
4.32
6.02
16.2
na
4.2
15.62
na
na
0.14
1.14
na
23.2
162.5
1332.
3870.
4035.
6451.
NOTES;
"Outliers" from Tables 4 through 6 were not used in calculations.
"nd" indicates none detected.
"na" indicates no analyses.
263
-------
TABLE 10
CALCULATED CONCENTRATIONS OF SELECTED
POLLUTANTS IN LEACHATE
Pollutant
Arsenic
Selenium
Sulfate
Maximum
Concentration
(mg/L)
0.83
0.04
1600
Average
Concentration
(mg/L)
0.25
0.01
135
EPA Drinking
Water Standard
Cmg/L)
0.05
0.01
250
NOTE:
Values are based on an average annual precipitation of 43 inches and 2%
infiltration rate into the ash pile.
264
-------
drinking water standards appear to represent the .greatest potential
problem. Interestingly, the Scenario 1 data also indicates that these
same constituents are not likely to be significantly attenuated by the
site soil, which suggests the need to intercept the leachate for possible
treatment or dilution with the other waste streams emanating from the
landfill, such as runoff.
265
-------
Runoff Simulations and Modeling
Stormwater runoff from the proposed landfill was simulated in the labora-
tory by using a modification of a technique developed by IU Conversion
Systems, Inc. (8). The purpose of the simulations was to provide input
to the computer model on runoff quality and flow. The IUC technique
basically involves packing a metal box with the ash to approximate field
conditions, preparing the surface (rough, smooth, baked with a heat lamp
to simulate prolonged exposure to intense sunlight, and so on), and then
spraying reagent grade water over the surface. The resulting runoff is
caught and recirculated for one hour. By using a constant flow rate
over a known surface area, each simulation represents a one hour rainfall
of a certain intensity running off a unit area of ash. The result is a
family curves relating inches of rain to linear feet of runoff for
different application rates (storm intensities). These curves are
presented in Figure 6.
An average Greenwich and average Benjamin ash (equal-volume grab samples
from the inlet and outlet hoppers) were used for the simulations. pH
and conductivity were monitored continously in the recirculated runoff.
The average pH of all the runoff water for both ash types was 6.0,
although the Greenwich runoff had an average pH of about 6.1 and the
Benjamin Runoff an average pH of 5.8. This contrasts with the drastically
different pH values found for the two ash types in the variability
tests. During each simulation, the initial runoff pH was about 5.5 and
then slowly increased to neutral or slightly above. Conductivity followed
the same pattern, starting at about 40 umhos and ending at about 800.
The results of the analyses performed on the final runoff for the other
parameters of interest are given in Table 11. Generally, they are
consistent with the results from the variability tests and so do not
deserve further discussion here.
The computer model selected for the second phase of the runoff analysis
is a modification of EPA's Short Stormwater Management Model (SSWMM).
The modification was necessary to allow consideration of the pollutants
in the runoff, rather than just the runoff flow. The model addresses
washoff, surface detention, depression storage, and infiltration, but
does not look at how this infiltration might change the runoff character-
istics. Instead, it focuses on pollutant transport by overland flow and
channel routing and the exponential decay of pollutants through washoff.
While the runoff simulator was exposed to a constant rate and volume of
precipitation, the model processed storms with varying intensities and
total precipitation amounts. Five storm events representing a broad
range of storm types were selected for analysis. These storms are
characterized in Table 12. Total precipitation values for the area were
obtained from U.S. Weather Bureau data (9). Published statistical
hyetographs for 10% and 50% probability storms were used to show how
266
-------
N)
C^
--J
100
150
200
250
300
350
400
LINEftR FEET RUNOFF
6 INCHES HIDE
FIGURE 6
SIMULATED FLOH-RUNOFF
-------
TABLE 11
SELECTED RUNOFF SIMULATOR TEST RESULTS
Sample
Description
Source Hopper No.
Bearing Cap.
Thickness of
% Moisture
(tons/ft2)
Ash (in.)
Avg. Test Flow (gph)
Vol. of Water
Analyses:
Avg. pH
Avg. Cond.
TSS
TDS
Sett. Solids
to As
S Al
Ca
Fe
Cr
Ni
Se
Zn
B
F
Sb
Mo
Mn
so,
Retained (ml)
umho
mg/1
rag/1
mg/1
rag/1
rag/1
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
rag/1
mg/li.
mg/1
mg/i
mg/i
mg/1
Rough Compact
Benj amin
A1B1A8B8
1.25
6
20
11.9
1400
—
—
6564.
714.
—
.085 *
nd *
66.5 *
nd *
—
nd *
—
nc
nd *
—
0.08 *
—
0.75 *
0.22 *
nd *
0.91
0.05 *
493. *
Sun Dried
Benjamin
VlA8B8
1.41
6
20
15.2
1500
—
—
47864.
1076.
—
.100 *
nd *
58.5 *
nd *
—
nd *
—
nd
0.03
—
0.03 *
—
nd *
0.43 *
nd *
1.10 *
0.08 *
718. *
Compacted
Benjamin
AlBlV58
1.78
6
20
14.9
0
5.8
420
37568.
799.
38.38
6.19
399.
204.
nd *
114.
nd *
1.09
—
0.052*
0.14
0.04 *
2.34
questionable data
1.0
—
0.430
—
452.
Compacted
Greenwich
A1B1B8
3.69
4
20
16.7
100
6.0
458
40662.
1060.
78.78
0.66
36.2
475.
nd *
0.41
nd *
0.26
—
0 . 006
nd *
0.98
—
15.4
0.26
—
0.196
—
605.
Rough Compact
Greenwich
A1B1B8
1.25
4
20
17.8
0
6.6
240
24165.
556.
32.91
3.34
70.0
207.
—
12.1
—
0.80
—
0.019
—
0.65
—
19.6
0.27
—
0.313
—
252.
NOTES:
"*" indicates dissolved; otherwise, total.
"nd" indicates none detected.
-------
Storm Event
1-Year/30-Minute
1-Year/2-Hour
1-Year/6-Hour
l-Year/24-Hours
10-Year/24-Hours
TABLE 12
CHARACTERISTICS OF MODELED STORM EVENTS
Description
Summer thunder shower
Series of summer thunder showers
Heavy rain preceding frontal
passage (spring, fall)
Moderate, steady rain
Heavy, steady rain
Total Rainfall (in.)
0.8
1.3
1.8
2.3
4.5
269
-------
rainfall intensity varies with time for short- and long-duration storm
events, respectively.
The modified SSWMM model evaluated 13 pollutants; however, it could not
be calibrated independently for each. As a result, only two pollutants,
arsenic and selenium, were selected as the focus of the calibration.
The model was calibrated to 0.1 milligrams for these two pollutants and
within one order of magnitude for sulfate, although it consistently
predicted total dissolved solids at much lower levels than the measured
values of 700-1400 milligrams per liter.
Figures 7 and 8 show how the concentrations of selected pollutants
change for the runoff from Greenwich ash during a short and long storm
event. The curves for the Benjamin ash are very similar. Also shown is
how the runoff flow changes. Table 13 is a summary of the pollutant
concentrations in the total runoff, assuming all of it is collected.
Interestingly, these concentrations are not significantly different for
the two types of ash and do not vary at all with the storm event; however
the arsenic and selenium levels consistently exceed the EPA's primary
drinking water standards. This again suggests the need for some sort of
treatment or flow combination.
270.
-------
200
i«« 100 +
Ot
LEGEND
• SULFATE
* SELENIUM
• ARSENIC
• TDS
.55
.50
.45
.40
.35
.30 5?
cc
•0
.25 5
.20
.15
.10
.05
.55
.50
.45
.40 t
.35
g.30
uf -25
.20 t
.15
.10
.05 t
TIME (HOURS)
FIGURE 7
MODELED 1-YEAR/30-MINUTE STORM RUNOFF
CHARACTERISTICS FOR GREENWICH ASH
TIME (HOURS)
-------
120-'
LEGEND
4 SULFRTE
* SELENIUM
• ARSENIC
• TDS
14
TIME (HOURS)
FIGURE 8
MODELED 1-YEAR/24-HOUR STORM RUNOFF
CHARACTERISTICS FOR GREENWICH ASH
272
-------
TABLE 13
RESULTS OF MODIFIED 5SWMM MODEL FOR A 5-ACRE ASH AREA
Storm Event
Ash Runoff Volume
(sal)
As
1-year/ 30-minute
1 -year /2-hour
l-year/6-hour
1-year/ 24-hour
10-year/ 24-hour
Benjamin
Greenwich
Benjamin
Greenwich
Ben j amin
Greenwich
Benjamin
Greenwich
Benj amin
Greenwich
11,848 0.
0.
24,235 0.
0.
36,188 0.
0.
32,418 0.
0.
47,176 0.
0.
02
01
04
03
05
04
03
02
07
05
Total Pollutant Load
Washed Off Ash Pile
(Ibs)
Se
0.01
0.01
0.02
0.02
0.03
0.03
0.02
0.02
0.04
0.04
S°4
9
11
18
22
27
33
15
19
34
41
.1
.0
.4
.3
.5
.3
.9
.2
.1
.3
TDS
4
3
8
7
12
11
7
6
15
14
.2
.9
.5
.9
.7
.8
.3
.8
.8
.6
Pollutant Concentration in
Hypothetical Collection Pond
(rng/1)
As
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
18
14
18
14
19
14
18
13
18
13
Se
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
10
10
11
10
11
10
10
10
10
09
SO, TDS
4
90.8 42.0
109.8 38.9
90.6 41.9
109.5 38.8
91.1 42.1
110.2 39.1
87.1 40.2
105.3 37.3
86.4 40.0
104.4 37.0
EPA Drinking Water Standard
0.05 0.01 250. 500.
-------
Investigation of Site Geohydrology
The 170-acre site of the planned fly ash landfill is located about one-
half mile from the generating station (Figure 1). It is relatively
flat-lying, with the maximum relief being only about 55 feet. Several
intermittent streams cut across the site, although the majority of them
appear to be diversion channels constructed for the purpose of drainage
control. Prior to its being purchased by PP&L, the land was used for
the farming of crops.
Geohydrological conditions at the site were investigated by means of
test borings, test pits, hand auger holes, monitoring wells, and field
examinations. What emerged is a picture of a fairly simple ground water
aquifer system. This system is depicted in the generalized geologic
section of Figure 9. Basically, the majority of ground water flow is
confined to a rather narrow zone represented predominantly by a weathered
and highly fractured shale. Immediately above this zone is a uniform
silty clay of extremely low permeability, which acts as a cap on the
ground water aquifer. Below it is a very tight (few open fractures)
unweathered shale with nearly horizontal bedding, which acts as a bottom
confining layer on the aquifer.
The hydraulic gradient over the site is about 1:120 and the rate of
ground water flow is probably on the order of only 0.05 feet per year
(based on a calculation using an assumed value for the permeability of
the fractured shale). The water table occurs at a depth of 5 to 10 feet
and its configuration generally conforms to the site topography. Unlike
the drainage ditches, which are probably perched on top of the clayey
soil, the one large intermittent stream which traverses the site appears
to be connected to the ground water system, since the ground water
contours indicate a depression in that area. This is worth knowing,
because the stream is located along the down-gradient border of the site
and could therefore possibly be used as a hydraulic barrier against the
migration of any pollutants which might enter the ground water system
upstream.
Local wells in the area range in depth from about 30 feet for those that
were hand dug to nearly 300 feet for those that were drilled. The water
is used predominantly for human consumption and livestock watering. The
quality of the water is generally acceptable for these purposes, except
that it contains unusually high concentrations of sulfur and iron due to
the presence of iron sulfide minerals in the local bedrock. Thus, any
pollutant introduced into the ground water might have an impact which
should be considered in the overall design concept for the landfill.
274.
-------
FT.
0
10
15
20
25
30
35
U.QLJ
oz>
uo
LUOZ
za:
OC3
.........
r .— •-—.— •-r.7^. -r.-.T- -.r
.p*f.'.1T.'.t».'.VRr. I IL . j I r.'ffT.'.rr/...
''''''''
TO DEPTH
MOTTLED SILTY CLAYj NO STRUCTURE
SILTY CLAY SAPROLITEi SOME STRUCTURE
WEATHERED SHflLEj ABUNDANT OPEN,
STEEPLY DIPPING JOINTS* "RUSTING'
"TIGHT", UNWEATHERED MARCELLUS
SHALEj BEDDING NEARLY HORIZONTAL
FIGURE 9
GENERALIZED GEOLOGIC SECTION
2/5
-------
THE LANDFILL DESIGN CONCEPT
AND PLAN OF OPERATIONS
The laboratory simulations and modeling and the geohydrological investi-
gation provided valuable input into determining how the fly ash landfill
should be designed and operated.
First, the leaching simulations showed that certain pollutants, such as
arsenic, selenium, and possibly sulfate, would likely appear in the
leachate in concentrations above the EPA's drinking water standards for
protecting human health. They also showed that these same pollutants
could not be expected to be attenuated to any great extent by the site
soil. From the geohydrological investigation it was found that attenua-
tion by dilution could not be relied on either because of the slow rate
of ground water movement and the confined nature and limited vertical
extent of the ground water aquifer system.
Thus, it was decided to collect all of the leachate for possible treatment
or combination with other waste streams, either on site or back at the
generating station. What was especially appealing about this scheme was
that no special liner would be required, since the natural clays at the
site could serve this purpose with the need for only simple in-place
reworking to remove any remnant structures (open fractures) which might
still be present from weathering of the parent rock material. Physical
tests performed on the station's bottom ash also showed it to be quite
suitable as a drainage blanket material, except that a filter fabric of
some sort would have to be used on top of the blanket to prevent its
becoming clogged with the finer fly ash particles.
The results from the runoff simulations showed that, despite the storm
event involved, the collected runoff from the operating portions of the
landfill would again contain concentrations of arsenic and selenium in
excess of the drinking water standards. While it was uncertain what
discharge limitations might be placed on the outfall pipe from a collec-
tion facility, the available receiving streams in the area are so small
that no in-stream dilution could be relied upon to assure compliance
with water quality criteria. After all, it was the violation of these
criteria that necessitated the switch from a wet to a dry ash handling
system in the first place. To reduce the volume of contaminated runoff
likely to be produced by the landfill, the size of the active portion
could be reduced. This has the added benefit of also reducing the
potential fugitive dusting problem, as well as the amount of leachate.
A working cell size of 5 acres was finally selected as representing the
best compromise between these concerns and operating efficiency.
The problem of what to do with the leachate and "dirty" runoff (as
compared to the "clean" runoff from unworked or revegetated portions of
the landfill) is still being addressed. One alternative, which has
received additional study effort, involves combining the two waste
276
-------
streams with enough clean runoff to lower the pollutant concentrations
in the final discharge to an acceptable level. This scheme has a great
deal of appeal because it would eliminate the need to install a sophis-
ticated metals treatment system based exclusively on laboratory generated
data. A drawback of the scheme is that it would require a fairly large
collection pond. Calculations show, for instance, that it would take
about 50 acres of clean runoff to adequately lower the pollutant levels
in the leachate (assuming full site development and 5% infiltration of
the precipitation falling on the landfill). The problem with constructing
such a large basin is that the thinness of the site soil and the low
relief at the site may limit the capacity which can be achieved by
excavation or diking while still allowing gravity flow into the basin.
An alternative is to pump the leachate and dirty runoff back to the
generating station for combination in the waste detention pond, which
also has some pH control capability. A disadvantage of this scheme is
that it would require a very reliable pumping system; and should that
system fail, the collected wastes would have to be discharged with
essentially no reduction in pollutant concentrations. This may not be
acceptable under the EPA's developing Best Management Practices require-
ments. Also, there is the concern that the chemistry of the leachate
and runoff may not be compatible with the other wastes in the detention
pond and may create a new problem at that location, even with adequate
pH control.
As a means of verifying the laboratory data, another step being considered
by the company is to prepare and monitor a test plot at the landfill
site. Based on this data it should be possible to make a more definite
decision about the need for treatment and the advisability of the various
flow combination schemes being considered. The problem, of course, is
which ash samples and which arrangement of the different ash types will
produce the most representative results—one of the same limitations of
the laboratory effort.
277
-------
BIBLIOGRAPHY
1. Furr, A.K., et al., "National Survey of Elements and Radioactivity
in Fly Ashes," Env. Sci. & Tech., Vol. 11, No. 13, 1194-1201 (1977).
2. Coles, D.G., et al., "Chemical Studies of Stack Fly Ash from a
Coal-Fired Power Plant," Env. Sci. & Tech., Vol. 13, No. 4, 455-459
(1979).
3. Smith, R.D., et al., "Concentration Dependence upon Particle Size
of Volatilized Elements in Fly Ash," Env. Sci. & Tech., Vol. 13,
No. 5, 553-558 (1979).
4. Ondov, J.M., et al., "Emissions and Particle Size Distribution of
Minor and Trace Elements at Two Western Coal-Fired Power Plants
Equipped with Cold-Side Electrostatic Precipitators." Env. Sci. &
Tech., Vol. 13, No. 8, 946-953 (1979).
5. Ondov. J.M., et al., "Elemental Emissions from a Coal-Fired Power
Plant, A Comparison of a Venturi Wet Scrubber System with a Cold-
Side Electrostatic Precipitator, Env. Sci. & Tech., Vol. 13, No. 5
598-607 (1979).
6. Theis, T.L., and Wirth, J.L., "Sorptive Behaviour of Trace Metals
on Fly Ash in Aqueous Systems," Env. Sci. & Tech., Vol. 11, No. 12,
1096-1100 (1977).
7. Houle, T.L., and Long, D.E., "Accelerated Testing of Waste Leacha-
bility and Contaminant Movement in Soils," in Land Disposal of
Hazardous Wastes, Proc. of Fourth Annual Symposium, U.S. EPA,
Cincinnati, Ohio, 152-168 (1978).
8. Roberts, B.K., "Simulated Field Leaching Test for Evaluation of
Surface Runoff from Land Disposal of Waste Materials," IU Conversion
Systems, Inc., Report to ASTM D-19 Subcommittee, Nov. 3, 1977.
9. , "Technical Paper No. 40: Rainfall Frequency Atlas of
the United States for Durations from 30 Minutes to 24 Hours and
Return Periods from 1 to 100 Years", U.S. Dept. of Commerce, Weather
Bur., Wash., D.C., 61 p. (1961).
273
-------
MEASUREMENT OF FUGITIVE EMISSIONS
FROM INCO'S COPPER CLIFF SMELTER
PART III: REVERBERATORY FURNACES
A.D. Church
W.J. Middleton
P. Gatha
ABSTRACT
Superintendent, Smelter Process Technology,
Copper Cliff, Ontario.
Section Leader, Smelter Process Technology,
Copper Cliff, Ontario.
Project Leader, Smelter Process Technology,
Copper Cliff, Ontario.
The Copper Cliff Smelter of Inco Metals Company, a major primary nickel
and copper producer, is located in Ontario, Canada. The essential processing
equipment consists of Multi-Hearth Roasters (^30), Reverberatory Furnaces (6),
Oxygen Flash Furnace and Peirce-Smith Converters (18). A comprehensive
program to measure fugitive emissions of sulphur dioxide, sulphuric acid and
particulates from the Smelter is being undertaken. This paper describes the
reverberatory furnace section of this program.
Total fugitive emissions of sulphur dioxide, particulates and sulphuric
acid from the reverb furnaces were measured as 15.2, 0.19, and 0.12 kg/Mg of
bessemer matte equivalent respectively. Nearly 70% of the emissions occur
during matte tapping. Slag skimming and furnace "puffing" contribute 20% and
3% respectively. Total fugitive sulphur dioxide emissions were found to be
equal to those being emitted by all the converters (previously measured).
However, most of the fugitive gas (^90%) was captured at source by hood
systems before release to atmosphere at roof level.
279
-------
INTRODUCTION
A major program of measurement of fugitive emissions originating from
the Copper Cliff Smelter of Inco Metals Company has been undertaken. The
objectives of this program are: (a) identification of sources of fugitive
emissions, (b) determination of sulphur dioxide, particulates and where
possible, sulphuric acid concentrations in these emissions, (c) correlation
of the fugitive emissions to the unit operations and the throughput of the
Smelter and (d) to curtail the fugitive emissions.
Fugitive emissions originating from copper and nickel converters have
been previously determined and reported}'2 This paper describes the findings
pertinent to nickel reverberatory (reverb) furnace activities in the Smelter.
There are at present six nickel reverb furnaces in the Smelter of which
usually three or four are operating at any one time. Four activities
associated with reverb furnace operation produce significant quantities of
fugitive emissions. These are matte tapping, slag skimming, furnace "puffing"
and return of converter slag. Emissions from the last source had been
previously estimated.2 A general layout of a reverb furnace (No. 3) and
associated hoods and vents is shown in Figure 1. Emissions arising during
matte tapping are collected by the hoods above the matte tunnel and the matte
launder, gathered in a fan-forced vent and exhausted to atmosphere via a
"black ship" ventilator at roof elevation (^50m). Similarly, slag skimming
emissions are collected in another vent and released into a roof ventilator
above the furnace and thence to the atmosphere. Emissions due to furnace
"puffing" rise by convection and exit the building via the same roof
ventilator.
Consequently, the measurement program was divided into two phases.
Phase I consisted of the determination of sulphur dioxide, particulates and
sulphuric acid in gases in the vents at the matte tunnel, and matte and slag
launders, while Phase II involved measurement of the same in the roof
ventilator. In the latter case measured emissions represented a combination
of slag skimming and furnace "puffing" activities. Thus emissions due to
slag skimming were measured twice which was considered useful to verify the
results.
The measurement techniques were essentially the same as those employed
in previous work2, with the exception of a method for determining sulphuric
acid. Previous attempts to measure sulphuric acid using an isopropanol
capture system (EPA method 5) were unsuccessful. In the present study,
sulphuric acid serosols were collected together with particulate matter on
the same filter and were subsequently determined by chemical analysis.
280
-------
EXPERIMENTAL
Phase I
This phase, as mentioned earlier, involved measurement of emissions in
the vents from the matte tunnel, the matte launder and the slag launder.
Sampling points were installed at suitable locations in the three vents
and gas velocities were determined with an S-type pitot. Volume flows were
calculated and from velocity profiles, points representing average velocity
were determined.
A sampling train as shown in Figure 2 was assembled. The gas sample was
drawn continuously with a pump through a Balston filter* for particulate and
sulphuric acid collection and was then directed to hydrogen peroxide
solutions in impingers to caputre sulphur dioxide. Typically, sampling was
done continuously for an eight hour period at the end of which the filter was
removed, dried and weighed. Subsequent chemical analysis determined the
sulphuric acid content on the filter. The hydrogen peroxide solution was
titrated with sodium hydroxide (IN) and the quantity of sulphur dioxide was
calculated. Using values of previously determined volume flows in the vents,
total sulphur dioxide, particulate and sulphuric acid emissions for the eight
hour period were determined. These were then expressed in terms of quantity
per ladle of matte tapped or slag skimmed. In all, emissions were monitored
for about 56 hours at the matte tunnel, 16 hours at the matte launder and 48
hours at the slag launder, covering the tapping of some 86 ladles (VI700 Mg)
of matte and the skimming of 83 ladles (-v/2000 Mg) of slag.
Phase II
In this phase, emissions in the roof ventilator above the furnace were
measured.
The roof ventilator consisted ofa5.9mx7.6m section of transite
stack of approximately 26 m in height and starting about 29 m above the
furnace level. The ventilator was of the gravity flow type with non-uniform,
low velocity gas flow profiles (see Figure 1).
The methods used in this phase were the same as those described in the
previous study at nickel converters.2 The sampling and analysis systems are
shown in Figure 3 and 4 respectively. The sampling system was installed in
the ventilator and was connected to the analysis system. The anemometer vane
and thermocouple were fixed at positions representing average gas velocity and
temperature and were connected to recorders.
* Manufactured by Balston Filter Products, Lexington, Massachusetts, U.S.A.
281
-------
Actual monitoring covered 19 shifts (VI50 hours) divided into two
sessions. Two technicians operated the analysis system on M-floor and
one technician was located at the skimmer's platform. The latter recorded
beginning and ending times for each ladle skimmed and transmitted this
information via radio to the technician at the analysis station. Data was
thus recorded on a per "event" basis where an "event" was defined as either
a "skimming event" (representing emissions due to skimming and "puffing") or
a "non-skimming event" (representing "puffing" emissions only).
Data recorded for each "event" included (a) activity, (b) duration of
the activity (5-15 minutes), (c) sample gas volume (50-300 litres), (d)
titrant (0.1N NaOH) volume, (0.1 to 10 ml), (e) gas velocity (4-5 m/s), and
(f) gas temperature (10-20 C).
The staplex filter was changed every shift (eight hour period), dried
overnight in a desiccator and weighed to obtain the combined weight of
particulates and sulphuric acid (50-500 rug). The quantity of the acid was
subsequently determined by chemical analysis. For each filter this data was
recorded together with corresponding sample gas volume and the ventilator gas
velocity and temperature.
Furnace draft was recorded at 15 minute intervals with a viewpoint of
establishing a relationship between the draft and "puffing" emissions.
DATA ANALYSIS
Fugitive Emissions - Return Of Converter Slag
These emissions occur at the converter aisle end of the furnace and have
been calculated in a manner similar to that used in the previous study of
nickel converters.2 It was assumed that the quantity of sulphur dioxide
emitted during this process was the same as that when skimming a ladle of the
furnace slag (12.1 kg/ladle from this study).
Phase I
Total fugitive emissions per ladle for matte tapping activity were
calculated as the sum of the two separate components measured in the matte
tunnel and launder respectively, both of which later combined and exhausted
to the atmosphere. Total rate and average concentration of emissions were
similarly derived. Fugitive sulphur dioxide emissions per ladle for slag
skimming activitiy were also calculated. The latter were used for comparison
with the measurements made in the roof ventilator.
282
-------
Phase II
A computer program was developed to calculate for each shift both total
sulphur dioxide emission and rate and concentration of the emission on a per
"event" basis. A preliminary examination of results thus obtained indicated
that emissions during a non-skimming "event" immediately following (1-3
minutes) a skimming "event" were invariably higher than during other non-
skimming "events". This is shown in Figure 5. A closer examination of a
typical "skimming event" as shown in Figure 6 revealed that a portion of the
emissions actually attributable to the skimming activity was being counted as
non-skimming or "puffing" emission. This carry over of emissions occured
because a "skimming event" was considered terminated as soon as that activity
stopped at the skimmer's platform, while the resultant change in emissions
recorded at the analysis station (some 50 metres above the furnace) was not
instantaneous.
Therefore, while total sulphur dioxide emissions per shift calculated
by the computer were correct, distribution between skimming and "puffing"
activities was erroneous. The following approach was used to calculate
sulphur dioxide emissions from skimming and "puffing" activities:
1. Total S02 (kg) per shift obtained from computer printout - Et
2. Average rate of emission (kg/min) during non-skimming
"events" calculated from computer output, excluding
"events" immediately following a skimming "event" - R
3. Emission due to "puffing" - E (kg) = R x time (minutes)
4. Emission due to skimming - E (kg) = E. - E
5 , Up
5. Emission (kg) ladle = ES/NO. of Ladles
The particulate and sulphuric acid emissions were also calculated in
terms of kilograms per ladle from the data obtained for each filter. Owing
to very low concentrations of the particulates, the filter was changed
approximately every eight hours. Therefore, the particulate and sulphuric
acid emissions were not correlated with skimming and "puffing" activities
separately, rather the results represented a combination of the two
activities.
Finally, fugitive emissions were expressed in terms of unit production
of bessemer matte* using relevant production data,
* Bessemer matte is the final output of the nickel circuit in the Smelter
and contains Ni (50%), Cu (25%), and S (22%).
283
-------
FUGITIVE SULPHUR DIOXIDE EMISSIONS
Total fugitive sulphur dioxide emissions from reverb furnaces were
determined as 15.5 kg/Mg bessemer matte. Of these about 70% originated
from matte tapping activity while slag skimming contributed 20% of the
total. The remainder was attributed to furnace "puffing" and the process
of returning of converter slag to the furnace. Results are shown in Table 1.
On a per ladle basis the sulphur dioxide emissions were 64 - 12 kg
during tapping and 12.1 * 2.9 kg during skimming. The latter figure compared
well with the value of 11.2 ± 4.2 kg obtained during Phase I.
The average rates of sulphur dioxide emissions during tapping, skimming
and "puffing" were found to be 5.1, 1.4, and .06 kg/min., respectively. The
average maximum rate of emission was therefore about 6.5 kg/min. The rate of
"puffing" emissions showed no clear relationship with the furnace draft.
However, the observed variation in the draft during the entire test was"very
small which presumably effected a change in sulphur dioxide emissions too
small to be detected by the present technique.
The average sulphur dioxide concentration in the matte tapping emissions
was 7400 mg/m3 (2470 ppm) while that in the skimming emissions was 180 mg/m3
(63 ppm). These were the concentrations in the emissions on entry to the
atmosphere. The slag skimming emissions were diluted some twenty fold in the
roof ventilator prior to release to the atmosphere, therefore, the concentra-
tion of sulphur dioxide was much lower. The average sulphur dioxide concen-
tration in the "puffing" emissions was found to be 8 mg/m3 (2.8 ppm).
FUGITIVE PARTICIPATE EMISSIONS
Total patticulate emissions from reverb furnaces were 0.19 kg/Mg
bessemer matte of which like sulphur dioxide, nearly 70% were attributable
to matte tapping while the rest was contributed by slag skimming and
"puffing" activities. Results are summarized in Table 2.
The average rate and concentration of particulate emission from
matte tapping were .06 kg/min. and 81 mg/m3 respectively. The mean combined
rate of particulate emissions due to slag skimming and furnace "puffing" was
.009 kg/min. The corresponding concentration was about 1.2 mg/m3.
The main constituents of the particulates from matte tapping were
sulphur (13%) and iron (5%) with copper and nickel being found in lower
amounts. Particulates collected at the roof ventilator consisted mainly
of iron (30%), copper (5%), and nickel (7%).
234
-------
FUGITIVE SULPHURIC ACID EMISSIONS
Total sulphuric acid emissions from reverb furnaces were measured as
0.12 kg/Mg bessemer matte. Distribution of the emissions between matte
tapping, slag skimming and "puffing" was much the same as for sulphur
dioxide and particulates. Results are shown in Table 3.
Average rates and concentrations of the emissions during tapping were
.04 kg/min. and 52 mg/m3 respectively. Those attributable to slag skimming
and "puffing" were .006 kg/min. and about 0.8 mg/m3 respectively.
In general, sulphuric acid emissions were proportional to sulphur
dioxide emitted. A plot indicating this relationship is shown in Figure 7.
Average sulphuric acid/sulphur dioxide ratio found in this study was
about .010 which was similar to the value obtained during recent tests
conducted on gases in the main chimney of the Smelter. This ratio was also
fairly constant, the relative standard deviation being ± 25%, indicating
that capture of sulphuric acid on the filter was probably quantitive. This
was considered indirect support for the technique used for the measurement
of sulphuric acid.
DISCUSSION
Precision and Accuracy
In previous studies of copper and nickel converters no attempt was made
to determine the precision of the measurements. This was mainly due to the
complexity of the converter operations, particularly in the case of nickel
converters where interaction of emissions from two or more converters
occured prior to reaching the measurement system. Consequently only a sub-
jective estimate of precision was made, which was - 10% for the copper
converters and ± 90% for the nickel converters. :'2
In the present study on the nickel reverb furnace, the system was found
to be more exact, with little or no interaction between emissions from
different units and also with a good separation of emissions from various
activities of the furnace. Thus precision (S.D./mean x 100) could be
determined on per ladle basis.
For sulphur dioxide emissions this was found to be - 20% for matte
tapping and ± 24% for slag skimming. It should be emphasized that these
values represented the overall fluctuations in the emissions, the precision
of the measurement technique itself was considered to be much better.
235
-------
In addition, slag skimming emissions were measured at two locations
employing different techniques. The results are compared in Table 5. The
sulphur dioxide emissions were measured as 11.2 kg/ladle in the vent from the
slag launder (Phase I) which showed a good agreement with the value of 12.1
kg/ladle obtained in the roof ventilator (Phase II). Similarly the rate of
sulphur dioxide emission of 1.5 kg/min. in the vent compared well with the
rate of 1.4 kg/min. measured in the ventilator. Finally, the concentration
of sulphur dioxide in the emissions was found to be 180 mg/m3 in the
ventilator. The same measured in the vent was 3080 mg/m3 which was sub-
sequently diluted 17 times in the roof ventilator to give a value of
181 mg/m3. These comparisons were indicative of the high accuracy of the
measurements.
GENERAL
Total sulphur dioxide fugitive emissions from reverb furnaces were
quite similar to the total converter fugitive emissions. This is shown
in Table 5. The particulate emissions from reverb furnaces were about
the same as those from nickel converters while copper converters emitted
much larger quantities of particulates. This was chiefly due to periodic
addition of very fine flotation concentrate to the copper converters. An
important difference between fugitive emissions from reverb furnaces and
converters was that essentially all of the emissions from furnaces (^902)
were captured at source by hoods before being released to the atmosphere.
The converter fugitive emissions were uncaptured and tended to wander before
leaving the building via roof ventilators. This difference was reflected in
the quality of work room air in the two areas. The average eight hour
reverb furnaces were found to be 0.9 ppm and 0.2 mg/m3 respectively,
compared to the corresponding average values of 2.5 ppm and 5.5 mg/m3 found
at the converters.
CONCLUSIONS
1. Measurement of fugitive emissions from nickel reverberatory
furnaces has been successfully completed. The precision of
the measurements was estimated as ± 20%.
2. Matte tapping activity was the major source of fugitive
emissions, contributing about 70% of the total with slag
skimming (20/0 and other activities contributing the remainder.
3. Essentially all the fugitive emissions (^90%) were captured at
source before being emitted to the atmosphere.
286
-------
4. Daily fugitive sulphur dioxide emissions from the reverb
furnaces (15.5 Mg) were similar to the total daily converter
fugitive emissions (13.8 Mg).
5. Judged by the criterion of the constancy of sulphuric acid/
sulphur dioxide ratio, the present technique of capturing
the acid by filtration seemed to be satisfactory.
REFERENCES
"Measurement of Fugitive Particulate and Sulphur Dioxide
Emissions at Inco's Copper Cliff Smelter", A.D. Church,
C. Landolt and F. Boniakowski; presented at the 108th
AIME Annual meeting in New Orleans, Louisiana, February
1979.
"Measurement of Sulphur Dioxide and Particulate Fugitive
Emissions From the Nickel and Copper Converter Operations
at Inco's Copper Cliff Smelter", A.D. Church, C. Landolt
and F. Boniakowski; presented at the 18th Annual Conference
of Metallurgists sponsored by Canadian Institute of Metallurgy
in Sudbury, Ontario, August 1979.
237
-------
Rgure 1
Vent Arrangements at N8 3 Reverb Furnace
i Himri SampHng Point
itaundv •
Psia« Laanatr
JS/SPT/SO
288
-------
Figure 2
Sampling Train Used in Phase I
00
IO
1st Impinge rrf BBi 2nd Impinger
-------
Figure 3
S02 and particulate monitoring system. (Phase ID
O
• — n —
1 — U —
Orifice
plate for
Dn<
Hurricane $;
_pump||t
;t
er
c
1 <
i
Ventilator.
3 Sets -Sample probes
for SOa and particulate
_ * Temperature recorder
».f^ne uolnritu rorrvrrfor
»< "1 Ci«fl4^K USllwA
-Rj— IT
O Draft indicator
measurement
-------
K!
VO
Figure 4
Analysis System (Phase n)
To Orifice
Event time = 1-15 minutes
Sampling rate ~ 14 litres/min.
Pump
Dust filter
From Vent
pH
Electrode
H202
Solut
Istlmplnger 2ndlmpinger Meter
-------
Figure 5
S02 Emissions during different events
l.2r
.
•
HH Skimming event
I I Non Skimming event
20
100
120
140
160
180
200
TIME (Minutes)
-------
Figure 6
S02 Emissions during Typical Skimming Event.
9 II
Time (Minutes)
Ts = Recorded duration of a SKIMMING EVENT
Tns= " " " NON SKIMMING EVENT
293
-------
Figure 7
Sulphur dioxide vs. Sulphuric acid.
NJ
IO
-P-
C 3
O
N
I
•Mean
IOO
2OO
300
4OO
S02(kg)
-------
TABLE 1
Fugitive SO? Emissions From Reverb Furnaces
Activity
Matte Tapping
Slag Skimming
Puffing
Return Of
Converter Slag
Total
S0£
(kg/Mg B.M.)***
10.80
3.07
0.41
1.20
15.5
Emission Rate
( kg/mi n)
5.08
1.41
0.06
N/A*
6.55**
Concentration
(mg/m3)
7,390
180
8
N/A*
-
* - Not Available
** - Maximum
*** - Bessemer Matte
295
-------
TABLE 2
Fugitive Particulate Emissions From Reverb Furnaces
Activity
Matte Tapping
Slag Skimming +
Puffing
Total
Participates
(kg/Mg B.M.)*
.119
.068
.187
Emission Rate
( kg/mi n)
.057
.009
.066**
Concentration
(mg/m3)
81
1.2
-
* - Bessemer Matte
** - Maximum
296
-------
TABLE 3
Fugitive Sulphuric Acid Emissions From Reverb Furnaces
Acti vi ty
Matte Tapping
Slag Skimming +
Puffing
Total
HzS04
(kg/Mg B.M.)*
0.077
0.043
0.120
Emission Rate
(kg/min)
0.037
0.006
0.043**
Concentration
(mg/m3)
52
0.8
-
* - Bessemer Matte
** - Maximum
297
-------
TABLE 4
Comparison of Slag Skimming Emissions Measured at Two Locations
Location
Vent From Slag
Launder (Phase I)
Roof Ventilator
(Phase II)
S02/Ladle
11.2 - 4.2
12.1 - 2.8
kg/mi n
1.5 - 0.4
1.4 - 0.3
mg/m3
181* 1 55
180 - 51
* Corrected for dilution in the roof ventilator.
293
-------
TABLE 5
Comparison of Fugitive Emissions From Converter and Reverb Furnaces
Source
Copper Converters
Nickel Converters
Total Converters
Nickel Reverb Furnaces
S02 (Mg/Day)
10.6
3.2
13.8
15.5
Particulates
(Mg/Day)
1.1
0.2
1.3
0.2
Sulphuric Acid
(Mg/Day)
N/A *
N/A *
N/A
0.1
* Not Available
299
-------
CONTROL OF FUGITIVE EMISSIONS FROM COAL STORAGE PILES
A. E. Veel
Research and Development
C. H. Carr
Sinter Plant and Docks
The Steel Company of Canada, Limited
Hamilton, Ontario, Canada
ABSTRACT
The Steel Company of Canada is a fully integrated steel producer with its
primary steelmaking facilities located in Hamilton, Ontario. The basic end of
this operation consists of four coke batteries, four blast furnaces, one OH
shop and one EOF shop.
During the period December to March when the Great Lakes and the St. Lawrence
Seaway are ice bound, shipping of raw materials is curtailed and both iron ore
pellets and coal must be stockpiled to allow for continuous operations. At
any one time approximately 83 acres of land is covered by up to 7 distinct
coals or coal blends. Because the percentage of fines can be as high as 15
percent, coal dust particle movement by the wind can be severe.
MHTR Engineering was retained to examine the situation and to recommend
measures to alleviate the existing conditions. This study was based on
constructing models of the site and testing them in a wind simulator to
determine the problem areas and to test the remedial solutions.
The study indicated that orientation and shaping of the coal piles were key
factors in controlling the turbulence which resulted in particle movement.
Suggested solutions which have been implemented have significantly reduced
the problems associated with storage of coal in open sites.
300
-------
CONTROL OF FUGITIVE EMISSIONS FROM COAL STORAGE PILES
Over the past few years environmental legislation has focused on control of
point sources of industrial emissions. While this control has led to improve-
ments in the levels of tsp in the ambient environment there has been a greater
recognition that fugitive emissions are also significant contributors to urban
dustfall levels. In an integrated steel plant a large source of potential
fugitive emissions is raw materials storage piles.
One of the world's major integrated steel companies, Stelco is ranked No 1 in
Canada, and No 9 in North America. Steel production at its primary facilities
in Hamilton, Ontario, at the western end of Lake Ontario, some 60 miles north-
west of Buffalo, New York, is approximately 6 million ingot tons per year.
When a greenfield expansion program, now in progress, is completed, it is
estimated that the Company's production will eventually double to 12 million
ingot tons per year. This greenfield site is located on the shore of Lake
Erie, opposite Erie, Pennsylvania.
Three basic materials — iron ore, coal and limestone — are needed to make
steel. Coal and iron ore are most economically shipped by water transport to
Stelco's docks at Hamilton. These facilities handle over 370 ships each year.
Since it takes 1 ton of iron ore, ^ ton of coal and miscellaneous flux
materials to produce 1 ton of steel, a company the size of Stelco must handle
approximately 10 million tons of bulk raw materials each year at its Hamilton
plant. At any one time the coal storage piles cover approximately 83 acres of
land (roughly 1/10 of the total area covered by these operations) and because
up to 7 different coals or coal blends are utilized, the coal storage area is
comprised of several separate piles rather than one large continuous pile.
As mentioned earlier the majority of raw materials are shipped through the
Great Lakes waterway system. Coal from Kentucky and West Virginia must cross
Lake Erie and move through the Welland Canal system, bypassing Niagara Falls,
in order to reach the facilities at Hamilton. However, due to the severe
Canadian climate this method of transport is not available from December to
March each year. As a consequence the storage piles are considerably larger
than those kept by most United States steel companies. Allowing for the four
months that the waterway is impassable and a one-month safety factor to ensure
a continuation of operations should strikes occur within any of the unions
involved in shipping, shore handling and mining, there is a requirement to
store approximately five months' supply of raw materials at the end of every
snipping season.
The coal arrives at Hamilton in self-unloading vessels. A conveyor deposits
tne coal in a conical pile on the dock. Water is continually sprayed on the
coal to limit dust blow-off during the unloading process. A bulldozer is
used to level this conical pile permitting large mobile scrapers to transfer
the coal to the storage piles, which may be up to 1100 ft in length, and
deposit the new coal in thin layers. This layering process is used to
partially blend the old and new coals. During the passage of the scraper a
great deal of coal dust is released into the air. A similar process in
301
-------
reverse is used during the ex-stocking process when the coal is removed from
the storage piles and taken to the Coke Ovens.
In periods of elevated wind speed fugitive emissions may be generated at the
f o1low ing t imes:
• when the coal drops from the conveyor to the conical pile,
• general wind erosion from the conical pile,
• during the period the scraper is traveling over the storage
piles or the arterial roads, and
• general wind erosion from the long-term storage piles
themselves.
Several of these problem areas were recognized many years ago and solutions
were implemented to control the problems. During unloading:
• water sprays were directed into the unloading coal, stream and
conical piles,
• the free-fall height of the coal was minimized by lowering the
conveyor as much as possible, and
• a protective enclosure was attached to the conveyor to minimize
wind access to the falling stream of coal.
While these precautions were satisfactory under normal atmospheric conditions
with extremely high winds they did not suffice and unloading operations had
to be terminated.
In order to minimize dusting while the scrapers move along the roads and
storage piles tanker trucks were used to spray water in the summer. In the
winter these same trucks were used to spray waste oil onto the roadways. The
storage piles were coated with a binding agent near the end of the shipping
season. This was only a limited success. Some of the problems associated
with wind loss from storage piles were reduced by the above measures. A key
step, however, was the reduction of the number of storage piles to six. This
allowed for both an orderly layout of the piles and contouring of the piles.
A final step in these early efforts was the adoption of a procedure whereby
coal was removed only from the end of the storage pile during winter months.
This eliminated disturbance of the top surface of the pile during the winter
when water spraying could not be utilized.
While these steps were successful to some degree, incidents of wind erosion
of the storage piles resulting in blowing coal dust continued to occur
especially during the extreme conditions experienced in winter. At this
time a consultant (Morrison, Hershfield, Theakston and Rowan), who had
considerable experience in the study of wind-blown materials as well as wind
tunnel and water flume testing was retained. The problem as presented to the
consultant was to determine the areas of coal loss from the storage piles and
to recommend methods for control. A second phase involved a study of surface
302
-------
binders to determine the best product to be used to coat the storage piles to
carry through the winter period.
On the first phase of the study the technique used consisted of a model
analysis by employing an open channel water flume especially designed for the
study of particle movement and control. This flume is 40 ft long, 4 ft wide
and 18in deep. By varying the depth of water a variety of wind velocity
conditions can be simulated. Particles of coal were simulated by silica sand
and wind movement was observed by injecting a dye into the flowing water to
indicate turbulence and direction.
A model was constructed of the complete site including the terrain, buildings,
conveyors, stacks, ore piles, coal piles, the unloading docks and surrounding
water area as well as separate models of the coal piles. The model of the
site was used to determine the general flow pattern of the wind and the
effect of the various components comprising the site. The models of the coal
piles were used in assessing, in detail, the effect of orientation and shape
of the piles.
Water Flume Test Results
Since wind direction is an important factor in the movement and entrainment
of coal particles, the water flume allows for the rotation of the model to
simulate changes in wind direction. In the Hamilton region south-west and
westerly winds are predominant. In all cases coal piles transverse to the
wind direction and even at a slight angle to the wind created turbulence on
the lee side of the pile. With consecutive piles a channel was created to
carry the particles at a higher velocity, due to the venturi action, to the
leeward side of the site. Coal piles located in line with the wind
eliminated the turbulence. Based on this result a decision was made to
reorient the axis of the storage piles 'so that they would be in line with
the prevailing wind. This required changing the entire field arrangement and
was carried out over one shipping season.
Shaping of the coal piles is an important factor in reducing turbulence due
to sharp edges and scour action. Reduction of sideslopes will reduce the
scouring at the top of the slopes and the optimum slope is 1:7. However,
steeper slopes will have progressively detrimental effects from 1:7 sideslopes
to 1:1 sideslopes.
While we were able to achieve the ideal slope at the leading edge of the pile
which faces the prevailing wind using that flatter slope for all sides would
have required a coal storage area many times the size available. We utilized
a rounded crown which was suggested as an alternative shaping program where a
1:7 optimum slope is not feasible.
yhen more than one coal pile with 1:1 sideslopes (angle of repose) is used,
turbulence is pronounced on the leeward side and in the channels formed by
adjacent piling. The 1:7 sideslopes used for optimizing wind flow patterns
eliminates the turbulence and eddy action.
An alternative recommendation to the 1:7 sideslope consists of rounded
sideslopes or the employment of a porous fence along the windward portion and
303
-------
top of the pile. These alternatives are proposed only where space
limitations require a 1:1 sideslope.
The leeward end of the piles, where coal is ex-stocked in the winter time,
is an area of particular concern. The flume studies indicated that the
installation of a 12-ft high porous fence on top of the pile, approximately
150 ft from the leeward edge, would be effective in reducing turbulence. As
an alternative, three 4-ft high snow fences placed 60 ft apart parallel to
each other in line, and parallel to the leeward face, were shown to be
equivalent to one 12-ft high fence. In fact, both of these methods were
tried and both proved to be effective and the combination of both is used
in the field.
The final recommendation of the study was to install a snow fence near the
leeward edge of the entire field to catch dust which may be entrained along
the ground surface. This was done and has proved to be effective in
controlling dust not caught in other ways.
All these recommendations have been acted on and the practices are in
continuous use in the coal storage fields. The coal piles were reoriented
and reshaped as recommended by the consultants.
Wind Tunnel Test Results
The use of binding agents applied to coal piles is a current practice in
reducing the movement of coal particles from storage piles. In the second
phase of the study, samples of six types of coal used by Stelco were compacted
in small boxes to simulate a coal pile with a leading edge to the wind. They
were coated with various combinations of six different binders and placed in
the wind tunnel. These tests were supplemented by small sample piles in the
field at Stelco to test the weatherability of the various binding agents.
As a general rule the thickness of the crust formed was found to be a measure
of the effectiveness of the binding agent in reducing particulate movement.
The study was narrowed to two binding agents plus waste oil which is
generated internally at the plant. All three were tried in test applications
on the actual piles for a complete winter season. We have decided to use an
asphaltic product supplied by Flintcote as the main binding agent to be
applied to the piles at the end of each shipping season and to carry us
through to the spring. Waste oil is used for the active area at the end of
the pile and for road dust control during the winter time.
The binding agents are to be applied in November and again at the end of the
shipping season to ensure that snowfall does not prevent us from getting the
agent directly on the surface before the main onslaught of winter. During
the summer season, water is the main dust control agent on the piles and on
the roads.
An important feature of our control program has been the issuance of a
detailed list of operating instructions for the coal field. This involves
regular attention to a wind recording device as well as regular contact with
the local Weather Office for wind prediction on each shift. Action to be
taken on prediction or recording of increased wind speeds is clearly recorded
-------
By these actions we have recognized that continuous attention to a remedial
program designed to combat fugitive emissions from coal storage fields is
required. This prolonged study has shown that a combination of both physical
and chemical methods based on a scientific approach can result in significant
reductions in dust emissions from bulk storage sites.
305
-------
USE OF ROOF MOUNTED TYPE ESP'S IN IRON
AND STEEL INDUSTRIES IN JAPAN
Senichi MASUDA
Department of Electrical Engineering
Faculty of Engineering, University of Tokyo
7-3-1, Kongo, Bunkyo-ku, Tokyo, JAPAN
SUMMARY
The first ROOF MOUNTED TYPE ESP (REP) in Japan was installed in March
1972 at a steel converter shop (80 tons x 3 units) in KOBE WORKS of KOBE
STEEL CORPORATION, LTD. by the hand of SUMITOMO HEAVY INDUSTRIES, LTD.(SHI)
after a cooperative developmental effort made by both companies. This
first REP showed a very successful result, lowering the dust loading from
0.246 (g/Nm ) at its inlet down to 0.032 (g/Nm ) at its outlet (collection
efficieny: 87 %). Thus, it proved to be one of the most effective means of
fugitive emission control in the iron and steel industries. The first aim
of this development was to solve the problem of a limited installation site.
However, it has become clear after the start of its operation that this
solution using a thermal lift of hot dirty gas itself has several other
large advantages compared to an alternative solution using forced suction
and bag-house.
The first of these additional advantages is its inherent extremely energy
saving character resulted by elimination of a large capacity suction fan.
The total power consumption is only 5 - 10 % of that needed for the system
with forced suction and bag-house, even estimated from a very moderate
assumption that the dirty gas volume to be handled is the same for both
cases. Actually, however, the gas volume to be treated by REP is only the
net gas volume of dirty gas automatically subjecting to a thermal lift.
Whereas, in the latter case, a large quantity of ambient air is inevitably
mixed to the hot dirty gas, so that the total gas volume to be suctioned
becomes enormous, amounting to 1.5 - 2.5 times the actual dirty gas volume.
The second of additional advantages is its noiseless character also due
to the elimination of the suction fan. As a result of these remarkable
merits REP has been very positively accepted by the iron and steel
industries in Japan for fugitive emission control.
The major manufacturer of REP is SUMITOMO HEAVY INDUSTRIES, LTD.(SHI)
but SUMITOMO METAL 4 MINING CO., LTD.(SMM), HITACHI PLANT ENGINEERING 4
CONSTRUCTION CO., LTD.(HPC), and CHIYODA CHEMICAL ENGINEERING CO., LTD. are
also sharing to the REP market. The total number of REP's installed by
March 1980 is at least 38 units with 9 units for converter shops, 7 units
for casting floors of blast furnaces, and 22 units for electric furnaces of
various types.
In this paper is reported the present status of the ROOF MOUNTED TYPE
ESP's for fugitive emission control in the iron and steel industries in
Japan.
306
-------
1. Introduction
The ROOF MONTED TYPE ESP (REP: ref. (1) - (7)) has aquired, since its
first installation in March, 1972 at a steel converter shop (3 units x 80
tons; KOBE WORKS of KOBE STEEL CORPORATION, LTD.), an established position
in the Japanese iron and steel industries for the control of fugitive
emissions at various workshops including converter shops, catsing floors of
blast furnaces, and electric furnaces, A light-weight ESP is mounted on top
of the roof of a workshop, and the dirty hot gas emited from its sources is
passed through this ESP by the action of its thermal draft without the aid
of a suction fan. The first aim of considering this technology was to solve
the problem of a limited site for installation of a conventional dust
collecting facility. However, after its start of operation a quite
remarkable advantage of REP in its inherent energy saving character has
become clear. This is because it does not use a large capacity induced fan
which, otherweise, has to suction the hot dirty gas plus a substantial
quantity of surrounding air against a high pressure drop of a gas duct
system (150 - 250 mm H-0) and a bag filter (150 - 200 mm H^O). The total
gas volume to be handled in the forced suction system amounts to as high as
1.5 - 2.5 times that of the actual dirty gas (5). In addition, its
noiseless feature resulted by elimination of the fan also provides a large
merit.
As a result REP's are being used in increasing number in the iron and
steel industries in Japan, amounting to as many as at least 38 units in
total at the end of March, 1980: 9 units for converter shops, 7 units for
casting floors of blast furnaces, and 22 units for electric furnaces of
different kinds. The largest share-holder among the manufacturers of REP
in Japan is SUMITOMO HEAVY INDUSTRIES, LTD.(SHI), whereas SUMITOMO METAL &
MINING CO., LTD.(SMM), HITACHI PLANT ENGINEERING & CONSTRUCTION CO.,
LTD.(HPC) and CHIYODA CHEMICAL ENGINEERING CO., LTD. are also manufacturing
REP's.
In this paper is reported the present status of the ROOF MOUNTED TYPE
ESP's in Japan as applied in the iron and steel industries, including its
construction, design problems, application fields, and its specific
advantages, together with its operation data.
2. Construction
Fig. 1 illustrates a system of the fugitive emission control equipped
with REP. There are slight differences in the construction of REP depending
upon the difference in manufacturer's design concept and specific conditions
of indivisual applications.
Figs. 2 illustrate the construction of SHI-type REP, and Fig. 3 shows its
photograph taken from outside. A light-weight ESP is mounted on top of the
roof of a workshop above the emission sources.
The collecting electrodes are out of thin stainless-steel plates, and the
discharge electrodes are stainless-steel wires or round piano-wires
supported by metal frames. The use of the frame construction provides
advantages in light-weight and ease of centering. No thermal deformation
307
-------
occurs in this particular case as the gas temperature at the precipitator
inlet is below 80 (deg C) (typically 40 - 60 deg C).
The whole electrode system is mounted inside a light-weight housing
having no roof on its top. The shell of the housing is out of a special
laminated plate consisting of a thin steel plate covered with plastic layers
so that sufficient durability can be obtained against corrosion and
weathering in spite of its small thickness.
The precipitator is of a semi-wet type with both collecting and discharge
electrodes being intermittently cleaned by water spray. The whole REP unit
is devided into a number of blocks, and washing is made block by block.
Each block is washed 1-4 times, a day, each time for _10 minutes with a
water feed rate of 0.1 - 0.6 (m /min) (typically 0.4 m /min). So, the
water feed rate needed is independent of REP capacity, and the total
quantity of spray water to be supplied per day is very small, and given by
Qw - (0.1 - 0.6 m3/min) x (10 min) x Nb x Nw (m3/d) (i)
where Nb « number of blocks, and Nw = number of washing per day. The use
of this semi-wet design provides the advantages such that both the back
discharge due to high resistivity dust and the dust reentrainment due to
mechanical rapping are effectively avoided, that the noise and vibration
problem in rapping operation is solved, and that elimination of the rapping
system with many moving parts contributes to the light-weight scheme and
ease in maintenance. In addition, the semi-wet scheme combined with the
open housing construction contributes much to avoid a danger of explosion to
be caused by coal dust and carbon monooxide.
The slurry is collected by a louver-shaped hopper located under each
block, and fed through troughs to a conventional water treatment system, as
indicated in Fig. 4. A portion of clear water is reused for spraying after
pH-adjustment.
The hot dirty gas from the emission sources rises with its thermal lift
up to the top of the workshop, and enters through the louver-hoppers into
the collection field of each block of REP. The louver-hoppers serve as a
gas distributer to equalize the gas velocity. The gas flows upward
vertically through the collecting ducts, and is emited directly into open
air from the top of REP.
The insulators supporting the frames of discharge electrodes are encased
in the insulator boxes located on top of the REP housing, and clean air is
purged into the boxes to prevent contamination of the insulators.
A high voltage power-pack with automatic voltage control is also mounted
on top of the roof of the workshop.
Table 1 indicates an example of the design data of a SHI-type REP
installed at a converter plant.
Figs. 5 shows a photograph of the SMM-type REP. In the SMM design, the
hot dirty gas is at first gathered in a monitor, and, then, fed horizontally
to REP mounted on top of the workshop roof adjacent to the monitor (Fig. 6).
A number of low-power roof-fans are installed at the outlet of the SMM-type
REP to suction the gas horizontally through its collection field. Both the
dry-type and wet-type REP's are manufactured by SMM.
The details of the constructions of REP's of SMM and other manufactureres
are not much ifferent from that of SHI, so that the further descriptions of
each design are omitted.
308
-------
3. Design Problems
There are several problems to be carefully considered in the planning and
design of REP. These are the estimation of gas volume to be handled by
REP, the properties of dust to be collected, the location of REP with
respect to the number and position of emission sources, and prediction of
wind.
3-1) Calculation of Gas Volume:
The ascending speed of hot dirty gas from emssion sources lowers with
height because of the decrease in its temperature due to mixing with ambient
air. Its final speed at the inlet of REP depends upon the roof height,
building volume and shape of the workshop as well as the number, capacity
and operating conditions of the emission sources. As a result, an accurate
estimation of the gas volume to be handled by REP is extremely difficult,
and requires much experience.
In the case when REP is to be installed on an existing workshop, it is
imperative to directly measure the gas volume in advance by measureing
velocity distribution of gas at each monitor under various operating
conditions. The measurement of dust loading must also be performed at the
monitor.
In the case when REP is to be installed on a new plant, the following
theoretical gas volume, Qz, used for design of a canopy hood provides a clue
for the gas volume estimation of REP (see Fig. 7):
Qz - 1.95 x Z3/2 x (H')1/3 (m3/min) (2)
H' = (1.45/60) x As x T4/3 (kcal/min) (3)
where Z = height of REP inlet from provisional position of heat source (m),
H1 " heat transmission of convection current fronu heat source (kcal/min),
As " surface area of heat source - ( 77/4) x Do (m ), T = temperature
difference between heat source and ambient (deg C).
The actual gas volume to be handled, Q, can be estimated by multiplying a
correction factor, K, as:
Q - K x Qz (m /min) (4)
The correction factor should be selected in the range of K = 1.5 - 2.5
from the experience at the similar installations.
Fig. 7 is a chart for estimation of Qz as a function of H and Do for the
case when Z = H + 2 Do (4).
3-2) Dust Properties:
The mass-loading, particle size distribution, chemical composition, and
electrical resistivity of the dust to be handled by REP must be carefully
studied. The inlet mass-loading of dust determines the quantity and
frequency of water spraying and the capacity of water treatment system. In
most cases, however, the inlet mass-loading of dust hardly exceeds the level
of 1 (g/Nm3).
309
-------
The frequency of water spraying must be raised for fine particles in
order to avoid the dust reentrainment. Table 2 gives the examples of the
particle size distribution of converter dusts in Japan.
The chemical composition of dust may affect the design of pH-adjustment
system for slurry treatment, the corrosion of electrodes, hoppers and a
piping system of REP, and probably the electrical resistivity of dust.
Table 3 provides the chemical composition of the dusts indicated in Table 2.
The electrical resistivity of dust may become very high at 60 - 90 (deg
C) to be encountered at the inlet of REP, as shown in Table 4. Then, a
severe back discharge may occur in the collection field of REP when the
quantity and frequency of water spraying is insufficient or the water spray
system fails.
3-3) Location of REP:
From an economical point of view REP should be located only at an area at
which dirty hot gas arrives. The location of REP must be determined,
considering the thermal current of the dirty gas, its diffusion and the
effect of wind.
4. Applications
4-1) Converter Shop
In this application the dirty gas volume is very large, so that its total
quantity is handled only by REP. The fluctuations in gas temperature, gas
velocity and dust loading at the inlet of REP are very large in this
application. Table 5 gives some of the design and operation data in this
application.
4-2) Casting Floor of Blast furnace
Again in this application the whole dirty gas is handled solely by REP
because of its very large volume.
4-3) Electric Furnace
In this application REP is used either alone or in combination with
monitors, depending upon the workshop conditions. The fluctuations in gas
temperature, gas velocity and dust loading at the REP inlet are also very
large. Table 6 provides again some of the design and operation data in
this application.
5. Advantages of ROOF MOUNTED TYPE ESP's
310
-------
REP's have the following inherent advantages:
(1) Saving of installation site.
(2) Stable operation and high collection performance.:
This is resulted by the semi-wet operation scheme where no back discharge
and dust reentrainment take place.
(3) Low installation costs:
The installation costs may become rather lower than the alternative
emission control system using forced suction and bag-house because of the
simple light-weight construction of REP as well as the elimination of a
large suction fan and a gas duct system..
(3) Very low power consumption:
The total electric power consumption of REP is only 5 - 10 % of that
needed for the system using forced suction and bag-house. This is resulted
by the eliminations of a large induced fan which has to suction a large gas
volume against the pressure drops at canopy hoods, gas duct (150 - 250 mm
H90) and filter bags (150 - 200 mm H^O). The pressure drop of REP is
almost the same as that of monitors. Tnis figure of energy saving is a very
moderate estimation assuming the gas volume to be the same for both cases.
However, the actual gas volume to be handled in the latter case amounts to
1.5 - 2.5 times that of REP which selectively handles the hot dirty gas
ascending by its inherent thermal lift.
(4) Operation without noise and vibrations.
(5) Low maintenance costs:
The maintence costs of REP are very small because it does not require the
exchange of bags and moving parts, and its water consumption is also small.
The amount of water feed rate needed for electrode spraying is only 0.1 -
0.6 m /min, and the used water is recycled after slurry treatment. As a
result the amont of water for replenishment is little. In addition, the
maintenance work is hardly needed because of its simple construction without
moving parts.
6. Energy Saving Feature of ROOF MOUNTED TYPE ESP's
As the contribution of REP to enery saving is quite remarkable, it
deserves a special investigation. A comparison is made of the power
consumption for the respective applications between a REP system and a
conventional system with forced suction and bag-house (7). The power
consumption of the REP systems is taken from the practical operation data,
whereas those of the systems with forced suction and bag-house are
calculated based on the same gas volume as previously described. Tables 7
g and 9 show respectively such a comparison in the applications for a
converter shop, a casting floor of blast furnace and an electric furnace.
It may be well understood that the contribution of REP to energy saving must
be given a special attention.
The case studies of comparison are further made between the two systems,
considering all the power consumption needed not only for operation of each
dust collecting equipment, but also for production of all the materials
necessary to construct each equipment (7). The calculated result for a
standard 30 ton/ch. electric furnace is given in Table 10 , whereas that for
a standard 160 ton/ch. converter in Table 11 . These results reveal that
311
-------
the saving of eneijgy resulted by the use of REP amounts in ten years to as
high as 188.8 x 10 (kcal) in the former case and 350.4 x 10 (kcal) in the
latter case. Based on these data the study is extended to all the electric
furnaces and converters existing in Japan in December, 1976 (Table 12). The
saving to be obtained by using REP's in ten years in all of the electric
furnaces will amount to. as high as 68.5 x 10 (kcal), while that in the
converters to 32.2 x 10 (kcal). This energy saving corresponds to ca. 1
million tons of crude oil per year.
7. Conclusion
The present status of ROOF MOUNTED TYPE ESP's in Japanese iron and steel
industries is described. The conclusions of this parer are as follows:
(1) REP's have aquired an established position in the Japanese iron and
steel industries.
(2) REP's have a number of inherent advantages over a system with forced
suction and bag-house. These are (a) saving in installation site, (b)
stable operation and high collection performance, (c) very small power
consumption, (d) low initial and maintenance costs, (e) operation without
noise and vibration, etc.
It is likely that ROOF MOUNTED TYPE ESP's will be more extensively used
in future in Japan, not only in the iron and steel industries but also in
the other industrial fields, primarily because of its very remarkable merit
of energy saving.
Reference
(1) Teruo WATANABE: Electric Precipitator on Roof to Prevent Emission Dust
from Shop Housing, Jounal of Powder Engineering Research Association Japan
Vol. 9, No. 6, pp. 418 - 421 (1972).
(2) Masakazu SAKAI: Roof-Mounted Type Electrostatic Precipitator, Technical
Journal of SUMITOMO HEAVY INDUSTRIES, Vol. 22, No. 65 (Aug., 1974).
(3) Tsutomu NOMURA and Masakazu SAKAI: Roof-Mounted Electrostatic Precipita-
tor for Casting Floor of Blast Furnace, Proceedings of Institute of
Electrostatics Japan, Vol. 1, No.2, pp. 82 - 87 (1977).
(4) SUMITOMO HEAVY INDUSTRIES, LTD.: R. Ep. - Roof-mounted Electrostatic
Precipitator, Technical Report of SHI (Oct., 1977).
(5) Shoji ITO, Shigeyuki NOSO, Masakazu SAKAI and Kiyoshi SAKAI:
Roof-Mounted Electrostatic Precipitator, Proc. of 1st EPA-Symposium on
Transfer and Utilization of Particulate Control Technology, Vol. 1, pp. 435
- 495 (July, 1978 in Denver).
(6) Ken TAKIMOTO: Wide-Spacing EP Is Available in Cleaning Exhaust Gas from
Industrial Sources, Proc. of 1st EPA-Symposium on Transfer and Utilization
of Particulate Control Technology, Vol. 1, pp. 297 - 305 (July, 1978 in
Denver).
(7) Shigeyuki NOSO, Masakazu SAKAI and Toru SASAKI: Roof-Mounted Electrostat-
ic Precipitator Saving Energy, Technical Journal of SUMITOMO HEAVY
INDUSTRIES, LTD., Vol. 26, No. 78, pp. 75 - 78 (Dec. 1978).
312
-------
Table 1 Design Data of SHI-Type REP for Converter
1. Dimension
2. Casing
3. Collecting Electrodes
4. Discharge Electrodes
5. Power Consumption
6. Water Feed Rate
7. Gas Volume
8. Gas Temperature
9. Inlet Dust Loading
10. Outlet Dust Loading
4800 mm height x 6800 mm width x 50000 nm length
# 24 color steel plate
stainless-steel plate
stainless-steel wires supported by steel frames
high voltage power pack: 90 kw
motors, etc.: 11 kw
0.6 m3/min. (intermittent cleaning)
17100 - 27800 m3/min.
35 deg C
0.270 g/Nm3
0.047 g/Nm3
Table 2 Example of Particle Size Distribution of Converter Dust
*^\^
A Co.
B Co.
C Co.
^ 10 urn
15.5
4
13.5
10 ^ 20 pm
19.5
11
30.5
20 % 50 ym
50
53
42.0
50 urn ^
11
32
14.0
true specific
.gravi ty
3.49
4.27
3.94
Table 3 Results of Chemical Analysis of Converter Dust
A Co.
B Co.
C Co.
T. Fe
29.6
45.07
32.92
Si02
8.5
8.76
6.92
A12°3
3.35
3.30
2.50
T. S
1.1
0.154
1.23
MgO
1.45
1.32
5.57
Na20
1.65
0.32
0.57
K20
0.17
0.17
0.16
CaO ZnO
22.85 0.
8.91 1.
13.88 0.
PbO
7 0.
99 0.
57 0.
T.C
15 4.85
36 9.93
09 2.16
MiO MnO
-
0.02 -
- 4.44
Table 4 Electrical Resistivity of Converter Dust
^\^
A Co.
B Co.
C Co.
30 deg C
6.78 x 10^1 n-cm
5.3 x 105 n-cm
7.31 x 1012 n-cm
60 deg C
6.83 x 1012 n-cm
9.6 x 105 n-cm
9.11 x 1012 si-cm
90 deg C
1.04 x 1Q13 fl-cm
5 x 10 6 fi-cm
1.07 x 1013 n-cm
difficulty in
collection
normal
easy
normal
313
-------
Table 5 Design and Operation Data of REP for Converter Shop
Installations
A
B
C
D
Design
Actual
Design
Actual
Design
Actual
Design
Actual
Total
Gas Vol.
m /min
24000
13600-
28600
43800
51000
30900
42400
27000
9800-
22800
Gas Vol. per
Converter
m /min
8000
4500-
9500
14600
17000
7700
10600
13500
4900-
11400
Gas Vel .
m/s
1.77
1.0-2.1
1.78
1.9
0.8
1.1
1.66
0.6-1.4
Inlet
Dust
Loading
q/Nm3
0.1
0.269
(max. )
0.4
0.33
0.4
0.35
0.4
1.09
(max.)
Outlet
Dust
Loading
q/Nm3
0.02
0.047
0.03
0.02
0.02
0.02
0.03
0.038
_
Remarks
LD
80 t/ch.x 3
LD 250 t/ch.
x 3
LD 160 t/ch.
x 3
Q-BOP 230 t/
ch. x 2
Remarks: SHI-Type REP's
Table 6 Design and Operation Data of REP for Electric Furnace
Installations
A
B
C
D
Design
Actual
Design
Actual
Design
Actual
Design
Actual
Total
Gas Vol.
m /min
8200
10700
4900
4600
5600
6800-
9000
4000
3700
Gas Vol. per
Electric F.
m /min
8200
10700
4900
4600
2800
3400-
4500
2000
1950
Gas Vel.
m/s
1.0
1.3
1.0
0.94
1.0
1.2-1.6
1.0
0.92
Inlet
Oust
Loadina
q/Nm3
0.2
0.2-0.4
0.3
0.15-0.29
0.3
0.16-0.42
0.3
0.1-0.16
Outlet
Dust
Loading
g/Nm3
0.02
0.01-0.03
0.03
0.01-0.02
0.03
0.01-0.03
0.03
0.01-0.015
Remarks
100 t/ch
x 1
,.,
30 t/ch.
x 1
•
18 t/ch.
x 2
• —
8 t/ch.
x 2
Remarks: SHI-Type REP's,
314
-------
Table 7
Comparison of Power Consumption between REP and Bag-House
(Converter Shop)
x
A Co.
B Co.
C Co.
D Co.
Gas Volume
3
(m /min)
24000
43800
30900
27000
Power Consumption (kw)
Bag-House
2800
5100
3600
3100
REP
41
150
170
90
Table 8
Comparison of Power Consumption between REP and Bag-House
(Casting Floor of Blast Furnace)
\
A Co.
Gas Volume
(m /min)
35000
Power Consumption (kw)
Ban-House
3500
REP
240
Table 9
Comparison of Power Consumption between REP and Bag-House
(Electric Furnace Shop)
\
A Co.
B Co.
C Co.
D Co.
E Co.
F Co.
G Co.
H Co.
I Co.
J Co.
Gas Volume
2
(m /min)
4000
8200
4900
2440
5600
2670
2450
1100
1400
38000
Power Consumption (kw)
Bag-'House
930
1900
1120
560
1290
620
560
1250
1250
8750
REP
50
100
63
40
38
40
45
25
35
330
315
-------
Table 10 Comparison of Total Energy Consumed in Ten Years
in Case A Bag-House Is Replaced by REP
(30 t/ch. Electric Furnace)
\^
Power Consumption
Production of Materials
Item
High Voltage
Powtr Supply
Suction Fan
Accessories
Total
Steel
(85 t)
Collector
Plastics
(13 t)
Steel for Reinforce-
ment- of Roof (39t)
Total
Bag-House
—
1100kwx7200h/yxlOy =
79,200,000 kw
10kwx7200h/yxlOy =
720,000 kw
79,920,000 kwh =
195.8 x 109 kcal
—
•
-
-
REP
10.7kwx7200h/yxlOy =
770,400 kwh
20kwx7200h/yxlOy =
1,440,000 kw
2.210,4000 kwh =
5.4 x 109 kcal
85txll.2xlOt)kcal/t=
952x1 O6 kcal
13txl3.0xlObkcal/t =
Ibyxlu6 kcal
39txll.2xl06kcal/t=
437x1 O6 kcal
1,558 x 106 kcal
Table 12 Number and Nominal Capacity of Electric Furnaces and
Converters in Japan (Dec. 1976)
Electric Furnace
Nominal Capacity (t)
5 - 18
20 - 35
40 - 60
70 -
Aver. Norn. Cap.
27 t
Number
185
92
73
28
Total
378
Converter
Nominal Capacity (t)
40 - 90
100 - 180
200 - 270
300 -
Aver. Norn. Cap.
152 t
Number
38
34
25
7
Total
104
316
-------
Table 11 Comparison of Total Energy Consumption in Ten Years
between Bag-House and REP
(160 t/ch. Converter)
Power Consumption
Production of Materials
Item
High Voltage
Power Supply
Suction Fan
Accessories
Total
Collector
Steel
Plastics
Suction Fan & Duct
System (126 t)
Steel for Reinforce-
ment of Roof (54 t)
Concrete (125 m3)
Total
Bag-House
—
1700kwx8400h/yxlOy =
142. 8x1 O6 kw
20kwx8400h/yxlOy =
1.63xl06 kw
144. 48x1 O6 kw =
354x1 O9 kcal
450txll.2xT06kcal/t =
5,040xl06 kcal
24txl3x106kcal/t =
312xl06 kcal
126txll.2xl06kcal/t =
I,411.2x106 kcal
-
125ra3x4.8xl06kcal/m3 =
600x1 O6 kcal
7,363.2xl06 kcal
REP
23kwx8400h/yxlOy =
1.9 32x1 O6 kw
—
20kwx8400h/yxlOy =
1.63xl06 kw
3. 61 2x1 O6 kw
8.8xl09 kcal
115txll.2xl06kcal/t =
1, 288x1 O6 kcal
20txl3xl06kcal/t =
260x1 O6 kcal
.
54txll.2xl06kcal/t =
604. 8x1 O6 kcal
-
2,152.8xl06 kcal
317
-------
t
MATER SUPPLY
REP
WATER TREATMENT
'/////s/////////
BAG FILTER
Fin. 1 System of ROOF MOUNTED TYPE ESP
t
Witer supply unit
Insulator ehtmbir
Spray norilt
Curtain
Drain
Detail of Hopper
F1g. 2 Construction of ROOF-MOUNTED TYPE ELECTROSTATIC
PRECIPITATOR (SHI-Type)
313
-------
Fig. 3 ROOF-MOUNTED TYPE ESP for Converter Shop
(SHI-Type)
1 waste water
11
9 10
2 chemical additive (pH-control)
3 stroage and neutralization 4 sedimentation tank
tank
5 chemical additive
7 recycling water for spray
9 sludge storage tank
11 cake
6 netralization tank
8 final effluent
10 hydroextractor
Fig. 4 Waste Mater Treatment for REP
319
-------
Fig. 5 ROOF-MOUNTED TYPE ESP (SMM-Type)
CLEAN GAS
GAS CONTROL VANE
MONITOR
COLLECTING ELECTRODE
DISCHARGE ELECTRODE
WORKSHOP
Fig. 6 Construction of ROOF-MOUNTED TYPE ESP (SMM-Type)
-------
£ 15,000 •
E
N
cr
i 10,000
(O
CJ
^ 5,000
u
-------
FUGITIVE HYDROCARBON EMISSIONS FROM AN
IN-SITU OIL SHALE RETORT
Gerald M. Rinaldi and David R. Tierney*
Monsanto Research Corporation
Station B, Box 8
Dayton, Ohio 45407
Oil shale has been recognized as a potentially substantial energy resource in
the United States for more than a century. An emerging technology for shale
oil production is in-situ processing, in which the shale bed is hydraulically
or explosively fractured and retorting is carried out underground. In order
to assess potential environmental impacts from in-situ processing, data were
collected on fugitive emissions from a pilot-scale retort producing 30 barrels
of crude shale oil daily. Fugitive gas seepage through the retort surface and
from around instrumentation well casings was measured using a specially designed
sampling system in conjunction with gas chromatographic and Orsat analysis.
Total hydrocarbons, Ci through C6 hydrocarbon fractions, and carbon monoxide
were quantified in the fugitive emission samples. Normalized fugitive hydro-
carbon emission rates due to seepage through the overburden ranged from 0.2 to
18 g/m2/hr, with an average of 5.5 g/m2/hr. The total hydrocarbon emission rate
due to ground seepage was calculated to be 5.7 kg/hr, using the retort surface
area of 1,043 m2; fugitive hydrocarbon emissions due to leakage around well
casings amounted to an additional 0.3 kg/hr.
*Present Address: Clarkson College of Technology,
Potsdam, New York.
322
-------
FUGITIVE HYDROCARBON EMISSIONS FROM AN
IN-SITU OIL SHALE RETORT
Introduction
Oil shale has been recognized as a potentially substantial energy resource in
the United States for more than 100 years. Rapidly escalating oil prices have
provided new incentive for shale oil recovery from deposits in Colorado, Utah,
and Wyoming. At least four U.S. firms (Colony Development Operation, Paraho
Development Corporation, Superior Oil Company, and Union Oil Company) have
developed surface retorting processes, in which oil shale is mined and crushed
prior to thermal processing in above-ground facilities. The costs associated
with mining and handling voluminous quantities of raw shale and the environ-
mental impact of spent shale disposal may limit commercial application of this
technology.1 Therefore, a good deal of research is currently being directed
toward the development of in-situ or modified in-situ retorting processes.
This technology, in which the shale bed is hydraulically or explosively frac-
tured and retorting is carried out underground, may offer a cost-effective solu-
tion to the shale handling and disposal problems associated with surface
retorting.1 In-situ or modified in-situ processes are now being developed by
Equity Oil Company, Geokinetics, Inc., Rio Blanco Oil Shale Company, and
Occidental Oil Shale, Inc.
Since in-situ oil shale retorting is an emerging energy technology, its poten-
tial environmental impacts are largely unknown. Under Contract No. 68-03-2550
for the U.S. .Environmental Protection Agency (EPA), a field sampling program
was conducted to identify and quantify emissions and effluents resulting from
horizontal in-situ oil shale retorting. The technical description and results
presented herein are concerned only with the measurement of fugitive emissions
from a pilot-scale retort, located on the Kamp Kerogen site in Uintah County,
Utah, and operated by Geokinetics, Inc. The shale bed was rubblized in May
1978 and ignited in April 1979; field sampling was conducted from July 16 to
July 26, during which time the crude shale oil production rate was approxi-
mately 30 barrels per day.
The Geokinetics horizontal in-situ retorting process begins with the drilling
of blast holes from the surface, through a maximum of 46 m of overburden, and
into the oil shale bed.2 The holes are then loaded with explosives, which
are fired to yield a rubblized mass of oil shale with increased permeability.
During fragmentation, the surface undergoes noticeable uplift. A slope is
created below the oil shale bed, allowing shale oil to drain to a sump for
recovery by production wells, as shown in Figure 1.
323
-------
AIR INJECTION WQi
LIQUID
PRODUCTION WELL
RETORT
•OFF-GASES
OIL/WATER
MIXTURE
SUMP
Figure 1. Sectional view of Geokinetics horizontal in-situ oil shale retort.
-------
Experimental Methods
The uplift of the retort overburden, due to shale deposit fragmentation,
causes ground cracks of various sizes. Plant personnel routinely seal major
cracks by filling with mud and tamping before retorting begins, but escape of
gases from the burning retort is not entirely eliminated. However, because
it was Geokinetics' understanding that the purpose of this fugitive emissions
study was characterization, not quantification, only those leaks that presented
an obvious health hazard were sealed during the sampling period. Additional
sources of fugitive emissions are created by drilling and casing instrumentation
wells into the oil shale bed. It has been estimated that one-third to one-half
of the total volume of gas injected into the in-situ retort is not recovered at
the outlet wells.3 A 30 kw (40 hp) vacuum blower, capable of reducing, if not
altogether preventing, such gas seepage, was not in operation during the sampling
period.
In order to obtain representative data, test locations for fugitive gas samp-
ling were randomly selected, with the constraint that an approximately equal
number of sites be placed in each of four quadrants of the surface area of the
retort. Locations for sampling ground seepage were selected by using a random
number table corresponding to consecutively numbered squares on a grid super-
imposed over a scale diagram of the retort. Twenty-seven of these potential
fugitive emission sources were tested, non-simultaneously, along with six
thermocouple wells and one of two shale oil production wells.
The sample collection method used in this effort was developed by combining
the applicable features of techniques previously used extensively in the
measurement of fugitive emissions from sources as diverse as vegetation and
petrochemical processing equipment.4'5 Prior to sampling, selected test
locations were enclosed using an aluminum box, 0.61 m square, having an open top
and bottom. The aluminum box was driven into the soil and anchored at all four
corners with metal rods; moist soil was used outside the bottom of the sampling
box to provide a seal to the retort surface. The top of the box was covered with
flexible plastic, attached to the sides by duct tape.
Figure 2 is a schematic diagram of the fugitive emission sampling system
used at the Geokinetics1 oil shale retort. The sample box outlet was
connected, via a three-way valve, to a 500-mL glass syringe and to a vacuum
pump. For approximately five minutes prior to sample collection, the pump
was used to evacuate the sample box, after which the pump was disconnected
and replaced with a Teflon sample bag. The three-way valve was then adjusted
so as to seal the sample box off from both the syringe and the sample bag.
The aluminum box was next filled with "zero-grade" air from a pressurized
cylinder at a measured rate of 10 liters per minute, causing the plastic cover
sheet to rise. When the box was partially inflated, a sample was obtained
by using the syringe to withdraw a portion of the gas containing fugitive
emissions and, after turning the three-way valve, to inject this into a Teflon
bag. Field "blanks", consisting of samples of zero-grade air which were
325
-------
Ul
NJ
Al
REGUL
1=3 GX
n
R 1
ATOR |
? u
I) 1 » 1
1
\^4
ORIFICE 'T
METER
AIR ROW
/'CONTROL VALVE
1
\
w.
MANOM
ETER
1
L|
I
i
1 FLEXIBLE
PLASTIC COVER
A/ A
"N /
SAMPLE BOX
ZERO-GRADE ^ RETORT SURFACE
IR CYLINDER
3-WAY
VALVE
A
1_
500 ml SYRINGE
... J I i
1 1 '
_ / /^AMPIFRAft
-•Q— ^^
VACUUM
PUMP
Figure 2. Fugitive emission sampling system.
-------
passed through the flow control valves, the orifice meter, and all connecting
lines, but not the sample box, were also collected using the glass syringe
and Teflon bags.
Samples contained in the Teflon bags were analyzed on-site for total hydro-
carbons and GI through Cg hydrocarbon fractions using a portable gas chromato-
graph with a flame ionization detector. Selected fugitive emission samples
were also tested for carbon dioxide (C02), carbon monoxide (CO), and oxygen
(02) by Orsat analysis.
Results
The rates of fugitive gas seepage from the retort surface and from around well
casings were determined using the following equation, which is based on the ideal
gas law (and which includes a numerical constant accounting for conversion of
mm Hg to atmospheres, liters to cm3, and minutes to hours):
E = 7.895 X 10"5 HP G(C - C, )/RT (1)
S S D S
where E = fugitive emission rate, g/hr
M = molecular weight of emitted material, g/g-mole
Ps = atmospheric pressure during sampling, mm Hg
G = cylinder gas flow rate, liters/min
C_ = concentration in sample bag, ppm
= concentration in field blank, ppm
= ideal gas constant =82.1 cm3«atm/g-mole* °K
T = ambient temperature during sampling, °K
The calculated hydrocarbon emission rates for tested sources are given in
Table I. Emission rates for ground seepage sources were normalized by
dividing by the cross-sectional area of the sample box, equal to 0.37 m2.
The average hydrocarbon emission rate for ground seepage was 5.5 g/m2/hr,
with a range of 0.2 to 18 g/m2/hr. In Figure 3, the distribution of measured
hydrocarbon emission rates over the surface of the retort is shown. The total
fugitive hydrocarbon emission rate of 6.0 kg/hr was calculated by adding the
product of the average ground seepage rate (5.5 g/m2/hr) and the retort surface
area (1,043 m2) to product of the average emission rate from wells (5.3 g/hr)
and the total number of wells (46).
Additional analyses were conducted to determine the general composition of
hydrocarbons in the fugitive emission samples. Table II gives the concentra-
tions of G! through C6 hydrocarbon fractions for selected bag samples. Orsat
analyses of selected samples are given in Table III.
327
-------
Table I. Fugitive hydrocarbon emission rates.'
Sample type
and
run number
Ground seepage:
G-l
G-2
G-3
G-4
G-5
G-6
G-7
G-8
G-9
G-10
G-ll
G-12
G-13
G-14
G-15
G-16
G-17
G-18
G-19
G-20
G-21
Hydrocarbon
emission rate
g/m2/nr
2.9
0.4
14
11
15
5.7
°'3b
0.7D
3.4
6.8
18
10
5.6
7.0
1.6
0.9
8.5
4.9
2.4
0.3
0.2
Sample type Hydrocarbon
and emission rate,
run number g/m2/hr
Ground seepage
(continued):
G-22
G-23
G-24
G-25
G-26
G-27
Liquid produc-
tion well:
P-l
Thermocouple
wells:
T-l
T-2
T-3
T-4
T-5
T-6
0.5
3.3
7.0
2.9
8.0
2.4
7.8b
c
4-8n
3-7?
6-*H
5-4n
3.1b
Emission rates per unit area determined by dividing hourly rate by
sample box area of 0.37 m2.
Emission rate given in g/hr (sample box not used).
thermocouple well sample No. T-l was not analyzed.
Discussion
The total fugitive hydrocarbon emission rate of 6.0 kg/hr is significant when
compared to the 17 to 19 kg/hr of hydrocarbon vapors contained in the retort's
off-gas outlet wells, the latter quantities having been determined in the
overall study of Geokinetics1 technology.6 In order to make a preliminary
assessment of the impact of fugitive hydrocarbon emissions on ambient air
quality in the vicinity of retort, dispersion modeling, according to the
method of Turner,7 was used to calculate ground-level hydrocarbon concentra-
tions. The following assumptions were made: 1) the emissions originate from
328
-------
10
vO
D
O
O
»
k
9 • ; a
A / APPROXIMATE
| DIRECT ION OF / BURN FRONT
BURN FRONT / LOCATION
MOVEMENT /
m \
\ ^ o
\e °
>
o
^RETORT BOUNDARY
METERS
H SYMBOLS:
< 0. 5 g/m/hr
«
0. 5 - 2. 5 g/m'/hr
i 6 - 5. 0 g/m/hr
5. 1 -
> 10 9/mZ/hr
Figure 3. Pictorial representation of magnitude of fugitive hydrocarbon
emissions due to seepage through retort surface.
-------
Table II. Concentrations of Ci-C6 hydrocarbon fractions
in selected fugitive emission samples.
Sample type
and
run number
Concentration, ppmv
C2
C6
Ground seepage:
G-4 1,700 360 100 20 50 5
G-ll 2,900 650 240 160 90 8
G-14 1,200 330 100 80 20 3
G-17 1,700 480 190 110 50 7
G-24 1,300 360 120 110 50 4
Liquid produc-
tion well:
P-l 740 240 90 40 20 1
Thermocouple
well:
T-2 810 200 60 50 20 3
Table III. Orsat analysis of selected fugitive emission samples.
M^ —— — •»_•• ^^^^^ ^.^^^ ^
Sample type
and
run number
Ground seepage:
G-93h
G-13a
G-183
G-24a
Concentration
Carbon
dioxide
(C02)
0.2
5.0
4.0
8.0
(dry basis)
Oxygen
(°2>
10.6
7.0
7.2
5.8
i , percent by volume
Carbon
monoxide
(CO)
1.0
1.1
0.8
0
Carbon monoxide
emission rate
q/m2/hr '
21
23
17
0
Thermocouple
well:
T-4b
3.6
7.4
1.8
Concentrations are averages of duplicate analyses of sample.
Concentrations are averages of triplicate analyses of sample.
cEmission rate given in g/hr (sample box not used).
330
-------
a "point source," not strictly correct in this case but proven by calculations
to be adequate for order-of-magnitude estimation,- 2) the atmospheric stability
is represented by Class C; and 3) the averaging time is three hours, that for
the ambient air quality guideline for hydrocarbons. In addition, a wind speed of
3 m/s, equivalent to the national average but corrected from a height of 7 m to
1 m, was utilized. The calculated results are meant to approximate concen-
trations, not to replace ambient air monitoring, because the accuracy of
Turner's method is limited to within a factor of two or three due to empirical
determination of dispersion constants.
Table IV is a matrix of downwind ground-level hydrocarbon concentrations
calculated using the total fugitive emission rate of 1.7 grams per second.
Assuming the "point source" of fugitive emissions is located at the mid-point
of the downwind boundary of the retort, Table IV indicates that, within a
limited area nearby, the hydrocarbon concentration exceeds the ambient air
quality guideline of 160 (jg/m3. During the three-hour averaging time, this
phenomenon occurs only within approximately 20 meters of the "plume" center-
line and up to approximately 300 meters downwind of the retort. Extrap-
olating from the pilot-scale retort to a commercial facility operated at a much
greater shale oil production rate, the mass flow rate of fugitive hydrocarbon
emissions is expected to be greater, and the ambient air quality effects
would thus extend to a larger area than that suggested by Table IV.
Table IV. Downwind ground-level concentrations due to fugitive
hydrocarbon emissions from Geokinetics retort.
Concentration3, (jq/m3
b
x,
meters
10
20
50
100
200
500
1,000
2,000
c
y, meters
0
69,000
20,000
3,700
1,100
300
60
20
5
10
_d
110
1,400
800
280
60
20
5
20
.
-
70
350
220
50
20
5
50
-
-
1
40
40
10
4
100
-
-
-
-
10
10
4
200
-
-
_
_
_
3
3
Calculated using the total fugitive hydro-
carbon emission rate of 1.7 g/s.
Downwind distance along the "plume"
centerline.
Horizontal distance from the "plume"
centerline.
T)ash indicates calculated concentration was
less than 1 (jg/m3.
331
-------
Acknowledgements
This research was funded by the U.S. Environmental Protection Agency (EPA)
under Contract No. 68-03-2550. The contents of this publication do not neces-
sarily reflect the views or policies of the U.S. Environmental Protection
Agency, nor does mention of trade names, commercial products, or organizations
imply endorsement by the U.S. Government. Both MRC and EPA would like to
acknowledge the contributions of Hilding Spradlin and Steve Mankowski, Geo-
kinetics1 environmental engineers, whose assistance was invaluable during
preparation for and implementation of this sampling and analysis program. The
role of David Farrier and Trudy Phillips of the Department of Energy's Laramie
Energy Technology Center (LETC) in allowing MRC to conduct field sampling at
the Kamp Kerogen site is also greatly appreciated.
332
-------
References
1. Baughman, G. L. Synthetic Fuels Data Handbook, Second Edition. Cameron
Engineers, Inc., Denver, Colorado, 1978. 438pp.
2. Lekas, M. A. Progress Report on the Geokinetics Horizontal In-Situ
Retorting Process. In: Twelfth Oil Shale Symposium Proceedings, Colorado
School of Mines, Golden, Colorado, August 1979. pp. 228-236.
3. S. G. Mankowski, Geokinetics, Inc., Kamp Kerogen, Utah, private
communication, 1979.
4. Zimmerman, P. Procedures for Conducting Hydrocarbon Emission Inventories
of Biogenic Sources and Some Results of Recent Investigations (paper pre-
sented at the EPA Emission Inventory/Factor Workshop, Raleigh, North
Carolina, September 1977). 28 pp.
5. Tierney, D. R., Z. S. Khan, and T. W. Hughes. Measurement of Fugitive
Hydrocarbon Emissions from Petrochemical Plants (paper presented at the
Third Symposium on Fugitive Emissions-. Measurement and Control, San
Francisco, California, October 23-25, 1978).
6. Rinaldi, G. M., J. L. Delaney, and W. H. Hedley. Environmental Characteri-
zation of Geokinetics1 In-Situ Oil Shale Retorting Technology (draft report)
U. S. Environmental Protection Agency, Cincinnati, Ohio, February 1980.
239 pp.
7. 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.
333
-------
A WIND TUNNEL STUDY OF FUGITIVE DUST FROM
TACONITE STORAGE PILES
Dr. Robert B. Jacko P.E.
Associate Professor of Environmental
Engineering and Health Sciences
School of Civil Engineering
George M. Palmer, Ae.E.
Associate Professor of Aeronautical
Engineering and Director of the
Aerospace Sciences Laboratory
School of Aeronautical and Astronautical
Engineering
Purdue University
West Lafayette, Indiana
ACKNOWLEDGMENT
Sponsorship of this project was provided by Mr. Richard A. Young of Young
Environmental Services, Glenview, Illinois for which we express our appreci-
ation.
ABSTRACT
A comparative wind tunnel study of the fugitive emissions from a taconite
pellet test surface conducted at Purdue University indicates that one of the
three commercially available dust suppressing agents tested at a dilution
ratio of 1 to 1000 was 33% more effective than the least effective agent.
Water only was also found to be as effective as the dust suppressing agents.
However, water and dust suppressing agents resulted in 3 times higher parti-
culate emissions than that of a dry pile. Both dry and wet transient testa
elucidated the nature of the particulate release. For the wetted pellets,
subsurface drying appears to "mobilize" interstitial and surface pellet mater-
ial which resulted in a higher release rate of particulate matter in time as
compared to a dry pellet.
334
-------
INTRODUCTION
This report summarizes the results of a Purdue University study which in-
vestigated the effectiveness of various binding agents to control fugitive
dust emissions from taconite surfaces. The three agents tested were composed
of hi-molecular weight polymers and synthetic resins. They were all diluted
1:1000 with Indiana tap water. At the outset, it should be understood that
this study is not a modeling study of a full size taconite storage pile.
Rather, this study is a purely comparative study with the objective of deter-
mining which of three chemical dust suppressing agents and also water do the
best job of reducing fugitive particulate emissions from taconite pellet sur-
faces. The study is in reality a comparative parametric study utilizing a
large wind tunnel and other closely controlled procedures. All taconite pre-
paration procedures and wind tunnel velocity were closely controlled in order
to discern the downwind particulate concentrations subsequent to application of
the various chemical dust suppressants.
WIND TUNNEL DESCRIPTION
The subsonic wind tunnel at the Purdue University, Aerospace Sciences Lab-
oratory is shown in schematic detail in Figure 1. The tunnel is a conventional
one with a 10-1/2 to 1 contraction ratio and is powered with a 400 H. P.
electric motor. Speed control is obtained by means of a rheostat-controlled
electric clutch permitting a tunnel speed range from 10 mph to 300 mph in the
high-speed, closed test section.
The present test program was conducted in the large open-throat test
section at the end of the 1st diffuser by means of a specifically designed and
constructed assembly described in Figures' 2 and 3 and shown photographically in
Figure 4. This provided a flow at a nominal 1/2 of the high-speed section so
speeds could be varied and controlled from approximately 5 mph to 150 mph as
needed.
The tests were conducted at a fixed velocity profile in the open-throat
section. The speed was monitored by means of a pitot-static tube mounted up-
stream of the test surface, see Figures 2, 3 and 4.
Figure 5 contains the vertical profiles of mean wind velocity and turbu-
lence intensity. Note that over the pile height the wind velocity varied from
40 mph at the top to 20 mph near the base. Turbulence intensity varied from
2% at the top of the pile to 5.5% near the base. These levels of velocity and
turbulence intensities are typical of those expected in relatively flat terrain
such as would be expected from off-shore breezes which do exist near the full
scale taconite piles.
335
-------
r
rT\
OJ
U-
c-
FIGURE 1. LAYOUT AND ELEVATION OF THE AEROSPACE SCIENCES LABORATORY SUBSONIC WIND TUNNEL TOP VIEW.
-------
(AIR SAMPLING
NOZZLE INLET
DISTANCE FROM
PILE CENTERLINE)
fM
AIR FLOW
PITOT-STATIC TUBE
ROUGHNESS ELEMENTS)
-26.5
(WIND TUNNEL FLOOR)
Figure 2. Schematic Cross Section of Test Assembly.
-------
Oo
fcO
CO
LU
-»- AIR SAMPLING NOZZLES
0 PITOT-STATIC TUBE
TEST
SAMPLE -
TRAY
AIR FLOW
Figure 3. Schematic Plan View of Test Assembly in Wind Tunnel.
-------
(a)
(b)
v'eure 4. (a) Looking Upwind at Taconite Test Surface from Rear Sampling
* ° »T 1
Nozzle
(b) Side View of Taconite Test Surface with Flow Left to Right
339
-------
2.0
o
tH
01
-------
TEST ASSEMBLY
In order to facilitate testing for taconite fugitive dust, an assembly for
into the wind tunnel was designed and constructed. Figures 2 and 3 are
schematics of the assembly installed in the large, low-speed, open test section
of the wind tunnel.
The assembly was constructed of plywood, masonite, and construction lum-
ber. The 30° rise, see Figure 2, was constructed from untempered masonite
fastened to 5 plywood ribs. This provided a stiff (non pulsing) surface with
a smoothly curved transition to the horizontal surface. The horizontal sur-
faces were constructed with sufficient strength and stiffness to permit 2 men
walking on them.
The two dashed lines label plus and minus 26.5° in Figure 2 show the
hinged panels in the two positions for flat-tray attitudes of plus and minus
26.5°. The back panel lays along the tunnel floor for the minus 26.5° po-
sition.
Both Figures 2 and 3 show the assembly to scale in the horizontal po-
sition with roughness elements, instruments, and test pile location.
Figure 4a is a top view looking upstream from the last air sampling noz-
zle. Figure 4b is a side view of the conical test surface. Flow is left to
right.
TEST PROCEDURES
Taconite Test Surface
The basic approach in this study was to present a taconite test surface to
a constant velocity air stream while measuring upstream and downstream partic-
ulate concentrations. Initially it was proposed that the test surface be com-
posed of a planar array of taconite pellets that could be presented at three
angles to the tunnel wind flow. The planar surface when presented at a +26.5°
angle to the wind would represent a section of the upwind face of a storage
pile. When presented at a -26.5° angle to the wind the planar surface would
represent the downwind face of the pile. At Cf angle the surface would repre-
sent top and edges of the storage pile. Preliminary wind tunnel tests indicat-
ed that the angle of incidence made little difference in the downwind particu-
late concentrations. All subsequent tests were therefore made in a horizontal
mode (0° angle to the wind). In addition, the absolute downwind particulate
concentration level was lower than desired for minimization of errors in the
gravimetric analysis. Therefore the test surface was changed from planar to
conical.
The test surface was made from taconite pellets supplied from the Michigan
Upper Peninsula. A size distribution was determined for the pellets upon
arrival at Purdue. It was found that virtually no sizes less than 1/4" were
341
-------
present in the received pellets. From previous taconite pellet size distribu-
tions it was known that copious fine dust does occur. Therefore sufficient
taconite was pulverized in the soils laboratory at Purdue so that 2.5% by
weight less than 149 micron dust could be added to the received pellets. A
two step procedure was utilized to form the fine dust. The pellets were first
placed in a ball mill for approximately one hour, then through a pulverizer.
The ball mill was a "Los Angeles Type Abrasion Machine" manufactured by F. M.
Welsh Engineering Service of Greenville, Ohio and the type UA pulverizer was
manufactured by F. W. Braun Co., of Los Angeles, California. Table 1 below
shows the resulting size distribution.
Table 1 Taconite Pellet Size Distribution
Particle Size (S) % by Weight
3/4" < s 0.670 (as received)
1/2" < S < 3/4" 34.440 " "
3/8" < S < 1/2" 61.400 " "
1/4" < S < 3/8" 0.960 " "
(Added to the pellets in the laboratory)
590y < S < 1/4" 0.008
296y < S < 590y 0.024
I49y < S < 296y 0.493
75y < S < I49y 0.261
S < 75y 1.750
Test Surface Preparation Procedure
The size distribution in Table 1 was followed in the making of all conical
test surfaces. Following is the exact procedure utilized:
1) Place 2' x 2' x 2" deep wood test tray on 800 Ib. capacity Toledo®
scale (31-1821-FD). Adjust scale to zero weight.
2) Use Fairbanks-Morse beam balance scale (0-250 Ibs., Model 505) to
weigh individual size fractions based on 100 Ibs. and pour onto test
tray. After weighing incremental fractions of 100 Ib. on smaller beam
scale, check larger Toledo scale weight to ensure a sample weight of
100 Ibs. has been achieved. However, do not yet disperse less than
149 micron dust on tray.
3) The pellets were then evenly spread over a specially built wooden
trough 18" wide x 10* long x 1.75" deep. The horizontal trough was
supported 11-1/2" above the floor. The purpose of the trough was tb~~
present each pellet for evenly distributing the less than 149 micron
dust as well as for spraying of suppressing agent.
4) Sieve the less than 149 micron size fraction of particles evenly over
the pellets in the trough.
5) Using a Binks^Model 370 external mix spray gun at an inlet delivery
pressure of 40 psig, evenly spray the pellets in the trough with 500
ml of the selected dust suppressing agent.
6) Using a square wooden stick 1" x 1" in cross-section and 12" long,
a furrow was ploughed through the pellets along the trough centerline
returning back along the trough between the outer wall and trough
342
-------
centerline on each side.
7) The pellets are once again spread evenly over the trough. This pro-
cedure exposes dry sides of pellets for more thorough wetting.
8) Proceed to spray another 500 ml of dust suppressing agent on the now
turned over pellets.
9) Using a 500 ml, polyethylene scalable bottle collect a sample of
pellets for moisture determination.
10) Using 8" wide x 10" long metal sheets (0.040" thick) scoop the pellets
and carefully pour them into the center of the test tray. In this
way, the pellets, evenly dispersed with less than 149 micron particles
and wetted seek their 26.5° angle of repose in the tray forming a
conical test surface.
11) Any fine particulate remaining in the trough is brushed into a sieve
and evenly spread over the conical test surface.
12) The conical taconite pellet test surface was then positioned in the
wind tunnel. The elapsed time between wetting and test startup was
kept consistently at 10 minutes (conical piling to test startup was
only 2 min.).
Particulate Sampling Procedures
In order to quantify the differences between the three chemical dust
suppressing agents, A, B, C and water as well as dry test surfaces of taconite,
isokinetic particulate samples were taken. During each test, except the trans-
ient tests, four simultaneous isokinetic samples were extracted using stainless
steel 1/4" inside diameter nozzles leading to 47 mm diameter filter holders.
Gelman type AE glass fiber filters were used for all tests. All filters were
handled and analyzed using standard gravimetric laboratory procedures.
One sampling nozzle was located upstream of the conical taconite test sur-
face centerline, refer to Figure 2 and 3. The upstream sample served as a
"blank" to correct the downstream samples for particulate concentration. All
samples were withdrawn isokinetically at 1.2 ft. 3/min. Adjacent to the up-
stream sampling nozzle a standard pitot-static tube was located for measuring
tunnel speed and maintaining isokinetic conditions. The pitot tube as well as
all sampling nozzles were offset (staggered) from the tunnel and conical test
surface centerline so none was in the direct downstream wake flow of the near-
est upstream nozzle.
Air Flow Instrumentation
The air samples were taken using 3 EPA Method 5 type control cases (Joy
Manufacturing Co.) and one diaphram pump and a rotameter. The method 5 type
control cases are equipped with a leak-free vacuum pump, dry gas meter, flow
control valves and an orifice rate meter. The orifice meter was used to set
up the isokinetic sampling rate. Each control case was previously calibrated
against wet test meters in the School of Civil Engineering's Air Pollution
Measurements Laboratory. The fourth pump and rotameter were also calibrated
prior to use. All air volumes measured were corrected to standard conditions.
343
-------
Specific Test Procedures - 60 min. Samples
With the wind tunnel having been previously "warmed up," the drive clutch
was disengaged and the tunnel was brought to its lowest possible speed of about
2 mph. The conical taconite test surface was placed in position and the sam-
pling system started. The tunnel clutch was then engaged and air speed was
brought to 40 mph at pitot tube height in approximately one minute. Sampling
continued for 60 minutes and therefore all data except the transient test data
are one hour time average samples. The filters were stored in glass petri
dishes and the nozzles were quantitatively backwashed with acetone. The total
particulate sample weight was therefore composed of the filter weight gain
portion and the nozzle backwash portion.
Transient Test Procedures
Everything for the transient tests were identical to the one hour time
average samples except the filter heads were exchanged at the second downwind
position with time. This involved recording volumes, recovering filters and
backwashing nozzles in a very closely coordinated effort. In this way, the
particulate concentration variation with time was determined.
RESULTS AND DISCUSSION
Steady State Measurements
The results from the one hour sampling time test series are presented in
Tables 2, 3 and 4. Each table presents the concentrations in milligrams per
cubic meter for each of the three downwind positions. The overall space and
test-wise average is shown in a box to the lower right of each table. Note
that copious replicate data is contained in Tables 2, 3 and 4. Nine parti-
culate concentration values went into the dry taconite surface test average
twenty-seven into the dust suppressent tests and 13 for the water tests.
In addition, the various tests were conducted such that systematic errors
would be minimized. For example, the test runs were mixed in time so that
systematic errors were minimized. The test run numbers increase with calen-
dar time. Test run one corresponds to Sunday, March 2, and run 18 to Friday
March 28, 1980.
In viewing the data, it becomes very apparent that the overall average
downwind particulate concentration is lowest for the dry conical test surface
The dry surface resulted in a 2.63 mg/m3 downwind average concentration while
the test surfaces treated with suppressing agents averaged 7.80 mg/m3. The
dust suppressing agents were all mixed at a dilution ratio with fresh tap water
of 1 in 1000. The amount of agent added to the pile resulted in a 1.1% pellet
moisture content by weight.
There appears to be a trend regarding the most effective suppressing
agent, however. Note in Table 3 that the use of agent "C" resulted in the
lowest downwind particulate concentration of 6.15 rag/nr. Agent "B" was
8.12 mg/m3 and agent "A" 9.13 mg/m3.
344
-------
TABLE 2 Downwind particulate concentration from a dry taconite test
surface
(Wind Speed = 40 mph)
(1 hr. sampling time average)
Test
No.
1
4
5
AVG.
1.6
o
mg/m
0.78
4.86
3.25
2.96
L/D3
2-7
mg/m
3.48
3.50
2.74
3.24
4.8
mg/m3 AVG .
1.52 1.93
2.30 3.55
1.28 2.42
1.7 2.63 b
a. L is downwind distance from pile centerline to sampling nozzle inlet.
D is pile diameter at the elevation of nozzle.
L/D is not a scaling parameter to be applied to a full scale situation,
it is being used for convenience.
b. All values corrected for background particulate which in most cases was
negligible.
345
-------
TABLE 3 Downwind particulate concentration from a taconite test surface
composed of pellets treated with dust suppressing agents
(Wind Speed = 40 mph)
(1 hr. sampling time average)
Dust
Test Suppressing
No. Agentb
6
15
16
3
7
14
8
17
18
AVG.
A
A
A
B
B
B
C
C
C
1.6
mg/m
9.35
11.55
6.45
8.60
8.02
7.12
7.36
5.28
4.19
7.55
L/D3
2-7
mg/m-*
12.40
11.25
7.08
9.57
8.82
11.35
9.21
6.86
4.62
9.02
4>83
mg/m
7.00
10.35
6.72
5.33
6.90
7.44
6.62
6.03
5.20
6.84
AVG.
9.59
11.05
6.75
7.80
7.91
8.64
7.73
6.06
4.67
Agent
Average
9.13
8.12
6.15
17.80 1
a. L is downwind distance from pile centerline to sampling nozzle inlet.
D is pile diameter at the elevation of nozzle.
L/D is not a scaling parameter to be applied to a full scale situation
it is being used for convenience.
b. Amount applied such that resultant pellet moisture content approxi-
mately 1% by weight. All suppressing agents diluted in fresh tap water
to 1 in 1000.
c. All values corrected for background particulate which in most cases was
negligible.
346
-------
TABLE 4 Downwind particulate concentration from a taconite test surface
using Indiana tap water0 as a dust suppressant.
(Wind Speed = 40 mph)
(1 hr. sampling time average)
L/Da
Test
No.
2
9
10
11
13
AVG.
mg/m
9.13
8.53
4.17
5.63
8.54
7.20
27 4.8
*•»'« *}
mg/m mg/m
11.2 9.1
8.90 8.57
8.83
9.66
10.06 7.8
9.73 8.49
AVG.
9.81
8.67
6.50
7.65
8.80
8.47
a. L is downwind distance from pile centerline to sampling nozzle inlet.
D is pile diameter at the elevation of nozzle.
L/D is not a scaling parameter to be applied to a full scale situation,
it is being used for convenience.
b. Amount applied such that resultant pellet moisture content approxi-
mately 1% by weight.
c. All values corrected for background particulate which in most cases
was negligible.
347
-------
From Table 4, the use of water only as a dust suppressing agent resulted
in a downwind particulate concentration of 8.47 mg/m . The amount of water
added and the procedure used to spread it over the pellets was identical to
that of the suppressing agents. For all water only tests, the average pellet
moisture content was 1.07%.
There appears to be a significant difference between a dry taconite sur-
face and a surface which has been treated with suppressing agent and water or
water only. The dry taconite surface fugitive dust concentration is signifi-
cantly lower than the suppressing agent surface or the water only treated sur-
face.
Further, it appears that the suppressing agent and water mixture (Table
3) are no better in reducing fugitive dust than a surface treated with water
only. The suppressing agent surface resulted in a downwind concentration
average of 7.80 mg/m while the water only treatment was 8.47 mg/m .
A graphical display of the same data presented in Tables 2, 3 and 4 is in
Figures 6, 7 and 8. From these plots the differences between the suppressing
agent and water only treated surfaces to the dry surface are clearly evident.
The variation with distance in downwind concentration is due to the nature of
the aerodynamic flow around a cone. The greater variability in concentration
at the 1st downwind measurement station (L/D - 1.6) as compared to the 2nd
(L/D = 2.7) and 3rd (L/D = 4.8) is also due to the nature of the aerodynamic
flow over and around a cone. The 1st downwind sampling nozzle is located in
the downwind aerodynamic cavity of the cone and the resulting turbulence and
vortex shedding causes the concentration variation to be greater than that at
the two further downwind stations.
Note also that the 2nd downwind station has slightly higher concentrations
than either the 1st or 3rd measuring station. This effect is due also to the
nature of the flow over and around a conical form and cannot necessarily be
extrapolated to a full scale taconite piles since they are in essence elongated
cones.
The foregoing results were not expected from these tests. Intuitively
it was expected that the dry pile would yield the highest downwind particulate
concentrations with the water only and suppressing agent mixed with water
following in decreasing concentration. However, the repeated concentration
values for each test series are supportive of the statements made in reference
to each test. Further testing was therefore undertaken to elucidate the nature
of the particulate release from the taconite surface. The further tests con-
ducted were transient as compared to steady state and support a particulate
release hypothesis that is covered in the next section.
Transient Measurements
A series of three transient tests were performed to determine the time
dependency of the downwind particulate concentrations from the taconite test
surface. From these tests three plausible theories are suggested to explain
the higher particulate release rates for wetted and then dried taconite
pellets.
348
-------
Figure 6. Downwind participate concentration from a dry taconite test sur-
face.
O>
6
oc
»-
LU
O
ee.
a.
12
10
0 8
HH "
TEST RUNS
01
03 5
A4
019 (DRY TRANSIENT TEST-TIME
WEIGHTED AVG)
A
(WIND TUNNEL SPEED = 40 mph)
(1 HR. SAMPLING TIME AVERAGE)
A
4.8 5.0
L/D, RATIO OF DOWNWIND DISTANCE TO CONICAL TEST SURFACE DIAMETE AT NOZZLE ELEVATION
-------
Figure 7. Downwind particulate concentration from a taconite test surface
composed of pellets treated with dust suppressing agents.
OJ
Ln
O
O>
e
V
O
O
O
12
10
8
TEST RUNS:
0 3
0 7
A6
O8
V14
017
a is
(WIND TUNNEL SPEED = 40 mph)
(1 HR. SAMPLING TIME AVERAGE)
I
1.0
1.6 2.0
2.7 3.0
4.0
4.8 5.0
L/D, RATIO OF DOWNWIND DISTANCE TO CONICAL TEST SURFACE DIAMETER AT NOZZLE ELEVATION
-------
Figure 8. Downwind particulate concentration from a taconite test surface
using Indiana tap water as a dust suppressant.
to
Ul
O»
6
s
O
UJ
C_)
O
O
UJ
12
10
8
0
1.0
J L
O
(WIND TUNNEL SPEED = 40 mph)
(1 HR. SAMPLING TIME AVERAGE)
_L
1.6 2.0
2.7 3.0
4.0
O
J i
4.8 5.0
L/D, RATIO OF DOWNWIND DISTANCE TO CONICAL TEST SURFACE DIAMETER AT NOZZLE ELEVATION
-------
Table 5 contains the downwind particulate concentration data and the
corresponding time into the one hour test. The test was conducted by placing
the wetted taconite test surface (1.1% moisture by weight) into the wind
tunnel, initiating particulate sampling, then bringing the tunnel up to 40
mph. The sampling nozzles were quickly interchanged at downwind sampling
sampling position no. 2 as time into the test increased. The tests were ter-
minated after 60 minutes. A plot of the data is seen in Figure 9. A sharp
increase in particulate concentration is seen within the first 1.5 minutes then
a rapid decrease at 4.5 minutes beyond which the concentration increases rapid-
ly once again at 5 minutes. The peak value of the second increase is much low-
er than the first. Beyond 5 minutes, the particulate concentration exponent-
ially decays until at 60 minutes into the test, the concentration appears
steady state and approaching that value seen earlier for the dry pile concen-
tration of 2.63
The rapid increase seen within the first 1.5 minutes is due to relatively
large material as examination of the filter pads revealed. This large parti-
culate is probably loosely held to the taconite pellet surface and is mechan-
ically removed by the action of the wind. Next, the surface of the pellet be-
gins to dry as the surface moisture is evaporated. Since the initial large
particles are quickly removed by wind action (within first 1.5 minutes at 40
mph), the concentration sharply decreases between 1.5 minutes and 4.5 minutes.
By this time (4.5) subsurface drying is now occuring with the water vapor
proceeding through the interstices of the pellet material. This subsurface
evaporation appears to trigger a rapid increase in particulate emissions be-
tween 4.5 minutes and 8.5 minutes into the test which may be due to the water
vapor entraining fine particulate matter. Beyond this time the particulate
concentrations exponentially decay out to 60 minutes. The decay approaches the
steady state dry pile test previously shown in Table 2 of 2.63 mg/m .
Subsequent to the initial release of the large particulate at 1.5 minutes
three theories are suggested for the particulate emission behavior.
1) As the subsurface drying occurs, water vapor off-gases through the
interstitial pellet material carrying with it fine particulate matter.
2) The absorption of the water chemically reacts with some unknown con-
stituent such as a calcium rich material which after firing could be
CaO (hydrated lime) which is water soluble. In essence the water
weakens the bond between pelletized particles causing higher parti-
culate release rates.
3) The drying of the pellet surface causes "micro-spalling" of the
pelletized material thus releasing more particulate.
The decay portion of the curve in Figure 9 is explained by a finite amount
of material that is being removed from the pellet. In the limit, therefore,
the curve should decay to the dry taconite surface downwind concentration.
Recall that the 60 minute time average concentration shown in Table 2 for dry
taconite was 2.63 mg/ra . The decay curve of Figure 9 is obviously approaching
this value in the limit.
Note also in Figure 9 that a curve is shown for a dry taconite surface.
This test was run to confirm that the dry taconite surface is lower in fugitive
emission at 40 mph than one which has been wetted and allowed to dry. Table 6
contains the dry transient data. As expected the initial release of loose sur-
352
-------
TABLE 5 Particulate concentration variation with time for
taconite test surface composed of pellets wetted
with Indiana tap water3.
Test Numbers
10, 11, 12
(Wind Speed = 40 mph)
Particulate
Concentration
at L/D = 2.4
o
mg/nr
Time into test
(midpoint of
sampling time
interval)
minutes
32.45
46.40
37.00
8.97
2.94
25.50
19.90
16.50
12.10
9.71
14.30
8.83
8.87
5.03
5.00
5.61
1.86
3.84
1.0
1.5
2.5
3.5
4.5
5.0
5.0
8.5
12.0
16.0
15.5
25.0
27.0
38.0
40.5
49.0
56.0
60.0
a. Amount of water added such that resultant pellet moisture content
approximately 1% by weight.
-------
TABLE 6 Particulate concentration variation with time
for a dry taconite test surface3.
(Wind Speed = 40 mph)
Particulate
Concentration
at L/D = 2.4
Time into test
(midpoint of
sampling time
interval)
Test Number mg/nr
19 56.73
25.31
25.02
4.74
0.83
0,24
3.15
0.71
a. No water or agent added to pellets
minutes
1.0
3.0
5.0
7.0
12.0
23.0
37.5
52.5
354
-------
Ui
^ 50
p 40
P, 30
g 20
at
Q.
O
10
RELEASE OF LARGE LOOSELY BOUND SURFACE PARTICLES
INITIAL SURFACE DRYING
NOTE: MOISTURE CONTENT OF PELLETS AT START OF TEST:
1X BY WEIGHT
RELEASE OF "SPALLED" SURFACE MATERIAL
TRANSIENT DRY PELLET TEST
PARTICLE EMISSION DECAY AS SPALLED MATERIAL
IS REMOVED FROM PELLET SURFACE
-©
10
20
30
TIME, MINUTES
40
50
60
Figure 9. Transient particulate emission characteristic from wetted tacon-
ite test surface at 40 mph and a dry taconite test surface.
-------
face particulate results in the high 56.73 mg/m^ concentration after only 1.0
minute. After 7 minutes, however, most of this fine particulate is gone and
the dust concentration drops to 4.74 rag/nr. Beyond this point, the concen-
tration is very low compared to the wetted and drying surface.
CONCLUSIONS
1) A dry taconite surface when exposed to a vertical wind profile from
20 to 40 mph results in a substantially lower downwind particulate
concentration than one which has been wetted and then allowed to dry.
2) Of the three commercially available dust suppressing agents tested at
a dilution ratio of 1 to 1000, one agent appeared to be more effect-
ive than the others. The more effective agent reduced the downwind
particulate concentration by 33% over the least effective agent.
3) Using three commercially available dust suppressing agents diluted 1
in 1000 with Indiana tap water appear to be no better than using
water only to suppress dust from a taconite pellet surface. Water
resulted in a dust concentration of 8.47 mg/m-* and suppressing agent
7.8 mg/m3.
4) The taconite surface when wetted either with water only or a suppres-
sing agent mixed with water at a dilution ratio of 1 in 1000 causes
about 3 times higher particulate concentrations than a dry pile with
no agent whatever applied.
5) The transient release of particles from a wetted surface appears to
be strongly related to the use of water. The water somehow causes
the surface of the pellet to become much more susceptible to the
release of particulate matter.
6) When water is used to control fugitive dust from a taconite surface,
the surface must not be allowed to dry but be continuously wetted.
356
-------
COMPUTING DESIGN CHARACTERISTICS
FOR COAL PILE DRAINAGE TREATMENT
Presented at the Fourth Symposium on Fugitive Emissions;
Measurement and Control
May 28-30, 1980
Monteleone Hotel, New Orleans, Louisiana
Pamela B. Katzf P.E., Principal Engineer
John A. Ripp, Project Engineer
TRC Environmental Consultants
ABSTRACT
Under Section 304(e) of the Clean Water Act, the U.S. EPA is developing
a program of Best Management Practices (BMP's) for control of toxic and
ha2ardous discharges from ancillary industrial sources. These sources
include plant-site runoff, spills and leakage, sludge or waste disposal, and
raw-material drainage.
For many electric utilities, BMP means evaluating coal storage piles
with regard to the quantity and quality of storm related and dry weather
drainage. Coal pile drainage is usually characterized as acidic with con-
centrations of trace metals, iron, sulfur, and solids.
TRC is developing a mathematical model to simulate coal pile drainage
for the design of appropriate treatment systems. The model will allow for
the consideration of various antecedent meteorological conditions including
rain and snow precipitation, freeze-thaw phenomena, and air temperature.
Other hydrologic phenomena the model will evaluate are:
o pile runoff
o snowmelt
o percolation through the pile
o infiltration into groundwater
o evaporation
The model will also simulate coal pile runoff quality. Components
included are framboidal pyrite oxidation, acid production, and the subse-
quent release of trace metals.
TRC will discuss how its model is one available method to characterize
coal pile runoff under varying meteorological conditions. TRC is currently
developing an extensive field program at a number of utilities to calibrate
and verify the model.
357
-------
1.0 INTRODUCTION
Current practice to compute design characteristics for coal pile runoff
treatment may be overly conservative and unnecessarily costly. Accepted
practice dictates that storage/treatment be designed to accomodate all of
the runoff from a 10 year-24-hour storm. EPA-sponsored^ studies com-
pleted by TRC in 1976 at utility power plants in Pennsylvania suggest that
while initial stormwater runoff contains relatively high levels of contami-
nants, the pollutant concentrations decrease rapidly as the storm pro-
gresses. Why, then not store and treat the initial "dirty" runoff and by-
pass the subsequent "clean" water if it meets NPDES limits?
This concept has captured the interest of the Utility Industry. The
industry through UWAG, EEI, and EPRI is presently participating in a joint
EPA-Utility Industry-TRC project to develop mathematical modeling and moni-
toring techniques that can be used to characterize coal pile runoff on a
dynamic basis for a storm event. This project has started where the pre-
vious EPA-TRC efforts left off.
This paper will discuss the conclusion of previous coal pile drainage
monitoring efforts, and the objectives of the current EPA-utility program
including the development of a mathematical model for coal pile drainage and
the related field program for calibration and verification of the model.
Brookman, Binder, Wade, "Sampling and Modeling of Non-point Sources at
a Coal-fired Utility", EPA - 600/2-77 - 1979, September 1977.
358
-------
2.0 PREVIOUS COAL PILE RUNOFF PROGRAMS
In 1976 coal pile runoff was investigated by TRC for two steam electric
generating stations in Pennsylvania-Warren Station of Pennsylvania Electric
Company and Portland station of Metropolitan Edison Company.1
At each site, a network of sampling stations was established to collect
coal pile runoff during the duration of several storms and the intervening
dry periods. Samples collected were analyzed for pH, sulfate, alkalinity,
total and dissolved solids, acidity, total iron, aluminum, and manganese.
Several interesting observations were made during the field program.
1) The seepage from the base of the coal pile after the storm event
was orders of magnitude more concentrated than the wet weather run-
off stream.
2) At the beginning of the storm, a "first flush" with higher pollu-
tant concentrations could be seen. Through the rainfall period
these values generally declined. For example, acidity levels on
the September 17, 1976 storm at Warren Station were initially 3200u
g/1 and declined throughout the storm. When the rain ended five
hours later the acidity was 500 Pg/1.
Tables 2-1 and 2-2 show analysis results from the Warren site. These
two phenomena can be seen in the results.
The results of the 1976 field program were used to calibrate the SSWMM-
Receiv II model developed by TRC for predicting the quantity and quality of
stormwater runoff and its impact on receiving waters. The model results
compared favorably with field measurements but identified several limita-
tions in the SSWMM-Receiv II model. These limitations include the lack of
capability to simulate storm erosion from material storage piles and to
simulate stormwater percolation through material storage piles. With these
limitations in mind, TRC outlined a scope of work to develop a model to bet-
ter simulate coal pile drainage and a field program to verify it.
Brookman, Binder, Wade, "Sampling and Modeling of Non-point Sources at
a Coal-fired Utility", EPA - 600/2-77 - 1979, September 1977.
359
-------
TABLE 2-1
CHARACTERISTICS OF COAL PILE RUNOFF
DURING SECOND STORM EVENT AT
WARREN STATION OF PENNSYLVANIA ELECTRIC CO., WARREN, PA.
17 SEPTEMBER 1976
LO
O
O
1000
1015
1030
1045
1100
1115
1130
1200
1230
1300
1330
TIME
- 1015 - Rain Start
- 1030
- 1045
- 1100
- 1115
- 1130
- 1200
- 1230
- 1300
- 1330
- 1500 - Rain End
POLLUTANT
TSS
9800
4200
6400
11400
5000
1700
1400
1600
1700
1700
23000
TDS
4600
3300
2400
2400
2500
3700
3800
3100
3000
-
500
SO,,
2300
2300
1600
1800
2100
2100
2700
1700
1000
-
200
COAL PILE
FE
900
-
700
1400
700
500
-
300
200
-
-
RUNOFF
AL
101
-
9
7
8
-
-
-
-
-
-
MN ACIDITY
40 3200
2600
10 3100
10 2000
10 2200
2900
-
2 500
DISCHARGE
FLOW RATE
1pm (gpm)
22 (5.8)
20 (5.3)
20 (5.3)
17 (4.5)
-------
TABLE 2-2
CHARACTERISTICS OF COAL PILE LEACHATE-DRY WEATHER
AT WARREN STATION OF PENNSYLVANIA ELECTRIC CO., UARREN, PA.
AUGUST - SEPTEMBER, 1976
lo
Date
8/25/76
8/27/76
9/16/76
Hours Since
Last Rain
250
17
505
Pollutant Concentration, mg/1
TSS
200
18,700
12
IDS
40.000
82,600
21,700
SO,,
57,000
45,000
25,000
Fe
23,500
14,000
9.700
Al
1,800
1,400
1.100
Mil
100
70
70
Acidity
18,000
27,000
37,600
pH
2.4
2.1
1.5
Discharge Flow Rate
Ipin (gpm)
1.5 (.39)
1.5 (.39)
1.4 (.39)
-------
3.0 OBJECTIVES OF THE CURRENT PROGRAM
As a part of the program to develop predictive measures and methodo-
logies for fugitive emissions, EPA's Industrial Environmental Research
Laboratory (IERL-RTP) has sponsored a study based on TRC's recommendations
to develop a predictive model for coal pile drainage. This project was also
of interest to utilities; hence, additional funding and support was provided
by the Edison Electric Institute (EEI) with special emphasis by its Utility
Water Action Group (UWAG). TRC has been working with a technical advisory
group consisting of utilities, EPA and technical organizations to ascertain
the utilities' needs in predictive methodologies for coal pile drainage
treatment design.
The EPA - EEI program has several components currently funded:
1) Research and evaluate existing models in related fields and specify
model modifications needed to simulate coal pile drainage.
2) Query the coal fired utilities as to the data available on coal
pile drainage. Assemble results of utility questionnaire for dis-
tribution.
3) Develop model, evaluate parameter sensitivity and model short-
comings.
4) Develop a field program design to calibrate and verify the model.
The above phases were completed in May, 1980.
In the upcoming phases, the field program will be conducted, the model
fine-tuned and modified, and a user's manual will be developed.
362
-------
4.0 DEVELOPMENT OF A COAL PILE MODEL
Using the experience of runoff modeling in the Pennsylvania field pro-
gram, and with input from the Technical Advisory Committee, TRC developed an
outline of the physical/chemical phenomena to be contained in a coal pile
drainage model. These are summarized below. The result of their inclusion
was a continuous simulation model to predict the dry and wet weather flow
quantities and characteristics.
4.1 Quantitative Phenomena
1) meteorology - hourly input of rain, snow, and air temperature
2) surface runoff
3) snowmelt
4) percolation through the coal pile, interflow from the sides of the
pile
A schematic diagram of the coal pile hydrology for the model is shown in
figure 4-1.
4.2 Qualitative Phenomena
1) pyrite oxidation, acid production
2) freeze-thaw cycles accelerating oxidation
3) gully erosion
4) coal pile washoff of pollutants
5) reactivity of trace metals
The methodology by which the phenomena are addressed in the model is
described in the following section.
363
-------
UPPER ZONE STORAGE = DEPRESSION STORAGE
DEPRESSION
STORAGE
EVAPORATION
NFILTRATION
RUNOFF
STREAM
1
DIRECT RUNOFF
LOWER ZONE
STORAGE
I
PILE MOISTURE
WATER TABLE
GROUND WATER
RUNOFF
STREAM
TO DEEP STORAGE
FIGURE 4-1: SCHEMATIC OF HYDROLOGJC CYCLE COAL STORAGE PILE
-------
5.0 MODEL DESCRIPTION
5.1 Base Model and Modifications
TRC investigated a number of runoff models to use as a base structure
for coal pile drainage. These included SWMM, STORM, ARM, Stanford Watershed
Model, and the Pyritic Systems model.
The Ohio State University version of the Stanford Watershed Model was
selected as the base model because it was a continuous simulation with
infiltration/percolation features. TRC utilized the 1977 version, made
available through the Extraction Technology Branch of the U.S. EPA.
The OSU watershed model simulates the hydrologic cycle of a rural area.
Features are included in the model for interception of rainfall by vegeta-
tion, transpiration, infiltration into the soil, percolation, evaporation,
overland flow, soil moisture, depression storage, and ground water flow to
deep storage.
OSU also developed a qualitative strip mine and refuse pile model to be
used in conjunction with the OSU version of the Stanford Watershed Model.
Both OSU co-models required extensive revisions to contain all the phy-
sical/chemical phenomena discussed in Section 4. TRC deleted from the model
a number of routines not applicable to coal pile drainage, such as those
relating to vegetation. In addition, a number of the features were added to
the model. These are listed below:
1) Addition of gully erosion to simulate the transport of solids from
the sides of the pile.
2) Simulation of pyrite oxidation and acid production with the subse-
quent release of dissolved iron, sulfate, and trace materials dur-
ing rainfall events.
3) Consideration of the acidity of rainfall and the alkalinity of
coals.
4) Consideration of the impact of freeze-thaw cycles on the active
coal surface area and subsequent pyrite oxidation.
5) Inclusion of a less complex snowmelt routine.
6) Modification of the meteorological input to allow the use of stan-
dard NOAA magnetic tapes of historical meteorological data.
5.1.2 Coal Pile Hydrology Model
TRC's coal pile hydrology model based on the OSU version of the Stanford
Watershed Model is called TRCH20.
365
-------
TRCH20 considers rainfall infiltration, gully erosion, runoff, snow ac-
cumulation and snowmelt at the coal pile surface. In addition, moisture
percolates through the pile to seep from the side of the pile, enter ground
water beneath the pile, or to be emitted as seepage from the base of the
pile.
TRCH20 uses a standard National Climatic Center magnetic data tape of
hourly weather observations for a selected station as meteorological input.
The data concerning the coal pile itself is input to the model by a small
card deck.
5.1.3 Coal Pile Qualitative Model
TRC's version of the OSU strip mine and coal refuse pile model is called
TRCCOAL.
TRCCOAL operates in either a dry day or wet day mode, depending on daily
rainfall. On dry days, framboidal pyrite in the coal is oxidized producing
acid, dissolved sulfate, and iron. The acid further reacts to dissolve
trace materials in the coal. The number of freeze-thaw cycles during the
preceding dry days accelerate the breakup of coal and the subsequent pyrite
oxidation. Moisture is emitted from the lower zone of the coal pile as see-
page.
During wet weather, dissolved materials are washed from the surface and
to the other zones of the pile. They are then washed out of the coal pile
to the runoff stream. The model then returns to its dry/wet day cycle.
The TRCCOAL model uses a tape created by TRCH20 for hydrologic data and
a small card deck describing the characteristics of the coal pile.
366
-------
6.0 MODEL OUTPUTS
6.1 TRCH20
The quantitative model, TRCH20, can be run without the qualitative com-
ponent. It prints a tabular output of precipitation, runoff flows, and pile
moisture data. The basic data is presented as daily flows and loadings for
the year. After data analysis the user can choose output intervals at hour-
ly or less intervals for a selected time interval. Some of the more detail-
ed data is presented as user selected options. The output of TRCH20 is
listed below:
Basic Outputs
1) A table of average daily rainfall in inches,
2) Summation of synthesized daily runoff rates, in cubic feet per
second, for each month, followed by the annual total.
3) Synthesized monthly and annual totals for each of overland runoff,
interflow, baseflow, and total flow, in inches.
4) Monthly and annual totals of liquid precipitation, in water equiva-
lent inches, developed from the meteorological input.
5) Monthly and annual totals of frozen precipitation, in water equiva-
lent inches, developed from the meteorological input.
6) End of month values, in inches, of pile, surface, and ground water
moisture.
7) End of month values for indexes of ground water slope and pile de-
pression storage.
8) End of month values for snowpack in water equivalent inches.
9) And annual balance, in inches, of unaccounted for moisture.
Optional Outputs
1) Echos all input data.
2) For one selected storm per year, prints values, in inches, of rain-
fall deposition, moisture storage, runoff origins and runoff out-
flow for each time interval.
3) Prints daily values of selected variables to give a better indica-
tion of the program interactions in the upper, lower, and deep
zones of the pile.
367
-------
4) Echos recorded runoff flows, in cubic feet per second, for compari-
son with simulated flows.
5) Records the twenty highest clockhour rainfall events in the water
year.
6) Prints daily pile moisture storage, in inches.
7) Prints daily values of the snowpack, snowmelt, and snowfall, in
water equivalent inches.
8) Prints out average daily temperature in °F.
9) Prints out number of freeze-thaw cycles in preceding dry days.
10) Prints out simulated gully erosion of solids to the base of pile
in-pounds per day.
11) Plots an arithmetic hydrograph for water year comparing recorded
and simulated flows.
12) Plots an arithmetic hydrograph and rainfall hyetograph for selected
data.
13) Prepares tape output file for use as input to TRCCOAL.
Plots of the annual runoff hydrograph are prepared by the model. An
example output showing simulated runoff flow over the water year is present-
ed in Figure 6-1. This hydrograph can also include recorded runoff flows
for a comparison with the modeled values. An optional plot of the rainfall
hyetograph and runoff hydrograph for a selected detailed storm is shown in
Figure 6-2.
6.2 TRCCOAL
The co-model TRCCOAL calculates the amount of acid, sulfates, iron and
trace materials in the runoff stream and summarizes it on a daily, monthly
and annual basis. Below are listed the tables produced by TRCCOAL:
1) Daily loadings of acid, iron, sulfate, and trace materials in the
direct runoff, interflow, and seepage, in pounds.
2) A calendar of simulated daily runoff flow volume, in cubic feet per
second.
As in the hydrologic model, printouts of hourly data can be chosen for a
selected time interval. In addition, the model produced plots of pollutant
loadings. An example plot is shown in Figure 6-3.
368
-------
00
u
a-
'
IT
C
in
•<*•
0
en
CO
c
er
t/t a
o
31 CD
Q.
< <=>
CC
O
o
cc co
Q «-;
^ O
o
J— • CO
IS o
»—
cc
°
0
CD
ca
I ,
FIGURE 6-1
ARITHMETIC HYOR06RAPH PLOT OF TRC H20 MODEL
•
•
1
•
j
I
l|
1 *
, ; \ :\
""^"*-^ '\J %- \^\f ^^
i "~" ~" ~ 7 "~ ~— ^ x. i i
-1 1 1
1
1
1
1
1
1
i
II
It
II
>l
II
II
II
II
1
i
!
ii
ii
h
n
i
1 I
'» ,
i \ }
i i .1
'
.
-
-
'
0.00 10.00 20.00
OCT
30.00 40.00 50.00 60.00 70.0
NOV
80.00
-------
TU
flBURE C-Z
DETAILED STOftM PIOT OF TRC HZO MODEL
U>
^J
o
s °
CO
CM
< O
co
u-
Cl
CFOO TOO 4?00 OO BiOO 10.00 12.00 14.00 16.00 TOO 20.00 22.00 24.00 26.00
3AM
9AM
3PM
9PM
3AM
-------
IT
UJ
or
o o
o
o
o
C\J
!).00 10.00 20.00 30.00 40.00 50.00 60.00 70.'00 360.00
OCT NOV
FIGURE 6-3: UAMPt.E PLOT Of QUAl ITAIIVE OU1PUT
o
o
--OUTFLOW STREAM (cfs)
o—ACID IN OUTFLOW STREAM (Ibs)
1316-005
-------
7.0 DESCRIPTION OF FIELD PROGRAM AND INTERFACE WITH MODEL
7.1 Outline of Field Program
Following initial model development TRC will begin a phased field pro-
gram for coal pile runoff from utility sites. The objectives of this pro-
gram are:
1) To obtain representative data to fine tune and modify the coal pile
runoff model.
2) To test the runoff model for a variety of climatic regimes, coal
pile configurations, and coal characteristics.
3) To generate a substantial data base on Che quantity and quality of
coal pile runoff from various utility sites throughout the country.
The proposed field program will be conducted at 12 different utility
sites nationwide over a 2 1/2 year period. The first two locations will
serve as test sites and the data will be used to fine tune and modify the
model. These two sites will also serve to refine the field program for the
remaining ten sites. The field survey at the ten sites will be used to test
the runoff model for a variety of climatic regions, coal pile configura-
tions, and coal characteristics as well as generating a substantial data
base on coal pile runoff. The field program for each site will vary from
one month to 9 months.
7.2 Parameters of Interest
During the field program at each site, TRC will take continuous measure-
ments of flow, pH, conductivity, and temperature.
In addition, samples will be taken of the runoff flow periodically and
analyzed for acidity, alkalinity, TSS, TDS, TOC, sulfate, hardness, iron
and 19 trace metals.
A small meteorological station will be erected at each site to measure
precipitation, air temperature, solar radiation and evaporation.
7.3 Data Reduction
All strip charts from the continuous monitors will be digitized to pro-
duce magnetic tapes of site data. The tape of the runoff flow volume can be
read into the model to plot a comparison of simulated and recorded flows.
All laboratory analysis will be reduced, evaluated statistically and placed
on magnetic tape.
372
-------
8.0 COMPUTING DESIGN CHARACTERISTICS
8.1 Data Bank
All of the data collected in the field program will become part of a
computerized data bank. The data can be used by utilities which do not
choose to simulate their coal pile runoff but rather would be interested in
evaluating data from a site with similar coal pile characteristics. The
data can be used to determine possible waste characteristics and volumes for
treatment system design.
8.2 Coal Pile Drainage Model
The model, calibrated and verified under a number of field conditions,
can be used to simulate different coal pile drainage situations. Using se-
veral years of historical meteorological data, the model can determine
"average" flows and pollutant loadings. The model also has the option to
read in selected precipitation data from a card deck. The user can create
different combinations of dry days and storm events to evaluate when the
highest pollutant loadings occur. In addition, the plots of detailed storms
will show when and if the "first flush" of pollutants occurs and when runoff
flow concentrations are low enough to be discharged without treatment.
From this analysis of both average and worst case conditions, collection
basins and treatment works can be more accurately designed and sized.
373
-------
9.0 SUMMARY OF UPCOMING ACTIONS
As of May 1980 TRC has completed initial model development, the field
program design, and a manual of field procedures. TRC is working with in-
terested utilities, EPRI, EEI, and EPA to locate test sites. Data from
these field programs will be used in model calibration and verification. By
the end of 1981 the model should be useable as a valuable predictive tool in
coal pile drainage characterization.
374
-------
EMISSIONS AND EFFLUENTS
FROM
RAIL AND TRUCK TANKCARS
T. R. Blackwood
Monsanto Company
800 N. Lindbergh Blvd.
St. Louis, MO 63166
ABSTRACT
As many as 700 different commodities are handled by rail or
truck tankcars. Approximately 37,000 railcars and 5,000,000
tanktrucks are cleaned each year by industry and service
companies. Air emissions of hydrocarbons can be as high as
2.4 kg for tankcars and 310g for tanktrucks. Viscosity and
volatility are the primary influencing factors on emissions.
If untreated, cleaning solutions from this process could ex-
ceed 2,500 metric tons/year of oil and grease in the wastewater,
Hydrocarbon emissions to the air could also exceed 620 metric
tons/year or about 0.0022% of U. S. emissions. Practical and
economically feasible control for1 air emissions does not exist
except for combustible gases and water-soluble vapors. State-
of-the-art technology for wastewater effluents does exist but
the effectiveness is widely variable and is very expensive.
This report describes the state-of-the-art practice in mobile
tank cleaning. Composition, estimated quantities, and rate of
emissions and pollutants are described along with control
methods and costs.
375
-------
INTRODUCTION
Various chemical and petroleum products are transported by rail
tankcars and tanktrucks. These shipping tanks must be cleaned
before being used to ship a different material, or contamination
of the new material could result. An experimental study (1)
was conducted by Monsanto Research Corporation to obtain prelim-
inary estimates on air emission and water effluent from the
cleaning process. This report assesses the impact and the
economics of one potential control technology system.
Rail tankcars and most tanktrucks are in dedicated service
(carrying one commodity only). Unless contaminated, they are
cleaned only prior to repair or testing. Nondedicated tank-
trucks, approximately 22% of the total tanktrucks in service,
are cleaned after every trip to prevent cross contamination.
The approximate total number of units cleaned per year are
37,000 rail tankcars and 5,010,000 tanktrucks.
Steaming, washing and/or flushing of the tanks result in air
emissions and wastewater effluents. These cleaning operations
are partially enclosed. Residual material is usually washed to
the wastewater collection system, and only small amounts of
material escape through the vents to the atmosphere.
Air emissions from cleaning of tanks are predominately organic
chemical vapors. As hydrocarbons, these emissions represent
less than .0022% of the national emissions of hydrocarbons.
Water pollutants from cleaning of tanks are primarily oil and
grease, suspended solids, and phenol. Uncontrolled water
effluent quantities from this source are summarized in Table 1.
TABLE 1. POTENTIAL WATER POLLUTANTS FROM
RAIL TANKCARS, TANKTRUCKS, AND
DRUMS (metric tons/yr)
(Estimated in reference #1)
Source type
(basis) Oil and grease Suspended solids Phenol
Rail tankcar cleaning
(37,220 cars/yr)
Tanktruck cleaning
(5,010,000 trucks/yr)
830
1,745
4,100
6,070
31
986
SOURCE DESCRIPTION
There are approximately 178,000 rail tankcars and 90,000 tank-
trucks in service in the United States. Most of the tanks are
owned privately and haul liquids. These privately owned tanks
are not cleaned as often since they generally can be used for
376
-------
dedicated service. Leased tanks, on the other hand, are cleaned
nearly every trip. The amount of cleaning that is conducted
depends upon the operational practice of the cleaning site, the
chemicals being removed, and the size of the tank. The quantities
used per tank truck vary from approximately 0.23 cubic meters
(sixty gallons) for steam operations to 20.9 cubic meters (5,500
gallons) for full flushing, with two cubic meters (500 gallons)
being considered the average. The average amount of material
cleaned from each tanktruck is estimated to be 100 kilograms.
Vapors from volatile materials are flared at a few terminals, but
the most common practice is to allow them to dissipate into the
atmosphere.
The amount of liquid used per rail tankcar varies from 0.23 cubic
meters for steam cleaning to 129 cubic meters (34,000 gallons)
for total flushing of a large tank car. The average amount of
residual material cleaned from a tankcar is estimated to be to 250
kilograms. Vapors from cleaning tank cars used to haul volatile
materials are sent to flares at some cleaning facilities. Vapors
of materials such as anhydrous ammonia and chlorine are dissolved
in water and become waste water constituents. Vapors not flared
or dissolved in water are dissipated into the atmosphere.
Tables 2 and 3 summarize the typical variety of chemicals cleaned
from tanktrucks during a one-month period along with the fre-
quency of the type of cleaning methods. The most common cleaning
practice is the use of steam, although most commercial cleaners
utilize a water rinse before and after the steam cleaning. As
shown in Table 3, the determination of emissions and emission
factors is complicated by the cleaning methods.
TABLE 2. TRAILER INTERNAL CLEANING GENERATION RATES
FOR ONE TERMINAL DURING ONE MONTH OF
OPERATION C2)
Method Water use,
Dumber Cleaning method m-y trailer
1
2
3
4
5
6
7
8
Cold water flush
Cold water flush — caustic/acid tank
Cold water flush — steam — cold water
rinse
Cold water flush — spin/detergent —
cold water rinse
MEK, MIBK,a or acetone solvent — cold
water rinse
Styrene solvent — cold water rinse
Cold water flush— steam— cold water
rinse — spin w/detergent— cold water rinse
Cold water flush w/Butterworth for
dry bulk trailer
0.57
8.3
3.0
1.1
0.57
0.57
3.6
5.7
Number of
trailers
84
321
316
123
0
0
68
78
Total water
use, ro-*
47.6
2,666.0
954.5
139.3
0
0
243.9
441.7
TOTALS 990 4,493.3
377
-------
TABLE 3. COMMODITY/TANKTRUCK DATA FOR ONE TERMINAL
DURING ONE MONTH OF OPERATION (2)
Cleaning
method
number
Commodity
No. of
trailers
cleaned
Uran fertilizer
PAPI—isozylate
Ethyl chloride
Alum
Water for glue
Water softener
Caustic soda (50%)
Silicate soda
Acetic acid
Phosphoric acid
Spent acid
Sulfuric acid
Hydrochloric acid
Corrosive liquid
Solvent
Toluene
Xylene
IPA—isopropyl alcohol
Sodium MET
EDA—ethylene diamine
DTA—diethylene triamine
Poly amines
Vinyl acetate
Cyescal
Phenol
Alcohol
Petroleum chemicals
Peroxide
Biphenyl
Sodium bichromate
Sodium methylate
PA-phthalic anhydride
Acetone
Adaline
Ferric chloride
TTA—Amine 220
AN—acrylonitrile
Protein feed supplement
Calcium chloride
Styrene
Methyl acrylate
Weed killer
Shell pan
DMK—dimethyl ketone
16
2
10
53
2
1
123
2
22
1
47
87
38
1
18
26
2
23
1
15
8
7
23
3
32
22
1
4
2
9
3
6
9
4
3
3
8
2
1
2
1
4
1
4
378
-------
TABLE 3 (Continued)
Cleaning
method
numbera
Commodity
No. of
trailers
cleaned
3(cont.) Benzene 1
Pentylamine 1
Ethylene glycol 3
MEK—methyl ethyl ketone 14
ITA 2
Mineral spirits 3
DAA—diacetone acrylonitrile 4
NBA—normal butyl alcohol 1
Methanol 3
Butyl cellosolve 1
Formaldehyde 24
Oxylene 1
Naphtha 1
MIBK—methyl isobutyl ketone 4
Demineralized water 1
Turpentine 2
Oxital—ethylene glycol monoethane ether 1
TRI Clean D 2
4 Glue 72
Paint 2
Resin 30
Water treating compound 5
Coastal pale oil 7
Petroleum oil 3
Cotton oil 2
Script set 2
7 Diesel oil 21
Petrolatum 16
Ink oil 2
Strip oil 13
Hi Boiler Oil 2
Tall Oil 5
Insulator oil (.new). 1
CPTIC—crude petroleum 8
8 Potash and fertilizers
Plastic pellets 78
a Cleaning method numbers correspond to those tested in Table 2
379
-------
AIR EMISSION
The great diversity of commodities (over 700 chemicals) carried
by rail tankcars and tanktrucks makes it nearly impossible to
sample and obtain emission factors for every possible material.
Sampling of emissions from the cleaning of selected commodities
was conducted to obtain an indication of the relative amounts of
various types of materials. In order to achieve a representative
picture, the organic chemicals were broken down into classes
cliaracterized by high, medium, and low viscosities and by high,
medium, and low vapor pressures at ambient temperature. Viscosity
affects the quantity of material which remains in the tank since
low viscosity materials would drain more easily than high vis-
cosity materials. It was anticipated that the higher the vapor
pressure the higher the air emission rate would be during clean-
ing.
It has been reported that virtually all the air and water pollu-
tants are removed from the tank during the first washing cycle.
This takes 45 minutes to one hour for tank trucks and one to
two hours for rail tankcars. Subsequent rinsing adds only small
(less than 2%) quantities of waste to the wash water. The
measured emissions from tankcar and tanktruck cleaning are
summarized in Table 4. Using these data and a Gaussian dispersion
model, the impact of a typical cleaning facility on ambient air
quality can be estimated. In its simplest form, the dispersion
model for the calculation of the maximum ground level concentra-
tion is (3) :
max
irH eU
TABLE 4. MEASURED EMISSIONS FROM TANK-
CAR AND TANKTRUCK CLEANING
Compound
Acetone
Perchloroethylene
Methyl methacrylate
Phenol
Propylene glycol
Ethylene glycol
Chlorobenzene
o-Dichlorobenzene
Creosote
Chemical
Vapor
pressure
High
High
Medium
Low
Low
Low
Medium
Low
Low
class
Viscosity
Low
Low
Medium
Low
High
High
Medium
Medium
High
emissions.
ga
311/truck
215/truck
32.4/truck
5.5/truck
1.07/truck
<0.32/car
15.7/car
75.4/car
2,350/car
(8-hr)
Measured
emission
concen-
tration,
mg/m3
654
526
79.1
14.0
4.3
<0.2
8.8
94.3
118
Total emissions = (emission rate) x (emission volume).
380
-------
where: Q = mass emission rate, g/s
U = U. S. average wind speed, 4.47 m/s
H = emission height, m
e = 2.72
IT = 3.14
Since the average emission height is about four meters, the
maximum ground level concentration for each chemical would be
as shown in Table 5. These maximums are based on cleaning
thirty tanktrucks or five and one-half tankcars each day. It
is rare for more than one truck or car to contain the same chem-
ical in any one day. Each maximum ground level concentration
shown in Table 5 represents the worst case for a typical facili-
ty (cleaning only one chemical over a long period of time). In
addition, this maximum can be shown to occur within twelve to
forty meters from the tank being cleaned.
The concentration at the plant boundary would be considerably
less than the maximum value since concentration decreases by the
inverse of the square of the distance from the source. Compared
to the ambient air quality standard for three hours of 160 micro-
grams per cubic meter, the typical cleaning facility is not
likely to result in a violation of the ambient air quality stan-
dard. A more rigorous analysis could be conducted utilizing the
distribution of commodities cleaned at each facility. However,
additional study was not warranted since the worst case analysis
showed that the present impact is very small.
TABLE 5. MAXIMUM GROUND LEVEL CONCENTRATIONS FOR
DIFFERENT EMISSIONS
Type of Xmax'
cleaning Emission yg/m3
Tank truck Acetone 359
Perchloroethylene 248
Methyl methacrylate 37
Phenol 6.35
Propylene glycol 1.24
Rail tank car Ethylene glycol <0.14
Chlorobenzene 7.14
o-Dichlorobenzene 34.3
Creosote 267
381
-------
From another perspective, if the highest emission factor for
tanktrucks and tankcars Cacetone) is applied to all of the tanks
cleaned, the total emissions of hydrocarbons for tanktrucks and
rail cars would be 538 and 87 metric tons per year, respectively.
This corresponds to less than .0022% of the total hydrocarbons
emissions in the U. S. Based upon this evidence, it has been
concluded that air emissions from the cleaning of rail tankcars
and tanktrucks are not significant emission sources.
WATER EFFLUENTS
Waste water collected in a tank cleaning facility is subject to a
great variability. Even with a good treatment system, it is diffi-
cult to treat to consistently acceptable levels. Table 6 sum-
marizes the effluent characteristics of a slightly larger than
average tanktruck cleaning facility. This facility is a forty-
five cubic meter per day (.12,000 gallons per day) demonstration
plant funded by the U. S. EPA (4). Also shown in Table 6 are
the approximate levels that have been achieved in the effluent
of this experimental system. A schematic of the system is shown
in Figure 1. Utilizing these data and other available infor-
mation, an analysis was conducted on the uncontrolled effluent
concentrations resulting from representative sources. A more
detailed description of this analysis is given in Reference 1.
The results, which are summarized in Table 7 show that the
effluent concentrations are especially high for oil and grease,
and phenol. Further analysis shows that the impact on typical
receiving waters is minimal.
To estimate the impact of an effluent on water quality, the
concept of source severity was developed. Source severity is
defined as the pollutant concentration to which aquatic life is
exposed, divided by an acceptable concentration. The exposure
concentration is the fully-diluted receiving water concentration
resulting from the source. The acceptable concentration is de-
fined as the concentration at which it is assumed that an incip-
ient adverse environmental impact occurs (.usually the water
quality criterion}. Mathematically expressed, this is:
B _ CW CD
O "~ ••—' i i i •" • '» »
where
CD = concentration in raw effluent, g/m
F = fresh water criteria, g/m
V = volumetric flow rate of discharge, ra /sec
o
VR = volumetric flow rate of receiving waters, m /sec
S = dimensionless
382
-------
TABLE 6. TREATMENT PLANT OPERATING RESULTS —
MATLACK, INC., SWEDESBORO, N.J. (4)
Parameter
PH
Color units
Turbidity, FTUa
COD, g/m3
BODs, g/m3
Oil and grease, g/m
Phenols, g/m3
Suspended solids, g/m3
Raw
10.5
Over
Over
1,800
600
110
1
300
feed
to
12.5
500
Effluent
6.5
10
500
to
to
to
to
to
11,000
2,000
350
250
1,300
125
20
0
to
t0u
_b
to
to
to
0,1
"
8.5
50
300
100
1
Forraazin turbidity units; a standard unit of turbidity based
upon a known chemical reaction
3Data not reported.
TABLE 7. EFFLUENT CONCENTRATIONS FOR
REPRESENTATIVE SOURCES
Source
Rail tank cars
Tank trucks
Effluent
Oil and
grease
2,005
230
concentrations, g/m3
Phenol
75
130
Suspended
solids
10,025
800
Flow,
m3/day
61
45
TABLE 8. SOURCE SEVERITIES FOR REPRESENTATIVE SOURCES
Source type
Oil and
grease
Source severity, s
Suspended
solids
Phenol
Rail tank car 0.16
Tank truck 0.014
0.00033
0.000020
0.062
0.080
333
-------
Table 8 is a summary of the severities for the effluents from
a representative cleaning facility. From this analysis, the
exposure concentration is found to be less than the fresh water
criteria (severity < 1.0) even from the uncontrolled effluents
into U. S. Average river flow (1).
The state total discharge rates of oil and grease, phenol, and
suspended solids, assuming no treatment, were calculated by:
dividing the production of the representative source into state
production totals and multiplying this figure by the discharge
rate per representative source. These quantities are given for
each state in Reference 1. Table 9 summarizes the estimated
uncontrolled emission burdens due to rail car and tank truck
cleaning for selected states.
RAW WASTE
SURFACE
SKIM OIL
RESALE)
FLOTATION
/CELL
pH PRESSURE f
1
ROLLINS ENV.
SERVICES
CLARIFIER
BIOLOGICAL
OXIDATION
P
V
ACTIVATED
CARBON
TRANSFER
TANK
ADSORBERS'
SPENT CARBON
DISCHARGE
RECYCLE
FIGURE 1. Wastewater Treatment System, Matlack, inc.
Swedesboro, New Jersey C4>. *'
384
-------
TABLE a. PROJECTED RAIL TANKCAR AND TANKTRUCK
CONTRIBUTION TO STATE.-EMISSION BURDENS
Rail Tankcars
State
Arkansas
California
Illinois
Indiana
Kansas
Louisiana
Ohio
Texas
U.S.
Totals
Cars/yr
3000
2000
2000
4750
5000
2500
1750
5500
37,220
Oil and
grease
tons/yr
67
44
44
106
110
56
39
120
830
Suspended
solids
. tons/yr
330
220
220
530
560
280
194
610
4100
Phenol
tons/yr
2.5
1.7
1.7
3.9
4.2
2.1
1.5
4.6
31
Tanks/yr
90,000
240,000
290,000
220,000
172,000
250,000
200,000
400,000
5,010,000
Tanktrucks
Oil and
grease
tons/yr
31
84
101
77
60
87
70
139
1745
Suspended
Solids '
tons/yr
109
291
351
266
208
303
242
484
6068
Phenol
tons/yr
18
47
57
43
34
49
39
79
986
CONTROL TECHNOLOGY
This preliminary study indicates that control of air emissions
from the cleaning of rail tankcars and tanktrucks is not necessary
except for combustible gases and water soluble vapors such as
ammonia and chlorine. Combustible vapors are usually sent to
flares and burned. Vapors and materials such as chlorine are
absorbed in water and sent to the wastewater stream.
Until the late 1960s, little attention was given to wastewater
treatment in this industry. This inattention resulted because
waters were generally low in volume, installations were small,
and environmental impacts were considered relatively small in
comparison to those of other sources. As shown in the source
assessment (.1) , the severity of the wastewater effluents is
quite small, especially in comparison to other industrial
sources.
In recent years, several rail tankcar and tanktruck cleaning
wastewater treatment facilities have been upgraded in an effort
to meet environmental regulations. However, no installation is
known to have a completely satisfactory treatment system. Treat-
ment technology applicable to rail and truck wastewaters is for
the most part well known. However, the wide diversity in the
materials entering the wastewaters prevents the use of a single
specific treatment system by all companies. For this reason,
tank car and tanktruck cleaning companies are approaching their
individual problems by using one or more combinations of conven-
tional wastewater treatment methods.
Industry and tankcar manufacturers have developed new tank de-
signs which avoid the necessity to clean the tank. Of particu-
lar interest is the recent development of diaphram tanks (.5)
which can carry a different cargo each way between plants. The
diaphram prevents product contamination.
385
-------
One tank truck cleaning facility has had some success in the
treatment of a wide variety of chemicals in its wastewaters.
However, their system, shown in Figure 1, has not been able to
get the biological treatment to work, and problems still exist
with the COD and BOD levels. The effluent concentrations ob-
tained by this experimental facility are even higher than those
obtainable with the same type of equipment on effluents from
petroleum refineries. This indicates that the job of achieving
comparable effluent concentrations will be much more costly and
require further developmental work.
The costs of operating a system similar to the one shown in
Figure 1 are summarized in Table 10. At present, this type of
effluent cleaning system would cost about $25.00 per tanktruck
and about $100.00 per rail tankcar.
TABLE 10. NET OPERATING COST C1975) FOR THE 45
CUBIC METER PER DAY (12,000 GPD) WASTEWATER
TREATMENT FACILITY SHOWN IN FIGURE 1 (4).
" $AOOO gal
Labor - 1 operator, 9 hr/day - 5 days/wk 4.93
Carbon - 30.77
Chemicals - 90% cation polymer 3.85
Power - 0.77
Depreciation - 4.81
Sludge disposal - 40 gals/1000 gal wastewater 2.56
Biological oxidation - 2.31
TOTAL 50.00
CONCLUSIONS
This preliminary assessment indicates that (.1) air emissions
from the cleaning of rail tankcars and tanktrucks are higher for
high vapor pressure. High viscosity materials will show higher
emissions due to the amount of residual material and the clean-
ing method. 02) The emissions contribution to ambient air
quality degradation is insignificant for the typical cleaning
facility. The source assessment study shows that C31 waste
effluents do not pose a major impact on the water quality, but
that a reduction will be required to be compatible with effluent
concentrations from other treatment processes. Finally, (4) the
resulting cost, with existing technology, is high compared to
the benefit. Wastewater treatment could be made potentially
386
-------
more effective by combining the stream to be treated with other
wastewater, equalizing load to the treatment system. In addi-
tion, the jase of diaphrams may reduce the amount of cleaning
which will further reduce the present emissions and effluents.
REFERENCES
(1) Early, D. E., K. M. Tackett, and T. R. Blackwood. "Source
Assessment: Rail Tank Car, Tank Truck, and Drum Cleaning,
State-of-the-Art", USEPA, EPA-600/2-78-004g, April 1978.
(2) "Development Document for Proposed Effluent Limitations,
Guidelines, and New Source Performance Standards for the
Trucking Segment of the Transportation Industry Point Source
Category"; Office of Enforcement and General Counsel,
National Field Investigations Center, Cincinnati, Ohio,
1975.
(3) Slade, D. H., "Meterology and Atomic Energy", U. S. Atomic
Energy Commission. TID-24190, 1968.
C4) O'Brien, J. E. "A Demonstration Plant for the Treatment
of Waste Waters from Truck Tank Cleanings. AIChE National
Meeting, Atlantic City, N.J. Sept. 1976.
(5) Personal Communication, D. E. Fogelsanger of Goodyear
Aerospace Corp., Sept. 15, 1978.
387
-------
AV-TP-80/541
A NEW CONCEPT FOR THE CONTROL OF
URBAN INHALABLE PARTICULATE BY THE USE OF CHARGED FOG
John Scott Kinsey
and
Carol E. Lyons
AeroVironment Inc.
Stuart A. Hoenig
University of Arizona
Dennis Drehmel
U.S. Environmental Protection Agency
Industrial Environmental Research Lab, Research Triangle Park
Presented at the EPA Fourth Symposium on Fugitive Emissions:
Measurement and Control
New Orleans, Louisiana, May 28-30, 1980
Mr. Kinsey is an Air Quality Engineer and Ms. Lyons is a Project Manager in the
Environmental Programs Division of AeroVironment Inc., 145 Vista Avenue, Pasadena,
California 91107. Dr. Hoenig is a Professor of Electrical Engineering at the University of
Arizona, Tucson, Arizona 85721. Dr. Drehmel is with the U.S. Environmental Protection
Agency's Industrial Environmental Research Laboratory, Research Triangle Park, North
Carolina 27711.
ABSTRACT
This paper is directed to those scientists and engineers concerned with the control of
fugitive particulates in the inhalable size range (<15 um) from non-traditional sources in
the urban environment. A new type of electrostatically augmented spray nozzle (fogger)
is described which is intended to achieve a target control efficiency of 90% for inhalable
particulates. The theory of operation is explained, along with a physical model for
predicting the control efficiency of the fogger. Results of preliminary wind tunnel tests
are presented showing the degree of control actually achieved by the fogger for various
operating modes. Finally, plans for additional modifications to the fogger are discussed
which incorporate the information gained from preliminary testing in the wind tunnel.
388
-------
Introduction
The use of electrostatic precipitation for the collection of fine particulate
matter is common engineering practice and is well-documented in the literature. ' More
recently, electrostatics has, been jjsed to augment traditional wet dust collectors
(scrubbers) of varying design »''». Since 1977, one investigator at the University of
Arizona has been studying a new approach for the application of electrostatically
augmented water sprays (charged fog) on a bench scale for the control of fugitive dust
generated by various industrial processes ' . Commercial devices (foggers) resulting from
this work are now available from a firm in Pennsylvania . These foggers offer
significant economic advantages over more traditional fugitive dust control technology,
but also have problems which inherently limit both their effectiveness to control fine
particulate and their applicability to control fugitive dust from certain types of sources.
A more detailed discussion of these problems will be presented later in this paper.
In recent months, the health-related problems associated with suspended
particulate matter in the inhalable size range (< 15 um aerodynamic diameter) have been
of rising concern. The U.S. Environmental Protection Agency plans to develop and
promulgate a size-specific ambient air quality standard to address the inhalable particu-
late problem in the near future. If such a revised ambient air quality standard is
promulgated, it will necessitate even greater degrees of control for nontraditional as well
as traditional sources of particulate emissions, especially for those sources in the urban
environment where the problem is most acute.
This paper will describe a new type of electrostatically augmented spray
nozzle (fogger) for the control of urban inhalable particulate. The new fogger has been
designed to eliminate the major operational problems associated with conventional
equipment and is intended to achieve a target control efficiency of 90% for particulates
in the inhalable size range. The Spinning Cup Fogger Thrower (SCFT), as it is called, will
eventually be used to control emissions from mobile sources representing the major
contributors to the inhalable particulate problem in the urban environment.
The following discussion presents the theory of operation and design of the
SCFT, along with the results of preliminary wind tunnel tests conducted on the original
prototype. A modified SCFT is currently under development which incorporates a number
of design changes based upon the experimental data reported below and will likewise
undergo detailed testing in the near future.
Theory of Operation
Experiments have shawn that most dust is electrically charged and that
generally the charge is negative . By producing a positively charged spray, the dust
particles are attracted to the water droplets, and the number of particles which the water
drops will collect by impaction without electrical effect is increased substantially when
the drops and particulate are of opposite charge. The agglomerated particles of mixed
droplets and particulates will settle out according to the Stokes relationship.
A relationship describing the overall collection efficiency of suspended
particulate matter by water droplets has been described by Calvert and later by Cheng
and can be expressed as:
389
-------
E - 1 - exp
where:
E - Overall collection efficiency
Q. = Liquid volumetric flow rate
Q = Gas volumetric flowrate
t & = Contact time (or characteristic length for the total capture
process) ^
D . = Sauter mean droplet diameter
TI - Single droplet target efficiency
From the above expression it can be seen that the single droplet efficiency, 77,
is the most important determining factor in the control of fine particulates. The single
droplet efficiency depends, to a large extent, on the unit mechanism, or mechanisms, by
which the particles are captured, and which of these dominate. Such unit mechanisms
include direct interception; inertial impaction; Brownian diffusion; and the electrostatic,
magnetic, diffusiophoretic, and thermophoretic forces on the particles ' ' .
A detailed physical model has been developed by Grover and other collabora-
tors for the scavenging of atmospheric aerosol particles by charged free-falling raindrops.
It was determined that this model best describes the singJe .droplet efficiency in terms of
all the various capture mechanisms and forces involved ' ' . The development of the
model is quite complex and thus is being included in this paper by reference only.
Description of the Spinning Cup Fog Thrower
The SCFT consists of three main components, each serving a particular
function. These components are a rotary atomizer, an ionizer, and a vane-axial blower.
The rotary atomizer is comprised of a small hollow-shaft motor and spinning cup. Water
from a low-pressure source is introduced into the hollow shaft and flows toward the other
end where the spinning cup is mounted. Upon entering the rear of the cup the water
stream strikes a rotating spider which deflects the water to the sides. A sheet of water
then flows toward the lip of the cup where droplets are formed by centrifugal force
normally leaving the edge of the cup tangentially. The droplets are then charged by a
stream of positive ions produced by an ionizer containing numerous small discharge
needles. Current to the ionizer ring is supplied by a vacuum tube high-voltage DC power
supply. The charged fog droplets are then deflected and projected forward by a flow of
air supplied by the vane-axial blower into which the ionizer and rotary atomizer have been
mounted. A diagram of the SCFT used during the wind tunnel tests described below is
shown in Figure 1.
As was previously stated, the commercially available foggers have certain
problems inherent in their design which not only limit their application to certain types of
sources, but can also severely limit the degree of control achieved on fine particles. The
first of these problems is that the existing foggers use either a standard pressure-type or
twin-fluid nozzle to atomize the water, thus necessitating a substantial supply of high
pressure water or air for proper atomization. In addition, these types of nozzles have a
tendency to clog if the water supply contains a high concentration of suspended solids.
This problem has been eliminated in the SCFT by a rotary atomizer which needs neither
390
-------
compressed air nor water pressure to generate a fog of fine water drops. In addition, a
rotary atomizer is insensitive to any solids found in the water supply.
The second and more severe problem with existing equipment involves the type
of charging mechanism being used. The existing foggers use an induction ring which tends
to charge only those droplets on the outside of the spray cone during heavy sprays with
the remaining droplets having little or no charge. Since the electrostatic attraction of
dust to water droplet is critical to the efficient capture and agglomeration of fine
particles, this is a significant problem. The SCFT solves the problem by using direct
charging which effectively charges most, if not all, of the droplets generated by the
atomizer, significantly increasing its potential to control particles in the inhalable size
range. By these two basic modifications, the SCFT is expected to have greatly reduced
water and power consumption over previous equipment, in addition to being able to more
effectively control emissions from mobile sources.
Experiment
In order to verify theoretical relationships with experimental data, a series of
tests were conducted in a 4-foot by 4-foot by 60-foot wind tunnel provided by the
University of Arizona. A standard dust (AC Coarse) of known particle size and density
was pneumatically fed into the entrance of the tunnel by a Vibra-Screw feeder and
compressed air which was homogeneously mixed with the incoming ambient air by a
diffuser cone. The SCFT was mounted approximately two equivalent diameters
downstream of the tunnel entrance with the fog and air in cocurrent flow. Plexiglas
viewing ports in both sides of the tunnel were provided at the discharge end of the fogger
so that the wet/dry interface could be more closely observed. Two sets of sampling ports
were provided in the front vertical wall of the tunnel; one set, approximately two
equivalent diameters downstream of the discharge end of the fogger, was for determining
charge/mass ratio and droplet size, and the other set at the far end of the tunnel was for
determining particulate mass concentration and size. An axial blade fan pulled
approximately 18,000 dscf/min of air through the tunnel. Figure 2 is a schematic
representation of the University of Arizona wind tunnel.
To determine the control efficiency of the fogger under various experimental
conditions, particulate samples were collected according to the test plan shown in
Figure 3. A sample train was especially designed for this purpose to isokinetically extract
a representative sample of particulate matter from the wind tunnel. The sample was
later analyzed gravimetrically to determine both the total mass concentration and
particle size distribution for each test scenario. This train consisted of a pitobe and
associated inclined manometer connected by a flexible sample line to a horizontal
elutriator. The elutriator was originally designed to have a 15 urn particle diameter cut-
point (50% efficiency) at a flow of 20 scfm. Behind the elutriator was mounted a standard
Sierra Instruments five-stage cascade impactor which discharged to a silica gel trap for
the determination of moisture content. The flow was measured by an NBS-traceable
Laminar Flow Element manufactured by the Meriam Instrument Company. The Laminar
Flow Element was located downstream of the silica gel trap and connected to a standard
Dwyer magnahelic gauge for reading differential pressure. The flow rate was metered by
a large capacity needle valve located downstream of the Laminar Flow Element followed
by a Cadillac centrifugal blower acting as prime mover. A photograph of the sampling
train with the various components is shown in Figure 4.
391
-------
For sample recovery, the front portion of the sample train and horizontal
elutriator were rinsed with de-ionized water into separate plastic sample containers for
transport back to the laboratory. There the liquid/particle mixture was quantitatively
transferred to tared beakers and evaporated to dryness to determine the total mass
collected. The samples were then resuspended in distilled water and deagglomerated
ultrasonically. The particle size was determined by a Leeds <5c Northrup Microtrac
Analyzer using a ruby laser and forward-scatter optical technique. The samples collected
on the various cascade impactor stages were also analyzed gravimetrically using a
standard analytical balance.
The charge-to-mass ratio was determined using a special sample train
developed by AeroVironment according to a method provided by Dr. Hoenig at the
University of Arizona. This train consisted of an insulated stainless steel probe tip
mounted on a standard glass midget impinger. The probe was connected electrically to
copper wool packing placed inside the midget impinger which also was connected by a
shielded cable to an electrometer (Hewlett-Packard Model <*25A) and associated strip-
chart recorder. The impinger was immersed in a Dewar flask containing dry ice. A
sample of fog droplets was then isokinetically extracted from the tunnel through the
sample ports at the discharged end of the fogger. As the droplets moved through the
impinger they were condensed out of the gas stream and frozen, thus transferring their
charge to the copper wool packing. The charge was then measured by the electrometer
and recorded by the chart recorder. The mass of water collected in the impinger was then
determined gravimetrically at the end of each test run using a portable triple-beam
balance. Isokinetic conditions were maintained by a Hastings mass flowmeter and needle
valve with a standard diaphragm pump acting as prime mover. A standard pitot tube and
associated inclined manometer were mounted next to the impinger/Dewar flask assembly
with adequate distance between them to avoid aerodynamic interference.
The size of the droplets was measured simultaneously with the charge-to-mass
ratio by a KLD Model DC-2 droplet counter. This unit uses a hot-wire technique to
measure-the size of the droplets in 1* incremental ranges from 1 um to greater than
450 um ' . The hot-wire probe was mounted a sufficient distance from the impinger to
avoid aerodynamic interference with it or the pitot tube used to measure the gas velocity.
Results
Data obtained from the first series of tests, as described above, of the total
mass concentration measured for various experimental conditions are shown in Figure 5
for a low dust feed rate to the tunnel, and in Figure 6 for a high dust feed rate. It was
also determined during the course of sample collection that larger, agglomerated particles
which could not settle out of the gas stream in the tunnel due to the high velocity were
being collected in the first bend of the sample probe. For this reason, Figures 5 and 6 also
show a second value, where available, for the total amount of particulate matter caught
by the sample train less the larger particles collected in the probe.
Conclusions
From the data presented in Figures 5 and 6, it can be seen that a significant
reduction in dust concentration can be achieved with the SCFT, but that it does not
currently meet the 90% target efficiency. It can also be determined that the fog droplets
392
-------
must have a high charge-to-mass ratio for effective control of fine paniculate and, as
would be expected, that uncharged fog or fog with little electrostatic charge does not
provide nearly as good results.
Based upon these conclusions, it was decided that certain modifications to the
design of the fogger or to the operation of the wind tunnel were necessary to determine
the optimal effectiveness of the SCFT under controlled conditions in relation to the
target results. For this reason, the velocity of the gas moving through the tunnel was
reduced to allow the larger particles to settle out and to increase the particle/droplet
contact time. It was also determined that major design changes to the ionizer would be
necessary to increase the charge-to-mass ratio.
A second prototype of the SCFT is currently being fabricated which will be
tested in the wind tunnel at University of Arizona in the near future. The modified
version will include a completely redesigned atomizer to provide more effective spray
pattern. It also incorporates a new method for charging the droplets which preliminary
data show to be markedly more effective. It is anticipated that the modified SCFT will
easily meet the required target efficiency during the upcoming tests.
Metric Conversion Factors for non-metric units used in text:
1 foot = 0.3048 m
1 dscf/min = 0.0283 dscm/min
1 scf m = 0.0283 scm/min
1 gr/dscf = 2.2883 gm/dscm
393
-------
References
1. H.J. White, Industrial Electrostatic Precipitation, Addison-Wesley, Reading, MA
(1963).
2. HJ. White, "Electrostatic precipitation of fly ash," 3APCA, 27 (1): 15-21; (2): 114-
120; (3): 206-217; (4): 308-318 (1977).
3. D.C. Drehmel, "Advanced electrostatic collection concepts," 3APCA, 27 (11): 1090-
1092 (1977).
it. M. Pilat and D.F. Meyer, "University of Washington electrostatic spray scrubber
evaluation," EPA-600/2-76-100 (NTI5 PB 252653), University of Washington, Seattle,
WA (1976).
5. C.W. Lear, "Charged droplet scrubber for fine particle control: Laboratory study,"
EPA-600/2-76-249a (NTIS PB 258823), TRW Systems Group, Redondo Beach, CA
(1976).
6. W.F. Krieve and J.M. Bell, "Charged droplet scrubber for fine particle control: pilot
demonstration," EPA-600/2-76-249b (NTIS PB 260474), TRW Defense and Space
Systems Group, Redondo Beach, CA (1976).
7. S. Calvert, 3. Goldschmid, D. Leith, and D. Mehta, "Wet scrubber system study
Volume 1, scrubber handbook," EPA-R2-72-118a (NTIS PB 213016), Ambient
Purification Technology, Inc., Riverside, CA (1972).
8. S.A. Hoenig, "Use of electrostatically charged fog for control of fugitive dust
emissions," EPA-600/7-77-131 (NTIS PB 276645), University of Arizona, Tucson, AZ
(1977).
9. S.A. Hoenig, "Fugitive and fine particle control using electrostatically charged fog "
EPA-600/7-79-078 (NTIS PB 298069), University of Arizona, Tucson, AZ (1979).
10. Ritten Corporation, Stop Dust Pollution, manufacturer's brochure, Ardmore, PA
(1980).
11. L. Cheng, "Collection of airborne dust by water sprays," Ind. Eng. Chem. Process
Des. Develop., 12 (3): 221-225 (1973).
12. S.N. Grover, H.R. Pruppacher, and A.E. Hamielec, "A numerical determination of
the efficiency with which spherical aerosol particles collide with spherical water
drops due to inertial impaction and phoretic and electrical forces," 3. Atmos. Sci
34: 1655-1663(1977).
13. K.V. Beard and S.N. Grover, "Numerical collision efficiencies for small raindrops
colliding with micron size particles," 3. Atmos. Sci., 31.: 543-550 (1974).
14. S.N. Grover and K.V. Beard, "A Numerical determination of the efficiencies with
which electrically charged cloud drops and small raindrops collide with electrically
394
-------
charged spherical particles of various densities," 3. Atmos. Sci., 32: 2156-2165
(1977).
15. H. Medecki, K.C. Wu, and D.E. Magnus, "Development of droplet sizing for the
evaluation of scrubbing systems," EPA-600/7-79-166 (NTIS PB 80-10857^), KLD
Associates, Inc., New York, (1979).
16. D.E. Magnus, H. Medecki, and K.C. Wu, "Sensing of droplet size and concentration in
pollution control equipment," Proceedings of the 4th Joint Conference on Sensing of
Environmental Pollutants, Amer. Chem. Soc., 605-609 (1978).
395
-------
0.6 H.P. MOTOR
VANE AXIAL FAN
HOLLOW SHAFT MOTOR
CHARGED
FOG
SPINNING
CUP
FLOW STRAIGHTENER
Figure 1. Schematic of spinning cup fog thrower.
-------
•*c
Air In
Pleni
F*^"
3^ DiffuserV
\Cone ^-
let
im
& / /
^
Spinning Cup
Fog Thrower
HT3
V
A
Viewing
Port
Viewing
Port
/ / g
0 n
Particulate y / Axial
O O Sampling Hy Blade
C/M Sampling Ports 1 \ Fan
Port o \s
/ / 1
/ / / /
Dust from
Vibrascrew Feeder
I .
10'
60'
Figure 2. Schematic of University of Arizona wind tunnel.
(I foot = 0.30^8 meters)
-------
Heavy Dust Loading
Ambient Relative
Humidity
1
High
Charge/
Mass
1
Low
Charge/
Mass
Light Dust Loading
Ambient Relative
Humidity
Figure 3. Test plan.
398
-------
Meriam Laminar
Flow Element
Probe, Pitot Tube,
Thermocouple
A
Horizontal
Elutriator
Sierra 5-Stage
Cascade Impactor
Silica Gel
Trap
Control
Valve
Figure k. Particulate sampling train.
-------
o
o
U.U13
1 0.010
m
L.
g
O
c
d
s 0.005
3
Q
n.ooo
-
—
-
—
—
-
| Total Catch
Q Total - Probe Catch
T
Uncontrolled Fog High L/G; Low L/G; High L/G; Low L/G;
No Charge High C/M Low C/M Low C/M High C/M
Figure 5. Summary of test data — light dust loading.
(1 gr/dscf = 2.2883 gm/dscm)
-------
0.03
0.025
0.020
2 0.015
§
0.010
0.005
0.000
I
|
v
y
gp£
U-U
Sf>jc
u=U
I Total Catch
n Total - Probe Catch
-2P
I
o
Figure 6. Summary of test data — heavy dust loading.
(1 gr/dscf = 2.2883 gm/dscm)
-------
CONTROL METHODS FOR FUGITIVE AREA SOURCES
Presented at the Fourth Symposium on Fugitive Emissions!
Measurement and Control
May 28-30, 1980
Monteleone Hotel, New Orleans, Louisiana
Dennis J. Martin, Edward Brookman and Les Hirsch
TRC Environmental Consultants, Inc.
Dr. Dennis Drehmel
U.S. Environmental Protection Agency
ABSTRACT
The prime control techniques for fugitive area sources are water sprays,
enclosures, wind screens and chemical stabilizers. A review of the
literature indicated that while cost data was readily available for these
methods, estimates of their efficiencies were not supportable from the given
data. Nonetheless, future work on wind screens and chemical stabilizers
appears warranted. An ongoing field study of a wind screen made of a
synthetic material indicates that efficiencies of over 60 percent should be
obtainable at a reasonable cost when applied to storage piles.
402
-------
1.0 INTRODUCTION
Due to the increased regulatory emphasis on fugitive emissions, interest
in the type and efficiency of controls applicable to sources generating
these emissions has been steadily growing. Prime among the sources of con-
cern are area sources, not only because of the magnitude of their emissions
but also since they are not amenable to types of controls with which
environmental engineers are most familiar. Furthermore, data on these types
of controls are scarce and scattered throughout the literature and, to date,
there has not been an attempt to compare the various control options on a
cost effectiveness basis to achieve a priority ranking for control strategy
development.
In order to rectify this situation the Particulate Technology Branch of
EPA's Industrial Environmental Research Laboratory (at Research Triangle
Park) is sponsoring a program being conducted by TRC-Environmental
Consultants. The purpose of the program is two-fold: first, a review of
the literature with respect to fugitive controls applicable to storage
piles, unpaved parking lots, and disturbed soil surfaces associated with
construction activities; and second, presently ongoing, laboratory studies
on the more promising of the identified controls, hopefully pointing the way
to future research.
This paper presents the results of the completed Phase I effort and
discusses some of the preliminary findings of the Phase II projects.
403
-------
2.0 PHASE I - LITERATURE REVIEW
2.1 Techniques Identified
Methods, identified from a literature search as prime control techniques
for fugitive emissions from area sources, were:
o Water Sprays.
o Chemical Stabilizers.
o Enclosures.
o Windscreens.
Other control methods (e.g., tailings ponds, vegetative stabilizers, and
housekeeping procedures) are not used as extensively as the four cited above
and are not dealt with in this paper although they will be covered in the
final report to EPA. Brief descriptions of the prime control methods are
given below.
2.2 Water Sprays
It was found that, while a number of chemicals have been used to reduce
emissions from aggregate storage piles and haul roads, plain water is still
the prime wetting agent used. Water use is attractive due to its relatively
low cost and ease of application. Effectiveness varies with amount applied,
the frequency of application, and meteorological parameters which have a
direct bearing on the evaporation rate of moisture. Prom the latter point
it can be inferred that the spray action itself is not an aspect of the
control; rather, the use of water to increase the apparent density of the
dust is the control mechanism. Sprays are effective only for very large
particles at the droplet sizes produced by most commercial spraying equip-
ment. Costs vary with the method of application (e.g., trucks, spray bars)
and the amount applied. For large storage piles, the cost of the equipment
alone can be high: large watering trucks cost about $250,000. Large coal
fields which require continuous watering also have high labor costs
associated with this method of control. The direct labor cost of watering
one 50 acre coal field can exceed $100,000 per year.
Prom the literature, a reasonable estimate of the accuracy of this tech-
nique can not be made due to the conflicting and inaccurate data found.
This point is discussed later.
Disadvantages to using water as a control for area sources are the
necessity of having an adequate water supply available, the need for
frequent reapplication, and the contribution to the leachate amount.
2.3 Chemical Stabilizers
Three basic chemical stabilizing agents are used, each operating to
different principles: wetting agents, hydroscopic salts, and surface
crusting agents.
404
-------
Wetting agents have been used to increase the effectiveness of water and
some of the chemical stabilizing agents by reducing surface tension and
thereby enabling the water (or chemicals) to be spread more evenly and over
a greater surface area then before. Disadvantages are the same as those for
water alone, with the addition of the leaching of the chemical into the
groundwater aquifer or water treatment system dealing with pile runoff.
Hydroscopic salts, by attracting the water out of the air, increase the
surficial moisture content of the dust, increasing its apparent density and
thereby reducing its emissions. The prime disadvantage to the use of such
materials is the fact that they are water soluble and can therefore be
easily washed away. A second disadvantage is that the salts (e.g., CaCl2)
are corrosive and may not be able to be used in all instances.
Surface crusting agents, third type of chemical stabilizers, can be
composed of various compounds? e.g., styrene/butadiene latices, acrylic
latices, vinyl compounds, synthetic polymers, lignosulfonates, and
petroleum-based resins. These chemicals are applied wet and when dry form a
hard crust on the surface of the material being treated. Costs for these
chemicals, even for similar types, vary widely (see Table 1). Latex
compounds can be combined with slack wax and filler materials (such as coal
dust) which, when applied at elevated temperatures, form a waterproof
covering of the pile. Costs for applying this material are reasonable (see
in Table 2 for various sized piles). Comparable costs for operations such
as compaction and latex crusting are shown in Table 3. The direct savings
referred to occur from a reduction in the oxidation and water uptake of coal
and are on a yearly basis. Total benefits are based on reductions in the
cost of treating the leachate and runoff of the pile, the elimination of the
need to monitor the pile's temperature, and other related functions. Cost
benefits of hot melt coatings apply only to degradable products such as coal
where waterproofing the pile yields these types of results. While these
compounds have never been tested for effectiveness it is reasonable to
assume that initial efficiencies approach 100 percent. The effective "life"
of these materials has not been documented.
2.4 Enclosures
Total enclosures should, by definition, eliminate wind-caused fugitive
emissions since the wind would no longer come into direct contact with the
source. However, emissions still occur since ventilation of the enclosure
is necessary to remove the dust induced by material transfer operations.
These emissions are subject to normal stationary source control techniques,
however, and are effectively controlled. The cost of this method can be
extremely high and varies according to the size and type of structure,
associated conveying equipment, etc. Of course, small piles (e.g., sand
used by road crews) are effectively and inexpensively controlled by
weighted-down tarpaulins.
2.5 Windscreen
Windscreens are porous windbreaks which reduce the velocity of the air
behind them. Figure 1 illustrates the performance of a screen having a
porosity of 50 percent. Solid (zero porosity) fences are not used as wind-
405
-------
TABLE 1
CHEMICAL STABILIZER COSTS
Styrene/Butadiene Latices $25 to $50/103 Meter2
Acrylic Latices $50 to $160/103 Meter2
Vinyl Compounds $12 to $50/103 Meter2
Synthetic Polymerts $150 to $420/103 Meter
Lignosulfonates $25 to $50/103 Meter
TABLE 2
HOT MELT COATING COSTS
Pile Size
Metric Tons
45,360
90,720
226,800
453,600
680,400
Treatment
Compaction
Latex Crust
Surface Area
Meters
6,782
13,006
32,515
65,030
97,545
COST SUMMARY FOR
Cost to Treat
$/Metric Ton
Cost to Treat/ $
2,623
5,137
12,788
25,466
38,255
TABLE 3
MATERIALS APPLIED
Direct Savings
$/Metric Ton
Cost $/Metric Ton
0.058
0.057
0.057
0.057
0.057
TO COAL PILES*
Total Benefits — Value
Equivalent $/Metric Ton
0.088
0.126
Hot-Melt Coating 0.144
0.303
0.716
1.15
0.89
1.20
2.09
* Assuming a coal pile of 250,000 metric tons at $20 per metric ton
406
-------
1QH 2QH 30H
DISTANCE IH FENCE HEIGHTS
VERTICAL SECTION (ALONG £ 3F FENCE)
5QH
6QH
DISTANCE IN FENCE HEIGHTS
GROUND PLAN
FIGURE 1: VJIND VELOCITY PATTERNS
4008-001
407
-------
breaks for large sources since the downstream turbulence created by the
breaks may actually induce more emissions than those which would have
occurred without the breaks. Figure 1 shows that a fugitive emission source
placed within 10 fence heights downstream of the fence would have the air
speed reduced to 40 percent of its upstream value in the vicinity of the
source. (Note that this does not mean that the emissions would be reduced
by 60 percent.) The emission reduction depends on the relationship between
the air speed and the amount of particulate matter emitted and has been
estimated to be anywhere from a straightforward E V1 to E V"9 and just
about every exponent in between. Also, note that, for every type of dust,
there is a critical threshold velocity at and below which only an insigni-
ficant amount of particulate matter would be removed. This threshold
velocity depends on such factors as particle size, adhesive forces par-
ticular to the dust, surface moisture content, and bulk density. The
overall efficiency of a windscreen, therefore, depends on the threshold
velocity of the dust, the relationship between emission and velocity and, of
course, local meteorological parameters.
For example, suppose that 50 percent of the time the wind at a given
location comes from a 30° sector and that a storage pile can be shielded
from its effect by a semicircular fence. Air velocity readings from the
noted direction can be broken down as: - 10 percent of the time, less than
10 kilometers per hour; 60 percent of the time, 10-20 kilometers per hour:
20 percent, 20-30 kilometers per hour and 10 percent, averaging 40
kilometers per hour. If the threshold velocity of the dust is about 10
kilo- meters per hour and if the emissions are proportional to kv3 then,
before control, emissions can be described as:
Emission rate - k (0.6) 203 + ka(0.2)(30)3 + k(0.1)(40)3
= k (4800 + 5400 + 6400)
= k 16,600
when the highest value of the velocity range is used as the average. After
installing the fence, only the last two ranges will be beyond the threshold
and the emission would be characterized as:
Emission rate = k (0.2) (15)3 + k (0.1) (20)3
- k (675 + 800)
= k 1475
There is a reduction over the uncontrolled wind from the one sector of about
91 percent. To determine the overall efficiency of the fence, the emissions
from the unshielded section of the piles would also have to be determined.
If these emissions were about 30 percent of the total (k 7115) , then the
overall efficiency of the fence would be about 64 percent. Naturally, the
closer the fence is to the pile, and the more unidirectional the wind is,
then the greater is the efficiency that can be achieved.
Windscreens can be made of various materials and can be arranged in
various configurations. Materials of construction include structural steel
406
-------
wooden slats in a cyclone fence, and a synthetic material of the type used
for glare screens and snow fences. Cost varies with material, design, and
height. A movable 4 meter high fence could cost about $3,280 per linear
meter if made of structural steel to enable it to be picked up and moved by
forklifts. A 25 meter high"permanent screen could cost the same, if made of
snow fence material.
409
-------
3.0 COST-EFFECTIVENESS FACTOR
The prime end result of the Phase I literature review was to have been
the development of a cost-effectiveness factor for each control method
identified.
In general, although cost data were relatively easy to obtain from the
literature, chemical manufacturers, etc., efficiency data were, for all
practicable purposes, non-existent. This is understandable for a number of
reasons. There are just not that many reports containing hard data on
control efficiency for fugitive sources. At the beginning of the study, a
computerized search of APTIC, NTIS, Engineering Index, and Pollution
Abstracts yielded fewer than 200 potential sources of information. Most of
the publications examined did not contain usable data. Many articles were
oriented toward control of fugitive dust in order to reduce nuisance com-
plaints. Typically these reports described the use of a type of control,
housekeeping procedure, or combination thereof which reduced the incidence
of the complaints. While these studies were useful in showing the state-of-
the-art control, actual data test were, for the most part, not given or
available. • In the relatively few instances where control efficiency data
were obtained or noted in the publication, they did not meet the preset
criteria for acceptance. These criteria were:
o Tests had to be performed upon the sources using acceptable method-
ology and instrumentation.
o Judgements on the effectiveness of a control under different
conditions than those tested were not permitted unless strong evi-
dence of a trend was noted and a basis in theory established for
the phenomeon.
o Test data and conclusions were to be considered absolutely valid
only for the source tested. Extension to similar sources required
that those factors affecting particle transport be quantified
during the tests and that data on these factors be available for
the similar sources. For example, particle size estimates would
have to have been made for the pile tested and data available
indicating that the size is similar for similar sources. it was
not considered necessary that particle size data be available on
the similar sources (although this would have been the ideal
situation); rather, engineering judgement could be used to infer
that the size was the same.
o Data from tests performed on one type of source could not be
extended to dissimilar sources (e.g., test results on sand and
gravel operations could not be applied to coal piles) without
extensive supportive evidence.
410
-------
o Wind tunnel testing could only be used when compared with field
results or where various control options were tested under similar
conditions.
Applying these criteria to the publications found eliminated, for all
practical purposes, all of the data from consideration. This should not
have been a surprise for a number of reasons. As noted before, available
data are not extensive. Most efficiencies of available data are derived,
rather than calculated directly. Also, to confuse the issue somewhat, some
numbers quoted in the literature appeared to increase each time they were
referenced. As an example of how this can occur, one reference stated:
"A study on the effect of watering on construction sites in-
dicates that extensive wetting of the soil may reduce
emissions from existing construction operations up to 60 or 70
percent. The study suggested that wetting of access roads
twice a day with an application of .5 gal. of water per square
yard will suppress dust emissions from existing baseline con-
struction practices by 50 percent."
The study cited actually stated:
"Watering on construction sites produced a wide variation in
apparent control efficiencies, due in part to the highly vari-
able nature of the emission sources. Activity logs kept at
the construction sites showed that some sampling periods with
extensive watering were accompanied by hi-vol readings 60 to
70 percent lower than anticipated with no watering, while on
other days the apparent effect of the watering was negli-
gible. The same variations were noted in analyzing data from
sampling periods with rainfall. With daily watering and com-
plete coverage, average control efficiency is about thirty
percent. This value is partially verified by another study
indicating a 30 percent reduction in dust emissions over con-
tinuously traveled gravel and dirt roads on days when their
surface was moist. However, with watering twice a day at the
same application rate, a reduction of 50 percent appears
feasible."
As can be seen, these studies did not establish the effect of
watering, but rather, extrapolated extremely limited data to
produce undependable results.
In conclusion, evaluation of the literature indicated that
there is virtually no accurate or usable data relating to control
efficiency for fugitive area sources and that, as a result, this
data must be generated.
411
-------
4.0 PHASE II PROGRAMS
The Phase I results indicated that the present state of knowledge with
respect to fugitive emission generation and transport is not extensive.
Available information on control technique effectiveness is practically
nonexistent. To alleviate this situation somewhat, Phase II was oriented
toward a better understanding of the transport phenomena and a first-cut
approach to quantifying the effectiveness of the two most promising tech-
niques, chemical stabilizers and windscreens. While Phase II is still
ongoing, some partial results are encouraging and are presented here for
review.
4.1 Particle Transport
While much has been established on a theoretical basis describing the
behavior of particle transport, agreement between theory and field results
has been poor. Part of the difficulty lies in the oversimplification of the
approach in field studies undertaken to date. For example, only recently
has surficial moisture content and the size distribution of the dust avail-
able for transport been measured during tests at storage piles. Attempting
to force-fit data to a given transport model has resulted in simplistic
equations good only for the pile on which the data were taken. To resolve
this, a wind tunnel study is being performed to quantify the effects of pile
shape, moisture content, particle size, and turbulence intensity on the
quantity of emissions generated. To date, only the results on pile shape
are available. These data indicate that the oncoming wind is accelerated up
the slope of the pile such that the air speed at the top of the pile is
increased about 50 percent over ground air speed. The amount of the
increase depends somewhat on slope. Slopes with leading angles equal to
about 10° produce less than half the increase in velocity at the top of
the pile than slopes with angles greater than 20°. The high velocities at
the top of the pile induce flow patterns as shown in Figure 2. While this
schematic is for a block, the same phenomenon holds true for a pile with a
sloping edge. Therefore, in predicting the emissions from a pile, the air
speed used in a predictive equation may not be amenable to a single value;
rather, emissions from different sections of the pile can be referenced to
the approach ground speed only if different equations are used for each
section. Ongoing work in TRC's wind tunnel will be used to obtain the form
of these equations so that future field studies can use them to set up their
testing programs and evaluate their results.
4.2 Windscreen Studies
One of the more promising control alternatives identified in the liter-
ature was windscreens. A major drawback to their use, however, is the
materials of construction; usually, wood slats in a cyclone fence or snow
fence type materials are used. While such materials are acceptable for
fences of limited height, protecting large piles could require fences of 15
meters or more in height. Long stretches of haul roads, while probably not
412
-------
•s.
X.
V, ^
V,>V,>V3
TOP VIEW
SIDE VIEW
FIGURE 2: FLOW PATTERNS ACROSS A BLOCK PILE
4008-002
413
-------
needing a high fence, would require one of sufficiently reasonable cost to
be applied to long stretches of road. TRC's ongoing investigation includes
an evaluation of a synthetic fence material manufactured by Julius Koch,
Inc., of New Bedford, Massachusetts. Preliminary results indicate that
performance of this material, available in varying porosities, is equivalent
to those of other fence materials. Since material costs, including con-
necting hardware, are approximately $19.7 per linear meter foe the 1 meter
wide section tested, this material is a preferred option in most instances.
Ongoing studies are being performed to evaluate the use of a semicircular
and a circular fence.
414
-------
5.0 CONCLUSIONS
An evaluation of the available literature indicated that stated effi-
ciencies for the prime control techniques applicable to fugitive area
sources were based on unreliable data of limited quantity. For this reason,
future testing must be directed toward developing this data. Ongoing wind
tunnel studies at TRC indicate that emissions from a storage pile depends on
pile configuration and that seperate equations for different pile sections
may have to be developed to accommodate this phenomenon.
Full-scale field testing of a synthetic fiber windscreen indicates that
the screen performs equivalently to other materials and that, due to its
cost, this material is preferable to others. Ongoing programs are eval-
uating semicircular and circular screens.
415
-------
CIVIL ENGINEERING FABRICS APPLIED
TO FUGITIVE DUST CONTROL PROBLEMS
Barry Levene
Air and Hazardous Materials Division
EPA Region VIII
Denver, Colorado
Dennis C. Drehmel
Environmental Protection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina
ABSTRACT
Fugitive dust sources contribute substantially to both total
suspended particulate (TSP) and inhalable particulate. Emission rates
from unpaved roads, construction activity, and other area sources have
been estimated at 320 x 106, 27 x 106, and 4 x 106 tonsVyear, respectively.
This may be compared to 14 x 106 tons/year, the emission rate for all
point sources. Within a given industry, the contribution of fugitive
emissions can be more than 50% with the added problem that these particles
may contain toxic or hazardous substances.
Civil engineering fabrics are in common use for a number of purposes
including: 1) ground stabilization; 2) subsurface drainage; 3) railroad
construction and maintenance; 4) sediment control; and 5) erosion control.
Fabrics are available from Celanese, DuPont, Monsanto, and Philips Fibers
under the trade names of Typar, Bidim, Supac, and Mirafi. Use of these
fabrics for air pollution control is a recent development and the subject
of recent and future field tests.
Reduction of emissions from unpaved roads is achieved by covering the
road first with the fabric and then with coarse aggregate. Unpaved
parking lots, inactive piles, and construction sites could be controlled
in the same way. During tests in November 1979 on an unpaved road
constructed with Bidim, the average reduction in TSP was 58% and in
inhalable particle concentrations was 46%. More tests are planned at
both eastern and western sites in the spring of 1980 and a year-long
monitoring at a western site will begin in the summer of 1980.
*To convert tons to kilograms, multiply by 907.
416
-------
INTRODUCTION
In the arid regions of the western United States, local sources of
fugitive dust appear to be the most significant contributors to ambient
levels of total suspended particulate (TSP). These sources also contribute
to the inhalable fraction although to a lesser degree. Vehicles travelling
on unpaved roads are a major portion of these emissions, ranging from
54% of the total annual ISP emissions in Phoenix to 9% in Denver and
48% in Colorado Springs. An integral part of a particulate control
plan is identifying fugitive emission control measures for unpaved
roads. In addition to unpaved roads, construction sites, haul roads,
and open areas within industrial sites are important contributors to TSP
and may be sources of undesirable species such as lead.
Traditional control techniques to reduce unpaved road emissions,
such as paving, chemical stabilizers, watering, gravel ing,gand traffic
controls, have been analyzed in several studies. * * ' ' ' However,
the use of a civil engineering fabric medium underlying the aggregate
roadbed is a promising new technique which, after further study, may
have application as an alternative control strategy in the western
United States. It may, in addition, have other benefits such as reduced
road maintenance. Fabric under aggregate may also be used to control
unpaved parking lots, construction sites, and disturbed open areas. For
these reasons, The Environmental Protection Agency (EPA) Regional Office
in Denver is assisting EPA's Industrial Environmental Research Laboratory
at Research Triangle Park in North Carolina in planning a study of the
effectiveness of this type of road treatment on an unpaved road segment
on the Ft. Carson Army Base near Colorado Springs, Colorado. This paper
describes the use of civil engineering fabrics and their role in the
context of the EPA particulate control program.
BACKGROUND
Even with the application of Reasonably Available Control Technology
(RACT) on all stationary sources, many TSP nonattainment areas are
projected to exceed the NAAQS after the Clean Air Act's 1982 deadline.
Control of the remaining nontraditional sources has presented a formidable
planning task to those responsible for State Implementation Plans (SIP's).
The problem has been complicated by the lack of definitive information
on the cost-effectiveness of the various control measures and the like-
lihood that a new inhalable particulate standard is in the offing. Few
States or local governments are prepared to commit large amounts of
public funds to projects of uncertain value in achieving the new
inhalable particulate standards. The Colorado Air Quality Control
Commission, in fact, has stated in policy paper 78-1 that it Is not
reasonable to gear a control program to meet a questionable standard
and that it is the policy of the Commission to direct control efforts at
the respirable fraction.
The EPA Administrator recognized these problems in his February 24, 1978,
guidance on SIP's for nonattainment areas. This guidance states that after
RACT for traditional sources, States could conduct demonstration projects
and studies in lieu of adopted specific measures as long as the data on
417
-------
costs and benefits could be analyzed and strategies adopted and implemented
in time to meet the 1982 deadline.
As SIP's were submitted in the months following January 1, 1979, it
became apparent that most states had opted for the demonstration project
alternative, although a few areas, such as Phoenix and Colorado Springs,
committed to extensive paving programs, which have since been delayed.
As a result EPA has provided further guidance on what should be the
emphasis areas in 1980 while the standards review is in progress. The
first priority is to assure that RACT is applied to all traditional sources.
The second is for those areas which have not yet done so to complete a
problem assessment of sources, emission, and nature of particulate (size
distribution in particular). EPA will be supporting several major studies
of available control measures to determine source/receptor relationships,
effectiveness, and air quality impacts.
EPA's Denver Regional Office has supported many such studies which
may eventually lead to widespread implementation. Through a grant to the
State of Colorado, a study design is now being developed to conduct an
evaluation of alternative snow control practices. The demonstration study
of civil engineering fabric in unpaved roads fits in well with EPA guidance
by providing an ideal opportunity to evaluate promising new technology.
FABRIC DESCRIPTION
At least four major manufacturers of civil engineering fabrics market
in the United States. These companies are Celanese, duPont, Monsanto, and
Philips Fibers. A description of some of their products is shown in Table 1.
Most manufacturers offer a range of properties according to thickness.
Data in Table 1 should be interpreted as showing that generally the same
range of properties is covered by all manufacturers. Usually civil engineer-
ing fabrics are sold for road stabilization, support, drainage, erosion
control or reinforcement. Such fabrics have been used under railroad
tracks, rip-rap, and paved and unpaved roads. Fabrics are available in
strengths to take loads from haul trucks and railroad trains, are rot
resistant, and have been estimated to have a useful lifetime of 12 years.
The approximate 1980 cost is $1.00 per square yard.
FIELD TEST SITE DESCRIPTION
The test section is a 610 meter (2,000 foot) segment of unpaved road
located at the northern boundary of the Ft. Carson military installation
south of Colorado Springs, Colorado. There were several reasons for the
selection of Ft. Carson as the site. First, Ft. Carson officials were
interested in cooperating in such an experiment. As a training center
for ground-based artillery, the base contains long stretches of unpaved
roads and cleared areas which are subject to fugitive dust emissions.
For sometime Ft. Carson has been considering conducting experiments with
various dust palliatives to reduce these emissions. Ft. Carson officials
were most cooperative in assisting in site selection, providing monitoring
equipment, and installing the fabric itself. In addition, the fabric
418
-------
VC
Table 1. PROPERTIES OF CIVIL ENGINEERING FABRICS
Brand Name
Composition
Structure
Thickness, mils
Weight, oz/yd2
Grab Strength, Ib
Burst Strength, psi
Water Permeability,
Bidim
polyester
spunbonded
60-190
—
115-610
225-850
0.03
Mirafi
polypropylene
woven
25
4
200
325
-__
Supac
polypropylene
heat bonded
50
5.3 (up to 16)
150
(cross direction)
300
0.05
Typar
polypropylene
spunbonded
15
4
130
cm/sec
To convert to metric equivalents
Multiply by_ to convert to
mils ,
oz/yd'
Ib
psi
0.00254
25.9
453.6
703.1
cm ?
9/m
gram,,
kg/nT
-------
had not yet been tested at a western site.
The road extends in an approximate north-south direction and provides
access to several residential areas on the post. It is not in the training
area, but is used mainly by Army personnel in light-duty vehicles enroute
from or to these residences. An occasional track vehicle uses the road
as well. Open fields extending for 100 meters or more lie on both sides
of the road. The test section slopes gently upward toward the south end.
The majority of the traffic traverses the entire length of the segment, but
an occasional vehicle either enters or exits at two infrequently used
intersections.
In November 1979, a Monsanto fabric was installed in the southernmost
305 meter (1,000 foot) section by Ft. Carson construction crews under
Monsanto supervision. The northern section is being used as a control
section.
TEST RESULTS
Field tests of ambient particulate concentrations around the test
road were conducted in November 1979 and April 1980. The test apparatus
consisted of upwind and downwind stations equipped with Hi-Vols and
dichotomous samplers in November and Hi-Vols and size selective Hi-Vols
in April. The data derived from the Hi-Vol measurements are shown in
Table 2. The lower increase in ambient concentration downwind of the
fabric-plus-aggregate road is shown by comparison to the increase across
the control section. The fabric-plus-aggregate road gave an average of
47% lower emissions within a range of 30 to 70% lower emissions. Wind
speeds during November were lower than during April which corresponded
to greater emissions reductions during November than during April.
Data for the control of inhalable particles (those with diameters
less than 15 micrometers) are shown in Table 3. As noted above, the
November data was derived from dichotomous sampler measurements and the
April data from size selective Hi-Vol measurements. The percentage
reduction shown in the table is again the lower increase in emissions
from the fabric/aggregate road as compared to the control road. The
reduction averaged 43% within a range of 26 to 53%. Again the reduction
was slightly greater with lower wind speeds.
These results may be used to compare the cost effectiveness of civil
engineering fabrics for unpaved road control to other methods as shown
in Table 4. Based on cost estimates by Cooper9 and Blackwood.10 the
other methods have a cost of at least $220 to $300 for each kilometer
of road per year per % reduction in emissions. For fabric/aggregate
roads, the cost is $140 per km/yr/% reduction. Thus, the fabric/aggregate
road is more cost-effective and avoids the problem of availability of
water in arid regions for watering or that of oil runoff for oiling.
FUTURE MONITORING PLAN
The current plan calls for a 1-year study to collect data on the
following parameters: total suspended particulate, size fraction under 15
micrometers, vehicle counts, precipitation, and wind speed/direction.
420
-------
Table 2. CONTROL EFFICIENCY OF ROAD CARPET (BIDIM)
FOR TOTAL SUSPENDED PARTICLES
Date Wind Speed, mph % Reduction
11/7/79 0.6 60
11/9/79 2.4 59
11/14/79 0.3 30
11/16/79 0.2 70
4/28/80 7 35
4/28/80 7 44
4/29/80 8 36
4/29/80 9 39
Average 47
To convert mph to cm/sec multiply by 44.7.
421
-------
Table 3. CONTROL EFFICIENCY OF ROAD CARPET (BIDIM)
FOR INHALABLE PARTICLES
Date Mind Speed, mph % Reduction
11/7/79 0.6 53
11/9/79 2.4 31
11/13/79 1.5 49
11/16/79 0.2 48
4/28/80 7 45
4/28/80 7 39
4/29/80 8 26
4/29/80 9 49
Average 43
To convert mph to cm/sec multiply by 44.7
422
-------
Table 4. COSTS OF UNPAVED ROAD CONTROL
Expected
Cost Effectiveness
Method
Oiling
Watering
CaCl2
Road Carpet
Efficiency, %
50-98
40
60
45
by Cooper"
up to 180,000
25,000-39,000
18,000
—
by Blackwood1"
11,000
11,500
—
6,000
$km-yr-%
220-1800
290-1000
300
140
-------
Five pairs of Hi-Vols/Hi-Vol with size selective inlets are to be used.
Because the predominant winds are drainage winds which are diurnal in
nature, from the northwest in the morning and southeast in the afternoon,
the monitor pairs are located north of the center point of each test
section and should measure mainly upslope afternoon emissions. A background
site will be located far enough away so as not to be impacted by road
emissions.
Instruments will be maintained and filters weighed by the local
health department which conducts the TSP monitoring for the State and is
cognizant of EPA quality assurance procedures. Samples will be taken
every 6 days to correspond to the sampling schedule for other Hi-Vols in
the area. Because of the heavy filter loadings, 4 or 8 hour samples may
be taken instead of the normal 24-hour sample.
All data collected will be submitted to the EPA Regional Office for
analysis and publication of results.
Data collection should begin in June or July and continue for 1
year. A preliminary report should be available, based on 6 months of
data, by April 1981, and a final report should be available late in
1981.
REFERENCES
1. "An Implementation Plan for Suspended Particulate Matter in
the Phoenix Area, Vol. II, Emission Inventory," TRW, Inc., June 1977,
EPA-450/3-77-021b (NTIS PB 278178).
2. Colorado State Implementation Plan, as submitted to EPA in
January 1979.
3. "An Implementation Plan for Suspended Particulate Matter in
the Phoenix Area, Vol. IV, Control Strategy Formulation," TRW, Inc.,
June 1977, EPA-450/3-77-021d.
4. Blackwood, T. R. and R. A. Wachter. Source Assessment: Coal
Storage Piles. Prepared by Monsanto Research Corporation for the Industrial
Environmental Research Laboratory, U. S. Environmental Protection Agency,
Cincinnati, Ohio, May 1975.
5. Chalekode, P.K., T. R. Blackwood, and S. R. Archer. Source
Assessment: Crushed Limestone, State-of-the-Art. Prepared by Monsanto
Research Corporation for the Industrial Environmental Research Laboratory,
U. S. Environmental Protection Agency, Cincinnati, Ohio, April 1978.
6. Metzger, C. L., Dust Suppression and Drilling with Foaming
Agents. Pit and Quarry Magazine. March 1976.
7. Harwood, C. F., and P. K. Ase. Field Testing of Emission
Controls for Asbestos Manufacturing Waste Piles. Illinois Institute of
Technology Research Institute, May 1977, EPA-600/2-77-098 (NTIS PB
270081). May 1977.
424
-------
8. Kromrey, R. V., R. Naismith, R. S. Scheffee, and R. S. Valentine.
"Development of coatings to reduce fugitive emissions from coal stockpiles."
In Third Symposium on Fugitive Emissions Measurement and Control. EPA-
600/7-79-182 (NTIS PB 80-130891).
9. Cooper, D. W., J. S. Sullivan, M. Quinn, R. C. Antonelli, and
M. Schneider. Setting Priorities for Control of Fugitive Particulate
Emissions from Open Sources. Harvard University School of Public Health,
August 1979, EPA-600/7-79-186 (NTIS PB 80-108962).
10. Blackwood, T. R. Assessment of Road Carpet for Control of
Fugitive Emissions from Unpaved Roads. Monsanto Research Corporation,
May 1979, EPA-600/7-79-115 (NTIS PB 298874).600/7-79-115. May 1979.
425
-------
FUGITIVE PARTICLE EMISSION CONTROL USING
SPRAY CHARGING AND TRAPPING SCRUBBER
Dennis C. Drehmel
U.S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
Shui-Chow Yung, Richard Parker, Sudarshan Bhutra, and Seymour Calvert
Air Pollution Technology, Inc.
San Diego, California 92117
Abstract
The control of fugitive process emissions (FPEs) with a Spray Charging and
Trapping (SCAT) scrubber was evaluated both theoretically and experimentally.
The SCAT system uses an air curtain and/or jets to contain, convey, and divert
the FPEs into a charged spray scrubber.
3
Experiments were performed on an 3.8 m /s bench-scale spray scrubber
to verify the theory and feasibility of collecting fugitive particles with
charged water spray. The effects of charge levels on drops and particles,
nozzle type, drop size, gas velocity, and liquid/gas ratio on collection
efficiency were determined experimentally. The results of the experiments
and the comparison between theory and data are presented.
An air curtain was developed for conveying the FPEs to the spray
scrubber, deflecting the crosswind, and containing hot buoyant plume.
The design and air flow field for the air curtain are presented.
The SCAT system is also being applied to improve the capability of
street sweepers for collecting inhalable particles. Preliminary design
approaches are presented.
42G
-------
FUGITIVE PARTICLE EMISSION CONTROL USING
SPRAY CHARGING AND TRAPPING SCRUBBER
Introduction
Fugitive emissions constitute a large percentage of total particulate
emissions. Therefore, control of fugitive process emissions and fugitive
dust will be an essential ingredient in control strategies for particles.
EPA classified fugitive emissions into two categories: fugitive dust
emissions and fugitive process emissions. Fugitive dust emissions originate
from non-industrial activities; e.g., wind blown dust, entrained street dust,
agriculture, and construction and demolition. Fugitive process emissions
are any non-ducted emissions due to industrial activities such as loading
and unloading raw materials, unpaved road in a plant, metal pouring, and
material charging.
Fugitive emissions are diffuse and typically come from many small
sources as opposed to a single large emitter. Present control methods use
secondary hooding at the local source of emissions, total building
enclosure and evacuation, or water sprays. Secondary hooding and total
building enclosure and evacuation require high energy and capital invest-
ments. In addition, secondary hooding may not be efficient in gathering
fugitive emissions.
Spraying with water could be effective for some processes and materials
which can tolerate water, such as transfer of raw materials on conveyor
belts. The main function of the spray in this application is to stabilize
the dust to keep it from dispersing. If the dusts are already airborne,
spraying with water alone will not do much good. The water drops can catch
some particles. However, the drops and collected particles will re-disperse
in the air after the water has evaporated.
A much simpler and cheaper method for controlling fugitive process
emissions is by diverting the fugitive particle emission (FPE) into a
control device located near the source. The particles are then removed
for the air and disposed of. The SCAT (Spray Charging and Trapping)
scrubber is such a system. It uses air curtains and/or jets to
contain, convey, and divert the FPEs into a charged spray scrubber. This
control method minimizes the apparatus required to contain, convey, and
control the fugitive emissions.
SCAT System
Figure 1 shows an example of the SCAT system. It has two sections.
One section contains the air curtain and air push jet assembly. The other
section consists of a particle control device. The SCAT system uses water
sprays, which could be charged for particle removal.
The two sections are arranged in a push/pull arrangement with the
fugitive particle emission source situated between the two sections.
The fugitive particles are contained with air curtains and are pushed away
427
-------
from the source into the spray scrubber. The scrubber has a low pressure
drop entrainment separator (e.g., zigzag baffles) to remove the spray drops.
The water from the entrainment separator can be 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. Alternatively, a blow-
down stream of dirty liquid may be directed to a disposal system.
The major SCAT system feature which suits it to FPE control is the use of
air curtains and/or air push jets to contain and convey FPEs. The use of
air curtains will minimize the requirement of solid enclosure (hooding). It
will be easy to build and unobstrusive. The air curtains could also be used
to deflect the wind to minimize the dispersion and the volume of air to be
cleaned and be used to contain FPEs from hot sources. For example, the
Naoshima Smelter of Mitsubishi Metal Corporation at Naoshima Island uses
a system similar to SCAT to control the fugitive emissions from a
copper converter (Uchida, et al. 1979). In that system, fumes from the
converter are contained and then are ducted to a central particle control
device.
A SCAT system could have the following components.
1. Air curtain and/or air push jets,
2. Water sprays (could be charged),
3. Entrainment separator,
4. Fan and pump, and
5. Water treatment.
It is not necessary for a SCAT system to have all the components. There
is no one fixed design and configuration. Its design depends on specific
sites and applications. Of the above five components, air curtains and water
sprays required further study. To generate design information quickly, the
air curtain and the spray scrubber were studied individually in separate
bench scale devices before being combined into one prototype SCAT unit. The
prototype was then used to determine the effect of crosswind, hot source,
and particle containment efficiency.
Air Curtain Study
An air curtain is a sheet of air blown either by a jet or out of a slot
at high speed. It has been widely used in industrial and commercial plants,
mainly to provide constant access or to isolate a warm interior from the
cold outdoors or vice versa.
Most of the published information on air curtain performance and design
has been related to air conditioning and ventilation. There is little
published literature on its dust containment capability even though it has
been used for this purpose in industry. Only one report has been published
on the application of air curtains to block or deflect crosswinds.
Design of the SCAT system requires information on several parameters of
the air curtain jet stream such as the jet expansion angle, air entrainment
423
-------
ratio, mixing of particles in the curtain, effect of crosswind, and the
effect of hot sources. The jet expansion angle and particle mixing deter-
mine the overall cross-sectional dimensions of the spray scrubber or the
receiving hood. The air entrainment ratio determines the volumetric
flowrate. The crosswind and heat effect dictate the placement of air
curtains and the sprays. These parameters are being studied.
In the SCAT system, the air curtain should have a small expansion
angle, small air entrainment ratio, and a uniform velocity distribution.
It is generally known that for uniform velocity distribution the distri-
bution manifold or duct should have a constant static pressure from end
to end. To obtain constant static pressure, the manifold should have
either a very large cross-sectional area compared to the total nozzle
area or a decreasing cross-sectional area to keep the velocity head in the
manifold constant.
Early attempts to simplify construction of an air curtain manifold
by drilling holes along the duct did not yield uniform flow. The main
reason was that the duct cross-sectional area was not much larger than the
total hole area. A better distribution manifold shown in Figure 2 was
built. The manifold was of rectangular cross-section 35.6 cm x 30.5 cm.
An adjustable plate was placed in the duct so that the flow area decreased
along with duct. Ideally the plate should be curvilinear in shape. But
for the short duct used in the present study, about 3 m in length, the plate
may be straight. The use of straight plate greatly simplified the
construction.
The air jet nozzle was a continuous slot 2.1 m (7 ft) long. The slot
was formed by two parallel plates which protrude 22.9 cm (9 in.) from one
side of the duct. The distance between the plates, which is the slot
width, could be adjusted. The slot was divided by thin cross-plates,
7.6 cm (3 in.) apart, so that the air would discharge perpendicularly to
the longitudinal axis of the duct.
Velocity profile was measured with the slot oriented vertically.
Linear velocity was measured with a hot wire anemometer for three vertical
levels at three downstream locations from the slot. Figure 3 shows the
measured velocity profile. As can be seen the velocity distributions for
the three vertical levels are virtually identical. Therefore, the air
curtain flow field is uniform.
The air curtain flow field was measured for several slot widths and
air slot exit velocities. The results were compared with the equations
suggested by McElroy (1943) for a two-dimensional jet exhausting into still
surroundings and jets with two-sided expansion.
Figures 4 and 5 show the measured center!ine axial velocity decay
and entrainment ratio, respectively. The agreement between the measured
and predicted centerline velocity decay is good. The measured entrainment
ratio was lower than that calculated.
429
-------
The lower measured entrainment ratio could be due to a smaller inclu-
sive jet expansion angle. The measured jet expansion angle varied between
20° and 28°, depending on air exit velocity. This is smaller than that
reported by Tuve and Priester (1944) of 32 - 34° and that suggested by
Hemeon (1963) of 30 - 40°.
The jet expansion angle and entrainment ratio were also measured for
the first air curtain manifold (duct with drilled holes) we had tested.
The jet expansion angle was 35°, which was close to Tuve and Priester
results and within the range recommended by Hemeon. The measured entrain-
ment ratio for this air curtain agrees with that calculated from theory.
Charged Spray Scrubber Study
Bench scale experiments were performed to determine the minimum water
requirements, to aid the selection of spray nozzles, to study various
drop charging methods to determine the effect of drop and/or particle
charging on particle collection efficiency, and to verify published
theories.
Figure 6 shows a sketch of the charged spray scrubber system. In
order to enable us to make good measurements, it is much more elaborate
than the actual SCAT will be.
The system was made in sections jointed by flanges. The scrubber was
made of thin aluminum sheets and supported on PVC frames in order to
permit electrical current measurements. The cross-section of the scrubber
is 0.91 m x 0.91 m (3 ft x 3 ft). The overall length, including the
blower, is about 11 m (36 ft).
The scrubber system consists of a flow straightening section, an
inlet particle sampling section, a particle charging section, a spray
section, an entrainment separator, and an outlet sampling section. It
should be noted that all sections except the spray and entrainment
separator section will not be included in an industrial installable
SCAT system. The inclusion of the sections in the bench scale device is
for the purpose of studying various combinations of components and to
provide for good particle sampling and electrical measurement.
The particle charging section consisted of two rows of corona wires.
Wire diameter was 0.18 mm (0.007 in.). The spacing between wires within
the same row was 6.5 cm (2.5 in.). The ground electrodes were 1.3 cm
(0.5 in.) diameter aluminum tubing.
The overall length of the spray section was 2.44 m (8 ft) including
two spray banks. The water was charged by induction. The nozzles and
water feed lines were kept at ground potential. A high voltage grid
assembly was placed in front of the nozzles to charge the water drops.
This arrangement not only simplified construction, by eliminating a
complicated electrical isolation system, but also gave a high charge level
on drops.
430
-------
The scrubber was operated for four conditions:
1. Uncharged particles/neutral drops,
2. Charged particles/neutral drops,
3. Uncharged particles/charged drops, and
4. Charged particles/charged drops.
Figures 7 and 8 show the data for nozzles "A" and "B", respectively.
The measured mass median drop diameter, spray angle, and discharge
coefficient for nozzle "A" (hook type) was 240 ym, 100°, and 0.63,
respectively. Only one spray bank was used in all experiments and nozzle
pressure was maintained at 432 kPa (48 psig). One power supply was used
to charge both the particles and the drops. The applied voltage was
-10 kVDC. The measured charge/mass ratio was +5.8 x 10~5 C/g and -1.5
x 10"5 C/g for drops and particles, respectively.
As can be seen from Figures 7 and 8 the collection efficiency of the
spray scrubber improves by charging either the water or the particles.
Further improvement was measured when the water and particles were
oppositely charged. The improvement is more with submicron particles.
For particles with diameters larger than 5 ymA, charging the water and/or
particles has little effect on efficiency.
The scrubber with nozzle "A" has better collection efficiency at a
lower liquid/gas ratio than that with nozzle "B". A possible reason is
that nozzle "A" produced finer drops.
Comparison Between Data and Theory
The model by Calvert (1970, 1975) along with the drop trajectory
formula given by Giffen and Murozew (in Walton and Woolcock, 1960) and the
appropriate single drop collection efficiency were used to predict the
spray scrubber performance. Figure 9-shows the predicted and the measured
grade penetration for the uncharged particle/neutral drop condition.
As can be seen, the agreement is good.
For electrostatically augmented conditions, i.e. either one or
both of particle and drops were charged, Nielsen's (1974) graph for single
drop collection efficiency was used in the calculation. The results were
identical to that for the uncharged condition. This means there should
be no improvement in efficiency with electrostatic augmentation. This is
contrary to experimental findings. The discrepancy could be due to the
assumption that drops were of uniform size and carried the same amount of
charge.
Collection in a vertical direction was not included in the calculations,
While collection in a vertical direction is negligible for the uncharged
particle/uncharged drop condition, it could contribute to particle
collection for the charged system due to the low relative velocity between
the drops and the particles.
431
-------
Prototype SCAT Scrubber
A prototype SCAT system was designed and built based on the
information generated in air curtain and spray scrubber studies. Its
configuration is similar to that shown in Figure 1. It had two separate
sections. One section housed the charged spray scrubber; the other section
had three air curtains and one air push jet. Both sections were mounted on
casters; therefore, the distance between the air curtains and the spray
scrubber could be adjusted.
The charged spray scrubber had a cross-section of 2.44 m x 1.83 m
(8 ft x 6 ft). The bottom 0.61 m (2 ft) was the scrubber sump. There-
fore, the active scrubber was made of aluminum plates.
There were 36 pigtail type spray nozzles at the scrubber front surface.
The sprays are cocurrent with the gas flow. The capacity of each spray
nozzle is 9.8 £/nrin (2.6 6PM) at a nozzle pressure of 377 kPa (40 psig).
The nozzle discharge coefficient is 0.6 and the drop diameter is 200-250 urn.
About 1 m downstream from the sprays were four rows of zigzag baffles
to remove the water drops. There was a blower at the scrubber outlet. The
blower is rated at 963 m3/min at 2.5 cm W.C. pressure drop (34,000 CFM
at 1 in. W.C.).
The other section had three independent air curtains and a propeller
fan. Two of the air curtains were mounted vertically and about 1.83 m
(6 ft) apart. The other curtain was mounted horizontally 2.7 m (9 ft)
above the ground. The air curtains could be swivelled as needed to deflect
crosswinds and contain buoyant smoke plumes. The propeller fan was located
at the center. The fan was for producing a "push" jet.
The prototype SCAT system was used in crosswind, hot source, and
containment efficiency measurements.
Hot Source Experiments
We are conducting experiments on the prototype SCAT system operating on a
simulated FPE hot source. Figure 10 illustrates the test setup. A
furnace is used to generate a hot buoyant plume. Smoke is injected in the
plume as a tracer. We will monitor the air curtain angle, flow rate, and
slot geometry required to divert the hot plume into the scrubber. No data
are yet available.
-------
Application of SCAT Scrubbing on Street Sweepers
As an application of the SCAT approach, Air Pollution Technology, Inc.
(A.P.T.) has undertaken a program to develop a special purpose street sweeper.
At present street sweepers are not designed to collect inhalable particles
dispersed during street cleaning.
The objective of the study is to retrofit existing vacuum sweepers
with an air pollution control system to collect 90 percent of the inhalable
particles dispersed by the sweeper.
The study is divided into two phases. Phase I includes a theoretical
and experimental study to define and verify estimates of power and water
requirements as well as cost and performance of the add-on system. In
Phase II a conventional vacuum sweeper will be modified and evaluated
under various street conditions.
A.P.T. will modify a regenerative air type vacuum sweeper to collect
inhalable particles dispersed during street sweeping. A schematic diagram
of the sweeper operation is given in Figure 11. The gutterbroom transfers
the dirt from the curb to the pick-up head. Air flow in the pick-up
head conveys the dirt to the hopper where the coarse particles are removed.
The air is then recirculated to the pick-up head. The two sources of fine
dust dispersion are the gutterbroom and the pick-up head. Dust dispersed
by the gutterbroom will be conveyed to the spray scrubber and collected.
Further, a stream of air will be exhausted from the system to ensure a
positive intake of air at the pick-up head. This will reduce air "spills"
created by the air in the head. Inhalable particles will be removed from
the air stream before exhausting it into the atmosphere. Fine water sprays
will be used to collect and agglomerate suspended dust particles. Either
the drops or the particles (or both) may be electrostatically charged to
increase the particle collection efficiency. The dust collection system
will consist of a charger and water sprays followed by an entrainment
separator.
The primary design constraints for this system are:
1. Minimize water usage,
2. Maximize collection efficiency for inhalable particles through
electrostatic augmentation,
3. Optimize entrainment separator to collect drops and coarse
particles, and
4. Minimize power required for air recirculation.
43J
-------
Conclusions
The SCAT system is a simple, versatile and effective technique for controlling
fugitive emissions. It involves the use of air curtains and air jets to
contain and convey the emissions into a nearby spray scrubber.
The collection efficiency of a single spray bank scrubber was
investigated experimentally. The collection efficiency could be improved
by charging the water and/or the particles. The measured particle
penetration agrees with theoretical predictions. For charged water and/or
charged particle condition, the measured penetration is lower than the
predicted.
Air curtains have been used in industries to contain dust but no
carefully performed study has been reported in the literature. The air
curtain developed in the present study can achieve a smaller expansion
angle and a lower entrainment ratio than those reported in the literature.
Small expansion angle and entrainment ratio are beneficial to the control
of fugitive process emissions with the SCAT system.
A prototype SCAT system has been built by A.P.T. to study the effects of
crosswind, hot buoyant plume, and containment efficiency. Experiments are
currently underway.
The SCAT concept is also being applied to urban street sweepers to
improve their effectiveness in controlling emissions of inhalable particles
from paved streets. Preliminary analysis indicates that the system is
very promising and can have a significant impact on urban air quality.
434
-------
REFERENCES
Calvert, S. Venturi and Other Atomizing Scrubbers, Efficiency and Pressure
Drop. AIChE J. 16.: 392-396, 1970.
Calvert, S., et al. Study of Flux Force/Condensation Scrubbing of Fine
Particles. EPA-600/2-75-018 (NTIS PB 249297), August, 1975.
Hemeon, W.C.L. Plant and Process Ventilation. Second Edition. The
Industrial Press, New York, 1963.
McElroy, G.E. Air Flow at Discharge of Fan-Pipe Lines in Mines. Part II.
U.S. Bureau of Mines Report of Investigation 3730, November 1943.
Nielsen, K.A. Effect of Electrical Forces on Target Efficiencies for
Spheres. Eng. Research Inst. Tech. Report 74127, Iowa State Univ., 1974.
Tuve, G.L. and G. B. Priester. Control of Air-Streams in Large Spaces.
Heating, Piping and Air Cond., ASHVE Journal, January 1944.
Uchida, H., et al. Processing of Copper Smelting Gases at Naoshima
Smelter. Paper presented at the Control of Particulate Emissions in the
Primary Nonferrous Metals Industries Symposium. Monterey, California,
March 1979.
Walton, W.H. and A. Woolcock. Aerodynamic Capture of Particles. E.G.
Richardson, editor. Pergamon Press, Oxford, England, 1960.
435
-------
AIR
CURTAIN
PUSH FAN
OR
PUSH JET
AIR
CURTAIN
SPRAY
SCRUBBER
PULL
FAN
ENTRAPMENT
SEPARATOR
Figure 1. Example of SCAT system arrangement.
-------
AIR IN
210 cm
Figure 2. Air curtain distribution duct.
437
-------
Ul
,
i )
f. <
to
• ;
0
DISTANCE DOWNSTREAM FROM SLOT = 4.6 m
DISTANCE DOWNSTREAM
FROM SLOT = 3.1 m
.
DISTANCE DOWNSTREAM
. FROM SLOT = 1.53 m
0 *
I SLOT GAS EXIT VELOCITY = 23 m/s
SLOT WIDTH = 1.3 cm
\ SLOT LENGTH = 183 cm
O MEASURED AT 137 cm FROM TOP
A MEASURED AT 91 cm FROfl TOP
Q MEASURED At 46 cm FROM TOP
0
•
. 1 , m W. . I . a M . . I
-i
0
+1
DISTANCE FROM CENTER OF SLOT, m
Figure 3. Velocity profile of an air curtain.
438
-------
01
"c
o
[/i
U
13
i O 1.27 cm
j A 2.54 cm
El 3.81 cm
500
1000
—, dimensionless
Figure 4. Measured and predicted centerline axial velocity decay.
-------
s
LiJ
• «
:
i
. •
• i
m
rrrr:.:...:....t:.
10
100
x/w, m/m
500
1000
Figure 5. Measured and predicted air entrainment ratio.
440
-------
POWER SUPPLY
--
--
BLOWER
FLOW
STRAIGHTENING
SECTION
\
INLET
SAMPLING
SECTION
SPRAY
SECTION
PARTICLE
CHARGING'
SECTION
i • H
SUMP
OUTLET
SAMPLING
SECTION
,ENTRAINDENT
SEPARATOR
VENT
Figure 6. Experimental apparatus for studying charged spray section of
SCAT scrubber.
-------
2
S
i.o
0.5
0.1
0.05
0.01
'. UNCHARGED PARTICLE/NEUTRAL
4-U-l
. . .
. . : . ,
: UNCHARGED PARTICLE/CHARGED DROP'
".I....I
= 2.9 m/s
QL/QG = 4 x 10 mVra
Nozzle A
1 Spray bank
.
ill
0.1
0.5
1
10
AERODYNAMIC PARTICLE DIAMETER, ymA
Figure 7. Measured spray scrubber penetration, nozzle A.
442
-------
UNCHARGED PART./NEUTRAL DRO
ART./NEUTP.AL...D
1 1 l l l l l i fr i i i i . IT 1111IIII I II II I )L J
UNCHARGED PART./CHARGED
-CHARGED PART./CHARGED DROP
= 2.9 m/s
Q, /Qr = 7 x 10'4 mVm3
L'XG
NOZZLE i;
TEST PARTICLES - HYDRATED LIME
AERODYNAMIC PARTICLE DIAMETER, ymA
Figure 8. Measured SCAT spray scrubber performance, nozzle B.
443
-------
u
2
MEASURED^
UG = 2.9 m/s
QL/QG = 7 x 10
1 SPRAY BATIK
NOZZLE = B.
NOZZLE PRESSURE = 423 kPa
UNCHARGED PARTICLE/NEUTRAL DROP
0.05
0.01
0.5 1
AERODYNAMIC PARTICLE DIAMETER, ymA
Figure 9 • Predicted and measured spray scrubber penetration, nozzle B.
444
-------
x
s
Air
Curtain
SCAT Scrubber
FURNACE
Figure 10. Hot source experiment.
-------
CONTAIN/CONVEY DUST
GUTTER
BROOM
AIR INTAKE
i
DUST
PICK-UP
HOPPER
CHARGED
SPRAY
EXHAUST
RECIRCULATED AIR
FIGURE 11. Improved Street Sweeper
-------
|