EPA-600/2-76-136b
May 1976
Environmental Protection Technology Series
PROCEEDINGS:
SYMPOSIUM ON FLUE GAS DESULFURIZATION
NEW ORLEANS, MARCH 1976
Volume I
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service', Springfield, Virginia 22161.
-------
EPA-600/2-76-l36b
May 1976
PROCEEDINGS:
SYMPOSIUM ON FLUE GAS DE S UL F URIZ A TION
NEW ORLEANS, MARCH 1976
VOLUME II
Program Element No. EHE624
Chairman: Richard D. Stern
Vice-Chairmen: Wade H. Ponder and Roger C. Christman
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
PREFACE
More than half of all "man-made" sulfur dioxide (SC^) is
emitted by electric power plants, and the use of sulfur-containing
fiossil fuels to generate electricity is predicted to increase by
50 percent by 1985. As a result, the development of SCL control
technologies is one of the most important goals of the U.S.
Environmental Protection Agency (EPA). Flue gas desulfurization
(FGD) is the most promising technique for control of SC>2 that will be
available for widespread application to fossil fuel-fired electric
power plants for at least the next decade.
The Industrial Environmental Research Laboratory - Research
Triangle Park (IERL-RTP) of EPA's Office of Research and
Development sponsors symposia for the transfer of information
regarding FGD research, development and application activities
with the objective of further accelerating the development and
commercialization of this technology. These symposia provide
an opportunity for users and developers to discuss their
experiences and the status of development and application of
FGD technology.
The March 1976 symposium addressed full-scale FGD process
applications in the U.S. and Japan as well as laboratory, pilot,
ill
-------
and prototype research and development efforts. The symposium
also provided an opportunity for the announcement of data and
results which were previously unreported or not widely publicized.
The economics of FGD and the disposal, utilization, and marketing
of FGD system by-products were also discussed. The symposium
papers were presented by a cross-section of those concerned
with FGD including users, government and private developers, and
vendors. The electric utility industry—the principal user of FGD-
participated extensively in the symposium program. More
than 650 people attended the symposium.
These Proceedings are comprised of copies of the participating
authors' papers as received. As supplies permit, copies of the
Proceedings are available free of charge and may be obtained by
contacting lERL-RTP's Technical Information Coordinator,
Environmental Protection Agency, Research Triangle Park, North
Carolina 27711.
iv
-------
CONTENTS
TITLE PAGE
VOLUME I
OPENING SESSION
Keynote Address - SULFUR OXIDE CONTROL AND ELECTRICITY
PRODUCTION
R. E. Train, Environmental Protection Agency,
Washington, D. C
REMARKS
S. J. Gage, Environmental Protection Agency,
Washington, D. C
STATUS OF FLUE GAS DESULFURIZATION SYSTEMS IN THE UNITED
STATES
T. W. Devitt, G. A. Isaacs, and B. A. Laseke, PEDCo-
Environmental Specialists, Inc., Cincinnati, Ohio 13
STATUS OF FLUE GAS DESULFURIZATION AND SIMULTANEOUS REMOVAL
OF S00 AND NO IN JAPAN
2 x
J. Ando, Chuo University, Tokyo, Japan 53
FLUE GAS DESULFURIZATION ECONOMICS
G. G. McGlamery, H. L. Faucett, R. L. Torstrick, and
L. J. Henson, Tennessee Valley Authority, Muscle
Shoals, Alabama 79
STATUS OF THE EPRI FLUE GAS DESULFURIZATION DEVELOPMENT
PROGRAM
L. W. Nannen and K. E. Yeager, Electric Power Research
Institute, Palo Alto, California 101
NON-REGENERABLE PROCESSES SESSION 115
IERL-RTP SCRUBBER STUDIES RELATED TO FORCED OXIDATION
R. H. Borgwardt, Environmental Protection Agency,
Research Triangle Park, North Carolina 117
RESULTS OF MIST ELIMINATION AND ALKALI UTILIZATION TESTING
AT THE EPA ALKALI SCRUBBING TEST FACILITY
M. Epstein, II. N. Head, S. C. Wang, and D. A. Burbank,
Bechtel Corporation, San Francisco, California 145
v
-------
TITLE PAGE
DUQUESNE LIGHT COMPANY ELRAMA AND PHILLIPS POWER STATIONS
LIME SCRUBBING FACILITIES
R. G. Knight and S. L. Pernick, Duquesne Light
Company, Pittsburgh, Pennsylvania 205
OPERATIONAL STATUS AND PERFORMANCE OF THE COMMONWEALTH EDISON
COMPANY WILL COUNTY LIMESTONE SCRUBBER
W. G. Stober, Commonwealth Edison Company, Chicago,
Illinois 219
Mill FLUE GAS DESULFURIZATION SYSTEMS APPLIED TO SEVERAL
EMISSION SOURCES
M. Hirai, M. Atsukawa, A. Tatani, and K. Kondo,
Mitsubishi Heavy Industries, Ltd., Tokyo, Japan 249
STATUS OF FLUE GAS DESULFURIZATION USING ALKALINE FLY ASH
FROM WESTERN COALS
H. M. Ness, E. A. Sondreal, and P. H. Tufte, U. S.
Energy Research and Development Administration, Grand
Forks, North Dakota 269
RESULTS OF THE 170 MW TEST MODULES PROGRAM, MOHAVE GENERATING
STATION, SOUTHERN CALIFORNIA EDISON COMPANY
A. Weir, Jr., L. T. Papay, D. G. Jones, J. M. Johnson,
and W. C. Martin, Southern California Edison Company,
Rosemead, California 325
LA CYGNE STATION UNIT NO. 1 WET SCRUBBER OPERATING EXPERIENCE
C. F. McDaniel, Kansas City Power and Light Company,
La Cygne, Kansas 355
RECENT SCRUBBER EXPERIENCE AT THE LAWRENCE ENERGY CENTER,
THE KANSAS POWER AND LIGHT COMPANY
D. M. Miller, Kansas Power and Light Company, Topeka,
Kansas 373
INTRODUCTION TO DOUBLE ALKALI FLUE GAS DESULFURIZATION
TECHNOLOGY
N. Kaplan, Environmental Protection Agency, Research
Triangle Park, North Carolina 387
OPERATING EXPERIENCE--CEA/ADL DUAL ALKALI PROTOTYPE SYSTEM
AT GULF POWER/SOUTHERN SERVICES, INC.
C. R. LaMantia and R. R. Lunt, Arthur D. Little, Inc.,
Cambridge, Massachusetts; R. E. Rush, Southern Services,
Inc., Birmingham, Alabama; T. M. Frank, Combustion
Equipment Associates, Inc., New York, New York; and
N. Kaplan, Environmental Protection Agency, Research
Triangle Park, North Carolina 423
VI
-------
TITLE PAGE
THE FMC CONCENTRATED DOUBLE-ALKALI PROCESS
L. K. Legatski, K. E. Johnson, and L. Y. Lee,
FMC Corporation, Glen Ellyn, Illinois 471
OPERATING EXPERIENCE WITH THE ZURN DOUBLE ALKALI FLUE GAS
DESULFURIZATION PROCESS
P. M. Lewis, Zurn Industries, Inc., Birmingham,
Alabama 503
KUREHA FLUE GAS DESULFURIZATION "SODIUM ACETATE-GYPSUM
PROCESS"
S. Saito, T Morita,' and S. Suzuki, Kureha Chemical
Industry Company, Ltd., Tokyo, Japan 515
THE BUELL DOUBLE-ALKALI S00 CONTROL PROCESS
11. E. Bloss, Buel1-Envirotech, Lebanon, Pennsylvania;
J. Wilhelm, EIMCO-Envirotech, Salt Lake City, Utah;
and W. J. Ilolhut, Central Illinois Public Service
Company. Springfield, Illinois 545
VOLUME II
BY-PRODUCT DISPOSAL/UTILIZATION SESSION XI
STATUS AND PLANS FOR WASTE DISPOSAL FROM UTILITY APPLICATIONS
OF FLUE GAS DESULFURIZATION SYSTEMS
J. L. Crowe, Tennessee Valley Authority, Chattanooga,
Tennessee; and H. W. Elder, Tennessee Valley Authority,
Muscle Shoals, Alabama 565
RESEARCH AND DEVELOPMENT FOR CONTROL OF WASTE AND WATER
POLLUTION FROM FLUE GAS CLEANING SYSTEMS
J. W. Jones, Environmental Protection Agency,
Research Triangle Park, North Carolina 579
FLUE GAS CLEANING WASTE DISPOSAL - EPA SHAWNEE FIELD
EVALUATION
J. Rossoff and R. C. Rossi, The Aerospace Corporation,
El Scgundo, California 605
CHEMICAL FIXATION OF FGD SLUDGES - PHYSICAL AND CHEMICAL
PROPERTIES
J. L. Mali loch, U. S. Army Engineer Waterways Experiment
Station, Vicksburg, Mississippi 627
VII
-------
TITLE PAGE
POTENTIAL UTILIZATION OF CONTROLLED SO EMISSIONS FROM
POWER PLANTS IN EASTERN UNITED STATES X
J. I. Bucy, J. L. Nevins, P. A. Corrigan, and A. G.
Melicks, Tennessee Valley Authority, Muscle Shoals,
Alabama 647
REGENERABLE PROCESSES SESSION 701
STATUS OF DEMONSTRATION OF THE WELLMAN-LORD/ALLIED FGD SYSTEM
AT NIPSCO'S D. H. MITCHELL GENERATING STATION
PART I: BACKGROUND AND OVERVIEW
E. L. Mann, Northern Indiana Public Service Company,
Michigan City, Indiana; and R. C. Christman, TRW
Environmental Engineering Division, Vienna, Virginia ... 703
PART II: CURRENT STATUS AND OPERATING PLAN
S. F. Lakatos, A. W. Michener, Jr., and W. D. Hunter,
Jr., Allied Chemical Corporation, Morristown,
New Jersey 709
AN UPDATE OF THE WELLMAN-LORD FLUE GAS DESULFURIZATION
PROCESS
R. I. Pedroso, Davy Powergas, Inc., Lakeland, Florida .. 719
SUMMARY OF OPERATIONS OF THE CHEMICO-BASIC MgO FGD SYSTEM
AT THE PEPCO DICKERSON GENERATING STATION
R. B. Taylor and P. R. Gambarani, Chemico Air Pollution
Control Co., New York, New York; and D. Erdman, Potomac
Electric Power Company, Washington, D. C. 735
MAGNESIUM OXIDE SCRUBBING AT PHILADELPHIA ELECTRIC'S
EDDYSTONE STATION
J. A. Gille, Philadelphia Electric Company-,
Philadelphia, Pennsylvania 749
INTERIM REPORT ON CHIYODA THOROUGHBRED 101 COAL APPLICATION
PLANT AT GULF POWER'S SCHOLZ PLANT
R. B. Dakan, Chiyoda International Corporation, Seattle,
Washington; and R. A. Edwards and R. E. Rush, Southern
Services, Inc., Birmingham, Alabama 75 ^
ADVANCED PROCESSES
785
Vlll
-------
TITLE PAGE
STATUS AND ECONOMICS OF THE ATOMICS INTERNATIONAL AQUEOUS
CARBONATE FLUE GAS DESULFURIZATTON PROCESS
D. C. Gchri and R. I). Oldenkamp, Atomics International,
Canoga Park, California . . . . . . 787
ENERGY REQUIREMENTS FOR SHELL FCD PROCESS
F. A. Vicari and J B. Pohlcnz, HOP Inc., DCS Plaines,
Illinois 817
THE DOWA'S BASIC ALUMINUM SULFATE-GYPSUM FLUE GAS
DESULFURIZATION PROCESS
Y. Ynmamichi and J. Nagao, The Dowa Mining Co., Ltd.,
Okayama, Japan 833
CITRATE PROCESS FOR FLUE GAS DESULFURIZATION, A STATUS REPORT
W. I Nisscn, D. A. Elkins, and W. A. McKinney, Bureau
of Mines, Salt Lake City, Utah 843
APCI/IFP REGENERATIVE FGD AMMONIA SCRUBBING PROCESS
C. E. Ennis, Catalytic, Inc., Philadelphia, Pennsylvania 865
BF DRY ADSORPTION SYSTEM 877
PART I: FW-BF GULF POWER DEMONSTRATION UNIT INTERIM
RESULTS
J. Strum and W. F Bischoff, Foster Wheeler Energy
Corporation, Livingston, New Jersey; and R. E. Rush,
Southern Services, Inc., Birmingham, Alabama 879
PART II: BF-STEAG DEMONSTRATION UNIT OPERATIONAL
EXPERIENCE AND PERFORMANCE
K. Knoblauch, Bergbau-Forschung GMBH, West Germany;
and K. Goldschmidt, STEAG Aktengesellschaft, West
Germany 899
THE CONSOL FGD PROCESS
R. T Struck, E. Gorin, and W. E. Clark, Conoco Coal
Development Company, Library, Pennsylvania 913
UNPRESENTED PAPERS 931
INFORMATION TRANSFER PROGRAM
T. W. Devitt and T C. Ponder, Jr., PEDCo-Environmcntal
Specialists, Inc., Cincinnati, Ohio 933
IX
-------
TITLE PAGE
CATALYTIC/WESTVACO DESULFURIZATION PROCESS PROTOTYPE
DEMONSTRATION PROGRAM
Catalytic, Inc., Philadelphia, Pennsylvania 945
THE SPRING-NOBEL HOECHST PROCESS FOR SULFUR DIOXIDE
RECOVERY FROM STACK GASES
W. H. Stark, A. A. Syme, and J. C. H. Chu, Spring
Chemicals, Ltd., Toronto, Canada 981
STARTUP OF AMERICAN AIR FILTER'S SULFUR DIOXIDE REMOVAL
SYSTEM AT THE KENTUCKY UTILITIES COMPANY'S GREEN RIVER
STATION
A. H. Berst, American Air Filter Co., Inc., Louisville,
Kentucky; and J. Reisinger, Kentucky Utilities Co.,
Central City, Kentucky 991
TCA SPHERE DEVELOPMENT AND EVALUATION
P. Sorcnson, N. E. Takvoryan, and R. J. Jaworowski,
UOP, Inc. , Daricn, Connecticut 999
-------
BY-PRODUCT DISPOSAL/UTILIZATION SESSION
Chairman: H. William Elder
Director, Stack Gas Emissions Studies
Tennessee Valley Authority
Muscle Shoals, Alabama
xi
-------
STATUS AND PLANS FOR WASTE DISPOSAL FROM UTILITY APPLICATIONS
OF FLUE GAS DESULFURIZATION SYSTEMS
J. L. Crowe, Office of Power
H. W. Elder, Office of Agricultural £ Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
ABSTRACT
Approximately 10,000 MW of electricity generating capacity will
be equipped with calcium-based scrubbing systems when those under
construction are completed. Disposal of the waste solids produced will
present formidable problems. The alternative disposal practices (storage
in unlined ponds, use of liners, and fixation) in use or planned are
discussed. Also, research in the utility industry on methods for dis-
posal or utilization is summarized. Stringent requirements for final
storage will make recovery processes (those that produce salable
products) more competitive with the throwaway processes.
565
-------
STATUS AND PLANS FOR WASTE DISPOSAL FROM UTILITY APPLICATIONS
OF FLUE GAS DESULFURIZATION SYSTEMS
INTRODUCTION
One might assume that a process that depends on solid and Liquid
phase concentration and composition, settled density,, filterability,
addition of secret ingredients, aging, and use of specially designed
containers would result in a product that would be dispensed from a square
bottle with a black label. Unfortunately, the product that is the subject
for discussion is not nearly so popular but does affect those dealing with
it in a variety of ways. The topic of course is sludge or as the Germans
say in a more descriptive way, schlum.
The most recent count of systems operating or under construction
shows that approximately 11,000 MW of generating capacity have or will have
a flue gas desulfurization (FGO) system. Approximately 10,000 MW of the
total are based on processes that produce calcium-based waste products.
The result is that the utility industry is faced with an enormous solids
waste handling and disposal problem. Utilities, government agencies, and
research organizations have begun a sizable effort to identify and solve
the problems associated with disposal or utilization of sludge. Much of
this effort is described in subsequent papers. However, a summary of the
current practice and plans for utility use of the technology should help to
illustrate the variety of approaches presently being tried by those who
have had to make decisions in the absence of clear guidelines.
The intent of this paper is to not only provide a compilation
of the utility experience but to also encourage communication between the
researchers and the users so that a more consolidated effort can be established,
CURRENT PRACTICE AND PLANS--FULL-SCALE INSTALLATION
The earliest installation of FGD systems generally incorporated the
simplest and least expensive method of disposal, discharge to an unlined pond.
Because of concerns about possible (but unestablished) effects of leaching
on ground water quality, lined ponds have been provided at some installations.
Most recently, chemical fixation of sludge has been chosen for some instal-
lations because of the improved structural properties of the waste solid and
the increased resistance to leaching.
Unlined Ponds
A brief description of facilities with unlined ponds include the
following:
566
-------
Kansas Fewer and Light Company Lawrence No. 4 and 5
Operation of the FGD system on unit k (l25 MW) began in December 1968
and unit 5 (^00 MW) started up in November 1971- Both units are
equipped to burn gas when it is available and therefore sludge pro-
duction has not been continuous. Waste solids, primarily a mixture
of flyash and calcium sulfite, are discharged from both systems into
a series of ponds with a combined surface area of 52 acres; the depth
is unreported. Clarified liquor from the pond is recycled to the
scrubbers and cooling tower blowdown is used for makeup. Good settling
has been reported and the settled solids apparently solidify to some
extent. No ground water degradation has been reported for this instal-
lation or for those listed below.
Kansas City Power and Light Company - Hawthorne No. J and k
Operation of the scrubbing systems for these two units, 1^0 MW and
100 MW, respectively, began in 1972- The availability has been low and
therefore the amount of solids discharged has also been low relative to
the generating capacity.
The scrubber effluent, including flyash, is discharged to a clarifier;
the overflow is recycled and the underflow from both systems is pumped
to a 160-acre pond. The pond also is used for ash disposal from other
units not equipped with scrubbers. Clarified liquor from the pond is
not returned to the scrubbing systems.
Kansas City Power and Light Company - La Cygne No. 1
This new 820-MW unit began operation, along with its FGD system in mid
1973. Early availability was low, but recent operating time has improved
at reduced load. Both flyash and S02 are removed in the absorbers and
sludge is pumped to a settling pond with a capacity to accommodate about
5 yr of operation. Supernatant liquor is returned to the scrubbing system.
Adequate settling has been obtained and the returned liquor is clear.
Plans for disposal beyond the initial 5 Yr have not been finalized.
Louisville Gas and Electric Company - Paddy's Run No. 6
The scrubbing system on this relatively small (70 MW) unit began operation
in April 197J. The system has been used intermittently and for short
periods. Flyash is removed ahead of the absorber. The sludge, almost
entirely calcium sulfite, is collected in a clarifier where a flocculent
is used to improve settling. Initially, the thickener underflow was de-
watered on drum filters and trucked to a nearby borrow pit. Recently,
the underflow (about 2J/o solids) has been treated with lime at the rate
567
-------
of about 80 Ib/ton of dry solids to stabilize the sludge. The treated
waste solids are trucked to a landfill dump. Future sludge disposal
studies to be conducted by Louisville Gas and Electric and sponsored
by EPA are reported in another paper. Methods for disposal of waste
solids from other planned Louisville Gas and Electric FGD systems have
not been disclosed except for general planning.
City of Key West - Stock Island Unit
Operation of this small (37 MW) oil-fired unit and its scrubbing system
began in the spring of 1973- Tne system is presently shut down,, but
during the period of operation, the waste solids containing a relatively
large proportion of limestone (coral) were settled in a storage pond
and the clarified seawater was discharged to the sea.
Arizona Public Service Company - Cholla No. 1
This 115-MW limestone scrubbing system was started up in late 1973-
On-stream time has been high. Sludge disposal consists of pumping a
sidestream from the scrubber directly to a settling pond. The same
transport system and pond are used for ash disposal. No pond water is
recycled to the scrubber system.
Nevada Power Company - Reid Gardner No. 1 and 2
These two 125-MW units were equipped with sodium scrubbing systems and
began operation in 197^-• They are listed here because they are operated
as throwaway systems. The scrubber effluent from both units is pumped
to a two-stage pond. The solids are settled out in the first pond and
the overflow goes to a larger ^5-acre evaporation pond. No liquor is
recycled.
Public Service Company of Colorado Valmont No. ^
The gas cleaning system on this unit treats a portion of the gas equiva-
lent to 100 MW and was installed initially for particulate collection.
In early 1975? addition of limestone was begun. Because of low inlet S02
concentration, the relative oxidation level is high and the sludge is
composed mainly of calcium sulfate and fly ash. Solids separation takes
place in a settling pond. Particle size segregation has been noted in
the settled solids; near the point of discharge, the solids content is
about 80/0 while sludge near the clarified liquor outfall contained only
solids.
568
-------
Montana Power Company - CoIs trip No. 1
Operation of the new J^O-MW unit., equipped with lime scrubbing began
in late 1975- The alkalinity from the flyash is utilized for S02 re-
moval while the addition of supplemental lime is used for pH control in
the scrubber (a developing trend,, when feasible,, for low-sulfur Western
coals). Much like Valmont No. 5? the low inlet S02 concentration yields
high relative oxidation and results in a sludge composed mainly of
calcium sulfate.
Montana Power has chosen ponding as a method of disposal. They had first
wanted to place the sludge back into the mine but the state would not
permit this method of disposal.
The following approach is used. The flyash-calcium-sulfur product mixture
is sluiced to a 15-acre temporary holding pond (operared on a closed-loop
basis) which will be reclaimed by dredging. The resulting dredged slurry
will be piped (l5$> solids) to an ultimate disposal site approximately
3 miles from the holding area. Excess water will be returned to the hold-
ing pond forming a closed-loop between the temporary and ultimate disposal
area.
Monitoring wells are to be constructed around the pond site. After the
disposal sites are filled., they will be covered and reclaimed.
Detroit Edison Company - St. Glair No. 6
A limestone scrubbing system on this IJO-MW unit began operation in late
19T5• Most of Detroit Edison's study of sludge disposal is a result of
limestone scrubbing pilot-plant testing at the 1-MW level. The sludge
produced in the Detroit Edison pilot plant was unique in that even though
the flue gas had a high S02 content, the resulting sludge was high in
calcium sulfate (approximately 90^> CaS04). The normal sulfate content
for their sludge should have been about 10-2.0%. The reason for this high
degree of oxidation is still not clear.
Detroit Edison in conjunction with Marston Associates have studied a
process that would pelletize sludge by utilizing the "plaster of paris"
reaction. The process as tested required the conversion of some of the
gypsum in the sludge into calcium sulfate hemihydrate (CaS04-1/2H20),
which in turn was used in a drum granulator for pellet conversion. The
end product was being viewed as a disposal product. Future status of
this process is not clear.
At Detroit Edison's St. Clair Station,, a 12-acre sludge disposal pond
will be used to take care of 2 yr of sludge produced from the full-scale
scrubber. The unit after 2 yr is scheduled to be converted over to low-
sulfur fuel with the scrubber being used only for flyash control.
569
-------
Kentucky Utilities Company - Green River No. 1, 2, and 3
The flue gas from three small (22 MW) boilers is treated in a lime
scrubbing system; operation began in late 1975- The sludge contains
a relatively large amount of calcium sulfate (kO$> oxidation) for a
high-sulfur coal and a small quantity of ash. Solids are removed from
the system in a diked settling pond with a storage capacity of about
5 yr. Clarified liquor is returned to the scrubber. During the short
period of operation to date,, settling has been good.
In addition to this listing of operating systems, 16 installations
based on lime or limestone scrubbing are under construction. The method of
disposal has not been published for many of these, but the number of those
using unlined ponds for disposal of untreated sludge is likely to grow.
Lined Ponds
Information is available for only one installation in the United
States that will utilize a lined pond, and it is not yet in operation.
Northern States Power Company - Sherburne No. 1 and 2
Two new 620-MW generating units are being equipped with limestone
scrubbing systems; startup is planned in mid-1976 for the first unit
and about a year later for the second. Low-sulfur Western coal will be
burned and ash will be collected in the scrubbing system. Sludge (clari-
fier underflow) disposal will be in a 65-acre pond with k-0-ft dikes and
lined with 18 in. of clay. Clarified liquor will be returned to the
scrubbing system. The pond is designed for a 10-yr life and additional
capacity of similar construction will be added as needed for these two
units. Future expansion at Sherburne County will be a problem because
no additional pond space is available. To make matters worse, it will
be difficult to find enough clay to line any additional ponds.
Tennessee Valley Authority - Widows Creek No. 8
A limestone scrubbing system is being installed on a 550-MW unit and
operation will begin in early 1977- The pond is not actually lined, but
has been constructed from local clay that would be suitable for use as a
liner; the compacted material will result in low permeability. The
100-acre pond with JO-ft dikes is designed with a divider dike to permit
maximum liquor retention time. Scrubber effluent with about 15/0 solids
will be pumped to the pond and clarified liquor will be recycled. The
estimated life of the pond is 7 yr. Additional capacity will be installed
later.
570
-------
Northwest German Power Company - Wilhelmshafen
Although this obviously is not a U.S. installation, the situation is
interesting. A new 7^-0-MW coal-fired unit will be completed in mid-
1976- Gas volume equivalent to Ik-5 MW will be treated with the Bischoff
lime scrubbing process. This process was developed in a 35~^ prototype
near Lunen where sludge storage leaching tests indicated that ground
water quality was not affected. A clay-lined pond is being constructed
and it will be operated with closed-liquor loop. However, the government
has not yet approved a permit for its use.
Sludge Fixation
A sizable amount of development work by vendors and utilities has
gone into methods for stabilizing sludge. The situation regarding use of
this technology is as follows:
Commonwealth Edison Company - Will County No. 1
At Will County flue gas from a 167-MW cyclone boiler is cleaned by lime-
stone scrubbers which in turn produce 200-250 tons of dry solids under
full-load conditions. These solids are typical of this method of scrubbing
in that they are composed of unreacted limestone, calcium sulfite, calcium
sulfate, and flyash. After the scrubber, the sludge is pumped from the
power station to a clarifier where the solids content is increased from
8 to 12% to approximately 351/0 solids. After dewatering, the sludge is
mixed with additives for fixation purposes.
Two fixation recipes have been used over the past few years. One con-
sisted of adding 100 Ib of portland cement and k-00 Ib of flyash per
2000 Ib of dry sludge. The other consisted of adding 200 Ib of lime and
k-00 Ib of fly ash per ton of dry sludge solids. Initially, the sludge
stabilization mixture was transported by mix trucks to a settling pond
at the plant where it was allowed to fix. At present, the latter fixation
recipe is used and the mixture is transported by redi-mix trucks to an
offsite disposal area (under EPA permit) owned by the transporting con-
tractor. The fixed sludge becomes strong enough to support construction
machinery and trucks after minimal compaction.
The original ponded material has been removed from the initial settling
pond. This material had fixed sufficiently to allow handling with front
end loaders for loading into dump trucks.
Future disposal will consist of hauling part of the fixed sludge to a
nearby site which is being developed as a golf course. An EPA disposal
permit was also obtained for this disposal site.
571
-------
The cost of this process is quoted at $l8-50/dry ton of sludge solids.
The cost includes labor, flyash, cement or lime handling, redi-mix
trucks, front end loaders, and maintenance. It is reported that this
operating cost could be reduced by $4 to $5/ton of dry solids if a
pugmill were used.
Duquesne Light Company - Phillips Station
Duquesne Light has been using the Dravo stabilizing agent, Calcilox,
since early 19J4 to fixate sludge produced by the lime flue gas desulfuri-
zation system located at their 1|-10-MW Phillips Station.
The sludge is first dewatered in a clarifier to about 35-^0% solids. The
clarifier underflow is then mixed with 200 Ib Calcilox per dry ton of
solids and sent to one of three 6,500' yd3 ponds for curing which requires
approximately 30 days. The purpose of having three ponds is to have one
for receiving, one for curing, and the last for excavating. Excavation
is required to remove the fixed sludge from the pond for transportation
to the final disposal site where it is mixed and compacted with dry fly-
ash from the precipitator. The station's ash disposal area, located
about a mile from the plant, is used for this purpose. Areas that become
filled are covered with top soil and seeded.
Based on the capacity of the present disposal site, it will be completely
filled in 2 yr. To solve this problem, Duquesne has requested proposals
from Dravo, IU Conversion Systems, and U.S. Utilities Service Corporation,
a company currently involved with the present ash removal and sludge
treating activities. No firm proposals had been received as of November
1975, although U.S. Utilities has indicated it will propose a method of
handling both the flyash and the sludge.
IU Conversion Systems has an 18-mo contract with Duquesne to handle the
sludge produced by a 200-MW lime scrubber located at Duquesne Light's
Elrama Station. The equipment used to make synthetic aggregate at
Southern California Edison's Mohave Station has been brought in for the
test periods.
Southern California Edison Company - Mohave Station
Much of the SCE full-scale sludge-disposal-related research was carried
out at the Mohave scrubber installation. The sludge produced here was
similar to other sludges which are a product of scrubbing flue gas with
a low S02 inlet concentration; it was composed mainly of calcium sulfate.
Therefore the Mohave sludge dewatered well (60-JO/o solids) and was fairly
stable without further treatment.
572
-------
Primary treatment of this sludge consisted of dewatering in a thickener
where both cationic and anionic flocculents were added, one for the
flyash fraction, the other for the sulfur products. After settling,
the underflow was subjected to fixation procedures. SCE was under con-
tract with both Dravo and IU Conversion Systems to perform fixation
studies.
The Dravo process took the thickener underflow, added Calcilox, and
sometimes lime, and placed the mixture in a pond for hardening; some
hydrogen sulfide odor was detected in the vicinity of this pond.
The IUCS work led to the production of a synthetic aggregate. Most of
the work was carried out in secrecy (tests conducted at an isolated area
surrounded by a fence) and details of the process are not available. It
was learned that control of the production step presented problems in
that the recipe for producing a good quality aggregate was critical.
Under certain conditions, the aggregate was fine and was easily blown
away, while with only a small change in process variables, the product
became very sticky. The process was however optimized and a paved parking
lot demonstration test using this aggregate was made in Las Vegas.
SCE will probably not use a fixation process at Mohave for sludge dis-
posal since calcium sulfate is somewhat stable and enough land area
exists at the Mohave site to take all the sludge produced for the life
of the plant (estimated to require 4,000 acre-ft).
Ohio Edison Company Bruce Mansfield No. 1 and 2
Two new ooO-MW units are being equipped with lime scrubbing systems;
one unit recently started up but the disposal system is not fully
operational. Dravo has the contract for waste disposal. Clarifier under-
flow containing about JOfo solids will be mixed with Calcilox and pumped
to the disposal area approximately 7 miles from the power station. The
site consists of a large landfill ravine behind a Ij-^O-ft dam constructed
for the project. The estimated life of the disposal area is JO yr. Clari-
fied liquor will be pumped back to the scrubbing system.
Other Installations
Other planned installations of fixation processes include:
Installation Startup date
Louisville Gas and Electric
Cane Run No. k, 1?8 MW June 1976
Louisville Gas and Electric
Mill Creek No. 5, k-25 MW June 1977
Columbus and Southern Ohio Electric
Conesville No. 5 and 6, 400 MW each 1976 and 1978
Indianapolis Power and Light
Petersburg No. J, 5JO MW 1977
573
-------
RESEARCH AND DEVELOPMENT BY UTILITIES
In addition to application of disposal technology to operating
or planned flue gas desulfurization systems, several utilities are conduct-
ing experimental work on different methods for handling the waste problem.
The following listing is not intended to be all-inclusive. Many utilities
have conducted pilot-plant studies, mainly in conjunction with vendors, and
several have also investigated stabilization.
Tennessee Valley Authority
Over the past several years, TVA has been involved extensively in sludge
disposal studies, both for its own program and in support of the EPA
program. Operation of a 1-MW limestone pilot plant to provide design
information for the full-scale Widows Creek unit provided an opportunity
to become intimately familiar with waste solids. This experience has
been reported in previous symposia in this series. A major portion of
the EPA scrubber sludge disposal program is being carried out at the
TVA Shawnee plant and is reported in another paper.
Additional studies are under way as part of the research effort funded
by energy-related pass-through funds from EPA to TVA. The main elements
of this program include:
1. A technical, economic, and environmental impact evaluation for the
production of granular fertilizer from scrubber product sludge.
2. A study to determine the range of variability of the solids from
the scrubbers operated at the Shawnee test facility and a corre-
lation of this variability with plant and scrubber operating
conditions.
3. An economic-engineering evaluation of alternative sludge disposal
methods.
TVA is also under contract with EPRI to provide solids characterization
and scrubber correlation at several full-scale scrubber installations
now in operation in the utility industry.
Southern California Edison
SCE is operating a 10-MW lime scrubber at its Highgrove plant. The main
purpose is to evaluate their horizontal scrubber efficiency when using
high-sulfur fuel oil. Because of the high S02 content in the flue gas,
the oxidized sulfur constitutes a small percentage of the total sulfur
absorbed (200 ppm out of 2,000 oxidized); thus the sludge is composed
mainly of calcium sulfite (90/0 CaS03), the opposite of what was seen at
Mohave. Edison has tested and developed a process to oxidize the calcium
sulfite to calcium sulfate that yields 97fo oxidation with air oxidation
and pH control.
574
-------
SCE hopes to develop a market for the oxidized sludge in the wallboard
industry and will probably make a 50-ton full-scale wallboard test with
Kaiser Cement Company.
Another interest of SCE is a method of sludge utilization where the
oxidized sludge is used in agriculture. Gypsum promotes growth of de-
nitrifying bacteria in soils containing high concentration of nitrates.
It has also been found to promote the growth of alfalfa and peanuts and
is a useful additive on any sulfur-deficient soil.
Other testing has included dewatering studies using thickeners,, filters.,
and centrifuges and drying studies using spray dryers,, thermal disc
dryers, and kiln dryers. Some study of pelletizing has also been made.
SCE is also studying the feasibility of mixing scrubber sludge with ash
from the coal-fired Kaiparowits Generator Station as a lining for reser-
voirs and for a structural fill or road subbase.
Southern Services
As a result of testing three flue gas desulfurization processes at Plant
Scholz (20 MW each), Southern Services has become involved in both dis-
posal and utilization studies on waste products from the Chiyoda gypsum
process and the CEA-ADC dual alkali process.
Southern Services has contracted with both National Gypsum and United
States Gypsum for bench-scale testing to determine the suitability for
using Chiyoda gypsum in wallboard plants. Southern Services is also
interested in the agricultural applications of using this gypsum, especially
in the area of peanut farming.
At Plant Scholz the sludges are being ponded in two ponds which are equipped
with underdrains so as to collect leachate samples for subsequent analyses
of both major constituents and trace elements.
Fixation work is also being conducted on the dual alkali sludge. Flyash,
Portland cement, and quicklime are being added to the sludge in various
concentrations and combinations so as to define the more promising fixation
scheme. IUCS and Amax have also had an opportunity to test their processes
on the sludge material.
Southern Services also plans to conduct economic and engineering evaluations
for large-scale sludge-handling systems, such as pumping, conveyors, etc.
Plans have been made for prototype and demonstration testing of the phases
of study that prove to be promising.
575
-------
Ontario Hydro
Ontario Hydro has been involved in the study of flue gas desulfurization
systems for several years and has looked at almost every aspect of
dealing with the resulting sludge disposal problem. Research at Hydro
has been broad and in some of the major areas of concern, quite inten-
sive. Test work has covered sludge dewatering techniques such as thick-
eners, vacuum filters, centrifuges and various combinations of each.
Various fixation schemes using flyash, quicklime, and portland cement
have also been evaluated, including defining their effect on permeabili-
ties and strength. Advanced studies have included bench-scale work
dealing with the regeneration of the product sludges to produce a salable
sulfur product and lime or limestone for reuse.
In the area of sludge dewatering, this progressive utility has studied
the effect of polyelectrolytes on settling, the mechanics involved
during the thickening, and methods of improving the efficiencies of the
thickener. At the pilot scale, they have evaluated the thickener-vacuum
filtration scheme and a solid bowl continuous centrifuge to determine
efficiencies and for comparison purposes.
The regeneration studies have centered on reductive roasting of lime or
limestone product sludges to form calcium sulfide with subsequent carbo-
nation of the aqueous slurry of calcium sulfide to recover calcium car-
bonate and to produce hydrogen sulfide for sulfur recovery. As a result
of this testing, much has been learned about the thermodynamics and
kinetics of the main reaction. Based on these results, Hydro has prepared
preliminary flowsheets and material and energy balances for the regeneration
reactions. Estimates of the capital and operating costs have also been
made.
Southwestern Public Service Company
Southwestern Public Service Company in conjunction with Combustion Engineer-
ing is also involved in studying the feasibility of producing structural
landfill material from sludge by mixing scrubber solids with local soils
found around the plant site. Their tests demonstrated that mixtures of
soil and sludge could produce a landfill material with significant load-
bearing strength. This type or method of approach looks more promising
in areas that have a more arid climate.
Empire State Electric Energy Research Corporation (ESEERCp)
A possibility exists that ESEERCO and the State of New York (Office of
Environmental Planning) may become involved in a joint program whereby
the feasibility of using sludge stabilized by the IUCS process for the
construction of fish reefs off the coast of Long Island will be studied.
No decision for this program has been made.
576
-------
CONCLUSIONS
The only firm conclusion that can be drawn with the current
situation is that this paper has come to an end--yet the story has just
begun. A few subjective generalizations probably won't do any harm.
Disposal of untreated sludge in unlined ponds has been in progress
for up to 8 yr with no apparent adverse effect on surface or ground water
quality. However, the extent of evaluation of effects is unclear. Ob-
viously., this is the least cost approach to disposal provided that storage
areas are available. Major unquantified factors with this method are:
(l) potential hazard from the quicksand nature of many sludges, (2.) commit-
ment of land to nonproductive use, (3) potential long-term effects of leach-
ing if and when attenuation capacity is exceeded, and (k-) likelihood of
solid waste regulations whether they are needed or not.
The problem with uncertain effects of leaching can be dealt with
by providing, at some increased cost, a low permeability lining. The speci
fications for alternative materials regarding method of placement, chemical
resistance, mechanical durability, and resistance to varmints of the above-
ground and underground types need to be established before meaningful
economic trade-offs can be considered.
Perhaps the land use problem can be handled best by conversion,
at some increased cost, of calcium sulfite to calcium sulfate (gypsum) so
that the waste solids can be used for landfill as long as convenient holes
are available. Studies under way should provide some answers to questions
(including the cost and whether there is a leaching problem) associated with
this method.
EPA has made its position fairly clear that chemical fixation
or stabilization is the preferred method. It will undoubtedly be the most
expensive. With this approach, all of the risks are reduced substantially,
but it would be good to know if it is worth the cost. Any good insurance
man will tell you that you can be overinsured.
One advantage to fixation is that the added cost to throwaway
processes will make recovery processes more competitive and likely the
economic choice in many situations. The advantage is that sulfur would
be recovered in a useful form and our natural reserves would be conserved.
However, one can't help but be sympathetic with the poor fellow who would
be stuck with the disposable variety.
577
-------
RESEARCH AND DEVELOPMENT FOR CONTROL OF WASTE AND WATER
POLLUTION FROM FLUE GAS CLEANING SYSTEMS
Julian W. Jones
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina
ABSTRACT
The EPA Program for Control of Waste and Water Pollution from Flue
Gas Cleaning (FGC) Systems is designed to evaluate, develop, demonstrate
and recommend environmentally acceptable, cost-effective techniques for
disposal and utilization of FGC wastes, with emphasis on Flue Gas Desul-
furization (FGD) sludge, and to evaluate and demonstrate systems for
maximizing power plant water reuse/recycle. The program currently consists
of nineteen projects, each covering one of six areas of interest--(1) environ-
mental assessment of FGC waste disposal/utilization processes and other power
plant effluents, (2) assessment of the technology of these processes
and development of new technology, (3) studies of the economics of these
processes, (4) development of alternative FGC waste disposal methods,
(5) development of new FGC waste utilization methods, and (6) development
of methods for improving overall power plant water use.
The environmental assessment efforts include FGC waste characterization
studies; laboratory and pilot field studies of disposal techniques for
chemically treated FGD sludges; characterization of coal pile drainage,
coal ash, and other power plant effluents; and studies of attenuation
of FGC waste leachate by soils. The technology assessment and develop-
ment efforts include field studies of untreated and chemically treated
FGC wastes; FGC waste leachate-disposal site liner compatibility
studies; studies to correlate waste solid characteristics with scrubber
operating conditions; and dewatering equipment design studies. Alternate
disposal method studies include both mine and ocean disposal assessments.
The economic studies include cost estimates of current disposal practices
(e.g., ponding, landfill) and by-product marketing studies. Utilization
efforts include development of a process for FGC waste conversion (to
sulfur and calcium carbonate); pilot studies of fertilizer production
(using the waste as a filler material and a source of sulfur); use
of FGD gypsum in portland cement manufacture; and FGC waste beneficiation
studies. The power plant water use effort is a single study to minimize
water use and waste water discharges.
Results from this program can accelerate increasing use of domestic
coal supplies in power plants.
579
-------
This paper has been reviewed by the Environmental Protection
Agency and approved for presentation. Mention of trade names,
commercial products, or commercial processes does not constitute
endorsement or recommendation for use.
580
-------
ACKNOWLEDGEMENTS
The author wishes to express his appreciation to the EPA
Project Officers and the contractor or agency Project Directors
and their associates for their highly professional efforts in the
formulation and implementation of the complex program described
in this paper. The author is also appreciative of the patience
and skill of Mrs. Lynn Pendergraft in typing the manuscript.
581
-------
1.0 INTRODUCTION
A major consideration inherent in plans for installing a flue gas
cleaning (FGC) system is the necessity for disposing of or utilizing
the by-products. This is true for all coal-fired boilers, especially
those with flue gas desulfurization (FGD) systems. Application of FGD
systems in the United States is accelerating; the majority of these are
lime/limestone wet scrubbing systems which produce a calcium/sulfur by-
product. Because of environmental and economic concerns related to the
disposition of this FGD waste, there have been considerable governmental
and industrial research, development, and demonstration activities.
Modest EPA efforts in this area were begun in the late 1960's in support
of the limestone scrubbing program. In late 1972, major research and
development (R&D) efforts were initiated; these efforts were described by
the author in a previous paper. In late 1974, plans were formulated to
greatly expand these R&D efforts as part of EPA's Energy Research Program.
These efforts were aimed at determining pertinent environmental parameters,
reducing costs, investigating alternative strategies, and encouraging by-
product usage. Although the major emphasis in these new efforts was on FGD
wastes, they involved consideration of overall power plant waste and water
problems, including the disposal and utilization of coal ash. For this
reason, the new program was entitled "Control of Waste and Water Pollution
from Flue Gas Cleaning (FGC) Systems". In this paper, the program will be
referred to as the FGC Waste and Water Program.
The FGC Waste and Water Program developed in Fiscal Year (FY) 1975
consisted of the expansion and/or extension of four existing projects and
the addition of twelve new projects. In FY 1976, three new projects are
being added, for a total of nineteen. Each of the projects is listed in
Table 1; the contractor or agency performing the project is listed, along
with funding by Fiscal Year.
2.0 PROGRAM PHILOSOPHY
The objectives of the FGC Waste and Water Program are to evaluate,
develop, demonstrate and recommend environmentally acceptable, cost-effective
582
-------
Table 1: Projects in FCC Waste and Water Program
Project Title
FGC Waste Characterization,
Disposal Evaluation, and
Transfer of FGC Waste
Disposal Technology
Lab and Field Evaluation of
1st and 2nd Generation FGC
Waste Treatment Processes
Characterization of Effluents
from Coal-Fired Power Plants
(Waste and Water only)a
Ash Characterization and
Disposal
Studies of Attenuation of
FGC Waste Leachate by Soils
Establishment of Data Base
for FGC Waste Disposal
Standards Development
Shawnee FGD Waste Disposal
Field Evaluation
Louisville Gas and Electric
Evaluation of FGD Waste
Disposal Options
FGD Waste Leachate-Liner
Compatibility Studies
Lime/Limestone Wet Scrubbing
Waste Characterization
Dewatering Principles and
Equipment Design Studies
Conceptual Design/Cost
Study of Alternative Methods
for Lime/Limestone Scrubbing
Waste Disposal
Gypsum By-Product Marketing
Studies
Evaluation of Alternative
FGD Waste Disposal Sites
Lime/Limestone Scrubbing
Waste Conversion Pilot
Studies
Fertilizer Production
Using Lime/Limestone
Scrubbing Wastes
Use of FGD Gypsum in
Portland Cement Manufacture
FGD Waste/Fly Ash Beneficiation
Studies
Assess/Demonstrate Power
Plant Water Reuse/Recycle
Con tractor/Agency
The Aerospace Corporation
U.S. Army Corps of Engineers
Waterways Experiment Station
Tennessee Valley Authority
Tennessee Valley Authority
U.S. Army Materiel Command
Dugway Proving Ground
Stearns, Conrad and Schmidt
Consulting Engineers, Inc.
(SCS Engineers)
Tennessee Valley Authority/
The Aerospace Corporation
Louisville Gas and Electric Company
(Subcontractor: Combustion Engineering
U.S. Army Corps of Engineers
Waterways Experiment Station
Tennessee Valley Authority
Auburn University
Tennessee Valley Authority
Tennessee Valley Authority
Arthur D. Little, Inc.
M. U. Kellogg Company
Tennessee Valley Authority
Contractor Selection Pending
TRW, Inc.
Radian Corporation
Prior to
FY 75
596
110
75
(203)
200
Funding, $1000
FY 75 FY 76
500
/N
325
300
300
300
150
(250)b
750
100
40
100
100
250
92.4
110
FY 77-80
(est.)
Totals
/\
\/
Totals 781 3617.4 2549 4685
emission characterization efforts; additional funds (not shown here) are provided for those
3This project also includes gaseous
bFunds for this project included in Shawnee lime/limestone scrubbing project funds, therefore not included in program totals.
11,632.4
efforts.
-------
techniques for disposal and utilization of FGC wastes, with emphasis
on FGD wastes, and to evaluate and demonstrate systems for maximizing
power plant water reuse/recycle.
To accomplish these objectives, projects were undertaken covering a
broad range of activities to thoroughly evaluate existing technology and
explore new, not-yet-commercial concepts. Six areas of interest were
identified as requiring attention; these areas reflect the basic philosophy
behind the FGC Waste and Water Program. The areas of interest are:
(1) Environmental assessment of FGC waste disposal/utilization
processes and other power plant effluents
(2) Assessment of the technology of these processes and
development of new technology
(3) Studies of the economics of these processes
(4) Development of alternative FGC waste disposal methods
(5) Development of new FGC waste utilization methods
(6) Development of methods for improving overall power
plant water use
Table 2 shows the relationship of each of the projects to the areas
of interest listed above. It is readily apparent that only'four of the
projects are aimed at investigating one specific area; most of the projects
cover several areas of interest. This is to be expected, since it would
be difficult, for example, to assess the technology of a process without
examining both the economics and the environmental effects. However, each
of the projects does have, as its main function, a single area of interest.
With this in mind, each of the projects in the FGC Waste and Water Program
is described in the next section under the heading of the main area of
interest addressed in the project.
3.0 FGC WASTE AND WATER PROJECTS
In this Section, each of the projects is briefly described, and the
current status of the project is discussed. For readers interested in
more detailed information, the EPA Project Officer and the Contractor or
Agency Project Director are also indicated.
584
-------
Table 2: Relationship of Projects to Areas of Interest
in FGC Waste and Waste Program
Areas of Interest
Project
FGC Waste Characterization, etc.
(Aerospace)
1st and 2nd Generation Waste
Treatment Processes
(Corps of Engineers)
Characterization of Power
Plant Effluents (TVA)
Ash Characterization (TVA)
Leachate-Soil Attenuation
(Army Materiel Command)
Establishment of Data Base
For Standards (SCS Engineers)
Shawnee Field Evaluation
(TVA/Aerospace)
Louisville Waste Disposal Evaluation
(LG&E/Combustion Engineering)
Leachate-Liner Compatibility
Studies (Corps of Engineers)
Scrubber Waste Characterization (TVA)
Dewatering Studies
(Auburn University)
Conceptual Design/Cost Studies for
FGC Waste Disposal (TVA)
Gypsum By-Product Marketing Studies
(TVA)
Evaluation of Alternative Disposal
Methods (A.D. Little)
Scrubbing Waste Conversion Studies
(M.W. Kellogg)
Scrubbing Waste in Fertilizer (TVA)
Gypsum in Portland Cement
(Not Selected)
FGD Waste/Fly Ash Beneficiation (TRW)
Assess/Demonstrate Power Plant
Water Reuse/Recycle (Radian)
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
585
-------
3.1 Environmental Assessment
3.1.1 FGC Waste Characterization, Disposal Evaluation, and Transfer
of FGC Waste Disposal Technology -
This is a contract (68-02-1010) with Aerospace Corporation, under
which a broad-based, on-going study is being performed to (1) identify
environmental problems associated with FGC waste disposal by comparing
chemical and physical characteristics of the waste with current, proposed,
or potentially applicable environmental standards; (2) assess current FGC
waste disposal methods, including feasibility, performance, and costs—by
conducting laboratory studies of wastes as they would be disposed, providing
engineering support and conducting chemical/physical analyses for the Shawnee
field evaluation (see below), making an evaluation of other available field
data, and conducting engineering cost studies of disposal methods; (3) make
recommendations regarding alternate disposal approaches based on (1) and (2);
and (4) assemble, assess, and report all FGC waste-related R&D activities in
EPA, TVA, and private industry. This project is the key effort in the EPA
FGC Waste and Water Program.
Formal reports on the Shawnee field evaluation are planned annually;
integrated R&D reports of government and industry activities are planned
annually; and technical papers will be presented approximately every 18
months at EPA FGD symposia. Initial results from this contract have been
123
reported. ' ' Additional results of Aerospace laboratory studies will
be released in a report within the next few months. The first annual
integrated R&D report, to be released in the Spring of 1976, will summarize
the results from all the projects described in this paper.
Currently, most of the efforts under this project, with the exception
of the Shawnee field evaluation support and the R&D integration effort, are
coming to a conclusion. The project has generated significant information,
examples of which follow:
(1) Chemical analysis of FGD waste liquors has revealed the need
for protection of ground and surface waters because of excess
(over drinking water standards) of chloride, sulfate, and
several of the trace metals.
586
-------
(2) Physical characterization of FGD wastes has quantified
the effect of moisture on physical properties; i.e.,
approximately 65-70 percent solids content is required
for landfill operation.
(3) Chemical fixation of FGD wastes can drastically reduce
the rate of pollutant release to the environment by
decreasing the solubility and permeability of the waste
and by physically stabilizing the material so that it
4
can be placed in a landfill.
(4) Costs of chemical fixation processes depend to a great
extent on local plant conditions (e.g., land availability),
but a reasonable total disposal cost appears to be in the
range of $9-10/metric ton (dry solids basis), or about 1.1-
1.2 mills/Kwhr for a typical plant.
The EPA Project Officer for this project is the author of this paper.
The Aerospace Project Director is Jerry Rossoff.
3.1.2 Lab and Field Evaluation of 1st and 2nd Generation FGC Waste Treat-
ment Processes -
This is an Interagency Agreement (EPA-IAG-D4-0569) with the U.S. Army
Corps of Engineers' Waterways Experiment Station (WES) under which laboratory
and field studies are being conducted to evaluate current commercial (1st
generation) and new, not-yet-commercial (2nd generation) processes for
treatment (e.g., chemical fixation) of several industrial wastes, including
FGC wastes. The 1st generation processes will be evaluated through laboratory
leaching column studies, physical testing of samples of several FGC wastes
treated by chemical fixation processors, and small field studies of selected
fixation processes. Initial efforts in this project are similar to the
Aerospace laboratory studies of chemical fixation processes, except that
additional commercial processes are being evaluated. In addition, visits
will be made to existing full-scale disposal sites for complete evaluation
of current disposal practices, including coring of soil beneath disposal
sites.
587
-------
The 2nd generation processes will be identified through a literature
survey and will be restricted to those which are chemically and operation-
ally defined. Candidate processes will be screened via an economic anal-
ysis, laboratory physical testing, and laboratory testing to determine
resistance to pollutant leaching. Based on the screening results, no more
than five processes will be selected for comprehensive evaluation, con-
sisting of studies similar to those conducted for the 1st generation processes.
An interim report on physical testing and leaching tests of the 1st
generation processes is expected to be issued in April 1976. In addition,
a paper summarizing this information is being presented at this symposium.
For this reason, a status report on this project will not be given here.
Final reports on the 1st and 2nd generation studies are expected in mid-1977
and mid-1978, respectively.
The EPA Project Officer for the WES efforts is Robert E. Landreth, a
Sanitary Engineer with the Municipal Environmental Research Laboratory in
Cincinnati, Ohio. The WES Project Director is Dr. Jerry L. Mahloch.
3.1.3 Characterization of Effluents from Coal-Fired Power Plants -
One of several tasks being conducted by the Tennessee Valley Authority
(TVA) under an Interagency Agreement (EPA-IAG-D5-E-721), this project
involves efforts to (1) characterize and quantify the chemical parameters
of coal pile drainage, (2) assess and quantify the chemical and physical
composition of ash pond effluent after adjustment of pH and suspended solids
reduction to meet effluent standards, (3) evaluate an ash pond monitoring
program to determine the sampling and analyses necessary to obtain repre-
sentative information, (4) assess, characterize, and quantify the effects
of coal ash leachate on ground water quality, and (5) evaluate and quantify
the chlorinated effluent in the discharge canal from once-through cooling
systems.* Information from this project will be supplemented by TVA fly ash
*This project also includes a major effort to characterize and quantify
the gaseous and particulate emissions from at least three typical coal-
fired boilers (including tangential and wall-fired) and at least one scrubber.
Although the emission characterization effort is not considered part of the
FGC waste and water program, the emission data will be combined with FGC
waste data to attempt an overall emissions/effluent material balance.
588
-------
characterization efforts (see below). Reports on each of these efforts
will be issued as the specific effort is completed. The final report is
planned to be released in late 1978.
The experimental design for the coal pile drainage study (which involves
two TVA plants) has recently been completed; sampling and flow measurement
equipment has been ordered. The coal pile drainage test program is expected
to be underway soon. In the ash pond effluent study, a computer model has
been developed to simulate the batch settling characteristics of fly ash
and bottom ash in a continuous ash pond system. Preliminary ash pond field
data indicate that cenospheres may be the major factor contributing to high
suspended solids in the ash pond effluent and that pH adjustment may be
somewhat more difficult than originally envisioned. In examining the effects
of coal ash leachate, the experimental design for soil coring and ground
water studies is complete and is expected to be underway soon. A plant has
been selected for the chlorinated cooling system effluent study; i.e., to
determine the minimum effective chlorine dosage to control biological fouling.
The EPA Project Officers for this project are Ronald A. Venezia and
the author of this paper. The TVA Project Directors are Drs. Billy G. McKinney
and Hollis B. Flora, II of TVA's Power Research Staff, Chattanooga, Tennessee.
3.1.4 Ash Characterization and Disposal -
One of several tasks being conducted by TVA under an Interagency Agree-
ment (EPA-IAG-D5-E-721), this project involves efforts to (1) summarize and
evaluate existing data on the characteristics of coal ash and ash effluents
from in-house TVA studies and from studies made by other organizations;
(2) perform chemical and physical analyses on coal, coal ashes, and ash
effluents to obtain a complete characterization of these materials as a
function of variation in boiler design and operation, as well as coal type;
(3) evaluate various methods for disposal and utilization of fly ash;
(4) summarize information on methods of ash sluice water treatment for reuse;
(5) conduct conceptual design studies of dry and wet ash handling systems;
and (6) recommend the most promising systems for ash handling and disposal/
utilization. Reports on each of these efforts will be issued as the
589
-------
specific effort is completed. The final report on the project is planned
for January 1979.
The summary of available data on coal and ash characterization is
essentially complete. The major experimental effort under this project,
item (2) in the preceding paragraph, is being coordinated with the gaseous
emission studies in the project described in 3.1.3, above, and is not
expected to be in "high gear" until the late spring of 1976. However, a
"dry run" to verify the rather complicated sampling and analytical program
has already been successfully accomplished.
The EPA Project Officer for this project is the author of this paper.
The TVA Project Director is Ms. Shirley S. Ray of TVA's Power Research
Staff.
3.1.5 Studies of Attenuation of FGC Waste Leachate by Soils -
This is an Interagency Agreement (EPA-IAG-D4-0443) with the U.S. Army
Materiel Command's Dugway Proving Ground under which a study is being
conducted to determine the extent to which heavy metals and other chemical
constituents in several industrial wastes, including FGC wastes, can migrate
through the soil in land disposal sites. At least six scrubber wastes and
three coal fly ashes will be tested under this project. The project will
consist of the following efforts: (1) physical and chemical characterization
of the wastes; (2) leachate studies in columns with the wastes applied to a
variety of U.S. soil types; (3) long term permeability tests with selected
clays; and (4) interpretation of data from the tests in (1) and (2) to
identify soil attenuation mechanisms and to develop empirical "attenuation
coefficients" for specific chemical substances. The final report for the
project is expected to be issued in late 1977.
Characterization of three of the FGC wastes has been completed;
permeability tests with these wastes are currently underway. The leachate
columns have been designed; several have been fabricated. The leaching
studies are expected to begin in March 1976. All of the scrubber wastes
and soils should be under test by the end of April. A preliminary progress
report on the leaching tests should be available by late 1976.
590
-------
The EPA Project Officer for this project is Dr. Mike Roulier, a
Research Soils Scientist with the Municipal Environmental Research Laboratory
in Cincinnati, Ohio. The Dugway Project Director is Dr. Martin Houle.
3.1.6 Establishment of Data Base for FGC Waste Disposal Standards
Development -
This is a contract (68-03-2352) with SCS Engineers which includes
efforts to: (1) establish the criteria necessary to evaluate FGC waste
disposal options; (2) compile any proposed or existing standards/regulations
which have been or could be applied to the land disposal of FGC wastes;
(3) evaluate the impact of applying existing and proposed standards/
regulations; (4) determine the interrelationships between environmental
effects and regulatory approaches, considering prevention of adverse
health, ecological, and aesthetic effects; (5) develop a technical basis
for future standards development; (6) recommend R&D to supply any additional
information and/or technology needed to implement the standards. Initiated
in November 1975, this effort is expected to be completed by late 1976;
however, because of emerging technology in the area of FGC waste disposal,
initial results are expected to require subsequent updating.
The EPA Project Officer for this project is Donald E. Banning, a
Research Chemist with the Municipal Environmental Research Laboratory in
Cincinnati, Ohio. The SCS Project Director is C. J. Schmidt.
3.2 Technology Assessment and Development
3.2.1 Shawnee FGD Waste Disposal Field Evaluation -
This is part of an Interagency Agreement (EPA-IAG-D5-E-721) with TVA
for lime/limestone scrubber tests at Shawnee. Under the current program,
the Chemfix, Dravo, and IU Conversion Systems processes for chemical fixation
scrubber wastes are being evaluated in three separate disposal ponds.
Untreated lime and limestone wastes are placed in two additional ponds.
Leachate, run-off and ground water samples as well as core samples of the
wastes and soil are being collected and analyzed to evaluate environmental
effects. Both Aerospace Corporation and TVA are performing selected analyses;
Aerospace is responsible for data evaluation and reporting. Future plans
591
-------
call for evaluation of other disposal approaches, including gypsum
disposal. Preliminary results from this effort are being reported by
Aerospace in a separate paper at this symposium. For this reason,
a status report will not be given here. The first formal report on
this project is expected to be released later this month. The project
is scheduled to be completed by September 1977.
The EPA Project Officer for this project is the author of this
paper. The Aerospace Project Director is Jerry Rossoff; the TVA
Project Director is James J. Schultz of TVA's Office of Agricultural
and Chemical Development, Muscle Shoals, Alabama. The program is
coordinated with the Shawnee scrubber test program directed by Bechtel
Corporation for EPA.
3.2.2 Louisville Gas and Electric Evaluation of FGD Waste Disposal Options -
This is part of a contract (68-02-2143) with Louisville Gas and Electric
Company (LG&E) to conduct a program of carbide and commercial lime scrubbing
tests and an extensive evaluation of scrubber waste treatment/disposal
options. Laboratory studies of non-chemical and chemical (fixation)
processes for stabilization of scrubber sludge will be conducted using
carbide lime and commercial lime scrubbing wastes; samples will be mixed
with fly ash alone or fly ash and one of several additives (e.g., Portland
cement). Results of testing of the stabilized samples will be used to
determine those mixtures to be field tested. The field studies will consist
of (1) small-scale (about 19 cubic meter) impoundment tests in which leachate,
run-off and physical stability tests of unstabilized and stabilized waste
material will be conducted in above-ground "swimming pools", and (2) larger
scale (about 76 cubic meter) landfill tests in which leachate migration,
run-off, and physical stability tests of unstabilized and stabilized waste
material will be conducted in landfill areas with in-situ clay and synthetic
linings.
An interim report will be issued after completion of the laboratory
(and part of the field) studies. The final report is expected to provide
significant information on the technical and environmental acceptability of
several scrubber waste disposal methods. The recently initiated project is
expected to be completed by mid-1977; the interim report is expected to
be issued in late 1976.
592
-------
The EPA Project Officer for this project is the author of this paper.
The LG&E Project Director is Robert P. Van Ness.
3.2.3 FGC Waste Leachate-Liner Compatibility Studies -
This is an Interagency Agreement (EPA-IAG-D5-0785) with the U.S.
Army Corps of Engineers' Waterways Experiment Station (WES) under which
a program is being conducted to (1) determine the compatibility of 18
liner materials with flue gas cleaning (FGC) wastes and associated liquors
and leachates; (2) estimate the length of life for the liners, and (3)
assess the economics involved with purchase and placement (including disposal
area construction) of various liner materials. A review of the literature
will be made to evaluate potential liner materials, including admixed
materials (e.g., soil cements, asphalt cements, stabilized FGC waste),
flexible materials (e.g., polyvinyl chloride, polyethylene), and sprayed-
on materials (e.g., plastics, asphalts, sulfur). At least 18 materials
will be selected for tests in exposure cells designed to simulate a depth
of sludge/liquor of at least 6 meters. The liner materials will be
subjected to physical tests to determine whether the exposure to the FGC
wastes causes deterioration in the strength and/or permeability of the
material; these will be conducted after zero (baseline testing), 12, and
24 months of exposure. The final report for this project is expected to
be issued in the spring of 1978.
Ten chemicals that are admixed into soil, six chemicals that are sprayed
onto soil, and two flexible materials that are placed over soil have been
selected by WES for testing. The eighteen liner materials will be placed
inside a 38 cm (15 in;)diameter test chamber which will be pressurized to
simulate a 6 meter (20 ft) sludge/liquor head. The soil to be used in the
-4
admixes and base filler is a silty sand with a permeability of about 5 x 10
cm/sec. Baseline tests of the flexible materials are underway; optimum
admixture compositions are currently being defined.
The EPA Project Officer for this project is Robert E. Landreth of the
Municipal Environmental Research Laboratory in Cincinnati, Ohio. The WES
Project Officer is Z. B. Fry.
593
-------
3.2.4 Lime/Limestone Wet Scrubbing Waste Characterization -
One of several tasks being conducted by TVA under an Interagency
Agreement (EPA-IAG-D5-0721), this project involves the physical and
chemical characterization of lime/limestone wastes as a function of
scrubber operating conditions. These studies are a continuation of
efforts conducted during 1974 with Colbert pilot plant wastes. The
Colbert waste solid crystals varied greatly depending on the presence
of fly ash. Oxidation is another major factor affecting the crystal
size and structure. Under these studies, TVA will (1) characterize waste
materials from the Shawnee facility* and correlate the physical and
chemical properties of the materials with the scrubber operating conditions
(e.g., boiler excess air, liquor and gas flow rates, inlet and outlet
scrubber liquor pH, and inlet SO,, concentration), and (2) suggest, if
feasible, a means of controlling waste characteristics, particularly
the physical properties (such as ease of dewatering) to improve disposal
or utilization economics. The initial Shawnee studies are expected to
be completed by mid-1976; the final Shawnee studies should be completed
by mid-1977.
So far, no distinct mathematical relationships have been shown
between the properties of the waste solids and the scrubber system
operational parameters. However, samples taken during high limestone
utilization tests conducted in November 1975 at the Shawnee facility
inferred the possibility of an inverse relationship between stoich-
iometric ratio and calcium sulfite hemihydrate crystal size. Greatly
improved dewatering characteristics were evident at stoichiometric
ratios near 1.0. Photomicrographs taken of samples from a test run at a
stoichiometric ratio of 1.03 showed crystals several times larger than
those from a test run at a stoichiometric ratio of 1.43. This relationship
will be studied further.
*Under a separate study funded by the Electric Power Research Institute
(EPRI), waste materials from other (full-scale) facilities will be obtained,
and the physical/chemical properties will be correlated with scrubber
operating conditions.
594
-------
The EPA Project Officer for this project is the author of this
paper. The TVA Project Director is James L. Crowe of TVA's Power
Research Staff.
3.2.5 Dewatering Principles and Equipment Design Studies
This project, for which a grant application has recently been
received from Auburn University, will consist of the following efforts
to improve current FGC waste dewatering equipment: (1) an examination
of current dewatering equipment design principles to determine their
applicability to FGC wastes; (2) laboratory settling and other tests to
determine the physical properties and behavior of FGC wastes as a basis
for dewatering equipment design studies; (3) analytical design studies
to develop dewatering equipment designs based on FGC waste physical
properties and behavior (these efforts will continue and will be updated
based on subsequent bench scale testing); and (4) laboratory tests of
dewatering equipment design concepts. Further testing of the concepts
developed may be conducted if there is sufficient interest expressed by
private industry. This project offers the potential of cost savings
through reduction in the size of dewatering equipment and the volume
of the disposal site, as well as through reduction in the amount of
chemicaJ fixation additive (if used) required. The project is expected
to be initiated in mid-1976 and should be completed by the fall of 1978.
The EPA Project Officer for this project is the author of this
paper. The Auburn University Project Director is Professor James C. Warman.
3.3 Economic Studies^
3.3.1 Conceptual Design/Cost Study of Alternative Methods for Lime/Lime-
stone Scrubbing Waste Disposal -
This project is one of several tasks which make up the economic studies
of major FGD processes being conducted by TVA under an Interagency Agreement
(EPA-IAG-D5-E-721) (results of other economic studies have been reported in
Q
another paper at this symposium). In this project several FGD waste
disposal methods and FGD system design and operating premises will be
selected for a detailed economic evaluation of FGD waste disposal. Currently
595
-------
available information, such as engineering cost estimates from the Aerospace
contract and fixation vendor estimates from the Shawnee field evaluation,
will be used in the initial efforts, with updating as additional information
becomes available. Alternatives will very likely include variations in
mechanical dewatering equipment, variations in treatment (e.g., oxidation
to gypsum, chemical fixation), and variations in ultimate disposal (e.g.,
ponding, landfill). The final report on this two-year effort is expected in
late 1977-
The EPA Project Officer for the TVA economic studies is Richard D. Stern
of the Industrial Environmental Research Laboratory in Research Triangle Park,
N. C. The TVA Project Director is H. Lewis Faucett of TVA's Office of
Agricultural and Chemical Development, Muscle Shoals, Alabama.
3.3.2 Gypsum By-Product Marketing Studies -
This project is one of several tasks which make up the FGD by-product
marketing studies being conducted by TVA under an Interagency Agreement
(EPA-IAG-D5-E-721) (results of the sulfuric acid and sulfur marketing studies
9
will be reported in another paper at this symposium). A preliminary study
conducted by TVA during early 1974 indicated the possibility that production
and sale of abatement gypsum might offer a substantial economic advantage
over FGD waste disposal. These new studies include a thorough economic
evaluation of gypsum producing processes (e.g., Chiyoda, carbon absorption,
CaSO,, oxidation) and a detailed U.S. marketing study of abatement gypsum
for wallboard. A report on this study is expected in late 1976. Future plans
include studies of abatement gypsum for use in portland cement manufacture.
The EPA Project Officer for the TVA marketing studies is Dr. Charles
J. Chatlynne of the Industrial Environmental Research Laboratory in Research
Triangle Park, N.C. The TVA Project Director is John I. Bucy of TVA's
Office of Agricultural and Chemical Development, Muscle Shoals, Alabama.
3.4 Alternate Disposal Methods Development
This effort consists of a;t contract (68-03-2334) with Arthur D. -Little,
Inc. (ADL) under which a program is being conducted to identify, assess, and
596
-------
demonstrate, on a pilot scale, alternate FGC waste disposal methods
(other than local ponding and landfilling). The present program calls
for the demonstration to be limited to coal mine and ocean disposal.
The initial effort will consist of an evaluation and assessment of the
compatibility, capability, and adequacy of deep and surface mines,
oceans and other potential disposal sites for handling and disposal of
untreated and/or chemically treated FGC wastes (sludges). The use of
these wastes as soil amendments will also be assessed. Although environ-
mental effects and operational safety will be the major initial consider-
ation, the assessment will also include a study of the economics of the
alternate disposal methods, as well as a study of applicable Federal and
state regulations. Based on all of the initial efforts, recommendations
and conceptual designs for the pilot demonstrations will be made.
The pilot demonstrations will be conducted at such a scale that
design data for full-scale operations can be obtained. The mine alternative
will probably consist of tests on small plots at an existing mine; the
ocean alternative will consist of a tank which can simulate the ocean
environment. The mine study will include monitoring for pollutants,
characterizing pollutants, identifying limits of physical and chemical
characteristics of wastes for disposal, and waste/mine interactions. The
ocean study will include settling and dissolution characteristics and,
in general, any parameters which would stress the ecosystem.
The initial assessment effort is nearing completion; a report on this
effort is expected to be released by mid-1976. This assessment is based upon
information, drawn from the published literature and from personal technical
communications, on the chemical, physical, and mechanical (engineering)
properties of both treated and untreated wastes. A key source of data has
been the results of work conducted for EPA by Aerospace and the Corps of
Engineers (WES). This data has been used to set a range of properties of
wastes generated by lime, limestone and dual alkali scrubbing systems.
In the evaluation of FGC waste disposal in mines, an initial screening
of the various types of mines in the U.S. was performed to group the mines
into categories that would be amenable to a general assessment effort and
597
-------
to identify the most promising candidates. The screening process resulted
in the selection of six general categories of mines: active area strip
coal mines; active underground coal mines; inactive or mined-out portions
of active underground limestone mines; inactive lead-zinc mines; mined-
out iron ore pits; and inactive or mined-out portions of salt mines.
Characteristics of mines in each of these categories were reviewed to
develop a set of typical environmental and operational conditions including
hydrology-, geology, climate, mine size and method of mining. As appropriate,
these general categories have been further subdivided according to their
general characteristics. Each group of mines was reviewed with regard to
the alternatives for FGC waste placement, the physical properties of wastes
that would be suitable, the capacity for FGC waste and the environmental/
technical impacts of FGC waste disposal. Attention is also being given to
the potential use of FGC waste (either treated or untreated) for subsidence
control, as a tailings (soil) amendment and for neutralization or prevention
of acid mine drainage.
The evaluation of the ocean disposal of FGC waste has focused on the
Atlantic and Gulf Coast areas, since these regions are the most likely for FGC
waste, generation. The assessment includes both shallow (near shore) and
deep ocean disposal with dispersed and/or bottom dumping of FGC waste. As
in the evaluation of the impact of mine disposal, the ocean disposal assess-
ment involves the determination of potential physical and chemical impacts.
Special emphasis is being placed on the biological impact of ocean disposal,
since the waste substrate and contaminants can be more readily available to
marine organisms.
The EPA Project Officer for the project is the author of this paper.
The ADL Project Director is Dr. Richard R. Lunt.
3.5 Utilization Technology Development
3.5.1 Lime/Limestone Scrubbing Waste Conversion Pilot Studies -
This project involves a cost-shared contract (currently being negotiated)
with the M.W. Kellogg Company to conduct pilot studies of two key process
steps in the "Kel-S" process for conversion of lime/limestone scrubbing waste
598
-------
to elemental sulfur with recovery of the calcium in the waste as calcium
carbonate. The effort includes (1) reduction of lime/limestone waste
to calcium sulfide in a continuously rotating kiln, (2) dissolution of
the kiln-produced calcium sulfide to calcium hydrosulfide using hydrogen
sulfide from the calcium recovery unit, and (3) recovery (precipitation)
of the calcium as calcium carbonate, using carbon dioxide-rich gas (avail-
able from the drying kiln), with simultaneous release of hydrogen sulfide,
which can be converted to elemental sulfur (for disposal or sale) in a
Glaus plant. Separate pilot tests of the dissolution and recovery contactors
will be made, followed by an integrated, continuous test using both con-
tactors. Design data will be generated to allow scale-up to a large (pro-
totype) test unit for a power plant. Assuming successful completion of
the pilot testing and reasonable economics, this process would present a
viable alternative to lime/limestone scrubbing waste disposal. The project,
expected to get underway in the spring of 1976, will last approximately
11 months.
The EPA Project Officer for this project is the author of this
paper. The Kellogg Project Director will be A. Glenn Sliger.
3.5.2 Fertilizer Production Using Lime/Limestone Scrubbing Wastes -
One of several tasks being conducted by TVA under an Interagency
Agreement (EPA-IAG-D5-E-721), this project involves the use of lime/lime-
stone scrubbing wastes as a filler material and a source of sulfur for
fertilizer. This study is a continuation and expansion of previous bench-
scale laboratory production tests and small field plot application tests
with rye grass. Under the current effort, TVA will (1) conduct pilot
plant tests of production of fertilizer using lime/limestone scrubbing
wastes, (2) conduct tests of compatibility factors involved in storage
and mixing of this fertilizer material with conventional fertilizer,
(3) conduct field plot tests using fertilizer from the pilot plant, (4)
conduct economic/marketing studies, and (5) determine the amounts of
trace and/or toxic elements in this fertilizer and compare them with
those in conventional fertilizer. Initial pilot plant production and
storage/compatibility tests are expected to be completed by mid-1976.
The project is expected to be completed by mid-1979.
599
-------
In the initial pilot production tests, severe chemical difficulties
were encountered when scrubbing waste was introduced into the preneutralizer
portion of the fertilizer granulation unit. (The preneutralizer is the
equipment into which ammonia and phosphoric acid are introduced,) When
the waste was fed to the drum granulator (the preneutralizer outlet stream,
or ammonium phosphate melt, was fed to the granulator via a separate system),
the pilot plant still did not operate satisfactorily because of the excess
moisture in the waste. It is felt that development and testing of a new
preneutralizer design will be necessary to alleviate the problems encountered
so far; these efforts are currently underway.
The EPA Project Officer for this project is the author of this paper.
The TVA Project Director is James L. Crowe.
3.5.3 Use of FGD Gypsum in Portland Cement Manufacture -
This project will consist of the following efforts: (1) laboratory
tests at a portland cement manufacturing plant to determine the possible
range of variability in the chemical quality and physical characteristics
of FGD gypsum which can be used, so that recommendations can be made for
operational changes either in the FGD system or the cement plant; (2) pilot
testing of FGD gypsum physical and/or chemical conditioning and portland
cement manufacture using the gypsum product; and (3) a test program
involving full-scale FGD equipment (possibly including FGD waste oxidation
equipment) at a coal-fired utility power plant and a full-scale portland
cement manufacturing facility. Reports will be issued after completion of
each of the efforts described. The project is expected to get underway by
mid-1976, and is expected to be completed by mid-1979.
The EPA Project Officer for this project is the author of this paper.
The contractor for this effort will be announced in the near future.
3.5.4 FGD Waste/Fly Ash Beneficiation Studies -
This project will consist of the following efforts: (1) a conceptual
design/cost study of a TRW-conceived, proprietary process (for which two
related patents have been issued) for producing sulfur, alumina, and dicalcium
600
-------
silicate for FGD waste and fly ash (and/or clay), including the development
of a preliminary process design (including material and energy balances)
and an estimate of capital and operating costs to determine the economic
viability of the process; (2) if the economics of the process appear
favorable, a bench-scale laboratory investigation to determine probable
ranges of operating conditions for each of the major processing steps
(reduction, calcination, leaching, separation) and to determine (with
actual FGD waste) probable yields and purity of products; (3) assuming
the process still appears viable, possible pilot scale testing of the
process steps to obtain design data for large-scale equipment. Reports
will be issued after each of the efforts described above.
The conceptual design/cost study is expected to be underway in the
very near future; this effort will take approximately 4-6 months to
complete. The bench scale tests should be initiated in early 1977.
The EPA Project Officer for this project is the author of this paper;
the TRW Project Director will be Dr. Jack Blumenthal.
3.6 Power Plant Water Use
This effort consists of a contract (68-03-2339) with Radian Corporation,
who will conduct a study on minimizing water use and waste water discharges
from coal-fired steam-electric power plants. The study will consist of six
tasks: (1) Plant selection and characterization—selection of three or more
specific plants for detailed analysis; collection of detailed data on make-up.
process, and effluent waters, plant design, operation modes, coal composition
and climate for each plant. (2) Process model preparation—preparation of
computer models to simulate make-up, process, and effluent water streams,
and chemical equilibria of the processes for each of the specific plants
selected for detailed study. (3) Simulation of existing plant operations—
verification of process computer models by comparing existing plant chemical
and operating data with data predicted by the models. (4) Technical assess-
ment of recycle/reuse options—formulation of a number of water recycle/re-
use options to minimize plant water requirements and discharges for the
specific plants selected for study; evaluation of at least one option (via
601
-------
process simulation) for each plant. (5) Cost estimates—preparation of
capital and operating cost estimates for each viable water recycle/reuse
option. (6) Recycle/reuse assessment report—detailed presentation of
program results, including recommendations of the recycle/reuse options
to be used at each of the plants studied. Current plans call for completion
of this effort in late 1976, followed by pilot plant testing of one or more
of the recycle/reuse options at a coal-fired power plant.
Currently, three plants are being studied under the project—Four
Corners (Arizona Public Service), Comanche (Colorado Public Service) and
Bowen (Georgia Power). Water samples from selected streams have been
taken at each of the plants; chemical analysis of these samples has recently
been completed. Simulation of existing plant operations will be made in
the near future, after computer program modifications have made to simulate
the equipment arrangements at specific plants.
The EPA Project Officer for this project is Fred A. Roberts of the
Industrial Environmental Research Laboratory, Research Triangle Park, N.C.
Mr. Roberts is currently located in Corvallis, Oregon. The Radian Project
Director is Dr. Delbert M. Ottmers.
4.0 CONCLUSIONS AND FUTURE PLANS
The main thrust of future activities under the FGC Waste and Water Program
is continuation of the efforts described above. In summary, this means that
the environmental/technological assessment and economic efforts will be
pursued to logical conclusions, and that development of promising FGC waste
disposal/utilization alternatives and power plant water use technologies will
also be pursued. However, because some of the technology in the area of FGC
waste disposal/utilization is only in the developmental stage, emphasis is
expected to shift to those projects with the greatest promise of success.
This shift in emphasis may create new projects in areas which are not not
being intensively studied: e.g., modifications within the FGD process to
change the physical/chemical characteristics of the waste solids.
The current FGC Waste and Water Program is a comprehensive program
designed to reduce the environmental and economic impact of power plant waste
602
-------
and water pollution controls. Results from efforts under this program
will provide alternative strategies which can be applied to a fairly broad
range of power plant situations. However, application of the results could
be strengthened and broadened if additional funding were applied to the
pilot/demonstration phases of several projects. Sources of these funds
could include private industry, the Electric Power Research Institute (EPRI),
and other government agencies. In this regard, discussions are now being
held with EPRI to determine those areas where joint financial support of
projects would be most fruitful. Although specific projects have not yet
been identified, the FGC waste utilization area is a likely candidate.
A main part of the significance of this program, in the author's
opinion, is that it sponsors research and development for which there is,
at least initially, little or no financial incentive to private industry.
More importantly, results from this program can accelerate the increasing
use of domestic coal supplies in fossil fuel-fired power plants; this is
an important part of the nation's energy objectives.
603
-------
5.0 REFERENCES
1. Jones, J.W., "Environmentally Acceptable Disposal of Flue Gas
Desulfurization Sludges: The EPA Research and Development
Program." In Proceedings: Symposium on Flue Gas Desulfurization—
Atlanta, 11/74, Volume II, EPA-650/2-74-126-b (NTIS No. PB 242-573/AS),
12/74.
2. Rossoff, J. and R.C. Rossi, The Aerospace Corporation, "Disposal
of By-Products from Non-Regenerable Flue Gas Desulfurization
Systems: Initial Report," EPA-650/2-74-037-a, (NTIS No.
PB 237-114/AS), 5/74.
3. Rossoff, J., et al., "Disposal of By-Products from Non-Regenerable
Flue Gas Desulfurization Systems: A Status Report." In Proceedings:
Symposium on Flue Gas Desulfurization, Atlanta, 11/74, Volume I,
EPA-650/2-74-126-a, (NTIS No. PB 242-572/AS), 12/74.
4. Personal Communication, J.- Rossoff, The Aerospace Corporation, 1/76.
5. Mahloch, J.L., U.S. Army Corps of Engineers, "Chemical Fixation of
Flue Gas Desulfurization Sludges - Physical and Chemical Properties,"
presented at Symposium on Flue Gas Desulfurization, New Orleans, 3/76.
6. Rossoff, J., and R.C. Rossi, The Aerospace Corporation, "EPA Shawnee
Flue Gas Desulfurization Waste Disposal Field Evaluation," presented
at Symposium on Flue Gas Desulfurization, New Orleans, 3/76.
7. Personal Communication, J.E. Williams, EPA, 11/75.
8. G.G. McGlamery, et al., Tennessee Valley Authority, "Flue Gas
Desulfurization Economics," presented at Symposium on Flue Gas
Desulfurization, New Orleans, 3/76.
9. Bucy, J.I., et al., Tennessee Valley Authority, "Marketability of
Abatement Sulfuric Acid and Sulfur from FGL Applied to Power Plants
in Eastern United States," presented at Symposium on Flue Gas Desul-
furization, New Orleans, 3/76.
604
-------
FLUE GAS CLEANING WASTE DISPOSAL
EPA SHAWNEE FIELD EVALUATION
J. Rossoff and R. C. Rossi
The Aerospace Corporation
Environment and Energy Conservation Division
El Segundo, California
ABSTRACT
This paper summarizes the first year's results (1974-1975) of the
EPA's on-going Flue Gas Cleaning (FGC) Waste Disposal Field Evaluation
Project at the TVA Shawnee Steam Plant in Paducah, Kentucky. In this
project, five different disposal ponds containing lime or limestone flue
gas scrubbing waste from two 10-MW (equivalent) prototype scrubber sys-
tems (i.e., a Chemico venturi/spray tower and a UOP Turbulent Contact
Absorber) are being evaluated. Two of these ponds contain untreated
wastes, and each of the other three contains waste treated by one of
three different commercial chemical fixation processors: Chemfix,
Dravo, and IU Conversion Systems. This paper discusses initial results
of analyses of leachate, supernate, and ground waters from all ponds,
and physical properties of fixed waste cores taken from the disposal
site. Based on data obtained from these evaluations and from correlations
of these data with long-term laboratory analyses at Aerospace, projections
are made that show orders-of-magnitude reductions of leachate total
dissolved solids as a result of reducing the solubility and permeability
of the FGC wastes and of minimizing the recharge of water to the
disposal site. Additionally, total fixation-disposal cost studies
were made, resulting in estimates of $7.30 to $11.40 per ton of waste
(on a dry basis).
605
-------
FLUE GAS CLEANING WASTE DISPOSAL -
EPA SHAWNEE FIELD EVALUATION
1. 0 INTRODUCTION
The Evnironmental Protection Agency (EPA) has initiated a power
plant site field evaluation of the disposal of untreated and treated flue gas
cleaning (FGC) wastes for the purpose of verifying the effects of several
disposal techniques and scrubbing operations, soil interactions, and field
operation procedures on the environmental acceptability of the disposal site.
The program began in September 1974 and is scheduled to continue to mid-
1977. The principal objectives of this program are as follows: (1) evaluate
current disposal techniques under representative field operating conditions;
(2) evaluate the environmental acceptability of current disposal technology
through periodic sampling, analysis, and assessment of water, soil, and
waste cores; and (3) develop engineering cost estimates for alternative dis-
posal methods on an operational basis.
The Tennessee Valley Authority (TVA) Shawnee Power Station at
Paducah, Kentucky, was chosen as the site for the evaluation. Two different
scrubber systems (i.e. , a UOP Turbulent Contact Absorber and a Chemico
venturi followed by a spray tower, using limestone and lime, respectively,
as the SO£ absorbent) upstream of fly ash collection are being operated at
this station as an EPA/TVA test facility. Each scrubber system is capable
of independently treating up to 10 MW (equivalent) of flue gas from one boiler
(approximately 30, 000 ft /min at 300°F). Wastes from these scrubbers are
being used in the disposal evaluation and are undergoing analysis in several
laboratories under EPA sponsorship. 1~5 Related laboratory analyses are
being conducted under EPA contract by Mahloch et al. °
The disposal evaluation site consists of five disposal ponds, each
occupying approximately 0. 1 acre near the Shawnee plant and approximately
one mile south of the Ohio River (see Figure 1). All are six feet deep and
have been filled with FGC wastes to a depth of approximately three feet. Two
of the ponds contain untreated wastes; each of the remaining ponds contains
wastes chemically treated by one of the three commercial contractors. The
surfaces of the three treated ponds are sloped to create a wet section con-
sisting of a combination of liquor and rainwater, and a potential dry section
(depending on weather conditions) for the observation of physical conditions
of dry material. Within each pond is a four-inch-diameter plastic pipe
leachate well that collects water samples at the waste-soil interface. A
ground water well is located on a berm of each pond, between the sludge and
the river, and a ground water well is located approximately 100 feet south of
each pond (away from the river) to provide monitoring of background water
quality. The disposal sites are being monitored periodically for leachate,
supernate, and ground water quality; soil chemistry changes; and treated
waste chemical and physical qualities.
It should be noted that each of the three treated ponds was conditioned
and filled by the respective contractors in response to program requirements
so that various operational conditions and the effects of weather could be
evaluated. These operations, therefore, do not represent identical disposal
conditions or necessarily the operational methods that would be employed by
any of the contractors. For example, different input materials were supplied
to the contractors; the ponds are not provided with drainage so that the effects
606
-------
Figure 1. Overview of disposal site
of trapped water can be observed; one treated material (Chemfix) was frac-
tured and contoured by a back hoe; one (Dravo) represents curing and dis-
posal underwater where high strength is not necessarily required; and one
(IUCS) was not compacted by placement vehicles that would be used in a full-
scale operation. Therefore, comparisons between processes are neither
attempted nor implied in these evaluations. The principal environmental
factors being observed are the quality of seepage water, and the strength and
permeabilities of the fixed materials. Correlations are being made between
laboratory-prepared samples and those created under field operating condi-
tions. Additionally, results obtained are being used to form a basic under-
standing of the major characteristics of field operations, estimate total costs,
define procedures for planning additional evaluation at this site, and assist in
the development of other EPA-sponsored field evaluations.
This program is managed by the EPA Industrial Environmental
Research Laboratory, Research Triangle Park, North Carolina. The
Aerospace Corporation provides program coordination, program plans,
selected analyses, analytical results, costing estimates, and reporting. TVA
provides for all construction, filling of untreated ponds, supplying wastes to
fixation processors at the site, maintenance, sampling and analyses, sample
distribution, climatological and hydraulic data collection, and photographic
documentation. The waste fixation processors are Chemfix, Inc., Pittsburgh,
Pennsylvania; Dravo Corporation, Pittsburgh, Pennsylvania; and IU
Conversion Systems, Inc. (IUCS), Philadelphia, Pennsylvania. The Bechtel
607
-------
Corporation (the scrubber facility test director) provides the technical
interface relating the scrubber test facility to the disposal evaluation.
A detailed description of this program, and results of the first year's
work are given elsewhere. Highlights of these results are presented herein.
2. 0
POND FILLING AND FIXATION
The five disposal ponds were filled between 7 October 1974 and
23 April 1975; the characteristics of the fill material are as shown in Table 1.
Table 1. CHARACTERISTICS OF POND FILL MATERIALS
Pond
A
B
C
D
E
Scrubber
Type
Venturi/ Spray
Tower
Turbulent Con-
tact Absorber
Venturi /Spray
Tower
Turbulent Con-
tact Absorber
Turbulent Con-
tact Absorber
Waste
Absorbent /Source
Lime /Filter Cake
Limestone /Clarifier
Lime /Centrifuge
Cake
Limestone /Clarifier
Underflow
Lime stone /Clarifier
Underflow
Solids
Content
(Untreated),
wt%
46
38
55
38
38
Treatment
Contractor
Untreated
Dravo
IUCS
Untreated
Chemfix
The operation associated with the filling of each pond is described in the
following sections.
2. 1
UNTREATED PONDS (A and D)
2. 1. 1 Pond A. Pond A was filled between 24 September and 8 October 1974
with untreated waste from the venturi/spray tower scrubber in which lime was
used as the absorbent. The waste had been dewatered by filtering and had a
solids content of 46 wt% when it was placed in the pond. Ash constituted
approximately 43 wt% of the solids. A rotary drum mixing truck was used to
haul the waste to the pond in order to maintain a homogenous mix during
loading and transportation. Dispersal of the waste in the pond was achieved by
dumping the waste at various locations in the pond from which it was allowed
to settle and seek its natural level. The condition of the waste in Pond A one
month after filling is shown in Figure 2. Rainwater accumulated on the pond
over the next few weeks, as shown in Figure 3.
608
-------
Figure 2. Pond A one month after filling
Figure 3. Pond A two months after filling
609
-------
2. 1. 2 Pond D. Pond D was filled twice, i. e. , from 11 to 20 October 1974
and from 13 January to 5 February 1975. The material used in the first filling
was subsequently transferred to Pond E; during the transfer the material was
chemically treated by Chemfix. The waste used for both fillings was clarifier
underflow from the Turbulent Contact Absorber, with limestone as the absor-
bent and a solids content of 38 wt%. Likewise, on both occasions, ash repre-
sented approximately 38 wt% of the solids. For both fillings, a rotary drum
mixing truck was used to transport the waste to the pond, and dispersal was
as described for Pond A. The condition of Pond D two months after the
second filling is shown in Figure 4.
2. 2
Figure 4. Pond D two months after second filling
TREATED PONDS (B, C, and E)
The materials used in the evaluation of chemical fixation represented
various disposal operating conditions. Pond B was filled using clarifier under-
flow chemically treated by Dravo and placed in the pond under conditions
approximating disposal behind a dam. Pond C was filled using waste that had
been dewatered by centrifuging, fixed by IUCS, and stored in the pond under
conditions representing a landfill. Pond E was filled with clarifier underflow
chemically treated by Chemfix and placed in the pond under conditions repre-
senting a landfill. Liquor on the surface and rainwater were allowed to re-
main in these ponds to represent a low spot in a landfill from which water
does not readily drain. The operations associated with the filling of each of
these three ponds were as follows.
610
-------
2. 2. 1 Pond B (Dravo). Pond B was filled from 7 to 15 April 1975. The
effluent delivered to Dravo was limestond clarifier underflow from the
Turbulent Contact Absorber. The waste was 38 wt% solids, and the solids
contained 40 wt% ash. Dravo received the effluent from the clarifier; used
a rotary drum mix truck for transportation, and added the Dravo proprietary
additive (Calcilox®) to each truck load from 55-gal drums through the use of
a fork lift. The amount of Calcilox® added represented approximately 11 wt%
of the dry solids being treated. The waste settled to approximately 45 wt%
solids and stabilized at a rate controlled by the Calcilox® content.
Because of the small scale and temporary nature of the project, the
treated waste was dumped directly into Pond B and allowed to settle and cure
under the supernate and subsequent rainwater. Under normal field conditions,
the supernate would be returned to the scrubber loop. Views of Pond B six
days and two months after filling are shown in Figures 5 and 6, respectively.
Solids can be seen in the foreground in Figure 5.
The purpose of this process is to treat -wastes of various solids
content. It produces a material that resembles cemented soil in consistency
for use as a material for landfill, either by mechanical compaction or by
stabilization behind a dam. (Dravo recommends Ref. 7 for further informa-
tion on their process. ) Pond B simulates the latter condition, except, as
noted, there is no recirculation of the supernate in the Shawnee evaluation.
In this evaluation, Calcilox® was prepackaged for 10 to 11 percent additive
based on an expected -waste solids content of 3 5 wt% and diluted further by
pump seal -water. Based on the experience gained at Shawnee and their
laboratory data, Dravo recommend a 7. 5 wt% (dry basis) additive for a full-
scale operation using Shawnee-type waste, with an assumed average of
38 wt% solids.
2. 2. 2 Pond C (IUCS). Pond C was filled from 31 March to 23 April 1975.
The -waste delivered to IUCS -was from the venturi/spray tower scrubber
(using lime as the absorbent) and was clarifier underflow dewatered by a
centrifuge. The average solids content was 55 wt%, and the ash comprised
45 wt% of the solids. The centrifuged waste was conveyed to an IUCS-
operated rotary drum mixer truck and transported to the pond site -where
additive -was mixed -with the -waste prior to discharge into the pond.
The IUCS process produces a material identified as Poz-O-Tec®,
-which has applications as landfill, as -well as variations of this material for
other structural applications. °-ll In the Shawnee field evaluation, IUCS used
a lime additive premixed with fly ash. (The quantity of lime is dependent on
the moisture content of the -waste and the reactivity of the fly ash already
contained in the FGC waste. ) The waste solids content ranged from 47. 5 to
59 wt%, and the average quantity of lime additive was approximately 4. 8 wt%
of the dry solids being treated plus approximately an equal amount of fly
ash. Delivery of the treated waste to the disposal site by truck is generally
the transportation mode recommended by IUCS for a full-scale disposal opera-
tion under conditions similar to those of the Shawnee evaluation.
A significant benefit claimed for this method of fixation is low
permeability (as well as reduced solubility). In this evaluation, the waste
was dispersed by manual raking; some degree of compaction was achieved by
this process. Dispersal and compacting methods appropriate to handling large
quantities -would be required in a full-scale operation. Figures 7 and 8 show
Pond C three days and three months after filling, respectively.
611
-------
Figure 5. Pond B six days after filling
Figure 6. Pond B three months after filling
612
-------
Figure 7. Pond C three days after filling
*-'
Figure 8. Pond C three months after filling
613
-------
IUCS has reported that the addition of dry fly ash is not mandatory,
nor is it planned for full-scale operations when fly ash is present in the wastes
as is the case at Shawnee. It was also reported by IUCS that, when their
process is used to treat an FGC waste such as that used at Shawnee and when
the treated waste is disposed of in a managed landfill, the lime additive will
be in the range of 1 to 4 wt% (dry basis).
2. 2. 3 Pond E (Chemfix). Pond E was filled between 3 and 7 December 1974
using the waste stored in Pond D as input material. This waste was clarifier
underflow from the Turbulent Contact Absorber in which limestone was used
as the absorbent. The solids content was 38 wt%, and ash constituted 38 wt%
of the solids. The waste stored in Pond D was thoroughly mixed before it
was pumped from the pond into a Chemfix processing trailer and then pumped
into Pond E (see Figure 9). The Chemfix process used the reaction of sodium
silicate and portland cement with the waste to stabilize it. Approximately
one month after filling, the material was contoured with a back hoe by TVA
field personnel so that it would more evenly cover the pond surface and so
that an evaluation could be made of a fixed material that had been fractured
and moved by heavy equipment (see Figure 10).
The Chemfix process is designed to handle -waste fixation over a
broad range of percent solids and produces a material having a soil-like
appearance. Furthermore, it-is not designed to prevent the percolation of
water but rather to bind the constituents chemically to accomplish pollution
control while also providing structurally stable properties.
In accordance with the specification provided Chemfix for the Shawnee
evaluation, they treated clarifier underflow. In their process, the amount of
additives is significantly affected by the moisture content of the wastes and
the degree of drainage of the landfill. For the Shawnee evaluation, with 38 to
40 wt% solids and the treated waste in an undrained landfill, Chemfix reported
that the additives required were 46 wt% of the dry solids content. Chemfix
also reported the following: (1) If the water in the pond above the treated
material were removed, the additive required to achieve an equivalent condi-
tion would be reduced to 39 wt%; (2) Dewatering of the -waste to 50 wt%
solids would reduce additive requirements to approximately 15 wt% of the dry
solids content; and (3) Dewatering to 55 wt% solids would further reduce the
additive requirements to about 9 wt% of the dry solids content.
3. 0 ANALYTICAL RESULTS
Considerable data have been determined from analyses of input FGC
wastes, supernate, leachate, ground water, and soil and fixed material cores.
The most significant of these data as regards environmental impact are the
quality of leachate and ground water, and the strength and permeability of the
fixed wastes; therefore, these analyses are emphasized in the following
sections. Data available are discussed, and as the project progresses, these
data will be expanded and correlated with other available laboratory and field
evaluation data.
3.1 UNTREATED WASTES
For Pond A, which contains untreated waste, the data obtained from
leachate samples to date show that the concentrations of the dissolved solids,
614
-------
Figure 9. Pond E during filling, before contouring
Figure 10. Pond E five months after contouring
-------
i. e. , Cl, SO4, and total dissolved solids (TDS), progressively increase with
time (see Figure 11). The data also indicate that the concentrations may
level off at approximately those measured in the input liquor. Simultaneously,
the concentrations of these same constituents in the pond supernate vary with
time. Neither the scrubber system waste solids nor liquors are replenished,
therefore, the supernate should become increasingly diluted with rainwater
as the program progresses. Some fluctuation in this trend can be expected
as a result of evaporation during dry periods. The detection of heavy metals
in the leachate and supernate of this pond shows trends similar to the major
species, however, concentration projections are not as easily made because
of the relatively small magnitude of the values (see Figure 12). Continued
field monitoring is expected to produce data that will define concentrations and
mass released to the soil with respect to time. As regards pH of the leachate,
the first year's monitoring of both untreated ponds has shown a decrease to 7
from the initial range of 8 to 9.
Pond D, the other untreated pond, was filled in October 1974 and
was used as the input source of material for the Chemfix operations at Pond E
in December 1974. At that time, approximately 75 percent of the waste from
Pond D was removed. The pond was refilled with a similar waste in February
1975. The removal and refilling of this pond have produced some discon-
tinuities in the initial data such that projections are not as easily made as for
Pond A. Leachate dissolved solids concentrations are shown in Figure 13 for
this pond starting with the second filling. The accumulation of additional data
will continue throughout the program to produce more accurate trends.
Analyses of samples of ground "water (approximately 12 feet below
Pond A and 40 feet below Pond D) have not shown any effect of seepage from
these ponds.
3.2 TREATED WASTES
The data from the ponds containing treated wastes, although sampled
over a shorter period of time, show trends similar to those of the untreated
waste, except the reductions of concentrations due to chemical fixation are
evident in the leachate analyses. Indications are that these concentrations
either start at or quickly build up to (depending on the amount of rainwater
present in the well initially) approximately 50 percent of the respective
concentrations in the liquor of the untreated input waste (see Figures 14, 15,
and 16). The long-term results of these field evaluations will be studied and
reported throughout this project.
A sample laboratory leachate analysis for long-term TDS concentra-
tions of an untreated Shawnee limestone scrubbing waste and of a core taken
from a treated pond at the Shawnee disposal site is shown in Figure 17. This
treated pond contains limestone scrubber waste and has been in operation for
the longest time period of the three treated ponds, approximately one year.
In these analyses, distilled water was allowed to seep through the pores of
the samples, and chemical analyses of the collected leachate were made
periodically. Figure 17 shows the variation of TDS concentration as a function
of pore volume displacement of leachate water (pore volume in this case is the
nonsolid volumetric portion of a given mass). These trends are generally
typical for the various constituents of the materials, including trace elements.4'
Measurement of the pH of leachate from the treated ponds has exhibited a
616
-------
- 10,000
INPUT LIQUOR IDS
AVERAGE = 8285 mg/l
10
(9/9/74)
20 30 40 50
ELAPSED TIME, weeks
Figure 11. Pond A leachate dissolved solids
0.001 -
0.0001
ELAPSED TIME, weeks 0 10 20 30 40 50 60
CALENDAR DATES 7-1-74 9-9-74 11-18-74 1-27-75 4-7-75 6-16-75 8-25-75
Figure 12. Minor constituents of Pond A supernate and leachate
617
-------
^ 5000
E
i 4000
or
\—
§ 3000
CD
CJ
s 2000
-A-
CD
oo
CD
OO
OO
1000
0
0
(1/27/75)
A IDS
INPUT LIQUOR IDS
AVERAGE = 5375 mg/l
10
20
30
40
50
60
ELAPSED TIME AFTER SECOND FILLING,
weeks
Figure 13. Pond D leachate dissolved solids
CJ
CD
CJ
oo
CD
CD
OO
CD
OO
OO
CD
4000
3000
2000
1000
0
(4/7/75)
INPUT LIQUOR TDS
AVERAGE = 5685 mg/l
10 15 20
ELAPSED TIME, weeks
25
30
Figure 14. Pond B leachate dissolved solids
618
-------
_ 5000
czn
E
4000
CD
CJ)
CD
GO
CD
DO
GO
3000
2000
1000
CD
E
CD
CJ
GO
CD
—I
CD
GO
CD
GO
GO
CD
INPUT LIQUOR TDS
AVERAGE = 9530 mg/l
(4/7/75)
10 15 20
ELAPSED TIME, weeks
25
Figure 15. Pond C leachate dissolved solids
4000
3000
2000
1000H
0
INPUT LIQUOR TDS
AVERAGED 6245 mg/l
TDS
(11/18/74)
10 20 30 40 50
ELAPSED TIME, weeks
30
60
Figure 16. Pond E leachate dissolved solids
619
-------
UNTREATED SHAWNEE LIMESTONE FGC WASTE
PONDE CORE
_L
j_
10 15 20
PORE VOLUME DISPLACEMENTS
25
Figure 17. Laboratory leachate analyses of treated and untreated
FGC wastes showing the effect of leachate pore volume
displacement on TDS concentrations
"100
10
A END OF 5TH PORE VOLUME
o 1.5 AT 1300 YEARS
J I I L
20
100
J_
YEARS
120 140
Figure 18. Mass loading of TDS to subsoil for various disposal
modes of treated and untreated FGC wastes
620
-------
decrease to values of 9 to 10 from a range of 11 to 1 2 during the first six-
month period. Based on laboratory column data, pH is expected to stabilize
at approximately 8, where it is believed to reflect the buffering action of the
carbonate ion.
Several significant characteristics are shown in the data presented
in Figure 17. First, both materials, treated and untreated, display a steep
drop in leachate TDS concentration during the first five pore volume displace-
ments. During this period, salts from the waste liquor trapped within the
pores are flushed out. Thereafter, solubility of the sulfite, sulfate, and
chloride within the solid mass becomes one of the controlling parameters.
Another point of interest is that the time to achieve five pore volume displace-
ments is not the same for the two materials. Because the permeability rate
of the treated material is at least one order of magnitude lower than that of
the untreated material, the time scale for that event is significantly retarded.
The effect of time is shown in Figure 18. A third significant point is that,
after the initial flushing period, the leachate and the TDS concentrations have
stabilized to about 1900 ppm for the untreated and 500 ppm for the treated;
this represents a reduction of concentration of the leachate from treated
material to about 25 percent of the untreated.
The combined effects of all these factors are shown in Figure 18. It
represents some initial results from an analysis being conducted to project
data such as that shown in Figure 17 to operational conditions and to time
periods well beyond the end of this project. Because a treated disposal site
releases leachate to the subsoil over a longer time period and at smaller con-
centrations than an untreated disposal site, the purpose of this analysis is to
make comparisions between treated and untreated materials in terms of the
mass per unit area of dissolved solids released to the subsoil as a function of
time. In Figure 18, five sample cases shown in Table 2 are plotted based on
Table 2. INPUT DATA FOR STUDY CASES
Case
1
2
3
4
5
Disposal
Mode
Lakea
Lakea
Pond
Pond
Landfill
Surface
Water
Constant
Supernate
Constant
Supernate
10 in. / yr
Recharge
10 in. /yr
Recharge
1 in . /yr
Recharge
FGC Waste, 5-Year Fill
Waste
Condition
Untreated
Treated
Untreated
Treated
Treated
Depth,
ft
30
30
30
30
30
Permeability
(Cp),b
cm/sec
io-4
ID'5
io-4
io-5
io-5
Fractional
Pore
Volume
0.67
0.67
0. 67
0.67
0. 67
Assumed maximum hydraulic head of six feet during filling, including
depth of wastes. One foot constant water cover thereafter.
b - 5
For all cases, subsoil C 10
P
621
-------
the data in Figure 17. The advantages of reducing the permeability and the
solubility of the wastes, and of preventing water from accumulating on the
surface, are highlighted in this figure. For example, for the first 100 years,
a treated material in a landfill with controlled runoff (Case 5) releases approx-
imately one-hundredth the mass of dissolved solids at a given rate to the sub-
soil as an untreated material in a lake (Case 1), and approximately one-tenth
as much as an untreated material in a pond (Case 3). Years to reach five
pore volume displacements (initial drainage period) are also shown.
Leachate analyses such as those discussed above are being continued;
more specific operational parameters will be applied, and the results related
to disposal criteria that may evolve during this project. Although some
methods of disposal produce much less release of mass to the subsoil than
others, this is not to indicate that the methods with the lower release rates
are the only acceptable approaches. Criteria for judging disposal methods
are being determined by the EPA and will be applied when they are available.
Some preliminary physical properties of the fixed wastes are given
in Table 3. Of particular interest are the permeabilities (6.9 X 10 ~5 to
5.5 X lO'7 cm/sec) and unconfined compressive strengths (3 to 39 ton/ft^).
Analyses for these properties will be made on cores taken from the treated
ponds periodically throughout this project. FGC wastes that have not been
treated or conditioned have a permeability coefficient of approximately 10~4
cm/sec and no practical structural strength. The effects of conditioning
untreated wastes, e.g. , dewatering and compaction, are being studied.
4.0 SOIL
As noted earlier, the ground waters show no evidence of altered
quality resulting from the filling of any of the five ponds. This result is in
agreement with expectations based upon the very low permeabilities measured
on clay soil samples from the floor of the ponds. Analyses conducted by TVA
show a typical permeability in the range of 10 "^ cm/sec for these soils. Thus
far, the waste leachate constituents would be expected to permeate to a depth
of less than 0. 5 inches. Laboratory analyses using an ion microprobe mass
analyzer are underway to detect the progress of the constituents in successive
soil cores in order to verify long-range analytical predictions over a relatively
short time period, i. e. , within the time span of the evaluation program. At
this point, the initial calibration runs have been completed on pond floor core
samples to provide background data for soils analyses.
5.0 TOTAL DISPOSAL COST ESTIMATES
In order to assess the economics of the cross section of chemical
treatment and disposal modes and variations in effluent conditions such as
ash and solids content, engineering estimates were requested from Chemfix,
Dravo, and IUCS for the estimated cost of full-scale operations. The Shawnee
field evaluation experience provided a basis to relate processing variables to
the costs of an operation treating and disposing of 125 ton/hr (dry basis) of
waste, including fly ash, from a 1000-MW power plant. Typically, the solids
in the wastes supplied to the processors consisted of approximately 45 wt%
fly ash, and 45 to 50 wt% sulfite and sulfate (in a 3 to 1 ratio), with the re-
mainder being unreacted limestone or precipitated calcium carbonate.
622
-------
Table 3. CHARACTERISTICS OF CORES FROM CHEMICALLY FIXED FGC WASTES
ON
Source
Pond EC
(2/27/75)
Pond Bd
(6/12/75)
Pond CC
(5/29/75)
(6/12/75)
Unconf ined
Compressive
Strength, psi
Weta
103-133
27-33
410-510
Dryb
95-165
40-46
470-540
3
Density, g/cm
Weta
1 .40-1 .46
1 . 36-1.44
1.67-1.70
Dryb
0. 69-0.73
0. 59-0. 62
1. 05-1.08
Water
Content,
wt%
51.0-51. 5
56.9-57. 8
36. 5-37.0
Estimated
Fractional
Pore
Volume
0.71-0.73
0.75-0.76
0.57-0 58
Water
Permeability,
cm /sec
1. 5-2.7 x 10~5
6.9 x iO"5
5. 5 x 10"^
5. 5 x 10
Wet: as received.
Dry: after oven drying.
Samples from Ponds E and C were taken from locations free of surface water.
Pond B material is kept underwater continuously as in the case of disposal upstream of a dam.
-------
Capital and operating costs for the full-scale projections were
presented by the fixation contractors, including items such as capital invest-
ment, additives, labor, processing, and transportation, with disposal of the
fixed material at sites both 0. 5 and 5. 0 miles from the power plant. Costs
in terms of dollars per ton of dry waste disposed were provided. Those costs
were evaluated and adjusted by Aerospace to produce estimates of total dis-
posal costs on a common basis as much as possible. Major steps taken were
as follows: (1) annualized 30-year average capital charges of 18 percent
were used in all cases, (2) land at $1000/acre and dewatering costs were
added as appropriate, (3) transportation and site preparation costs were ad-
justed as appropriate, and (4) all costs were adjusted for 30-year average
annual load factors of 50 and 65 percent.
Considering all three processes, the Aerospace-adjusted total cost
for fixation and disposal of the FGC wastes (using a 50 percent load factor and
a 5-mile disposal site) is in the range of $7. 30 to $11. 40/ton of waste (dry) in
1975 dollars. If it is assumed that coal is burned at a rate of 0. 88 Ib/kW-hr,
the total disposal costs for this Shawnee-type waste are $2.07 to $3. 24/ton of
coal. This cost equates to 0.9 to 1. 4 mil/kW-hr. An average cost reduction
of approximately 7 percent can be attained in this analysis by increasing the
average load factor from 50 to 65 percent, and approximately 9 percent by
using a disposal site 0. 5 miles from the plant instead of 5 miles. Detailed
background data related to these results are given elsewhere.
Because these cost data are based on types of plants and operations
that are not universally applicable, no attempt was made to rank the disposal
costs for the three processes being evaluated. As a result of this evaluation,
however, these data provide a range within which the total cost of FGC "waste
fixation and disposal may be expected.
5. 1 OTHER COST CONSIDERATIONS
The engineering estimates provided are considered to be represen-
tative of the cost of disposal by fixation. Some factors that could affect dis-
posal costs somewhat but are highly site-dependent were omitted. These
include access roads and rights-of-way whose costs may be offset by the
residual value of the land. Also, credit for the cost of fly ash disposal was
not applied in this study. Consideration will be given to that factor in follow-
on assessments of the cost impact of FGC waste disposal over current waste
disposal costs. Another approach that may have merit in reducing disposal
costs is the removal of fly ash prior to scrubbing and, in some cases, reintro-
ducing it after the scrubber waste is mechanically dewatered. An appreciable
increase in the percent solids would result, thereby reducing the following:
(1) fixation additive requirements, (2) the total mass of material to be treated
and handled, and (3) the acre-feet of disposal site required. Cost trade-off
studies to evaluate these effects will be made as this project progresses.
6.0 PLANS
Plans are being implemented to make two modifications to the Shawnee
Scrubber Program to ( 1) include an oxidation unit that can be used to convert
sulfite wastes to gypsum, and (2) incorporate a flue gas by-pass so that fly ash
can be collected upstream of the scrubbers and ash-free wastes can be produced,
Concurrently, plans are being made to expand the disposal project to include
two new impoundments for the evaluation of disposal of the gypsum, and the
624
-------
ash-free waste that is dewatered, mixed, and compacted at the disposal site
"with low-moisture -content fly ash. Linings and underdrainage are being
considered. These sites should be operating by the summer of 1976. Later
in this project, one or two of the four untreated disposal areas will be drained
of -water and air dried. These sites will then be capped with a clay cover that
will be contoured to drain rainwater and planted with vegetation. Monitoring
-will be continued for the quantity and characteristics of the leachate and for
strength of the materials throughout the remainder of the project.
7. 0 REFERENCES
1. Jones, J. W. , "Environmentally Acceptable Disposal of Flue Gas
Desulfurization Sludges: The EPA Research and Development
Program, " in Proceedings: Symposium on Flue Gas Desulfurization
Atlanta, NoyembeT 1974, Volume II, EPA -650/ 2 -74- 126-b (NTI5 No.
PB 242-573/AS), Environmental Protection Agency, Research
Triangle Park, N.C. (December 1974).
2. Rossoff, J., and R. C. Rossi, Disposal of By-Products from Non-
Regenerable Flue Gas Desulfurization Systems: Initial Report,
EPA-650/2-74-037-a (NTIS No. PB 237-114/AS), Environmental
Protection Agency, Research Triangle Park, N. C. (May 1974).
3. Rossoff, J. , et al. , "Disposal of By-Products from Non-Regenerable
Flue Gas Desulfurization Systems: A Status Report," in
Proceedings: Symposium on Flue Gas Desulfurization — Atlanta,
November 1974. Volume I. EPA -650/ 2 -74- 1 26-a (NTIS No.
PB 242-572/AS), Environmental Protection Agency, Research
Triangle Park, N.C. (December 1974).
4. Fling, R. B., et al. , Disposal of Flue Gas Gleaning Wastes: EPA
Shawnee Field Evaluation - Initial Report, EPA - 600/ 2 -7 6-070,
Environmental Protection Agency, Research Triangle Park,
N. C. (March 1976).
5. Rossoff, J. , et al. , Disposal of By-Products from Non-Regenerable
Flue Gas Desulfurization Systems : Interim Report, Environmental
Protection Agency, Research Triangle Park, N.C. (1976)
(to be published).
6. Mahloch, J. L. , et al. , Pollution Potential of Raw and Chemically
Fixed Hazardous Industrial Wastes and Elue Gas Desulfurization
Sludges: Interim Report, Environmental Protection Agency,
Research Triangle Park, N.C. (September 1975) (to be published).
7. Lord, W. H. , "FGD Sludge Fixation and Disposal, " in Proceedings:
Symposium on Flue Gas Desulfurization - Atlanta, November 1974,
Volume II, EPA-650/2-74-1 26-b, Environmental Protection Agency,
Research Triangle Park, N.C. (December 1974).
8. Kleiman, G. , "A Practical Approach to Handling Flue Gas Scrubber
Sludge, " Paper presented 37th Annual Meeting of the American
Power Conference, Chicago, April 1975.
625
-------
9. Minnick, L. J. , "Stabilization of Waste Materials Including
Pulverized Coal Ash, " Paper presented Meeting of the American
Institute of Chemical Engineers, Chicago, 8 May 1975.
10. Minnick, L. J. , "Environmental Considerations for Disposal of
Industrial By-Products, " Paper presented Annual Meeting of the
American Institute of Chemical Engineers, New York, 16-20
February 1975.
11. Minnick, L. J. , "Utilization of Fly Ash Sulfate Sludge Based
Synthetic Aggregate for Highway Construction Use, " Paper
presented Coal and Environment Technical Conference and
Equipment Exposition, Louisville, Ky. , 24 October 1974.
12. Conner, J. R. , "Ultimate Liquid Waste Disposal Methods, "
Plant Engineering ( 19 October 1972).
13. Conner, J. R. , "Ultimate Disposal of Liquid Wastes by Chemical
Fixation, " Paper presented 29th Annual Purdue Industrial Waste
Conference, Lafayette, Ind. , 7 May 1974.
14. Gowman, L. D. , "Chemical Stability of Metal Silicates vs
Metal Hydroxides in Ground Water Conditions, " Paper presented
Second National Conference on Complete WateReuse, Chicago,
4 May 1975.
15. Conner, J. R. , "Disposal of Liquid Wastes by Chemical Fixation,"
Waste Age 5 (6), (1974).
16. Conner, J. R. , "Ultimate Disposal of Liquid Residues by
Chemical Fixation, " Paper presented National Conference on
Management and Disposal of Residues from the Treatment of
Industrial Wastewaters, Washington, D. C. , 3-5 February 1975.
METRIC CONVERSION TABLE
EPA policy is to express all measurements in metric units; however,
for clarity, British units have been used in most instances in this paper. A
list of conversion factors for the British units used is as follows:
British Metric
1 acre 4047 m2
1 ft 0. 3048 m
1 g/ftz 10.76 g/m2
1 in. 2. 54 cm
1 lb 453. 6 g
1 mi 1. 609 km
* PPm 1 mg/1 (equiv. )
1 psi 0. 0703 kg/cm2
1 ton (short) 0. 9072 metric tons
1 ton/ft^ 9765 kg/m2
1 ydj 0. 7641 m3
626
-------
CHEMICAL FIXATION OF FGD SLUDGES
PHYSICAL AND CHEMICAL PROPERTIES
Jerome L. Mahlock
Environmental Effects Laboratory
U. S. Army Engineer Waterways Experiment Station
P. 0. Box 631
Vicksburg, Mississippi 39180
ABSTRACT
Five FGD sludges arising from three types of scrubbers and two
sulfur content coals were subjected to physical and chemical analysis.
Testing included the raw sludges and materials produced by chemical
fixation utilizing five processes. Physical analysis included deter-
mination of bulk density, porosity, permeability, and unconfined
compressive strength. Results of the physical tests were inter-
correlated, but highly dependent on the fixation process employed.
Chemical testing consisted of leaching experiments for both the raw
and fixed sludges, conductivity, and sulfate data were presented in
this paper. Little dependence between the results of the chemical
and the physical tests could be detected. Source of the sludge and
initial sulfur content of the coal did have an effect on leaching and
fixation performance. Pollutant migration rates may be compared between
fixed and raw products based on leaching mechanisms proposed in this
paper, and will probably be useful in disposal operation design.
627
-------
CHEMICAL FIXATION
OF
FGD SLUDGES - PHYSICAL AND CHEMICAL PROPERTIES
INTRODUCTION
The removal of sulfur oxides from flue gases to meet air quality standards is
usually accomplished by scrubbing with a fluid which absorbes the sulfur oxides
from stack gases. These scrubbers may be classified by the nature of the absorbant
(or precipitant) employed; lime, limestone, or double alkali. The predominance of
the existing scrubbers are nonregenerable; that is, the absorbant is removed from
the system after a certain contact period. The use of nonregenerable flue gas
desulfurization (FGD) systems by definition results in a residue or sludge for
disposal. The ultimate disposal of these residues is usually on land, and attention
must be directed at the potential environmental impact of such an activity.
Land disposal of sludges or residues may result in an adverse environmental
impact if pollutants migrate from the disposal sites at rates greater than those
compatible with acceptable environmental quality. The principle endpoint of this
migration is usually groundwater or surface water systems. The resultant pollution
of these systems is not compatible with proper disposal procedures; consequently.
procedures must be employed to reduce the rate of pollutant migration. The
techniques commonly employed to accomplish this objective include site selection/
operation, liners, chemical fixation, and a combination of the above.
Site selection is used to assess the suitability of a disposal site prior to
the actual disposal operation. Factors included within this analysis are geology,
hydrology, and soil properties which would favor reduced pollutant migration.
Liners are utilized as a barrier between the disposal site and the environment, and
reduce pollutant migration by preventing flow of leachates from the disposal site
to the surrounding environment. Chemical fixation is a process which alters the
physical and chemical properties of the sludges to retard pollutant migration.
The U. S. Army Engineer Waterways Experiment Station is presently studying
the chemical fixation of hazardous wastes and FGD sludges. The principle objective
of this study is to assess the technical feasibility of chemical fixation and to
specifically evaluate it based on physical and leaching tests. This paper will be
confined to discussing the current results of the testing program for FGD sludges
regarding the above tests. Discussion of the leaching test will be restricted to
those data from conductivity and sulfate analyses. Properties of the sludges and
their fixed products will be related to source of coal, and scrubber type.
628
-------
METHODS
Sources of Sludges
The FGD sludges selected for the chemical fixation testing program are listed
in Table 1. Included within Table 1 are the scrubber type, source of coal, sulfur
content of the coal, and a code number associated with each sludge tested. All
sludges originated from stationary power plants and were collected at the point
prior to disposal at the respective sites.
Sludge Fixation
Fixation of the sludges was performed at the Waterways Experiment Station by
the processors selected for inclusion in this study. The fixation processes
employed will be identified by a code letter and the fixed material will be identi-
fied by a code letter and a number indicating sludge source. The processes were
not universally employed to all FGD sludge types, since the decision to fix a
particular residue was made by the individual processors.
Physical Testing
The physical test results included in this paper are for bulk density, porosity,
permeability, and unconfined compressive strength. With the exception of the
permeability test, all methods conform to those of the American Society for Testing
and Materials (ASTM ). Permeability of the raw sludges was determined by use of
a falling head pereameter and by use of a triaxial testing machine for the fixed
residues.
Leach Testing
Leach testing was performed on the raw sludges and fixed materials. The methods
2 3
for leach testing have been documented in earlier project reports ' . All leach
testing was performed in triplicate and utilized two leaching solutions of differing
pH (4.7 and 7.7). The leachates were analyzed for specific conductance and sulfate
4
by procedures presented in Standard Methods .
RESULTS
The results for the physical testing of raw and fixed FGD sludges are presented
in Table 2. Fixation of the FGD sludges results in a consolidated material which is
generally of higher density, lower porosity and permeability, and demonstrates some
degree of structural strength. These results are generally confirmed by the physical
testing for processes A, E, and F. The results for processes B and G tend towards
629
-------
Table 1 FGD SLUDGE DESCRIPTION
Code Number Coal Source Percent Sulfur* Scrubber Type
Lime
Limestone
Double Alkali
Limestone
Double Alkali
*Based on coal analysis at the time of sludge sample collection.
100
400
500
600
1000
Eastern
Eastern
Eastern
Western
Western
1.67
1.82
2.00
0.60
0.46
630
-------
Table 2 PHYSICAL PROPERTIES OF RAW AND FIXED FGD SLUDGES
Sample ID
R-100
R-400
R-500
R-600
R-1000
A-100
A-400
A-500
A-600
A-1000
B-100
B-400
B-500
B-600
B-1000
E-100
E-400
E-500
E-600
E-1000
F-100
F-600
G-100
G-400
G-500
G-600
G-1000
Bulk Density
(Ib/cf)
51.7
63.1
52.3
89.0
47.4
100.1
108.3
95.5
109.3
96.6
77.0
89.0
90.8
79.6
81.5
101.1
82.7
99.3
110.9
82.7
81.0
62.7
52.5
56.9
68.1
Porosity
67,
64.
74.0
48.5
75.4
51.4
55.6
57.8
41.0
51.4
75.3
75.5
71.6
76.3
73.2
45.7
55.4
49.0
35.4
50.8
52.4
69.7
75.4
73.4
70.0
Permeability
(cm/sec)
1.070 x 10
7.784 x 10
2.505 x 10
1.439 x 10
6.536 x 10
-5
-5
-5
2.06 x 10
1.13 x 10
4.31 x 10
8.95 x 10
1.59 x 10
1.08 x
4.56 x
7.33
10
10
3.96 x 10
6.62 x 10
-7
-7
-7
-4
-5
-5
-5
-5
7.94 x 10
2.52 x 10
4.54 x 10
3.57 x 10
-4
-6
-11
10
-7
5.01 x 10
-6
5.24 x 10
1.39 x 10
1.22 x 10
4.05 x 10
-5
-4
-4
-5
Unconfined
Compressive Strength
(psi)
100.3
188.3
403.1
337.4
23.7
44.5
42.7
35.3
23.2
2570,0
720.0
2220.0
4486.0
1374.0
395.6
242.6
86.4
126.1
144.3
631
-------
the opposite of above, fixation resulting in a more porous material with higher
permeability and lower structural strength. The physical properties appear to be
dependent on a specific fixation process and are consistent with that process.
The results of the leaching tests for conductivity are presented in Figures
1-5, respectively. These data, as well as those for sulfate, will be grouped
according to FGD sludge type. The results presented are for the acidic leaching
solution and represent the mean for triplicate tests. A comparison of leachate
properties as a function of leach solution pH for a hazardous waste has been
presented previously , and demonstrated that leachate properties are roughly
independent of leaching solution pH for the range tested. These results appear to
hold in a likewise manner for the FGD sludges and their fixed products. The
conductivities for the leachates from the fixed material are generally equivalent
to, or greater than, those of the raw sludges. An exception to this appears for
*
sludge 500 in which the raw sludge leachate conductivity is significantly higher
than the fixed materials. The trends for the conductivity data appear to follow a
classical leaching phenomenon, a high initial concentration followed by a gradual
decrease to a steady-state condition. This trend appears to be more evident for
the double alkali scrubber sludges, and is particularly noticeable for residue 1000.
Historically, specific conductance has been used as a measure of dissolved
solids present in a solution. This relationship is dependent on the chemical
composition of any particular solution and should ideally be confirmed for a group
of solutions. The results presented in Figure 6 are paired data for conductivity
and dissolved solids on elutriates from all raw and fixed sludges utilized in this
study. These data demonstrate a strong linear relationship between these parameters.
This result may be used to interpret conductivity data derived from the leaching
experiments by inferring a relationship to dissolved solids. Utilizing the relation-
ship previously defined and the specific chemical analyses, a speciation of chemical
form as well as an appropriate check of analytical methods may be made.
The results for sulfate are grouped in a similar fashion to those of conductivity
and are presented in Figures 7-11, respectively. The leaching data for sulfate
appear to be well correlated with conductivity and follow similar trends. The rapid
decline in leachate sulfate concentration is more apparent for the double alkali
sludges as compared to a relatively constant sulfate concentration for leachates
from the remaining sludges. In the case of the double alkali sludges the leachates
from the fixed materials are generally of superior quality than those from the raw
632
-------
I 00,000
10,000
ON
W
00
100
PROCESS A
PROCESS B
PROCESS E
PROCESS F
PROCESS G
RAW SLUDGE
I 00,000
10,000
1000
80 12O 160
ELAPSED TIME, da-/s
100
LEGEND
A PROCESS A
A PROCESS B
O PROCESS E
• PROCESS G
0 RAW SLUDGE
J I L
80 120 16O
ELAPSED TIME, days
2OO
Figure 1
Leaching results, conductivity,
sludge no. 100-
Figure 2
Leaching results, conductivity
sludge no. 400-
-------
I 50,000
ON
"1 T
I 00,000
i 10,000
1000
100
Figure 3
LEGEND
PROCESS A
PROCESS B
PROCESS E
PROCESS G
RAW SLUDGE
I I I
I 00,000
10,000
100
LEGEND
A PROCESS A
A PROCESS B
O PROCESS E
• PROCESS F
• PROCESS G
D RAW SLUDGE
80 I2O 160
ELAPSED TIME, days
200
240
80 120 160
ELAPSED TIME, days
Leaching results, conduct-ivity ,
sludge no. 500.
Figure A Leaching results, conductivity,
sludge no. 600.
-------
I 50.0CO
100,000 -
1000
100
80 120 160
ELAPSED TIME, days
240
Figure 5 Leaching results, conductivity,
sludge no. 1000.
6500
6OOO
55OO
O 4500
Q-
w
2000
(500
1000
1000
200O 3000 4OOO
DISSOLVED SOLIDS, MG/l
50OO
6OOO
Figure 6 Conductivity vs. dissolved solids
-------
O-I
50,000
10,000
100
LEGEND
PROCESS A
PROCESS B
PROCESS E
PROCESS F
PROCESS G
RAW SLUDGE
1
0 40
Figure 7
200
24C
8C 120 160
ELAPSED TIME, days
Leaching results, sulfate,
sludge no. 100.
50,000
10,000
= V
1,000
lOO
LEGEND
PROCESS A
PROCESS B
PROCESS E
PROCESS G
RAW SLUDGE
40
200
240
Figure 8
80 120 160
ELAPSED TIME, doys
Leaching results, sulfate,
sludge no. 400.
-------
50,000
10,000
ON
CM
A PROCESS A
A PROCESS B
O PROCESS E
• PROCESS G
D RAW SLUDGE
50,000
10,000
1,000
100
LEGEND
A PROCESS A
A PROCESS B
O PROCESS E
• PROCESS F
> PROCESS G
D RAW SLUDGE
200
240
Figure 9
8C 120 160
ELAPSED TIME, days
Leaching results, sulfate,
sludge no. 500.
40
200
240
8C 120 160
ELAPSEC TIME, days
Figure 10 Leaching results, sulfate,
sludge no. 600.
-------
sludges. In contrast, the remaining sludges demonstrate an equivalent quality
between fixed and raw sludges and a significant presence of sulfates in all
leachates.
DISCUSSION OF RESULTS
One principle result of residue fixation is generally an alteration in the
physical properties of the material. These alterations are manifested by a
consolidation of the sludge into an agglomerate mass which demonstrates a certain
structural strength. From a material behavior with respect to leaching, the
consolidation of residue material will be reflected in a transition from a solubility
mechanism to a diffusion mechanism for pollutant migration. These mechanisms will
realistically reflect field conditions if the permeability of the fixed material is
significantly less than that of the surrounding material in which the fixed material
is placed. Initially, both materials will exhibit similar leaching behavior until
surface wash-off from the fixed materials has occurred and the diffusion mechanism
becomes dominant.
The conductivity data for the FGD sludges appear to support the above
observations. Leachates from the raw sludge are generally characterized by a
constant conductivity. Sludge 1000 demonstrated a rapid decline in conductivity
to a stable level. This would indicate a transition between quantities and types
of chemical forms available for solubilization within this material. The fixed
residues demonstrate an initial decline in leachate conductivity with a subsequent
stabilization in quality indicating the predominance of the diffusion mechanism.
It may be generally postulated that there is a relationship between physical
properties of a fixed material and the leachate quality produced. This relation-
ship will hold if there exists a link between physical properties of a fixed
material and the diffusion mechanism which characterizes leaching. This observation
must be qualified for a specific fixation process. Chemical alterations which affect
leaching behavior may not necessarily be reflected in the physical properties of the
material. To investigate this relationship a plot of unconfined compressive
strength versus relative leachate sulfate concentration is presented in Figure 12.
There appears to be no relationship between specimen strength and leachate quality
for the fixed FGD sludges. An increase in structural strength is related to
decreased sulfate concentrations in the leachates but appears to represent a lower
bound for sulfate leaching. It may be concluded from these data that alteration in
physical characteristics of the FGD sludges will increase leachate quality after a
638
-------
50,000
10,000
O\
CXI
IT
E
1000
100
PROCESS A
PROCESS B
PROCESS E
I PROCESS G
D RAW SLUDGE
80 120 I6O
ELAPSED TIME, days
Figure 11 Leaching results, sulfate,
sludge no. 1000.
to
0.
jf
STRENG
o ii
» c
3 C
3 C
PRESSIVE
j n
3 (.
3 C
> C
O
O
a
u
Z
3 C
:> c
T C
dNODNH
4
•
•
(
•
•
e <
4 4
1
• «
I •*
6 4
«
8 5
»
0 5
•
•
«r
2 5
9
4 5
SULFATE LEACHED, RELATIVE
Figure 12 Specimen strength vs.
sulfate leaching.
-------
certain degree of modification is obtained. The remainder of the variation of
sulfate quality for the leachates must be attributed to chemical modification in
fixed materials.
The selection of FGD sludge samples within this effort presents an excellent
opportunity to evaluate fixation performance as a function of sludge category. To
accomplish this analysis, the sulfate data from the limestone and double alkali
scrubbers for eastern and western coals and for processes A, B, E, and G were
analyzed. The procedure used for analysis involved canonical decomposition of the
data followed by graphical presentation of the results, A similar technique has
been used previously for the analysis of water quality data . The results of the
analysis for double alkali scrubbers is presented in Figure 13. The horizontal bars
representing sulfate leaching for the raw and fixed products are proportional to
the 5 percent confidence limits for the data, and distances between bars represent
differences between groups. This presentation is analogous to an analysis of
variance which is based on within and between group variation. The scale is
relative, but comparisons may be made between fixed and raw groups by noting that
displacement to the right indicates increasing sulfate leaching and displacement to
the left, the converse. The analysis is presented by coal origin which is related
to initial sulfur content, Table 1. It is interesting to note that the data for
one scrubber type (e.g., limestone or double alkali) were well correlated between
scrubber type. This result would tend to indicate that leaching of raw and fixed
materials (or sludges) is very dependent on the scrubber source, and this factor
may play a critical role in disposal.
The results of this statistical analysis for double alkali scrubber sludges is
presented in Figure 13. In this case a dramatic improvement in leachate quality may
be noted for both coal origins. This fact is related to the fixation process and
agglomeration of the residue to reduce the availability of the more soluble sodium
sulfate present in these sludges. The performance of all fixation processes appears
to be roughly equivalent and does not demonstrate the variability shown for the
limestone scrubber sludges, Figure 14.
A similar analysis for the limestone scrubbers, Figure 14, demonstrates a
significant effect of coal origin on the performance of fixation. The fixed materials
originating from the eastern coal scrubber show a deterioration in leachate quality
for sulfate, while fixed materials for western coals show a slight improvement.
Examination of the sulfate data, Figures 8 and 10, would confirm this statement.
640
-------
LEACHATE QUALITY
DOUBLE ALKALI SCRUBBER
RAW
FIXED
EASTERN COAL
SULFATE
"CONCENTRATION
RAW
FIXED
WESTERN COAL
__
B
SULFATE
CONCENTRATION
Figure 13 Fixation performance, double alkali scrubber sludges.
LEACHATE QUALITY
LIMESTONE SCRUBBER
RAW
FIXED
EASTERN COAL
RAW
B
SULFATE
CONCENTRATION
RAW
FIXED
WESTERN COAL
B
SULFATE
CONCENTRATION
Figure 14 Fixation performance, limestone scrubber sludges.
641
-------
The sulfate leaching from these samples, Figures 8 and 10, appears to be at the
solubility of calcium sulfate and differences between fixed materials may be a
function of sulfate availability between the two types of sludge.
It may be concluded from this analysis that the leaching and performance of
fixed FGD sludges is strongly dependent on the type of scrubber and, for the
limestone scrubbers, is dependent on the source of coal (initial sulfur content).
The double alkali scrubbers tend to produce more soluble sulfate compounds and thus
their leaching behavior is affected by fixation to a greater degree. The limestone
scrubbers produce less soluble sulfate compounds and fixation of these sludges
affects the leaching of sulfate to a minor degree. Within the limestone scrubber
sludges, the amount of sulfate available appears to be proportional to the initial
sulfur content of the coal utilized. The western coals, which have a lower sulfur
content, produce less available sulfate in the sludge and the fixed materials demon-
strate an improvement in leachate quality. The limestone scrubber sludges from
eastern coals have sufficient available sulfate such that agglomeration (fixation)
of this residue does not affect the leaching behavior of the material.
The interpretation of leaching data based on laboratory experiments must be
qualified to some degree by accounting for conditions which exist at a field disposal
site. The leaching of the raw FGD sludges is primarily a function of the solubility
of the pollutants present within the sludge. The fixed materials, because of their
agglomerate nature, leach primarily via a diffusion mechanism after initial wash-off
of surface bound pollutants has been attained. The performance of the fixed
materials in an actual disposal site is related to the available surface area
subject to leaching and ultimately to the stability of the agglomerate in the field.
An extrapolation of laboratory leaching tests to field conditions has been made for
hazardous wastes , and the data show a significant improvement in leachate quality
after fixation of the sludge.
The stability of a fixed material in a field disposal site is related to its
durability and resisitance to degradation. Destruction of the agglomerate will
increase the effective surface area which is subject to leaching and subsequently
increase pollutant migration rate. To test the effect of these parameters on the
leachability of a field disposal site an analysis was made based on the following
assumptions:
(a) Pollutant migration from a raw sludge disposal site (Volume = 1000 m3,
depth = 3 m) is a function of pollutant solubility and permeability of
the sludge material.
642
-------
3
(b) Pollutant migration from a fixed sludge disposal site (Volume = 1000 m ,
depth = 3 m) is a function of initial pollutant present (assumed to be
105 mg) and the effective diffusivity?,
The results of this analysis are summarized in Figure 15. Pollutant migration from
a raw sludge disposal site is presented in a family of curves covering a permeability
-3 -5
range from 1 x 10 to 1 x 10 cm/sec. Pollutant migration from a fixed sludge
disposal site is presented as a family of curves for effective diffusivities
-9 -11
ranging from 1 x 10 to 1 x 10 cm/sec. The curves are parallel indicating that
equivalent conditions between the two leaching mechanisms will give similar pollutant
migration rates. Given a leaching condition for a raw sludge and the effective
diffusivity for a fixed material, both of which may be experimentally determined,
an upper bound for effective surface area may be established. An increasing
effective surface area is synonomous with agglomerate destruction and can be
estimated from the properties of the fixed material.
The increase of effective surface area by a fixed material does have a bound
at which time the leaching mechanism will become solubility limited but not diffusion
limited. This transition probably would occur when the effective surface area
exceeds 100, The lower bound on effective surface area is theoretically zero, but
a practical limit of 0.1 has been assumed. The foregoing analysis is dependent
on the type of pollutant considered and will vary depending on interactions with
chemcial or biological mechanisms affecting pollutant migration. Determination of
the effective diffusivity may be made on certain specimen sizes and is assumed to
hold over a reasonable range. Chemical effects of fixation may alter the diffusivity
as a function of effective surface area and possibly will introduce an error into
the analysis. The above presentation will allow comparison of fixed and raw sludge
disposal operations, and knowing conditions of the site, may be used to select a
required diffusivity for a fixed material. Alternately- the analysis may be
coupled to soil movement and attenuation factors plus groundwater hydrology to give
an estimate of environmental impact for selected disposal methodology.
643
-------
SOLUBILI TY, MG/Z
10 100
1000
o
UJ
(/>
\
LJ
h-
<
o:
o
h-
OL
(D
\-
z
<
h-
D
O
a.
1 10
EFFECTIVE SURFACE AREA,
Figure 15 Disposal impact, sulfate, raw and
100
CM"1
fixed FGD sludges.
1000
-------
REFERENCES
1. Annual Book of ASTM Standards, Parts 11 and 12, American Society for Testing
and Materials, Philadelphia, PA 1973.
2. Mahloch, J. L., and D. E. Averett, "Pollutant Potential of Raw and Chemically
Fixed Hazardous Industrial Wastes and Flue Gas Desulfurization Sludges,"
Unpublished Interim Report, January, 1975.
3, Landreth, R. E., and J. L. Mahloch, "Stabilization of Hazardous Wastes and
•SOx Sludges," Proceedings of the National Conference on Management and Disposal
of Residues from the Treatment of Industrial Wastewaters, February 3-5, 1975,
Washington, D. C.
4. Standard Methods for the Examination of Water and Wastewater, 13th Edition,
American Public Health Association, Washington, D. C., 1971.
5. Mahloch, J. L., "Leachability and Physical Properties of Chemically Stabilized
Hazardous Industrial Wastes," presented at the Hazardous Waste Research
Symposium: Residual Management/Land Disposal, Tucson, Arizona, February 2-4,
1976,
6. Mahloch, J. L., "Graphical Interpretation of Water Quality Data," Water,.,Air,
and Soil Pollution, 3^: 217-236, 1974.
7. Godbee, H. W., and D. S. Joy, "Assessment of the Loss of Radioactive Isotopes
from Waste Solids to the Environment, Part I: Background and Theory," Oak
Ridge National Laboratory, Report No. ORNL-TM-4333, Oak Ridge, TN, 1974.
645
-------
POTENTIAL UTILIZATION OF CONTROLLED SO EMISSIONS
FROM POWER PLANTS IN EASTERN UNITEDXSTATES
J. I. Bucy, J. L. Nevins, P. A. Corrigan, and A. G. Melicks
Tennessee Valley Authority
Muscle Shoalsj Alabama
ABSTRACT
This EPA sponsored study (EPA-IAG-D4-0500) was initiated to
evaluate the potential marketing of abatement acid and sulfur from
fossil-fuel-fired power plant sources in eastern U. S. A systems
approach was selected to manage the data bases needed to identify
the social cost associated with the alternative choices for reducing
SO emissions in our market-oriented economy. Although conclusions
wifl not be published until data bases and State Implementation Plans
have been further refined and verified, several intermediate results
are worth noting. A possible Frasch sulfur shortage in the next ten
to twenty years was identified. Since sulfur is a vital resource in
the American economy, every effort should be made to recycle it into
productive use.
Total potential abatement production from 2,564 boilers in
eastern U. S. (22.3 million tons of 100% acid) could have been absorbed
in the existing market. However, the by-product market simulation
model derived for this study demonstrated in the first case that at
zero incentive cost, there would be no significant abatement production
at power plants. But as power plants are forced to reduce emissions,
the cost of alternative controls can be a credit to recovery processes
and thereby provide a basis for assigning a market adjustment value.
At the highest level of market adjustment analyzed, $65/ton H-SO.
(23
-------
A second case analyzed the throwaway sludge alternative assuming
that marketing revenue would have to exceed the cost differential
before acid production would be selected. Considering all boilers
(169) with cost less than $175/ton of H SO (61<£/million Btu), scrubbing
in competition with acid produced from $60/ton sulfur, 7.4 million
tons (152 boilers) of abatement acid (96% of total considered) could
be produced in lieu of throwaway sludge.
Conclusions thus far must be tempered with possible clean-fuel
market developments, future sulfur price levels, air pollution regula-
tions, and further verification and refinement of the data bases.
648
-------
POTENTIAL UTILIZATION OF CONTROLLED SOV EMISSIONS
A.
FROM POWER PLANTS IN EASTERN UNITED STATES
INTRODUCTION
A noted economist, Dr. Kenneth Boulding, suggests we move from
a throughput economy to the concept of spaceship earth. He infers that
our economic survival and comfort in the long run may depend on our
ability to recycle our waste products in a way that assures continual
replenishment of the raw material base. This concept agrees with the
fundamental physical law of the conservation of mass. That is, every
ton of material removed from the earth and transferred into goods still
remains to be disposed of when the goods in question are finally con-
sumed. Society has the choice of either disposing of the waste products
in an acceptable manner or recycling the abatement materials into
productive use. This EPA sponsored study attempts to identify the social
cost associated with the alternative choices for managing the SOX emissions
from fossil-fuel-fired power plants in our market-oriented economy.
The first phase of this work was reported at the last Flue -,
Gas Desulfurization Symposium (FGD), Atlanta, Georgia, November 1974.
The study focused on the TVA power system as an example of theoretical
production and distribution of abatement sulfuric acid. Although
hypothetical, it provided insight as to the impact that abatement sulfuric
acid could have on the existing market.
Phase II involved a preliminary market study of the potential
use of calcium-sulfate sludge by the wallboard fabrication industry.
This too was reported at the last FGD symposium.^
The current phase of this work, identified as Phase III,
expands the first phase to include elemental sulfur as well as sulfuric
acid as a potential byproduct from the FGD processes installed in power
plants located in the states served by the inland waterway system in
eastern U.S. The objectives were outlined to provide general and
practical information concerning the current production, distribution,
and use of sulfur and sulfuric acid in the U.S. The computer program-
ming model of Phase I was expanded to cover the power plant data base
and programmed to reflect pollution restrictions established by State
Implementation Plans (SIP). The model was designed for multiproduct
capability and ease of modification to accommodate the expanded data
bases.
As the data bases were acquired, verified, adapted, and
cataloged, it became apparent that available information was adequate
to assess all power plant and acid plant candidates in the U.S. Also,
the recent influence of smelters and sour gas sulfur sources on the
649
-------
Frasch sulfur market is reaching a point (25%) that cannot be ignored.
The need for increasing the accuracy of the cost screen for FGD systems
by use of a retrofit difficulty factor also became apparent. Complexity
of the model was so great that additional refinement was also desirable.
Due to these considerations, it was decided that an interim report
describing progress to date, design of model, and possible uses would
be in order, and a final report would be prepared later covering full
assessment of markets for abatement sulfur and acid from all sources in
the 48 contiguous states of the U.S.
Phase III Interim Repor.t
As the work progressed on this project, unique research
opportunities became apparent. The 1972 Federal Power Commission (FPC)
Form 67 data file contained power plant operating variables which
matched the detailed cost estimates of the five leading FGD processes
prepared by TVA in January 1975 (EPA 600/2-75-006).4 Availability of
these data presented the opportunity to develop a cost screen designed
to focus on the most promising power plant boiler candidates for abate-
ment byproduct production. It was also learned that the 1973 FPC
Form 67 data, which would not be available until late 1975 or early
1976, would contain 5- and 10-yr projections of proposed new power
plant installations. SIP standards for air pollution control effective
in March 1975 identified allowable emissions for each power plant in
the U.S. A Rail Transportation Rate Generation model being developed by
TVA had reached the point that with concerted effort, accurate trans-
portation rates for elemental sulfur and sulfuric acid could be
generated in this system for all origins and destinations in the rail
rate territories located east of the transcontinental territory. To
take advantage of these opportunities, the project was expanded to
provide a more meaningful research effort.
A systems approach was selected to combine data inputs needed
to expand the scope of study, refine the data bases, impose emission
restrictions, and accurately assess the nationwide market potential for
abatement byproducts. Five major data bases -- for (1) sulfuric acid
producers, (2) transportation-distribution options, (3) steam plant
boilers, (4) SIP data, and (5) TVA cost estimate data -- feed a market
simulation model through three cost generation modules. Estimated long-
run competitive equilibrium solutions based on realistic outputs of
abatement byproducts are achieved.
A flow diagram of the major system design requirements is
shown in Figure 1. As part of the model design, it is assumed that the
sulfuric acid market can be simulated as though all consumption occurred
at sulfuric acid plants, and that acid-producing firms would close
these plants and buy abatement acid if acid prices were below their
650
-------
01
TVA worldwide
fertilizer and
related products
data base
Sulfuric acid
and
elemental sulfur
data base
Production
cost
generator
Transportation
cost
generator
Market
simulation
model
I
Case study results
for decisionmaking
Power plant
boiler design
data base
I
Scrubbing
cost generator
Figure 1. Flow diagram for major system design requirements
-------
expected long-run average total cost* It is also assumed that steam
plants would produce sulfuric acid or sulfur if that were the least
costly (after credit for sales income) alternative for meeting clean
air requirements. Given these conditions, it is believed that long-
run competitive equilibrium market conditions can be simulated by
minimizing the cost to both the sulfuric acid and power industries. In
this framework, power plants are assumed free to produce or not produce
sulfur or sulfuric acid and to sell to any sulfuric acid plant in
competition with other power plants. Likewise, the sulfuric acid
plants are assumed free to continue buying sulfur from traditional
sources or from power plants, and free to buy acid in lieu of production
subject to competition in their respective industries. Product dif-
ferentiation is not assumed significant. Problems of stable, guaranteed
abatement supplies are ignored, but probably are solvable. Rail trans-
portation cost is assumed adequate for simulating competitive market
conditions, though barge-truck strategies could be considered in some
cases as more refined data analyses are made.
Model Complexity and Scope
As a methodological consideration, the computer model designed
is directed toward identifying major candidates for abatement byproduct
production and consumption.
From a data bank maintained at the TVA National Fertilizer
Development Center, design and operating inputs were cataloged in the
computer for existing U.S. contact process sulfuric acid plants.
Capital and operating costs for mining sulfur by the Frasch process,
marketing terminal storage, manufacturing acid by the contact process
with storage, and controlling acid plant tail gas emissions were calcu-
lated to determine the competitive costs of sulfur and acid production.
An extremely important component of the model is the trans-
portation data base for computing accurate shipping costs for sulfur and
acid. Since shipping cost is an essential element in the price of acid
and sulfur, a great deal of creative and complex effort was required
to derive usable values for the entire U.S»
By the use of power plant design and operating data provided
in the FPC Form 67, a data bank was constructed to accommodate key
operation parameters for all power plants in the U.S. For this effort,
only boilers burning coal or oil are of interest. Parameters such as
fuel type, sulfur in fuel, boiler heat rate, fuel consumed, on-stream
time, age of plant, etc., are vital. Using these data, possible out-
put of byproducts can be calculated for each power plant -- given the
level of S02 control designated by March 1975 SIP standards.
652
-------
THE EXISTING FRASCH SULFUR INDUSTRY
Most of the sulfur consumed in the U.S. is used to produce
sulfuric acid which in turn is mostly used in the manufacture of
fertilizers. Sulfur is also used to produce such items as carbon di-
sulfide, pulp and paper, insecticides, specialty steels, and leather
goods. A breakdown of the estimated domestic consumption of sulfur in
1975 is given in Table 1.
Table 1. CONSUMPTION OF SULFUR IN THE U.S.
1975 (estimated)
Amount
Use (OOP's long tons) % of total
Carbon disulfide 236 2.1
Pulp and paper 541 4.9
Rubber 39 0.4
Sulfur dioxide 59 0.5
Agriculture 89 0.8
Phosphorus pentasulfide 44 0.4
Other 205 2.6
Total nonacid 1,303 11.7
Acid 9,839 88.3
Total 11,142 100.0
Note; For metric system conversion factors see Appendix A.
Sulfur enters into the production of many products in varying
amounts, ranging from 18.090 tons of sulfur or sulfur equivalent per ton
of Uranium 235 to 0.0003 ton per ton of phenol-formaldehyde plastic
molding compound. The amount of sulfur consumed per ton of manufactured
product is shown in Table 2.
Frasch sulfur production is a mining operation. Wells are
sunk into a sulfur-bearing stratum, sulfur is melted by hot water in-
jected into the stratum, and the molten sulfur is pumped out. The
molten sulfur is pumped from the well to either heated tanks for storage
as a liquid or to vats where it cools and solidifies.
Domestic sulfur production in 1972 was 10.2 million long tons
mostly from mines in Texas and Louisiana. This was 247o of the estimated
653
-------
Table 2. TONS OF ELEMENTAL SULFUR OR EQUIVALENT ELEMENTAL SULFUR
REQUIRED IN MANUFACTURE OF ONE TON OF INDICATED PRODUCT
Short
tons of sulfur
or equivalent
sulfur per ton
Product
of product
Fertilizers
Diammonium phosphate (DAP) 18-46-0 grade
Granular triple superphosphate (GTSP) 0-46-0 grade
P205 in 54% P205 wet phosphoric acid
Wet phosphoric acid (54% P2°5)
Granulated ammonium polyphosphate (GAPP) 12-57-0
Normal superphosphate (NSP) 0-20-0 grade
Liquid fertilizer 11-37-0 grade
Sulfuric acid, 100%
Synthetic fiber intermediates
Hydrogen cyanide (Modacrylic fiber)
Caprolactan (nylon 6 fiber)
Acetate rayon (fibers, photographic film, etc.)
Synthetic rubber (SBR)
Vulcanized synthetic rubber (SBR)
Carbon disulfide (fibers, cellophane, other chemicals)
Paper pulp
Indigo dye
Pheno- formaldehyde plastic moulding compound
Phenol by sulfonation (plastics)
Explosives
Nitrocellulose
Black powder
Nitroglycerine
Lithopone paint pigment
Leather tanning
Vegetable tan
Chrome tan
Bordeau mixture (4-4-50) (fungicide)
Treflan (100%) (herbicide)
Alum, 17% Al203 (water treatment chemical)
Sodium dichromate (tanning, dyeing, paint pigments, etc)
Uranium 235
Sodium sulfate (100%)
Ammonium sulfate (100%)
°'311b
0.9^3
0.509b
0.538
121
0.646b
0.338°
0
0.081
1.019
0.034
0.005
0.012
0.936
0.109
0.297
0.0003
0.441
0.169
0.100
0.014
0.105
0.007
0.076
0.002
0.420
0.150
0.142
18.090
0.226s
0.243£
Equivalent
tons
of H2S04
1.355
0.951
2.885
1.557
1.646
0.370
1.976
1.000
0.248
3.117
0.104
0.015
0.037
2.863
0.333
0.909
0.001
1.349
0.517
0.306
0.043
0.321
0.021
0.232
0.006
1.285
0.459
0.434
55.341
0.691
0.743
values are from Shreve, R.N. Chemical Process Industries 3rd Edition,
McGraw-Hill Book Company, New York, 1967, unless otherwise noted.
bUnpublished TVA data.
cAverage of several published values.
dAnon. Chemical Week, June 26, 1974, p. 41.
eAssuming direct neutralization of sulfuric acid with sodium hydroxide
with no losses.
^Assuming direct neutralization of sulfuric acid with ammonium hydroxide
with no losses.
654
-------
world production of 43<,0 million tons. The same year Frasch sulfur
accounted for 72% of all domestic sulfur production. Recovered sulfur
accounted for 197<> of domestic production, byproduct sulfuric acid from
smelters, 5%, and other sources, mainly pyrites, accounted for the
balance, 4%,. About 747> of the domestic Frasch sulfur production was for
domestic consumption, 22% for export, and 470 for increases and adjust-
ments in producers stocks.
In 1972 the Frasch sulfur supply (7.29 million long tons)
was produced by five comparies operating 13 individual mines. The three
major companies were Duval Corporation, Freeport Minerals Company, and
Texas Gulf Incorporated. Operating 10 mines, they accounted for more
than 907o of the Frasch production and more than 657> of domestic
production of sulfur in all forms. The five largest mines owned by
these companies produced 737o of the F.rasch sulfur and 5270 of the total
domestic sulfur supply in all forms.
Field surveys conducted during this study revealed a unique
characteristic of the Frasch sulfur mining process: The development of
a sulfur dome can be compared to the punching of pins in a pin cushion.
Each well punched into a sulfur dome formation has an expected life of
1-2 yr. At any time the sulfur mine can have several wells operating
in parallel. The number of wells depends primarily on the short-run
market demand. As the mining process for a given dome reaches the
mature stage, operating costs increase at an increasing rate due to
the increased water and energy requirements. This incremental increase
causes the supply price of sulfur to rise to a point where it is no
longer economical to continue mining the sulfur dome.
Sulfur can be marketed either in the liquid or dry solidified
form. It is estimated that more than 95% of all domestic sulfur
deliveries to acid plants are in liquid form. March 1975 list prices
f.o.b. Port Sulphur, Louisiana, are about $60/long ton.
Most molten sulfur is shipped by water from the mines or
transshipment terminals on the Texas and Louisiana Gulf Coast to the
marketing terminals. The basing point for the Gulf Coast market is
Port Sulphur. Marketing terminals are strategically located either on
the inland waterway system or along the east coast adjacent to ports
served by deep-water vessels. From the marketing terminal, molten
sulfur is transported by barge, truck, or rail directly to the point of
consumption. March 1975 transport rates were used in this study."
The delivered cost of sulfur to each acid plant in the model
is based on $60/long ton for sulfur f.o.b. Port Sulphur plus trans-
portation cost. The sulfur transportation costs used are based on water
transportation from Port Sulphur through marketing terminals by either
truck or rail to the acid plant, whichever is lower cost. Average
655
-------
sulfur storage costs were established at $1.35/long ton for the program
model used to calculate the production cost of H2SC>4 in this study. The
least cost mode for transporting sulfur was extended to cover a relative
comparison of rail transportation from Port Sulphur versus water trans-
port - terminal throughput - truck or rail to the acid plant. This
least cost mode was selected in all cases for sulfur delivered to a
given acid plant.
The Long-Run Sulfur Supply in the U.S.
G.H.K. Pearse recently completed a study on the long-run
sulfur supply for North America.10 He projects the long-run sulfur
supply price gradually increasing from $10 to $40/ton f.o.b. mine in
1970 dollars (about $50 in 1974). The resulting sulfur supply price
increase will encourage the relatively inefficient Frasch and native
sulfur mines to become producers over time. He projects that the
current sulfur reserves in the U.S. from conventional sources (Frasch
process, native ore, petroleum, natural gas, sulfide ores, and pyrite)
in the amount of 290 million tons will be mined out at current pro-
duction rates by 2000 A.D. During the same period it is estimated that
110 million tons of sulfur will become available from oil shales and
coal gas, giving a total cumulative production of 400 million tons.
The cumulative demand in the U.S. is estimated to be 550 million tons
by 2000 A.D., leaving a deficit of 150 million tons of sulfur (513.95
million tons of H2SOA).
Based on data in Table 2, about 0.943 ton of sulfur is re-
quired as sulfuric acid per ton ^2^5 produced as wet-process phosphoric
acid. The U.S. is producing 6.248 million tons of ^2®$ (1974) as wet-
process acid. The TVA projection for 1980 U.S. production is 10.016
million tons of P2°5' Assuming that a median value of about 9.8
million Btu as natural gas (9800 ft^) are required for mining a short
ton of sulfur then it can be estimated that about 7 million Btu are re-
quired per short ton of ^2^5 to Pr°duce wet-process phosphoric acid
from Frasch sulfur. Recovering sulfur from fuels would therefore not
only conserve our natural sulfur reserves but would also reduce the
energy requirement needed for mining.
The Impact of Abatement Sulfur on the Sulfur Market
The Frasch sulfur industry is dominated by a small number of
large producers. The industry is concentrated and production is
centralized in a few mines located within a small region of the Texas
and Louisiana Gulf Coast. The marketing system is highly organized and
specialized. It has the flexibility of marketing either elemental
sulfur or sulfur in the form of sulfuric acid. It is recognized that
656
-------
expanded rates of extraction of sulfur reserves will come about only with
increasing supply costs. The sulfur producers are in an excellent
position to balance the rate of development of the reserves in the
Frasch mining industry with the production and marketing of abatement
sulfur byproducts. In the long run, sulfur reserves could be used to
smooth out short-run fluctuations in supply and demand.
Assuming the sulfur industry finds it profitable to absorb the
abatement products into its marketing system, then little disorder or
disequilibrium in the market would result. This is especially true
since the existing industry already serves a majority of the sulfuric
acid manufacturers in industrial chemicals as well as the fertilizer
industry. In this market, sulfur is traditionally sold on a contract
basis directly to industrial users where it is used predominantly in
the form of sulfuric acid. Current reserves provide reliability of
supply essential to the contract consumer.
If the sulfur producer does not choose to develop the market
for abatement sulfur byproducts, then an alternate strategy might be
the establishment of a formal commodity exchange. This would involve
considerable reorganization but would assure utilization of abatement
sulfur byproducts in a competitive market.
THE EXISTING SULFURIC ACID INDUSTRY
Sulfuric acid is produced by burning sulfur or sulfur-bearing
materials to form SC^. Sources of sulfur or S0£ for the manufacture of
sulfuric acid include (1) elemental sulfur, (2) pyrites, (3) gypsum,
(4) petroleum products, (5) smelter off-gases, and (6) waste gases from
burning fossil fuels. In 1973 a brimstone-based acid accounted for
82.5% of the total sulfuric acid production, followed by pyrite-based
acid at 3.17o and all other raw material sources at 14.4%. Thus far, this
study focuses on acid plants burning elemental sulfur.
The estimated end uses of sulfuric acid for 1975 are presented
in Table 3. These data indicate that the manufacture of phosphate
fertilizers requires more than half of the sulfuric acid used in the
U.S.
657
-------
Table 3. SULFURIC ACID-CONSUMING INDUSTRIES IN THE U.S.
(000's of net tons, 100% H2S04)
1975
Consuming industries Ton %_
Fertilizer industries
Phosphatic fertilizers 16,900 51
Ammonium sulfate or other 2,280 1
19,180 58
Other industries
Chemicals 3,830 12
Petroleum refining 1,880 6
Iron and steel 530 1
Other metals 1,300 4
Paints and pigments 1,160 3
Rayon and cellulose film 860 3
Miscellaneous 4,260 13
Total 33,000 100
U.S. production of sulfuric acid in 1973 was 31.7 million tons
(1007, HoSO') which is about 807> of the total production capacity in the
U.S. (47 million tons). About 59% of production capacity was committed
to captive use. Only about 12.9 million tons were externally marketed
from the 1973 production. Although there has been little growth in
new sulfuric acid production in the period 1966-1973, there has been a
shift in the geographical location of acid production. During this
period, production of new acid in the north central states declined from
4.5 to 3.7 million tons, but production of new acid in Florida and
Louisiana increased from 8.8 to 12.3 million tons. This shift in lo-
cation is due to the recent increase in phosphate fertilizer production
in the Gulf Coast area, especially in Florida and Louisiana.
The long-term growth in acid consumption in the U.S. has been
about 4-6%/yr and has been closely tied to the fertilizer growth pattern.
In recent years, however, the world market for sulfuric acid has increased
rapidly -- and further growth in international trade has been predicted. ^
The trend started in 1965 and by 1973 world trade had increased fivefold
-- to 2.5 million tons. Acid is fast becoming as important as pyrites
(sulfide ores) in terms of the overall world trade in sulfur-containing
materials. The trend is away from trade in pyrites toward trade in
pyrites-based sulfuric acid. New economies realized, in transport of
sulfuric acid by sea have been a key factor in the recent evolution of
the world acid trade pattern.
658
-------
Existing Contact Acid Plants in Eastern U.S.
TVA's computerized file of worldwide manufacturers of ferti-
lizer and related products contains a list of 209 U.S. sulfuric acid
plants producing acid by burning elemental sulfur. Total annual pro-
duction capacity exceeds 47 million tons. Only those acid plants
located east of the Rocky Mountains were considered as a potential mar-
ket for abatement acid. This reduced the list to 166 plants with an
annual production capacity of 42.9 million tons. The geographic
distribution of these plants by states is outlined in Figure 2 and in
Table 4.
Table 4. EASTERN U.S. SULFURIC ACID PLANT CAPACITY (1975)
(000's Short tons -- 330 Days per annum)
State Number
Alabama
Arkansas
Colorado
Delaware
Florida
Georgia
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
of Plants
5
4
1
1
22
10
15
3
3
3
2
11
1
3
1
4
Annual
Capacity
455
584
40
360
16,552
768
2,176
713
651
265
218
6,206
75
465
140
168
State Number
Minnesota
Mississippi
Missouri
New Jersey
New York
North Carolina
Ohio
Oklahoma
Pennsylvania
Rhode Island
South Carolina
Tennessee
Texas
Virginia
West Virginia
Wisconsin
Wyoming
of Plants
1
3
3
10
1
8
8
1
4
1
6
5
13
9
1
1
2
Annual
Capacity
107
1,250
581
2,300
6
1,885
844
90
437
20
134
578
3,876
733
135
17
100
Total plants 166
Total
annual capacity 42,929
The Impact of Abatement Acid on the Sulfuric Acid Industry
Sulfuric acid plants are widely scattered throughout the U.S.
chiefly because of the low bulk value of the acid, difficulties of
handling the acid in the bulk, and subsequent high cost of shipment as
compared to handling elemental sulfur. Therefore, acid has been
659
-------
I*] Port Sulfur
® Sulfur Terminal
A Sulfuric Acid Plant
ON
ON
O
Figure 2. Geographic distribution of sulfuric acid plants and marketing terminals (1975),
-------
traditionally produced by sulfur-burning plants near the point of con-
sumption in captive use. However, many existing plants are old and will
soon need replacing. Some will shut down in 1975 because compliance with
pollution control regulations may not be economical. Producers in this
situation will like the opportunity to buy abatement acid in lieu of
building a new acid plant.
Many power plants enjoy a unique location advantage for sup-
plying abatement acid in the existing market. This is especially true
for acid plants located in areas remote from traditional sulfur supplies.
Therefore, the most orderly way to incorporate the abatement acid would
be to replace first the capacity of sulfur-burning sulfuric acid plants
remotely located from sulfur sources. In this instance the relatively
high-cost producer is given the opportunity to close his plant and buy
abatement acid at a saving. The more efficient and better located
plants continue to produce in status quo.
Production Costs for Sulfuric Acid
To put realism in such a marketing approach, some estimate is
needed of the costs that could be saved or avoided by existing acid
producers. Such costs are delineated below:
Raw material Sulfur
Utilities Electric power
Cooling water
Processed water
Boiler-feed water
Operating expenses...Labor
Supervision
Capital costs ...Amortized costs for maintenance of
existing facilities and amortized
costs of new capital investment at
end of useful plant life
In the Phase I study, a computer algorithm was developed using
these inputs and others to calculate contact sulfuric acid production
costs. Details of this program are given in the report on Phase I. The
unique concept incorporated into this model relates to the method used
for handling existing plants as compared to the traditional static
analysis used to justify the investment in a new plant. For existing
plants, the initial capital expenditures are handled as a "sunken in-
661
-------
vestment" and, therefore, do not enter directly into the firm's decision
to discontinue present production and buy abatement sulfuric acid. Only
avoidable costs are considered in making this decision.
Annual costs are calculated in perpetuity using the discounted
cash flow analysis method. The outlay streams are then amortized or
averaged over all years in the firm's planning horizon. The cost streams
are composed of (1) constant annual expenditures for sulfur, utilities,
and operating expenses, (2) periodic expenditures for new plants, and
(3) maintenance of existing facilities which is assumed to grow at a
compound rate. The impact of inflation is not included in the analysis.
These cost streams for a new plant are presented in Figure 3.
The optimum useful life is identified as the minimum point on
the average total cost curve. At this point the added capital cost
savings enjoyed by increasing useful life by 1 yr equals the added main-
tenance saving from shortening useful life by 1 yr. The average capital
charge of 19.3%, identified in Figure 3, covers a range of from 23 to
36 yr. Possibly random effects, such as abrupt physical, economic,
technological, or environmental changes, play the dominant role during
this period with regard to the timing of plant replacement or shutdown.
Sulfuric acid plants built prior to 1960 were assumed to
average 95.5% conversion of sulfur to acid. Plants built between 1960
and 1975 are assumed more efficient with 97% conversion. These effi-
ciencies, however, are not representative of plants that must operate
after 1975 since emission limitations will require an efficiency of at
least 99.7%. As for the potential growth market, it will be necessary
to consider tail gas cleanup at a 99.7% efficiency level for new plants
(double absorption). Capital and operating cost estimates for acid
plant tail gas cleaning systems for existing plants have been completed
but are not yet incorporated accurately in the model. These costs, of
course, must be added to the avoidable costs of existing plants.
The major variables used in the acid plant cost generation
program are outlined in Table 5 which contains sample parameters
associated with each variable. The results of this program are plotted
as a demand curve for abatement acid in Figure 4. The avoidable costs
(theoretical) are calculated at each respective acid plant location
considered.
662
-------
O\
CO
O
o
£
Q_
O
cr
UJ
a.
\OPTIMAL
USEFUL LIFE
0
0
30 40
USEFUL LIFE (YEARS)
Figure 3. Amortized value of maintenance and capital outlays for new plants
(assuming 11% interest and 5% compound maintenance).
-------
Table 5. MAJOR PARAMETERS IN MODEL
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Description of variable
Tons of sulfur per ton H£S04 (before YEAR60)
Tons of sulfur per ton H2S04 (after YEAR60)
Year of technology change
Sulfur ic acid plant investment ($/ton-yr)
Capacity for this plant (000 's tons/yr)
Scale factor for determining investment for
other sized plants
Fixed conversion cost per ton ($/ton)
Fixed annual conversion cost ($/yr)
Taxes and insurance rate
Time preference rate for money
Compound maintenance rate
Economic useful life
Percent £[2804 concentration
Port Sulphur price ($ /short ton)
Steam plant H2S04 price ($/ton 1^504)
Proportion of 330 TPD capacity estimate
Number of steam plants
Number of acid plants
Number of years considered
Years considered
Example
value
.3053
.3006
60.
27.285
247.5
.734054
.47
116.620
.015
.11
.05
34.
98.
53.57
0.
1.
610.
166.
1.
75.
PRE60
POST60
YEAR60
EXPENDO
SIZED
FACTOR
AVC
FC
TIR
RATE I
RATEM
USELIFE
ACDCON
PS
PA
DEMAND
N
M
NYEARS
YEAR(I)
The Demand Curve for Abatement Acid
The March 1975 price for sulfuric acid (100% H^SO, f.o.b.) was
$44.95/ton.° This represents a 37% increase since April 1974 ($32.75).
Cost estimates used for each specific acid plant are depicted in
Figure 4. Costs of manufacture range from $28 to $45 depending upon
plant location, size, and age.
For illustrative purposes it is useful to consider the demand
curve for abatement acid that would be traced out by a uniform f.o.b.
acid plant supply cost if acid supply location in Figure 2 were ignored.
This demand curve is estimated by ranking all acid plants from highest
to lowest cost and accumulating demand quantities as shown in Figure 4.
At a very high supply cost from power plant producers, only a few plants
would be better off buying rather than producing sulfuric acid. These
plants tend to be old, low-volume producers far from sulfur supplies.
664
-------
SULFUR1C ACID DEMAND CURUE
o\
Ox
01
P
R
I
C
T
0
N
100
48
20
I I
I I
Q 10000 20000
ANNUAL CAPACITY (1000 TONS)
30000
DEMAND
40000
50008
Figure 4. Abatement byproduct sulfuric acid demand curve.
-------
As supply cost declines, very high-cost producers disappear
rapidly and the demand curve for abatement acid flattens. At low
supply costs from power plants, all but the largest, most modern acid
plants located near sulfur supplies would buy rather than continue
production. Small quantities of abatement acid could then be marketed
at high value but as the supply increases the value declines.
POTENTIAL ABATEMENT PRODUCTION OF SULFUR AND SULFURIC ACID FROM POWER PLANTS
In assessing potential abatement byproduct production, only
emissions from power plants have been evaluated. Figure 5 shows the
locations of major coal- and oil-fired power plants. Important design
and operating characteristics of these plants are available from FPC
through published reports.13,14 These data are compiled each year
from FPC Form 67 which the utilities file with FPC. With these data it
is possible to project the amount of sulfur byproducts that could have
been produced each year in the U.S. The latest data available from FPC
are for 1972; however,they can be updated annually. FPC has asked for
5- (1978) and 10-yr (1983) projections with the 1973 data items. TVA
has contracted with FPC to obtain these projections for use in this study.
Federal-State Air Quality Standards
Not all 2,564 power plant boilers in the U.S. require more air
pollution control measures. Under the Clean Air Act Amendments of 1970,
EPA has set both primary and secondary ambient air quality standards to
be met by mid-1975. There are also Federal emission standards for new
sources in operation after 1975. To meet these standards, the Federal
government has asked each state to propose its own implementation plans
for the 247 Air Quality Control Regions (AQCR). Each SIP would be re-
viewed by EPA, then approved or returned to the state with suggested re-
visions.
A list of the states and their strategies is given in Table 6.
Some states have more than one strategy listed; this occurs when special
situations such as a concentration of power plants and industries in one
AQCR make more than one strategy desirable.
A review of the status of SIP showed that the states have proposed
several methods for meeting the Federal ambient air quality standards.
These have been classified into 10 groups outlined in Table 7.
Only a few emission standards for SC>2 from fossil-fueled steam-
electric generating plants have been agreed upon by both the Federal and
state governments. Other SIP are being argued in the courts or are
being negotiated by the states and the Federal government. Some states
have not issued SIP.
666
-------
LEGEND
TOWER GENERATION SIZE - MEGAWATTS
oooooOOO
O- SOi- iGOi- i50i- 200(- 250i- 3001- 3501 - 1OOI-
500 lOOO 1500 JOOO 250O 3OOO 35OC 4OOO 5COO
5001- 6OOI- 7001 - BOO i - 9001- lO.OOi-
6000 7OOO BOOO 90OO 10,000 15, OOO
Figure 5. Location of major coal and oil fired power plants (1971).
-------
Table 6. INDEX OF STRATEGIES FOR CONTROL OF S02 EMISSIONS BY STATES
State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Strategy
group number
3
4
8
2 or 7
9
1 or 3
District of Columbia 2
Delaware
Florida
Georgia
Hawaii
Idaho
Illinois
^"Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
2 or 8
4
2
1
2
4
10
4
9 or 5
4
7
1
2
6
2
2
3
1 or 3
2
State
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
-Ohio
Oklahoma
Oregon
Pennsylvania
Puerto Rico
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Strategy**
group number
3
1,2,3, or 5
2 or 6
2 or 4
3
2 or 4
3
3
10
8 or 4
7 or 2
2 or 3
1
5 or 3
3
3
3 or 4
8
2
1
3 or 7
3 or 7
3
8 or 4
8
See Table 7 for description of strategies
*Note SIP not enforceable (because of court action or other reason in
March 1975).
Table 7. STRATEGIES FOR STATE IMPLEMENTATION PLANS
Type of state strategy proposed
1 7° S for all fuels combined
2 % S for an individual fuel type
3 lb S02/miHi°n Btu for all fuels combined
4 lb SOo/raiHi-011 Btu for an individual fuel type
5 lb S/mi11ion Btu for all fuels combined
6 lb S/mi11 ion Btu for an individual fuel type
7 ppm S0£ emission regulation
8 ppm S02 ambient air quality standard
9 no SIP
10 SIP not enforceable (because of court action
or other reason)
668
-------
SIP regulations can be determined accurately; however, this con-
straint is dynamic over time. Regulations can change or be in litiga-
tion. (Indiana and Ohio were in litigation in March 1975 and were
excluded from this study.)
By applying SIP, a reasonable assessment of the upper limit for
abatement byproduct production from power plants can be made; however,
another more stringent screen is needed to define rational quantities.
Selecting Most Promising Candidate Boilers for Abatement Production
With the large data base available from FPC Form 67, it should be
possible to create a more accurate screen to center on the most promising
candidates for abatement byproduct production. Earlier work by TVA for
EPA in preparing detailed cost estimates for the five leading FGD
processes^ offers the best potential as a basis for development of a
final screen. Completed in 1975, the referenced report contains well-
defined investment and operating cost estimates of the five leading FGD
systems based on key power plant design and operating parameters. The
five processes are listed as follows:
Limestone slurry scrubbing
Lime slurry scrubbing
Magnesia slurry scrubbing -
regeneration of H9SO/
Sodium solution scrubbing -
S02 regeneration and reduction to sulfur
Catalytic oxidation
In the EPA-TVA cost report, two of the processes (the magnesia slurry
scrubbing - regeneration process producing 98% sulfuric acid and the sodium
solution - SOo reduction process producing elemental sulfur) produce sale-
able byproducts covered in this study. The costs of these processes can
be compared with those of the throwaway limestone slurry scrubbing process
in deciding what scrubbing strategy is to be adopted for each boiler or
plant where appropriate. The detailed investment and operating cost
projections given and the method illustrated for scaling costs are
suitable for projection of costs of the processes at other capacities
based on comparable FPC operating characteristic data.
669
-------
The three scrubbing processes selected above for analysis in this
paper will be referred to as (1) limestone slurry scrubbing, (2) magnesia
(MgO) slurry scrubbing, and (3) sodium solution scrubbing.
To serve as an economic screen, power unit operating character-
istic data from the FPC file and cost data for the limestone, magnesia,
and sodium scrubbing processes were combined in a computer subprogram to
project the investment requirements and operating costs on a boiler-by-
boiler basis. The following data within the FPC file were used in this
procedure:
1. Power unit size, MW
2. Boiler startup, yr
3. Power unit heat rate, Btu/kWh
4. Design coal feed rate, tons/hr
5. Heat value of coal, Btu/lb
6. Sulfur content of coal, Ibs S/lb coal
o
7. Annual coal consumption, 10 tons/yr
8. Design oil feed rate, bbls/hr
9. Heat value of oil, Btu/gal
10. Sulfur content of oil, Ibs S/lb oil
3
11. Annual oil consumption, 10 bbls/yr
12. Total air rate to boiler, scfm
13. SO.-, recovery unit startup, yr
14. Power unit status, new or old
The program uses an operating capacity factor calculated from
FPC data for the 1972 calendar year.1^ Depreciation and capital charges
included in the annual operating cost projections are based on a total
depreciable power plant life of 30 yr as recommended by FPC. The S09
removal facilities for existing plants are, therefore, depreciated over
a period equal to the remaining life of the power unit. For plants more
670
-------
than 30 yr old, whose remaining life could be calculated to be negative,
depreciation and capital charges are based on a 1-yr depreciation period.
Waste disposal pond size and costs for the limestone slurry process are
scaled proportional to the remaining life of the power unit.
In projecting the investment and annual operating cost of
scrubbing systems on each power unit for each run of the program, the
results show the boilers for which S02 removal is the most economical in
order of increasing unit operating costs (either $/ton 1007, H2SC>4 or $/ton
elemental sulfur), and the relative difference in unit operating costs
for pollution abatement by the three processes, i.e., (1) H/?S04 pro-
duction, (2) elemental sulfur production, or (3) limestone throwaway.
Also, by incorporating the unit operating cost for each boiler in the
production-distribution model, a more realistic market potential for
each abatement product can be defined. One weakness of this screen yet
to be overcome is the need for a retrofit difficulty factor covering
each site specific application. Some power units are more difficult to
fit with scrubbers than others. The screen still enables a decision of
the best scrubbing option to be made on a boiler-by-boiler basis. But
each boiler's relative competitive position with others cannot yet be
made accurately.
Potential Abatement Production on Boiler Basis
Based on SIP regulations outlined above, total emissions can
be calculated for all plant boilers in the model exceeding the allowable
emissions. This amounts to 154 plants (610 boilers) emitting a total
of 10.83 million tons of equivalent 1007, I^SO^., If one calculates the
tons of H2S04 emissions allowed by existing (March 1975) SIP standards
(5.80 million tons HpSO^.) and subtracts this quantity from total emis-
sions, then the actual amount to be controlled is identified. This
amounts to 5.03 million tons (10.83 - 5.80) of 112804. However, once the
decision is made to scrub a boiler the technical requirements to
operate a flue gas scrubbing system are imposed. The model assumes that
the FGD system would remove 907, of the SOo emissions. Therefore, from
the total emission of 10.83 million tons of 112804, 9.75 million tons
of 112804 could be produced with installations of FGD systems on all
boilers exceeding SIP standards. This amounts to an over control of
4.72 million tons (9.75 - 5.03) of abatement 1^304 production.
Program Analysis at the Boiler Level
Results of the program run that identified the unit cost of
operating three different FGD systems on each of the 610 boilers are
outlined in Figures 6 and 7. In Figure 6, the unit costs vary from $0
to >$999,999.00. The reason for the extreme range is that the amount
671
-------
SUIFURIC ACID US UNIT COST
U
N
I
T
C
0
S
T
T
0
H
1MC0
Sludge
Sulfur
SULFUR1C ACID CAP (!•*») TONS
Figure 6 . Unit production cost for abatement products for eastern U.S.
-------
SULFLIRIC ACID US UNIT COST
ON
~-J
CM
u
N
I
T
C
0
S
T
S
/
T
0
N
250
200
150
100
0
0 2000
SULFURIC ACID CAP (1000) TONS
1
Lime
(Sludge)
(H2S04)
o Sodium (Sulfur )
MgO
4000
6000
8000
Figure 7. unit production costs for abatement products.
-------
of abatement acid which must be scrubbed from the boilers to meet SIP
standards ranges from 318,963 tons/yr for the lowest unit cost boiler
down to 1 ton of abatement acid for the highest unit cost boiler. The
model assumes that if a boiler exceeds SIP standards in any quantity,
then the cost of operating a scrubber is imposed. Instead of scrubbing,
the high cost boilers would probably turn to clean fuel substitution.
The comparative cost of the three scrubbing systems is depicted
on a smaller scale in Figure 7. The lowest unit cost boiler is identified
at $43.77/ton of H2S04 equivalent ($0.25/million Btu) for limestone
slurry, $53.05/ton of acid for magnesia, and $98.73/ton of acid
equivalent for sodium. (Multiply $/ton of H^SO^ equivalent by 1.531 to
get $/ton of S02 removed.)
The sodium solution process is at a disadvantage in the model
because of the relatively higher unit cost as compared with either lime-
stone slurry or magnesia slurry processes. Figure 7 shows that on the
average the sodium solution process for abatement sulfur is about $50 -
$60/ton higher cost on an equivalent basis of 100% H2SO, as compared to
the magnesia slurry process for abatement FUSO^. This prevents abate-
ment sulfur from competing with abatement t^SO, even at the optimum
location considered in the model not counting the additional cost of
converting the sulfur to li^SO,. This raises the question of a con-
servation of energy principle. Is it a rational decision to reduce S02
to sulfur, then market the sulfur in processes that require oxidizing it
back to SOA? This study indicates a tentative negative answer.
Supply Curve for Abatement Acid
A demand curve for abatement acid was presented in Figure 4.
For illustrative purposes it is useful to consider the supply curve that
would be traced out by different levels of a uniform f.o.b. steam plant
supply cost for sulfuric acid. Again, this ignores steam plant location
relative to acid plants. Such a curve can be estimated by ranking power
plant boilers from lowest to highest cost for producing sulfuric acid and
accumulating supply quantities shown in Figure 7. Low-cost boilers
would tend to be new, large plants which burn high-sulfur coal, while
high-cost boilers would tend to be old, small, and inefficient. The
only boilers shown in Figure 7 are those exceeding SIP standards since
others would probably not enter the supply picture. Note that the
range of supply costs is much greater than those for demand. While
about 9 million tons of acid are available at an infinite price,
supplies greater than about 8 million tons are unreasonable.
While a few large modern boilers are reasonable candidates
for scrubbing, supply costs increase rapidly as more and more boilers
674
-------
pursue this alternative. It is unreasonable for all boilers to use a
scrubbing strategy.
Summary of Operating Characteristics of Power Plants in Eastern U.S.
Data from FPC Form 67 were loaded on MRI System 2000 to provide
an improved data management system for the operating characteristics of
each power plant boiler. The scrubbing cost data calculated for each
boiler in the unit cost subprogram discussed above were also added to
System 2000 as well as an indication that a given boiler did or did not
meet the SIP.
Table 8 shows selected operating characteristics for all
fossil-fuel-fired boilers located in eastern U.S. in 1972. Column 2
contains all boilers having emissions exceeding SIP standards. Column
3 and 4 contain the boilers selected by the two scrubbing cost screens
used in model runs. Total abatement capacity for all plants is 22.292
million tons as compared with 9.748 million tons for the plants exceeding
SIP.
Boilers with emissions exceeding SIP standards tend to have
greater generating capacity, fuel consumption, average H/jSO, abatement
capacity, and plant boiler factors. The average sulfur content of
fuels is also greater. The age distribution, however, is uniform for
both groups on a percentage basis. The last two columns relating to
boiler scrubbing cost screen for model runs will be discussed in the
model analysis.
ABATEMENT PRODUCTION-DISTRIBUTION-TRANSPORTATION SYSTEM
To assess representative competitive costs, this market system
analysis must generate accurate sulfur freight rates from the Frasch
sulfur sources to the acid plants and sulfuric acid freight rates from
all power plants to all sulfuric acid plants. This represents more than
300,000 possible rates. Also, the overall market study series calls
for a possible expansion to evaluate calcium sulfate for wallboard,
ammonium sulfate, and several other fertilizer classifications.
Standard Point Location Code
The logistical linkage between the sulfuric acid and power
plant data bases and the rate generation system is the Standard Point
Location Code (SPLC). This is a transportation-oriented, six-digit
number prescribed by the National Motor Freight Traffic Association
under the guidance of the SPLC Policy Committee. The system is similar
675
-------
Table 8. OPERATING CHARACTERISTICS OF POWER PLANTS IN
EASTERN U.S. (SPLC <700000) (1972)
Boiler scrubbing cost
screen for model runs
No. of; Utilities
Plants
Boilers
I^SOA abatement
capacity, tons
Average H?SO, abatement
capacity^ cons
Average coal consumed,
1000 tons
Average oil consumed,
1000 bbls
Coal-sulfur content
average, %
Oil-sulfur content
average, %
Associated generating
capacity average, MW
Plant boiler factors
average, 70
Age of boilers, 70
0-5 yr
6-10 yr
11-15 yr
16-30 yr
Over 30 yr
All boilers
182
597
2,564
22,292,127
10,373
980
897
2.4
0.7
135
51.7
10
7
12
39
32
Boiler
emissions
>SIP
57
154
610
9,748,274
15,981
1,115
1,176
3.0
0.8
158
55.1
9
8
13
43
27
Unit cost MgO
<$100
18
25
38
3,285,710
86,466
1,835
21
3.6
0.3
403
64.1
42
34
21
3
0
Unit cost
limestone
<$175
37
73
169
7,749,216
45,853
1,567
529
3.1
0.4
263
63.8
23
21
27
29
0
676
-------
to U.S. mail zip codes. Figure 8 shows location areas to the first two-
digit level. The first digit indicates a region relating to major
traditional traffic associations. The first two digits uniquely identify
a state or a portion thereof. As more digits are added, smaller nested
areal units are identified. The third digit gives a cluster of counties,
the fourth digit a county, the fifth digit a cluster of points within a
county, and the sixth digit identifies all rail and truck specific points.
Distribution Cost Generation
A logic flow diagram of the freight rate generation system used
in this model is shown in Figure 9. It shows that an SPLC for a power plant
origin and one for a sulfuric acid plant destination are input to the
National Rate Basis Tariff 1-C (NRBT 1-C). This tariff determines for
rail rate purposes the basing points for the origin and destination.
While there are about 60,000 rail points, there are only 2,632 basing
points east of the Rocky Mountains (SPLC <700000).
Outputs from the NRBT 1-C block in Figure 9 are two sets of
codes used to define mileage and tariff rates between the byproduct
shipping origin and destination points. One set - Index I and Index 2 -
is used to determine the appropriate rate base mileage to be applied.
The two index values are pointers to a 3.5 million record triangular
mileage file compiled from 12 tariffs resulting from the landmark 1945
Interstate Commerce Commission hearing entitled Docket 28300. ^ The
file includes rail (and some truck) tariff mileage for shipments within
and between five major freight associations shown in Figure 10. Output
from the Docket 28300 block of Figure 9 is a rate base number.
The second group of codes output from the NRBT 1-C define the
mutually exclusive set (MES) numbers for the origin and destination
points and serve as input to the Tariff Rate Generator block to choose
the appropriate tariff. In Figure 10, locations of the nine major MES
numbers east of the Rockies are shown. These result from considerable
overlapping occurring among the five major freight associations. There
are 19 mutually exclusive sets of basing points in the Docket 28300
tariffs. Table 9 lists the basing points which define the remaining 10
MES numbers. Table 10 lists all MES numbers as rows and the 12 Docket
28300 tariff numbers as columns and indicates that each MES in each
tariff either (1) did not occur (blank), (2) occurred as a headline
point (H), (3) as a sideline point (S), or (4) as both a headline and
sideline point (B). Depending on the tariff association originating the
tariff, the headline point can be either the origin or destination of
shipment. Given the MES for both the origin and destination it can be
determined in which tariff(s) these points can be found, which is output
from the Tariff Generator block in Figure 9.
677
-------
Source; Transportation Data Coordinating Committee'
National Motor Freight Traffic Association
Figure 8. Geographic distribution of standard point location codes (SPLC).
-------
SPLCi —i
SPLC2
o\
~J
to
NRBT I-C
INDEX, INDEX.
I T
DOCKET
28300
I
RATE BASE MILEAGE
RATE
SEARCH
MINIMUM
RATE
MES,
TARIFF
GENERATOR
TARIFF NUMBER
Figure 9. Flow diagram of freight rate generation model,
-------
WESTERN TRUNK
TERRITORY
GENERAL FREIGHT TRAFFIC
COMMITTEE TERR
TRANS-CONTINENTAL
TERRITORY
SOUTHER
ASSN
SOUTHWESTERN
TERRITORY
SWL 5
RAILROAD RATE TERRITORIES
FIGURE
-------
Table 9. RECLASSIFICATION OF BASE POINTS
00
10
10
11
12
12
12
13
13
13
13
13
13
14
14
KY
KY
KY
DC
VA
WV
VA
VA
VA
VA
VA
VA
VA
VA
LEXINGTON
WINCHESTER
CHILESBURG
WASHINGTON
NORFOLK
CHARLESTON
ALBERTA (S)
ALTAVISTA (s)
BURKVILLE (S)
LYNCHBURG (S)
PETERSBURG (S)
SUFFOLK (S)
BRISTOL
NORTON
14
14
15
15
16
16
17
18
19
19
19
19
19
19
VA
TN
SC
sc
NC
TN
IL
NY
IN
IN
IN
IN
IN
WV
ST PAUL
BRISTOL
ANDERSON QUARRY
RION
BUNN
BROWNSVILLE
SPARTA
PULASKI
CANNELTON
CORYDON
FERDINAND
HUNTINGBURG
MARENGO
OLCOTT
-------
OO
to
Table 10. TARIFF GENERATOR
GIVEN TWO RATE BASIS POINTS FROM MUTUALLY EXCLUSIVE BASING POINT SETS MES AND MES
THEY CAN BE FOUND IN ONE OR MOPE OF THE TARIFFS REPRESENTED BY COLUMNS BELOW, IF
ONE APPEARS AS A HEADLINE (H) AND ONE AS A SIDELINE (S), WHERE BOTH (B) QUALIFIES
AS HEADLINE AND SIDELINE
TARIFF
[ES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
1000 1001
S
B
B H
B
S
B H
S
S
S
B
1002 1003 1004 1005
S
S
B H
B H
B
S
S B
S
S
S S
S
B
S
S
1006 1007 1008
S S
H
S
S
H H
S
S
S
S S
S
H
S
S
S S
S
H
1009
B
B
B
B
B
B
B
B
B
B
B
B
1010
H
S
H
H
H
H
H
1011
B
B
B
B
B
B
B
B
B
B
B
-------
Figure 11 shows sulfuric acid rail rates within four major
freight associations as a function of rate base numbers. A slight error
in mileage is not nearly as critical as knowing which tariff applies.
Also these are the only published sulfuric acid rates. Rates for the
other eight tariffs were generated by TVA's Navigation and Regional
Economics Branch from these using sound traffic legal arguments similar
to the negotiation process that would insue should large acid movements
become a reality. In Figure 9 rate base and tariff numbers are input
to the Rate Search block and the minimum rate is output for use in the
transportation cost model.
MARKET SIMUIATION MODEL THEORY, ANALYSIS, AND RESULTS
Objective of the overall marketing model is to simulate long-
run competitive equilibrium sulfur and sulfuric acid market conditions
in the U.S. as might be impacted by clean air restrictions and available
power plant scrubbing technologies which result in abatement acid or
sulfur. To simulate these conditions, total cost of both the sulfuric
acid and power industries is minimized subject to the condition that
acid production (demand) is still met, either from traditional sulfur
sources or with a partial substitution of abatement sulfuric acid.
The model is similar to the classic transportation model of
linear programming where demands represent sulfuric acid plant customers
and supplies represent either production at each of the commercial acid
plants or purchases from any power plant boiler capable of producing
acid. Incurred (transfer) costs represent either sulfuric acid pro-
duction cost using Port Sulphur sulfur in the first case or boiler
scrubbing cost plus transportation cost to the respective acid plants
in the second case.
Economic Theory
Through the use of several simplifying assumptions, the com-
plex process taking place in the model can be conceptualized in terms
of classical supply and demand curves.
It was explained earlier, and summarized in Figure 4, that
from the sulfuric acid plant data base and the long-run average acid
production cost generator, all commercial acid producers can be ranked
in terms of a demand curve. Assuming an average transportation cost,
one might also conceive the price (cost) level being shifted down a
certain number of dollars per ton to give a conceptual f.o.b. power
plant acid demand curve such as DD1 in Figure 12. At high abatement
supply price levels, only the smallest, oldest, most remotely located
acid plants would be interested in curtailing production and buying
683
-------
-------
Q
O
CJ
cr
D
CO
LU
O
CE
CL
0
D'
Q
QUANTITY OF SULFURIC ACID
FIGURE I2« CONCEPTUAL DEMAND CURVE FOR
SULFURIC ACID AND SUPPLY CURVE FOR
ABATEMENT PRODUCTION.
685
-------
abatement acid. As supply price declines, more acid producers buy until
even the largest, newest plants located near sulfur supply sources become
interested.
Likewise, as explained earlier and summarized in Figure 7,
from the power plant data base and scrubbing cost generator, boilers can
be ranked in terms of a supply curve, f.o.b. each power plant, such as
SS1 in Figure 12. The intersection of such a supply curve with the
conceptual demand in Figure 12 would represent an equilibrium position.
However, this intersection occurs at a very small tonnage (15,000 tons)
because the abatement supply price level is so high. This is shown by a
dotted line labeled SS^ in Figure 12. Applying a unit revenue adjustment
parameter drops the level of the abatement supply curve, thereby pro-
ducing a series of equilibrium intersections as increasing quantities
of abatement production are considered. The supply curve SS' can be
thought of as having a market adjustment (credit for pollution control
cost) applied. Again the important concept to retain is that the
supply curve slopes upward. At low demand prices only the largest,
newest boilers would scrub even with the assumed market adjustment. As
demand price increases, more marginal boilers become interested.
If spatial considerations could be ignored, the demand curve
in Figure 4 and the supply curve in Figure 7 could be plotted on the
same graph as in Figure 12; where they intersect would represent supply-
demand equilibrium.
Looking first at the demand curve DD1, those sulfuric acid
producers above equilibrium price P would find it profitable to buy acid
from steam plants rather than continuing production. Those less than
price P would find it more profitable to continue production.
Looking at the supply curve SS', boilers below price P would
find it feasible to use a scrubbing strategy that produces sulfuric
acid, while those above price P but less than price F would profit from
using a scrubbing strategy that produces and disposes calcium sludge.
Those above price F would find it more profitable to use a clean fuels
strategy in lieu of scrubbing.
Multidimensional Equilibrium Model
While the preceding simplified economic theory presents the
essence of the economic model, a more elaborate spatial equilibrium
model is required for realistic analysis. The problem is that the lowest
acid cost boiler could be close to or far from the highest cost com-
mercial acid plant. It, therefore, only makes sense to trace a demand
curve for a single source of supply. This could be done for every
supply point, but would be of no analytical significance unless there
686
-------
was only one supply point being considered. As soon as more than one
supply point is considered, they become competitors for lucrative demand
points. Hence, while traditional supply-demand concepts are helpful
in exploring the basic underlying economic structure of flue gas market-
ing alternatives, a much more elaborate multidimensional equilibrium
model is required before analytical conclusions can be drawn.
The model finally developed considers total cost of both the
sulfuric acid and power industries and chooses the set of alternatives
that minimize the total cost. Sulfuric acid producers are given a choice
of continuing production or buying acid from any steam plant. Steam
plants not meeting SIP are given the choice of (1) selecting a clean
fuels strategy, (2) selecting a calcium sludge scrubbing technology, or
(3) selecting a sulfuric acid or sulfur-producing scrubbing technology.
In the latter case, product would then be available to supply sulfuric
acid plants if they choose to buy it. The mix of abatement strategies
and marketing patterns resulting in the lowest possible cost to the
combined industries (society) is said to be optimal. Economic theory
also supports the proposition that such a solution simulates the result
of long-run competitive equilibrium solutions.
The model implicitly ranks potential acid buyers as described
earlier under The Demand Curve for Abatement Acid except thr t every
potential abatement acid producer is given his own window through which
to view the market with reference to his specific location. As the
model is being solved this view is dynamically changed to reflect the
bidding away of markets by other potential abatement producers. The
decision making process might be viewed in two stages: (1) the steam
plant is bidding for markets it might want at a given price and (2) it
is deciding if it wants the market at all. The results of this bidding
process interacts with the plant's abatement strategy decision. The
task of simulating this process from the viewpoint of 166 acid producers
and 610 steam plant boilers can be solved with modern computers and
linear programming techniques. The solution reveals not only which acid
producers would buy and which steam plants would sell sulfuric acid,
but also which steam plants would sell to which acid plants. Any
variation to this optimal solution would increase the total cost to
both industries.
Analysis of the Model
Without clean air restrictions, computer runs indicate that
only two sulfuric acid plants would be interested in buying 15,000
tons of abatement acid, given the March 1975 price of $60/ton of sulfur
f.o.b. Port Sulphur, but this would not support the output of even the
most likely boiler. In other words, without other considerations such
687
-------
as air quality control restrictions, it would not even pay the lowest
cost boiler to supply the highest cost acid plant unless the cost of
sulfur increased substantially,.
Air pollution control restrictions are real, however, and
some economic effect can be expected in meeting current and future
control requirements. At present, no single available option (low-
sulfur fuel, fuel desulfurization, stack gas scrubbing, etc.) appears
to satisfy the needs of all power plants; therefore, the option taken
can be expected to be the lowest cost one for each power plant. Based
on current data paying either the premium for low-sulfur fuel or lime-
limestone scrubbing costs appears to be the lowest cost economic options.
If a credit, tax, subsidy,, or opportunity cost equal to these options
could be absorbed at each power plant, the net effect on the price needed
to justify production of abatement acid would be to lower the abatement
cost curve SS1 in Figure 12, thereby, increasing the equilibrium quantity
of abatement acid. This could also be thought of as the revenue re-
quirement necessary to bring supply and demand into balance at any
desirable level of abatement acid production. Naturally, the "desirable
level" will depend heavily on how much society is willing to pay for
air pollution control.
If revenue such as a credit, market adjustment factor, penalty,
or opportunity cost equal to $50/ton acid were to be acknowledged by
the power company as the cost of pollution control, net reduction on
cost (price) at which abatement acid enters the market could be as much
as $50/ton. For example, a 500-MW power plant burning coal with 3.5%
sulfur could produce about 112,900 tons/yr of acid by magnesia scrubbing
for about $85/ton ($0.30/million Btu). If the lowest cost option for
pollution control was limestone scrubbing at $70/ton of acid equivalent
($0.25/million Btu), then the acid could enter the market for as little
as $15/ton and the power plant not incur any higher costs for manu-
facturing acid than limestone scrubbing. Since it costs most commercial
acid producers at least $25/ton to produce, many would be interested in
acid at this price.
Results from the Model
An actual run of the model is summarized in Figure 13. To
conserve computer costs only sulfuric acid plants east of the Rocky
Mountains were considered. Also those boilers whose sulfuric acid
production costs were above $100/ton of acid were screened out. As
shown in Table 6, only 38 boilers qualify at this screening level. The
system is designed so that the level of the screen can be changed
easily. Note that as the level of market adjustment is parametrically
increased, equilibrium quantities of abatement sulfuric acid increase.
688
-------
COST OF POLLUTION
CONTROL OPTIONS
ON
00
0.35-1 $100-)
0.28-
CD 0.21 -
UJ
0.14-
o
0.071
0J
80-
CO
-------
At a $65/ton market adjustment, all 38 boilers would scrub to produce
3.285 million tons of abatement acid. While including more boilers
might affect the upper end of this curve, the lower end would probably
not change.
From an opportunity cost point of view, it was considered
important to run a second case in which limestone scrubbing versus
sulfuric acid were compared, since from the scrubber cost study, EPA-
600/2-75-006, it was known that by ignoring market revenue the lime-
stone throwaway system was the least-cost process. In this run, the
market adjustment was raised to a maximum of $175/ton of sulfuric acid
equivalent for the limestone process so that 169 rather than 38 of the
best boiler candidates could be considered. Rather than a uniform
market adjustment, individual adjustments were placed on each boiler at
an opportunity cost level equal to the sulfuric acid equivalent for
limestone throwaway. At the current price of $60/ton of sulfur, 7.447
million tons of sulfuric acid were produced and marketed out of the total
7.749 million tons considered. Implications of such a solution rest
heavily on the assumption that retrofit difficulty for a given boiler
would be similar for both processes and the level of long-run sulfur
price is realistic.
Detailed listing of this particular optimal solution is not
ready for publication. However, a sample computer report format out-
lining results of the optimal solution is presented in Table 11. A
complete detailed listing of the results of analysis is provided in this
table. The results could be further depicted on a map providing the
geographic distribution of the power plants and the acid plants identified
in the abatement production-marketing pattern for the optimal solution.
The published version will also provide a computer printout
report of the optimal solution including the information outlined in
Table 11.
Another Use of the Model
Another important methodological use of the marketing model
is shown in Figure 14. This is the welfare econometric notion of social
cost as reflected through consumer and producer surplus. Figure 14 is
the classical presentation of these concepts. The argument is that the
area under the demand curve DD1 (DRQS) out to supply quantity Q is the
total gain to consumers (acid plants) from the purchase of Q tons of
sulfuric acid. (it should be recognized that marketers can't give
preferential treatment to certain customers because of antitrust laws;
therefore, in the optimum solution all customers pay the same price for
abatement acid.) Consumers pay a total of P $/ton times Q tons or the
690
-------
Table 11. SAMPLE COMPUTER REPORT OF OFJTTMAL SOLUTION
NO. FPC CODE BOIL LOCATION CAPACITY LIMESTONE MAGNESIUM SODIUM MARGIN
(1) (2) (3) (4) (5) (6) (7) (8) (9)
NO. ACID PLANT LOCATION DEMAND PURCHASES PRICES PORT SULF
(10) (11) (12) (13) (14) (15) (16)
(1) Power plant listed in a ranked order from lowest to highest scrubbing
cost (1...610 possible)
(2) FPC power plant code
(3) Boiler number
(4) Power plant location
(5) Abatement capacity (OOO's tons H2SO )
(6) Limestone slurry scrubbing cost, $/ton 10070 H?SO, equivalent
(7) Magnesia slurry scrubbing cost, $/ton 100% H2S04
(8) Sodium solution scrubbing cost, $/ton 10070 H2SO/ equivalent
(9) Margin represents the stability of the solution. It is the marginal
increase in sulfuric acid production cost or reduction in sales
revenue which could result without changing the optimal production-
distribution solution.
(10) Acid plant number (1...166 possible)
(11) Name of acid plant buying abatement acid from the above power plant
(12) Location of acid plant
(13) Acid plant production capacity (OOO's tons)
(14) Abatement acid purchased from steam plant (OOO's tons)
(15) Abatement acid price $/ton f.o.b. power plant
(16) Avoidable cost of acid production using elemental sulfur at
given f.o.b. Port Sulphur price.
691
-------
o
cc
UJ
o
cc
CL
CONSUMER SURPLUS
PRODUCER
SURPLUS
Q
QUANTITY OF SULFURIC ACID
FIGURE W: CONCEPTUAL DEMAND CURVE FOR
SULFURIC ACID AND SUPPLY CURVE FOR
ABATEMENT PRODUCTION, ILLUSTRATING
PRODUCERS 8 CONSUMERS SURPLUS.
692
-------
rectangle below P out to Q (PSQR). Consumer surplus is the difference
or the indicated area (DPR) in Figure 14. Similarly this same total
revenue goes to producing power plants whose total is the area (RSQ)
below the supply curve SS1 out to supply quantity Q. Producer surplus
is the area below P and above the supply curve, as indicated (PRS). Net
social gain, therefore, is the combination of consumer and producer
surplus. At equilibrium the most marginal consumer and producer neither
gain nor lose, but all others in the solution are better off. Net social
gain is the amount by which they are better off. Methodology used in
this study does not address specific division of net social gain
between producer and consumer. It is assumed this will be determined in
the market place.
With the above welfare econometric notions as background,
consider their operational interpretation in terms of the present model.
For example, with a parametric market adjustment of $65/ton of tUSO/,
total cost of acid production in both industries is $1.347 billion and
the cost to power consumers for the market adjustment on 3.286 million
tons of acid is $214 million (about 2.5 mills/kWh to the customers
served by the 38 subject boilers). The level of this market adjustment
is very important politically since this is the apparent change in cost
of power for resolving about 657» of their total excess SOX emission
problem. However, the cost to both industries (really the sulfuric
acid industry) before the adjustment was $1.418 billion, so that total
cost reduction of the combined industries is $71 million. Hence, the
real cost to society is only $143 million, with $71 million being the
combination of consumer and producer surplus.
The divergence between apparent market adjustment and real
social cost for eliminating specific levels of SOX emissions is shown in
Figure 15. This welfare econometric phenomenon is significant since it
is a readily overlooked aspect relating to the methodology of economic
comparisons of saleable versus throwaway abatement strategies. It
supports one's intuitive resource conservation feelings and becomes
more socially significant in the face of long-term Frasch sulfur
supply shortages and higher prices.
CONCLUSIONS AND RECOMMENDATIONS
Conclusions
The point has been passed where studies about hypothetical
plants and markets are relevant in environmental and energy related
research. Important decisions are being made which involve substantial
financial commitments. Decision makers at the power utility and national
693
-------
Figure 15: SOCIAL COST OF ABATEMENT BY-PRODUCT H2S04 PRODUCTION,
o\
TOTAL COST
(000,000'
220-
APPARENT MARKET ADJUSTMENT
REAL SOCIAL COST
1.4
1.8
2.2
2.6
3.0
TONS OF ABATEMENT H2S04
(000,000' TONS)
-------
policy level need the best possible information on which to base these
decisions.
Recognizing this need, a systems approach to potential abate-
ment production and marketing of byproduct elemental sulfur and sulfuric
acid in the U.S. was taken. It was concluded in this research phase
that adequate data bases do exist. However acquiring, organizing,
verifying, and updating them is still a significant problem to be re-
solved. A market simulation model was developed to use these data bases
which permit the evaluation of alternative production and marketing
strategies.
The system was used to make trial runs to (1) demonstrate
its validity and (2) pretest the approach to analysis as well as
presentation of results. The philosophy was that agreement should be
reached on methods of analysis before spending large sums on computer
runs. Also, final runs should be made as close to publication date as
possible since the data bases and policy considerations are so dynamic.
However, some preliminary conclusions can be drawn based on the trial
runs subject to the specific assumptions made, and the degree to which
simulations based on 1972 production data can be extrapolated to future
conditions.
Total potential abatement acid production based on total
emissions east of the Rocky Mountains in 1972 was 22.3 million tons from
2,564 boilers. Total acid consumption for the same area, excluding
existing abatement production, was slightly higher. Hence, total
potential abatement production could have been physically absorbed. The
purpose of the market simulation model is to estimate how much would
have been supplied in a competitive market environment; who would
produce it, and who would consume it.
Based on SIP requirements in March 1975, only 610 of these
boilers were out of compliance, representing 9.7 million tons of
potential abatement acid. Of these, 70% were more than 15 yr old,
indicating unfavorable scrubbing economics. Many of these boilers had
very high abatement costs while only a few new large boilers had
reasonable costs. A point is obviously reached where an alternative
such as clean fuel substitution is less costly than scrubbing.
While the level of alternative emissions control cost other
than scrubbing is outside the scope of this study, it was analyzed
parametrically to determine its impact on simulated equilibrium quanti-
ties of abatement acid supply and demand. At zero alternative cost
there was no significant power plant production. It can therefore be
concluded that when sulfur value is $60/ton there would be no power
plant production without regulations for emission control. As plants
are forced to take some action, the cost for alternative controls
can be considered as a credit to recovery processes and provides the
695
-------
basis for assigning a sulfuric acid market adjustment value. At the
highest level of market adjustment analyzed, $65/ton H^SC^ (23£/million
Btu), 38 boilers producing 3.3 million tons were in the equilibrium
solution. Definite conclusions cannot be drawn from this analysis with-
out an insight into the clean-fuel market situation, which is uncertain
at this time.
Conclusions must also be tempered with expected sulfur price
levels, which could fluctuate in the short run but may stabilize at
relatively high prices depending on energy costs and the severity of
the impending Frasch sulfur shortage.
For boilers that do not select an alternative strategy, such
as clean fuel substitution, the question remains as to whether they would
produce (1) sulfuric acid, (2) sulfur, or (3) a throwaway sludge. While
the sulfur alternative was modeled, production costs with current techno-
logy were high enough that the alternative did not require a computer
run to conclude that sulfur would not be chosen over acid at any boiler.
But if technology improves, sulfur should be subjected to closer analysis
by the model because of the potential savings in freight and storage costs
for sulfur over acid.
The throwaway sludge alternative was analyzed in detail since
it is the lowest cost scrubbing method. In this case marketing revenue
would have to exceed the cost differential for a given boiler before
acid production would be selected. For this analysis only boilers with
cost less than $175/ton of H/jSO^ (61^/million Btu) were analyzed; it
is recognized that this level is possibly high, but it was chosen to
study the market for large quantities of acid. Those having alternative
costs higher than this would definitely not choose a scrubbing techno-
logy. In competition with acid produced from $60/ton sulfur, 7.4
million tons of acid (967» of the total considered) were produced in
lieu of sludge. Conclusions based on this result are less dependent on
retrofit difficulty inaccuracies than a clean fuel alternative but are
dependent on sulfur price. This result should be analyzed in future
runs at alternate price levels.
An important finding was that while long-run competitive
equilibrium solutions predict what may happen in competitive markets,
they do not identify net social gain. For example, with a $65/ton
market adjustment value for sulfuric acid, total adjustment is $214
million/yr. However, $71 million of this was identified as savings to
the utility (producer surplus) and acid industries (consumer surplus),
though it is not possible to identify exactly who would benefit and by
how much. Society in general would gain, however, which is not the
case with a throwaway strategy.
696
-------
Although the trial results thus far are very preliminary,
they offer considerable encouragement that a significant amount of by-
product sulfuric acid could be moved within the present market structure
at a sales value high enough to make sulfuric acid FGD processes
competitive with throwaway sludge systems.
Specific conclusions at this preliminary stage should be
drawn with caution. Publication of the detailed analysis is delayed to
accommodate further refinement and updating of data bases. For example,
several boilers that required control in the data base are now in com-
pliance with regulations. Furthermore, Indiana and Ohio SIP were in
litigation and boilers in these states were not considered as scrubbing
candidates. However, the results to date indicate that marketing sulfuric
acid from more efficient boilers is a viable alternative.
Recommendat ions
Although the market simulation model outlined above is workable
and has been demonstrated as a highly effective way to address the air
pollution abatement problem, it needs further updating, verification,
and refinement before it can be used as an accurate decision making
tool. It is recommended that emphasis in the future should be placed
first on improving the existing data bases which support the systems
analysis approach to the evaluation of potential byproduct marketing of
various FGD abatement products in the 48 contiguous states of the U.S.
Then the elemental sulfur and acid markets can be accurately assessed,
and other abatement byproducts such as (NH/^SO^, CaS04, NaSO^., etc.,
can be considered. When finally developed, the data bases and the model
would be available on a time-sharing network accessible by any private
or public agency interested in abatement byproduct marketing.
Four major areas need further refinement and verification:
1. The scrubbing cost generator.
a. Include scrubber system retrofit difficulty factors
for specific power plants in the U.S.
b. Update the SIP standards for power plants in the U.S.
c. To derive more representative power plant operating
data, expand the 1972 FPC Form 67 master file for
operating characteristics of existing power plants in
the U.S. to include 1969, 1970, 1971, and 1973 data,
as well as new plant projections for 1978 and 1983.
697
-------
d. Convert the FPC master file to the System 2000 data
management system for all years outlined in £ above.
e. Refine FGD cost generation model to accommodate
economies to scale for plants with more than one
boiler selected for scrubbing.
2. Update and verify the production cost generator for
sulfuric acid and elemental sulfur, including tail gas
cleanup for new and existing plants.
3. The transportation-storage cost generator.
a. Update the transportation cost generator to include
current rates on rail, barge, and truck shipments.
b. Update the elemental sulfur transportation, terminal
storage, and delivery cost to acid plants.
4. Develop projections for 1978 power plant abatement pro-
duction estimates, commercial acid production, sulfur-
supply, and define other byproduct supplies of recovered
sulfur and byproduct acid.
698
-------
REFERENCES
1. Bucy, J. I., and P. A. Corrigan, TVA-EPA Study of the Marketability
of Abatement Sulfur Products, a paper presented at the Flue Gas
Desulfurization Symposium in Atlanta, Georgia, November 4-7, 1974.
2. Waitzman, D. A., et. al., Marketing I^SO^ from SC^ Abatement Sources -
The TVA Hypothesis, EPA-650/2-75-051, December 1973.
3. Corrigan, P. A., Preliminary Feasibility Study of Calcium-Sulfur Sludge
Utilization in the Wallboard Industry, June 1974.
4. McGlamery, G. G., et. al., Detailed Cost Estimates for Advanced Effluent
Desulfurization Processes, EPA publication 600/2-75-006(TVA Bulletin
Y-90), January 1975.
5. Riegel, Emil Raymond, Riegel's Handbook of Industrial Chemistry (seventh
edition) (ed. James A. Kent). Van Nostrand Reinhold Company, New York,
New York. 1974
6. United States Bureau of Mines, Mineral Industry Survey, Sulfur in 1972,
July 2, 1973.
7. Sulfur Commodity Statement, United States Bureau of Mines dated
December 6, 1973.
8. Chemical Marketing Reporter, March 18, 1975.
9. Bucy, John I., The Economic Feasibility of Producing Sulfuric Acid from
SC>2 in Power Plant Stack Gases, University of Nebraska, Ph.D., Thesis,
iy/5.
10. Pearse, G.H.K., Sulphur-Economics and New Uses, a paper presented at
the Canadian Sulfur Symposium, May 30-June 1, 1974. Industrial
Mineral Section, Minerals and Metals Division, Energy Mines and Resources,
Canada, Ottowa, Ontario.
11. Harre, E. A., The Supply Outlook for Phosphate Fertilizers, a paper
presented at the TVA Fertilizer Conference in Louisville, Kentucky,
July 29-31, 1975, pg. 42.
12. British Sulphur Corporation, Ltd., Sulfur No. 113, July-August, 1974.
13. Steam Electric Plant Air and Water Quality Control Data for year ending
December 31, 1972, based on FPC Form No. 67, Federal Power Commission,
Washington, D.C.
14. Machine Sensible Magnetic Tapes of FPC Form 67 Data (1972).
15. Short Line Rail Mileage published by all railroads operating east of
the trans-continental territory in Figure 10.
699
-------
APPENDIX A
GENERAL CONVERSION FACTORS
EPA policy is to express all measurements in Agency documents in metric
units. Values in this report are given in British units for the convenience
of engineers and other scientists accustomed to using the British system.
The following conversion factors may be used to provide metric equivalents.
CONVERSION FACTORS FOR METRIC EQUIVALENTS OF BRITISH UNITS
ac
bbl
Btu
°F
ft
ft2
ft3
ft/min
ft 3/min
gal
gpm
gr
gr/ft3
hp
in
Ib
lb/ft3
Ib/hr
mi
rpm
scfm
ton
ton, long
ton/hr
Elitist.
_M.etric.-
Multiply
acre
barrels of oil
British Thermal Unit
degrees Fahieaheit-32
feet
square feet
cubic feet
feet per minute
cubic feet per minute
gallons
gallons per minute
grains (troy)
grains per cubic foot
horsepower
inches
pounds
pounds per cubic foot
pounds per hour
miles
revolutions per minute
standard cubic feet
per minute (32° F)
tons (short)3
tons (long)3
tons per hour
-Ey_
0.405
158.97
252
0.5555
30.48
0.0929
0.02832
0.508
0.000472
3.785
0.06308
0.0648
2.288
0.7457
2.54
0.4536
16.02
0.126
1609.
0.1047
1.695
0.90718
1.016
0.252
To obtain
hectare
liters
gram-calories
degrees Centigrade
centimeters
square meters
cubic meters
centimeters per second
cubic meters per second
liters
Liters per second
grams
grams per cubic meters
kilowatts
centimeters
kilograms
kilograms per cubic meter
grams per second
meters
radians per second
normal cubic meters
per hour (0°C)
metric tons
metric tons
kilograms per second
ha
1
g-cal
°C
cm
mj
m3
cm/sec
m3 /sec
1
I/sec
g/m3
kW
cm
kg
kg/m3
g/sec
m
rad/sec
Nm3/hr
t
t
_kg/sec_
dAU tons are expressed in short tons in this report except sulfur which is expressed in long tons.
700
-------
REGENERABLE PROCESSES SESSION
Chairman: Richard D. Stern
Chief, Process Technology Branch
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina
701
-------
STATUS OF DEMONSTRATION OF THE WELLMAN-LORD/ALLIED
FGD SYSTEM AT NIPSCO'S D. H. MITCHELL GENERATING STATION
PART I
BACKGROUND AND OVERVIEW
E. L. Mann
Northern Indiana Public Service Company
Michigan City, Indiana
Roger C. Christman
TRW Environmental Engineering Division
Vienna, Virginia
ABSTRACT
In June of 1972, the Northern Indiana Public Service Company and
the U. S. Environmental Protection Agency entered into a cost-shared
contract for the construction and operation of a flue gas desulfurization
demonstration plant. The system selected for the project was a combination
of two processes which had been installed and operated in applications
other than coal-fired utility boilers. These are the Wellman-Lord SO
Recovery Process and the Allied Chemical S0_ Reduction Process. Used
in tandem, the two processes result in a regenerable FGD process which
recovers SO as marketable elemental sulfur. The Wellman-Lord/Allied
FGD system has been installed on NIPSCO's 115MW coal-fired Unit No. 11
at the D. H. Mitchell Station in Gary, Indiana.
The construction of the demonstration plant is essentially complete
with mechanical checkout underway. During the coming year the system
will be operated by Allied Chemical Corporation. The EPA has retained
the services of TRW's Environmental Engineering Division to conduct
a comprehensive test and evaluation of the Wellman-Lord/Allied demon-
stration plant. A final project report, to be prepared by TRW, will
include information on the pollution control performance, secondary
effects, economics and reliability of the system. It is anticipated
that the NIPSCO/EPA demonstration project will provide information
useful to the continuing evaluation of alternatives in dealing with
air pollution problems.
703
-------
PART I
BACKGROUND AND OVERVIEW
Northern Indiana Public Service Company is a combined natural gas and
electric utility operating in the northern third of Indiana. The service
area includes the highly industrialized and environmentally sensitive
southern tip of Lake Michigan. The Company operates three coal-fired gen-
erating stations along the lake shore and one of these, the D. H. Mitchell
Station in Gary, was selected by NIPSCO as a site for demonstration of a
flue gas cleaning process.
In the early 1970's the Company was involved in evaluating alternatives
for dealing with the SOz problem and the regulations established by the
Federal EPA and state and local authorities. In considering flue gas desul-
furization it was necessary to examine both the ability to meet S02 emission
requirements and the total environmental impact of the system. Limited avail-
ability and high cost of land in the immediate area of the NIPSCO plants meant
that the installation of a throw-away system would involve the overland haul-
ing of the sludge product for substantial distances. As a result, the Com-
pany's attention focused on regenerable processes which produced sulfuric acid
or elemental sulfur suitable for disposal by sale.
Simultaneous with NIPSCO's evaluation of FGD alternatives, the Industrial
Environmental Research Laboratory - RTP of the 1). S. Environmental Protection
Agency was also evaluating regenerable processes for possible demonstration at
the 100MW level. The EPA, having two demonstration projects underway which
involved the production of sulfuric acid, was seeking an FGD process capable
of producing marketable elemental sulfur. Among the candidate processes, the
Wellman-Lord S02 Recovery Process (developed and marketed by Davy Powergas of
Lakeland, Florida) appeared to be the most fully developed with a 70MW oil-
fired boiler application operating in Japan. Allied Chemical had demonstrated
a process for reduction of S02 to elemental sulfur. This plant, installed on
a non-ferrous smelter in Ontario, had a sulfur throughput rate which was
several-fold greater than the system subsequently constructed at NIPSCO. All
indications pointed to the fact that the Allied S02 Reduction Process could
be mated to the Wellman-Lord S02 Recovery Process, thus resulting in a regen-
erable FGD process with elemental sulfur as the major product.
The parallel interests of NIPSCO and EPA in the Wellman-Lord Process led
to initiation of negotiations involving EPA, Davy Powergas, Allied Chemical
Corporation, and NIPSCO early in 1972. The EPA and NIPSCO reached agreement
on a cost-shared contract in June 1972. The EPA-NIPSCO contract and, in turn,
the NIPSCO-Davy Powergas contract has requirements for an emission control
system which will operate at a minimum of 90% S02 removal from the flue gases
of the 115MW boiler firing coal and with a sulfur content of 3.5%. The process
also is guaranteed not to allow any emission in excess of 200 ppm by volume of
S02 in the exit gas with lower sulfur fuel. An existing electrostatic precip-
itator is expected to reduce the dust loading to 0.044 grains per actual cubic
704
-------
foot at the scrubber inlet. The Davy Powergas system has been designed with
a pre-scrubber to remove 0.2 grains per ACF with the capability of handling
considerably greater fly ash loadings for short periods. Ash content of the
Midwest coal is approximately 10% and the heat content is 11,000 BID per
pound.
The descriptions of the Wellman-Lord and Allied Chemical processes are
well reported elsewhere and will not be covered here. The current status of
the project is discussed below.
Construction of the plant began in the Fall of 1974, with pilings and
foundations completed in December of that year. The field activity was
suspended during the winter months except for receiving and off-loading of
equipment. Equipment delivery problems which were typical of the construc-
tion industry in 1974 threatened to drastically delay completion of the
project. These difficulties relaxed somewhat in late 1974 as the business
slump intensified and the construction proceeded in a fairly orderly fashion
through 1975. At this writing, the construction is essentially complete ex-
cept for minor painting and insulation. Mechanical checkout of equipment is
underway and it is anticipated that flue gas will be admitted to the absorber
in April of 1976. The procedures and timing of startup activities are dis-
cussed by Allied Chemical in Part II of this paper.
In order to evaluate the performance and economics of the Wellman-Lord/
Allied demonstration plant, the Industrial Environmental Research Laboratory -
RTP retained the services of TRW's Environmental Engineering Division. The
TRW activities include: the conduct of a baseline test of the host boiler
prior to installation of the FGD system (completed); conduct of an acceptance
test to determine if contract guarantees have been met; and conduct of a
one-year demonstration test during which time the long-term performance and
economics data will be collected and evaluated for presentation in a final
report. In preparation for the acceptance and one-year demonstration test
efforts, TRW has designed and installed an automated test measurements system
which includes simultaneous measurement and recording of both host boiler
data and FGD system performance data. The continuous measurements will be
checked and supplemented by periodic manual sampling tests. The continuous
monitoring and automatic data acquisition permits correlation of FGD per-
formance with changes and rates of change of host boiler operating condi-
tions. It will also be possible to evaluate the FGD system emission response
to changes in operating mode resulting from optimization efforts by the
Allied Chemical operating staff. The final project report to be prepared by
TRW will include data for and assessment of the pollution control performance
and secondary effects of the system, evaluation of the system reliability and
economics, discussion of the system's response to transient conditions (boiler
startup, shutdown, load changes, fuel changes, etc.), and an evaluation of the
range of utility population applicability of the WeiIman-Lord/Allied FGD
system. Data and information for the final report will be supplied by NIPSCO
and Allied Chemical in addition to the data collected by the TRW operated test
measurements system.
The projected cost of the demonstration system is shown on Table I. Cal-
culations were done on the basis of complete financing by NIPSCO although
approximately 50% of capital costs were actually covered by Federal funds.
705
-------
The costs presented are for a 115MW retrofit system. Costs for full-scale
or new systems would, of course, benefit from some economy of scale.
The Company is hopeful that the information and data which will be
developed in the coming year as a result of operation of this important FGD
demonstration system will be helpful in evaluating S02 alternatives for
NIPSCO and the utility industry in general. The Wellman-Lord/Allied demon-
stration represents only one route under investigation by NIPSCO. At the
present time, the S02 regulations require that the D. H. Mitchell Station
comply with the emission limitation of 1.2 pounds S02 per million BTU. The
four coal fired units, totaling approximately 500MW gross, have been burning
Western coal for a little over 12 months. The station is in S02 emission
compliance. Marked loss in precipitator efficiency has been rioted. Also,
ash hopper overflows and fly ash sluice piping systems are being plugged
frequently with deposits. Coal feed to the 115MW demonstration unit will
change from Western coal to Indiana-Illinois coal during the demonstration
year, while the other three units will remain on the Western coal. This
should permit some interesting comparisons of operating and maintenance re-
quirements.
S02 emission regulations in the State of Indiana have not yet jelled.
At present, NIPSCO's two other coal fired stations are using 3% sulfur coal.
A new coal fired station now under construction with a 500+ MW cyclone unit
and a 500+ MW pulverized unit are to be fired with Western coal to attain
emission limitation of 1.2 pounds S02 per million BTLJ. Northern Indiana
Public Service Company is cognizant of the difficulties experienced by others
in S02 scrubbing. This soon will be supplemented by more intimate experience.
The same comment applies to use of low sulfur coal. The course which NIPSCO
takes to satisfy S02 regulations is not expected to be a highway, rather a
path certainly not well defined at this time. Future action, if required,
will be site related, based on experience and economic evaluation.
706
-------
TABLE I
Capitalization (1976 dollars)
Including top charges^ ' $14,824,690
($129 per installed KW)
Excluding top charges $12,794,733
Operating Cost (After Demonstration Year)
(?}
Operation and Maintenance^ ' $ 2,513,000
Product Credits, sulfur at $32/1ong ton
purge salts at $50/ton (361,560)
Net Operating Cost $ 2,151,440
Fixed Charges^ (26% x $14,824,690) $ 3,854,419
Total Annual Cost $ 6,005,859
Total Annualized Cost^ 8.1 mils/KWH
Notes:
(1) Interest during construction, overhead.
(2) Operating supplies include $400/day for anti-oxidant agent
or $11,200/month. Early operation may show anti-oxidant
is unnecessary or needed in lesser amounts.
(3) Plant life of 10-15 years assumed.
(4) Annual load factor of 73.5% assumed.
707
-------
STATUS OF DEMONSTRATION
WELLMAN-LORD/ALLIED CHEMICAL FGD SYSTEMS
NIPSCO D. H. MITCHELL GENERATING STATION
PART II
CURRENT STATUS $ OPERATING PLAN
Stephen F. LaKatos, Aubrey W. Michener, Jr.,
and William D. Hunter^ Jr.
Allied Chemical Corporation
Sulfur Pollution Control Services
P. 0. Box 1139-R
Morristown, New Jersey 07960
ABSTRACT
Construction of the SO emission control facility at NIPSCO is
nearing completion and preparations for operation are now well underway.
Start-up specialists of Allied Chemical and Davy Powergas are presently
working with Allied1s permanent operating staff in completing mechanical
checkout and acceptance of the plant.
Following completion of a stringent 15 1/2 day performance test
program, operation of the facility for a demonstration period of about
one year's duration will be conducted to document the operating efficiency
and reliability of the FGD system over a wide range of boiler operating
conditions. Allied Chemical will operate the FGD system for NIPSCO
on a continuing basis and will market the elemental sulfur and sodium
salt products.
709
-------
Status of Demonstration of Wellman-Lord/Allied Chemical
FGD Systems at NIPSCO D. H. Mitchell Generating Station
Part lit Current Status & Operating Plan
The engineering and construction phases of the SC>2 emission
control project at Northern Indiana Public Service Company's
(NIPSCO) D. H. Mitchell Station, Gary, Indiana will be completed
this month. Preparations for start-up and operation of the
system are now well advanced.
The flue gas desulfurization (FGD) system which has
been installed is an integration of the Wellman-Lord S02 Recovery
Process of Davy Powergas (DPG) and Allied Chemical's SC>2
Reduction to Sulfur. DPG is the engineer and constructor of the
entire facility while Allied Chemical will operate the complete
system and market the products.
Several "firsts" are represented by this venture. It
will be the first regenerable FGD system to control the total
flue gas emission from a full-scale coal-fired generating unit,
as well as the first application of the Wellman-Lord process
on a coal-fired boiler. It will be the first FGD system to
recover elemental sulfur commercially from the flue gas emission
of an electric utility. It is also the first time that a
utility has engaged a chemical company to operate and maintain
the emission control facility and to market the chemical products.
Participation in this project by Allied Chemical is a
logical extension of pollution control services that Allied
has provided to others for many years as part of its sulfuric
acid and sulfur products operations and sales in the U.S. and
Canada. Such activities include control of SC>2 emissions from
metallurgical operations and reprocessing of waste acids from
petroleum refineries and other sources. The skills and
experience that are an integral part of Allied's own operations
are now being applied to the chemical process problem of SO2
emission from an electric utility.
710
-------
Experience gained in chemical plant start-ups over the
years has led to the belief in Allied that there are two key
elements in the successful start-up of any plant, whether it
be a new process or a new installation of an old, well estab-
lished process. These are:
-technology which provides a sound basis for
the process and equipment design, and
-an experienced operating staff, technically
oriented to the specific requirements of the
chemical process.
These elements are equally important. Success cannot be
assured with either one alone. It has been Allied's
objective to incorporate both into the NIPSCO project.
At the inception of this project, NIPSCO commissioned
Allied to provide the following services:
-Preoperational Activities
-Mechanical Checkout and Acceptance
-Start-up & Performance Testing
-Demonstration Period Operation
-Continuing FGD System Management
PREOPERATIONAL ACTIVITIES
The NIPSCO preoperational effort has been conceived and
carried out on the basis that the facility will be a long term
operation and that Allied will have continuing responsibility
for its management. The supervisors, especially selected from
within Allied's organization for this project, bring to their
permanent assignments at the NIPSCO facility many years of
experience, not only in start-ups, but in the continuing
administration and technical supervision of chemical plant
operations and maintenance. Operators and mechanics, recruited
locally as Allied employees, have just completed a formal
training program, and training of technicians in analytical
procedures is now in progress.
711
-------
Other prestart-up preparations now completed include
a risk analysis of each equipment item in the facility which
enabled Allied Chemical to procure for NIPSCO the essential
spare parts and operating supplies required to minimize down-
time and maintenance costs during start-up and continuing
operations. This initial provisioning is a type of insurance
vital to early completion of the start-up and achievement of
operating reliability.
MECHANICAL CHECKOUT S ACCEPTANCE
Allied's primary concern in this phase of the project
has been to assure the operability of the plant. Much of
the activity of the operating supervisors has involved
inspection and monitoring of construction, with the assistance
of chemical, mechanical and instrument specialists from the
home office.
Conditional acceptance of the plant by Allied Chemical
for NIPSCO precedes the actual start-up. Acceptance is con-
ditional at this point because certain items of rotating
equipment, such as pumps, blowers and agitators can only be
checked under actual process conditions. The plant will be
accepted in its entirety at such time as all of the equipment
is functioning properly.
START-UP AND PERFORMANCE TESTING
Overall direction of the operation through start-up and
completion of performance testing is the responsibility of
Davy Powergas, the engineer/constructor. Allied will assist
DPG in conducting the performance test program established
for the facility. Performance criteria for the plant are
summarized as follows:
(1) The test period will consist of 12 days'
operation at an average load of 92 mega-
watts followed by 3% days at an average
load of 110 megawatts. Coal with an
average sulfur content of 3.2% has been
stockpiled for consumption during this
test period.
712
-------
(2) Other performance criteria provide that:
(a) When operated with 3.15 to 3.5% sulfur in
the coal, the system must achieve 90%
sulfur removal from the flue gas, or
retain not more than 200 ppm of SC>2 in the
exit gas stream, whichever is lower. With
fuels containing more than 3.5% sulfur,
the absorber must achieve at least 90%
sulfur removal from the flue gas. With
fuels containing less than 3.15% sulfur,
the absorber exit gas must contain not
more than 200 ppm SO^.
(b) Sulfur recovery efficiency must be not
less than 90% across the SC>2 reduction
system, and the elemental sulfur re-
covered from the concentrated SC>2 must
be of a quality suitable for conversion
to sulfuric acid in standard sulfuric
acid plants.
(c) Aggregate consumption of utilities
electric power, steam, cooling water
and natural gas has been established
at a fixed total maximum value.
(d) The maximum allowable consumption of
soda ash, the principal chemical raw
material and a function of sodium sulfite
oxidation in the absorption system, is also
specified for operation at the 92 megawatt
rate.
In addition to the tests that will be performed by
Allied during performance testing, the operation will also
be monitored by TRW, Inc. under contract to the Environmental
Protection Agency -
The start-up supervisors and specialists that Allied
Chemical and DPG now have on site preparing for initial
operation will remain through completion of performance
testing, working with the permanent Allied supervisory and
hourly staff. The entire start-up organization is illustrated
in Exhibit 1.
713
-------
DEMONSTRATION PERIOD OPERATION
Following completion of the performance testing and
final acceptance of the plant, the supervisory and technical
staff will be reduced to ten people while the operation, con-
tinues for a demonstration period of approximately one year's,
duration. The organization during this phase of the project
is shown in Exhibit 2, attached.
A series of operational tests will be conducted
throughout this period and the detailed data will be con- ,
tinuously evaluated to enable adjustments to assure the
efficient performance of both the Wellman-Lord and Allied
Chemical processes under changing load and other conditions.
TRW will perform periodic tests to acquire data on -the
operation of NIPSCO Unit 11 and to determine efficiency of
particulate and S02 removal from the flue gas in the Wellman-
Lord absorption system. Tests will be carried out over a
range of boiler operating rates and conditions to document
the turndown and maximum operating capabilities of the complete
FGD system. Additional tests employing coals of different
sulfur contents may also be conducted. The extensive informa-
tion which will be developed during the demonstration period
will enable projection of meaningful capital and operating
costs for new and retrofit utility applications.
CONTINUING FGD SYSTEM MANAGEMENT
Following a decision by NIPSCO to extend operation of
the S02 emission control facility beyond the demonstration
period, Allied will perform FGD system management services
for Unit 11 at the Mitchell Station on a continuing basis.
The NIPSCO FGD facility will be afforded the administrative
and technical support that is provided by Allied to its own
national network of chemical operations. At that time, the
supervisory and technical personnel will be further reduced
to 8 with the staff of operators and mechanics remaining un-
changed in number. Personnel for the permanent operation
are shown in Exhibit 3. It is important to note that this
same staff, with some additional maintenance personnel, would
have the ability to operate and maintain a FGD facility
serving all four coal-fired generating units at Mitchell Station.
714
-------
Marketing of chemicals produced in the facility is
the responsibility of Allied Chemical. The elemental sulfur
product will be shipped from the site in tank trucks as a
liquid. It will be consumed initially at an Allied Chemical
plant in tae manufacture of sulfuric acid. Likewise, the
sodium salt recovered from the Wellman-Lord system as a dry
granular product will be channeled into appropriate markets,
primarily in the pulp and paper industry. NIPSCO will receive
full credit for the revenues from the sale of these products.
In conclusion, as NIPSCO approaches start-up of the FGD
facility at Mitchell Station, there is every expectation that
this regenerable system will demonstrate the efficiency and
reliability sought by the electric utility industry in coping
with the problem of S02 emissions from coal-fired generating
units.
715
-------
NIPSCO EMISSION CONTROL SYSTEM
Start - Up and Performance Testing
MAINTENANCE
SUPERVISOR
-------
NIPSCO EMISSION CONTROL SYSTEM
Demonstration Period Operation
-------
NIPSCO EMISSION CONTROL SYSTEM
Continuing FGD System Management
NIPSCO
STATION
MANAGER
TECHNICAL
SUPERVISOR
TECHNICAL
ANALYSTS
(2)
ALLIED
PLANT
MANAGER
OPERATING
SUPERVISOR
OPERATING
FOREMAN
OPERATORS
(14)
PLANT
CONTROLLER
MAINTENANCE
SUPERVISOR
MAINTENANCE
MEN (8)
-------
AN UPDATE OF THE WELLMAN-LORD
FLUE GAS DESULFURIZATION PROCESS
Roberto I. Pedroso
Davy Powergas Inc.
Lakeland, Florida
ABSTRACT
The Wellman-Lord SO Recovery Process and three different end
product plants for handling the recovered SO are described. A world-
wide listing of all Wellman-Lord installations in operation or in the
design or construction phase is given, and the performance or design
of some installations is described. Capital and operating costs for
a typical Wellman-Lord installation are presented, and major design
criteria and their impact on cost are discussed. Problems facing
the Wellman-Lord Process and action being taken to resolve them are
discussed.
719
-------
AN UPDATE OF THE WELLMAN-LORD
FLUE GAS DESULFURIZATION PROCESS
INTRODUCTION
The Wellman-Lord SO- Recovery Process was developed by Davy Powergas in
the late 1960's. It is Being used in, or designed or constructed for, thirty
commercial installations throughout the world. At every installation, the
Process has proven itself as operationally reliable, and has met or surpassed
governmental regulations regarding SO- emissions.
The Process can be applied to any flue gas containing S0~. It is equally
applicable to the flue gas from fossil fuel fired boilers, nonferrous smelters,
sulfuric acid plants, and Claus plants. The Process uses a recycle system and
yields an S09 gas suitable for conversion to sulfuric acid, elemental sulfur,
or liquid SO .
CHEMISTRY
The Process is based on the chemistry of the sodium sulfite/bisulfite
system. Flue gas containing S0« is scrubbed with a sodium sulfite solution
which absorbs SO-, converting sodium sulfite to sodium bisulfite:
(1)
The sodium bisulfite solution is regenerated by thermal decomposition.
Application of heat simply reverses equation 1:
2 NaHS03 Na2S03 + S02 + H20 (2)
The SO,-, is recovered in a concentrated stream.
PROCESS DESCRIPTION
Gas Pre-Treatment
Each plant's pre-treatment requirements must be studied individually. For
example, the flue gas from a fossil fuel fired boiler will require cooling and
particulate removal prior to absorption of the S0». In a Claus plant, the tail
gas must be incinerated to oxidize the sulfur compounds to S0?.
Absorber
After appropriate pre-treatment, the flue gas enters the SO- absorber, which
is a simple gas-liquid contacting device with two or more absorption stages. The
absorber can be designed to reduce S0« concentration to the required level, and
can accommodate a wide range of turndown conditions. If desired, separate ab-
sorbers can be connected to each S02 source, with all absorbers serviced by a
single regeneration system. The Wellman-Lord absorption step is free of scaling
problems .
720
-------
Chemical Regeneration System
Sodium bisulfite solution containing the S0? exits from the absorber into
a surge tank. It flows at a steady rate into a forced-circulation evaporator-
crystallizer, which can be heated by low pressure exhaust steam. The sodium
sulfite slurry produced in the evaporator is redissolved in a dissolving tank
with recycled condensate from the condenser system. The resulting lean solution
flows to another surge tank and is recycled to the absorber.
Solution Surge Tanks
Fluctuations in the absorber and regeneration sections are stabilized by
routing the solutions from each section through surge tanks. In addition, the
regeneration section can be remotely located from the absorber section. Further-
more, the regeneration section can be shut down completely for scheduled mainte-
nance without interrupting SO™ removal in the absorber section. Requirements
for expensive spare equipment are minimized without sacrificing basic pollution
control reliability.
Purge Treatment
Oxidation of sodium sulfite in the circulating solution to sodium sulfate can
occur whenever oxygen is present in the flue gas.
To control the level of inactive sodium sulfate in the solution, a small
sidestream is sent to a purge treatment section to remove the sulfate and return
active solution to the chemical section. The sodium sulfate is precipitated from
the solution by cooling in a chiller-crystallizer. With controlled crystalliza-
tion, the sulfate precipitates in a much greater proportion than the other
sodium compounds. Thus, the solid phase is enriched in sulfate while the liquid
phase becomes leaner. The solid phase is separated from the liquid, and the
latter is returned to the chemical section. The solid phase can be dried for
sale or for disposal. Or it can be further treated and discharged as an in-
nocuous effluent.
Product SO
Vapor from the evaporator is cooled to condense water and achieve the de-
sired S09 concentration. Wellman-Lord units treating flue gas from sulfuric acid
or Glaus plants recycle the S0? to the respective plant. The S0~ recovered from
boiler or smelter flue gases can be converted to sulfuric acid, elemental sulfur,
or liquid SO,,. A brief description for each of these processes follows:
Liquid SO . The S02 gas is first dried with silica gel. The dry SO
gas is then compressed and condensed. The liquid SO is collected in a pres-
surized storage tank.
Sulfuric Acid. The S0? gas and ambient air are first dried with 93%
sulfuric acid. The S09 then reacts with oxygen, in the presence of a vanadium
pentoxide catalyst, forming sulfur trioxide:
S0 + 1/2 0 S0 (3)
721
-------
The S0_ is absorbed from the gas mixture into 98% sulfuric acid, reacting
with the dilution water to form additional acid:
S03 + H20 H2S04 (4)
Tail gas from the acid plant is recycled to the Wellman-Lord absorber.
Elemental Sulfur. The S02 gas can be processed to high purity elemental
sulfur. This conversion is carried out in two steps. In the first step, a
portion of the S0? in the feed gas reacts with natural gas (methane), yielding
a mixture of elemental sulfur, hydrogen sulfide, carbon dioxide and water vapor:
2CH4 + 3S02 S + 2H2S + 2CC>2 + 2H2
-------
All Wellman-Lord installations currently operating in Japan also have an
SO™ removal efficiency greater than 90%, and an on-stream time greater than 98%
for the absorption area.
The Wellman-Lord installations on coal fired power plants in West Germany
have special significance. At both of these installations, only a portion of
the boiler flue gas will be treated in the desulfurization facilities. The
remainder will be mixed with desulfurized, saturated gas from the Wellman-Lord
absorbers. This will provide the reheat required to prevent a steam plume when
flue gas is discharged to the atmosphere. The amount of SO- contained in the
combined stack gasses will still be less than the amount allowed by governmental
SO,., emission regulations.
These applications required a flue gas desulfurization process with a com-
mercially proven record for high S0_ removal efficiency and operational re-
liability. The Wellman-Lord SO.-, Recovery Process was selected. A U.S. utility
is presently considering a proposal presented by Davy Powergas for such an in-
stallation.
UPDATED WELLMAN-LORD PROCESS FGD COSTS
Tables 4, 5 and 6 show installed capital and annual operating costs of a
Wellman-Lord installation. The costs are based on a 670 MW power plant operating
8,000 hr/yr burning coal with a heating value of 8,100 Btu/lb and a sulfur con-
tent of 1.3%. The end product is sulfuric acid, elemental sulfur, or liquid
S0?. The costs are expressed in January 1976 dollars.
The capital costs presented in Tables 4 and 5 are realistic. They reflect
actual Davy Powergas experience in projects which are presently in the design
and/or construction phase. The capital cost presented in Table 6 includes an
estimate for the liquid SO- plant. Therefore, it may not be as accurate as
that in Tables 4 and 5. It must be pointed out that these costs are for an
installation with specific parameters, i.e., power plant generating capacity,
sulfur content of the coal, etc- Each of these parameters will exert a sepa-
rate and distinct influence on capital cost. Therefore, each installation
must be considered separately.
The consumption of utilities presented in the tables is for an installa-
tion with specific parameters. Each of these parameters will exert a separate
and distinct influence on consumption of utilities. Geographical location may
also affect consumption of utilities. Furthermore, unit costs of utilities
will vary considerably with geographic location and with type of industry. Unit
costs used in the tables are believed to be conservative.
The raw material consumed and product produced again are dependent on
specific parameters of the installation. Prices for the raw material and pro-
ducts were obtained from a January 1976 issue of the Chemical Marketing Reporter.
The annual maintenance cost is expressed as a percent of installed capital
cost per year. It reflects actual Davy Powergas experience with this type of
plant. The overhead, miscellaneous, and capital recovery costs will vary from
industry to industry, and even within the same industry. This is due to indi-
vidual regulations or an individual company's operating and accounting philo-
sophies. Variations in the capital recovery factor used can easily add (or
subtract) several million dollars to the annual cost.
723
-------
Some of the flue gas characteristics which will influence installed capital
and annual operating costs are briefly discussed below.
Flue Gas Quantity
Of all the design parameters, the quantity of flue gas treated has the
largest effect on the capital cost. For this reason, Davy Powergas does not
require spare flue gas handling equipment. The omission of spare equipment is
supported by the trouble-free operation of Wellman-Lord absorbers.
Flue Gas S0? Concentration
The S0? inlet concentration has some effect on capital cost, and a marked
effect on the annual cost of utilities. As the SCU inlet concentration increases,
the number of required mass transfer stages may decrease. This is due to the
higher driving force of the more concentrated gas. For example, for 90% removal,
an inlet S0? concentration of 1,000 PPMV may require as many as five mas transfer
stages. For the same 90% removal, an inlet S0? concentration of 3,000 PPMV may
only require three mass transfer stages. Each mass transfer stage will add two
to five feet to the absorber height, and approximately three inches of water
pressure drop.
Furthermore, for a given amount of SO- to be recovered, as the inlet, concen-
tration in the flue gas decreases, the required amount of heating steam and
cooling water increases.
Flue Gas Oxygen Concentration
The formation of inactive sodium sulfate in the absorber solution varies
with the oxygen concentration in the flue gas. As the concentration of oxygen
increases, the size of the purge treatment section and the chemical make-up
requirement will increase.
Flue Gas Particulate Loading
A wet scrubber is used upstream of the Wellman-Lord absorber to saturate
the flue gas and remove its particulate content. In general, as the flue gas
particulate content increases, the energy (pressure drop) required to remove
particulate in the scrubber increases. The higher pressure drop will increase
the booster fan power consumption. Davy Powergas recommends an efficient, dry
particulate removal device upstream of the Wellman-Lord system. This will ob-
viate the need for excessive pressure drop across the wet scrubber.
PROBLEMS FACING THE WELLMAN-LORD PROCESS
AND ACTION BEING TAKEN TO RESOLVE THEM
Sulfate Purge
Oxidation of a small amount of the solution in the Wellman-Lord absorber
creates the need to purge a portion of the solution from the system. This
purge has two effects: a chemical make-up requirement and a potential waste
disposal problem.
724
-------
As the size (SCFM treated) of Wellman-Lord installations increased, the
purge stream also increased. An intensive testing program was initiated to
prevent solution oxidation by the addition of different antioxidants. This
approach proved unsuccessful and was eventually abandoned.
Efforts were then directed toward minimizing the purge stream and the
chemical make-up requirements. This was accomplished by a system that se-
lectively removes the oxidized part of the solution by crystallization. This
system achieved a five to six-fold decrease in the purge stream and the chemical
make-up requirements. It is in operation at eight Wellman-Lord installations,
and has been highly successful in each case.
The concentrated purge stream from this system can be dried for sale or for
disposal. Or it can be further treated and discharged as an innocuous effluent.
The final purge stream at NIPSCO will be dried for sale. Potential markets
exist for this material throughout the U.S. In some cases, marketing the product
may not be feasible.
Davy Powergas is engaged in the development of two processes to chemically
reduce sodium sulfate. The purge stream would thus be regenerated and returned
to the Wellman-Lord Process. One of these processes uses natural gas in the
reduction step, x<7hile the other one uses coal. Both have been successfully
tested on a pilot plant scale. While both processes must be optimized, Davy
Powergas believes that either can be included in future Wellman-Lord installa-
tions. The Process will then produce only sulfuric acid, elemental sulfur or
liquid S0~ as a product. The chemical make-up requirement would be reduced to
a minimum.
Capital and Operating Costs
All SO,., recovery processes suffer from high installed capital and/or annual
operating costs. Davy Powergas has always specified high quality proven equip-
ment and conservatively chosen materials of construction. Undoubtedly, this is
a major reason for the high reliability of the systems installed. Despite this,
the Wellman-Lord Process is competitive in price with other processes.
Davy Powergas has an active program to examine ways of reducing capital
and operating costs without sacrificing reliability. Several examples can be
noted.
Partial Flue Gas Treatment. As previously mentioned, this concept will be
applied at two West German installations and is also being considered by a U.S.
utility. A potential cost reduction by use of this concept is envisioned. Based
on the average of Davy Powergas projects to date, the flue gas handling area
represents approximately 40% of the capital cost. Power consumption by the
booster fan(s) represents approximately 70% of the annual power cost.
Sulfate Purge Concentration/Elimination. Efforts in the area of purge
elimination have already been discussed. In addition, Davy Powergas is also
studying a new method for sodium sulfate concentration. This method employs
high temperature separation of the sulfate and has been successfully tested on
a bench scale. One of the Wellman-Lord installations currently operating in
the U.S. is considering installation of a pilot plant for demonstration of the
method. High temperature separation of the sulfate would reduce annual cost
of utilities.
725
-------
Multiple Flue Gas Handling Modules. The current maximum size booster
fan, scrubber, absorber module which Davy Powergas recommends will handle
approximately 300,000 SCFM of flue gas. For large installations such as
PSCNM's Units #1 and 2, multiple modules are required. A highly innovative
concept was introduced to this application. It consists of rectangular
absorbers placed immediately adjacent to each other with a common wall shared
by every two absorbers. The walls will be constructed of tile lined concrete.
The tile is resistant to corrosion from the circulating solution. It also
provides structural support for the concrete during pouring.
SUMMARY
To summarize the preceding discussion, it is appropriate to highlight the
strengths of the Wellman-Lord S02 Recovery Process.
Desulfurization Efficiency
From any typical fossil fuel fired boiler, an SO,, removal greater than
90% can be achieved. Commercial units in operation consistently operate with
this efficiency.
Reliability
Full scale units have demonstrated availability approaching 100% over
several years of operation.
Commercially Proven
Seventeen Wellman-Lord installations in commercial service have logged an
aggregate total of more than thirty-seven years of successful operation.
Saleable Product
The product from a Wellman-Lord installation is sulfuric acid or elemental
sulfur. Both products are in high demand world-wide.
726
-------
~~j
to
^J
FIGURE I
WELLMAN-LORD S02 RECOVERY PROCESS
t
DESULFURIZED
FLUE GAS
.LEAN SOLUTION
ABSORBER
SODIUM MAKEUP SOLUTION
PURGE TREATMENT STEAM
REGENERATION
-------
TABLE 1
WELLMAN-LORD INSTALLATIONS IN THE UNITED STATES
Company and Location
Units on Stream
Olin Corp./Paulsboro, NJ
SOCAL/E1 Segundo, Gal.
Allied Chem./Calumet, 111.
Olin Corp./Curtis Bay, Md.
SOCAL/Richmond, Gal.
•
-------
TABLE 2
WELLMAN-LORD INSTALLATIONS IN JAPAN
Company and Location
Units on Stream
Japan Syn. Rubber/Chiba
Toa Nenryo/Kawasaki
Chubu Electric/Nagoya
Japan Syn. Rubber/Yokkaichi
Sumitomo Chem./Sodegaura
Kashima Oil/Kashima
Mitsubishi Chem./Mizushima
Toa Nenryo/Hatsushima
Toyo Rayon/Nagoya
Japan Nat. Railroad/Kawasaki
Kurashiki Rayon/Okayama
Fuji Film/Fujinomiya
Units in Design or Construction
Shin Daikyowa/Yokkaichi
Sumitomo Chem./Niihama
Mitsubishi Chem./Mizushima
Mitsubishi Chem./Kurosaki
Tohoku Electric/Niigata
Feed Gas Origin
Oil Fired Boiler
Claus Plant
220 MW Oil Fired Power Plant
Oil Fired Boiler
Oil Fired Boiler
Claus Plant
Oil Fired Boiler
Claus Plant
Oil Fired Boiler
200 MW Oil Fired Power Plant
Oil Fired Boiler
Oil Fired Boiler
Oil Fired Boiler
Oil Fired Boiler
Oil Fired Boiler
Oil Fired Boiler
100 MW Oil Fired Power Plant
SCFM Treated
124,000
41,000
390,000
280,000
225,000
20,000
373,000
10,000
218,000
435,000
248,000
89,000
253,000
91,000
390,000
330,000
236,000
Disposition of SO,.
Sulfuric Acid Plant
Recycle to Claus Plant
Sulfuric Acid Plant
Sulfuric Acid Plant
Sulfuric Acid Plant
Recycle to Claus Plant
Sulfuric Acid Plant
Recycle to Claus Plant
Sulfuric Acid Plant
Sulfuric Acid Plant
Sulfuric Acid Plant
Liquid SO
-------
TABLE 3
WELLMAN-LORD INSTALLATIONS IN WEST GERMANY
Company and Location
Units in Design or Construction
Confidential
Confidential
Feed Gas Origin
350 MW Coal Fired Power Plant
175 MW Coal Fired Power Plant
SCFM Treated
778,000
389,000
Disposition of SO,.
Liquid
Glaus Plant
-------
TABLE 4
JANUARY 1976 COST PROJECTIONS
WELLMAN-LORD FGD PROCESS
SULFURIC ACID BYPRODUCT
Basis: 670 MW Coal Fired Power Plant
8,100 Btu/lb, 1.3% Sulfur Coal
8,000 Hr/Yr Operation
Installed Capital Cost $55,500,000
Annual Operating Cost $ 1,000/Yr
Soda Ash $$54/Ton) 840
Steam ($1.00/1,000 Ib) 1,540
Water ($0.03/1,000 gal) 450
Power ($0.015/kWh) 2,680
Labor & Supervision (Labor @ $6.00/hr) 230
Payroll & Plant Overhead (140% of Labor) 325
Maintenance (4% of Capital/yr) 2,220
Miscellaneous (3% of Capital/yr) 1,665
Capital Charges (20% of Capital/yr) 11,100
Acid Credit ($24/Ton) (2,830)
Total $18,220
Installed Capital Cost $83/kW
Annual Operating 3.4 Mils/kWh
731
-------
TABLE 5
JANUARY 1976 COST PROJECTIONS
WELLMAN-LORD FGD PROCESS
ELEMENTAL SULFUR BYPRODUCT
Basis: 670 MW Coal Fired Power Plant
8,100 Btu/lb, 1.3% Sulfur Coal
8,000 Hr/Yr Operation
Installed Capital Cost $58,500,000
Annual Operating Cost $ 1,000/Yr
Soda Ash ($54/Ton) 840
Steam ($1.00/1,000 Ib) 1,380
Water ($0.03/1,000 gal) 420
Power ($0.015/kWh) 2,590
Natural Gas ($1.50/MM Btu) 790
Labor & Supervision (Labor @ $6.00/hr) 230
Payroll & Plant Overhead (140% of Labor) 325
Maintenance (4% of Capital/yr) 2,340
Miscellaneous (3% of Capital/yr) 1,755
Capital Charges (20% of Capital/yr) 11,700
Sulfur Credit ($60/Ton) (2,310)
Total $20,060
Installed Capital Cost $87/kW
Annual Operating Cost 3.7 Mils/kWh
732
-------
TABLE 6
JANUARY 1976 COST PROJECTIONS
WELLMAN-LORD FGD PROCESS
LIQUID SULFUR DIOXIDE BYPRODUCT
Basis: 670 MW Coal Fired Power Plant
8,100 Btu/lb, 1.3% Sulfur Coal
8,000 Hr/Yr Operation
Installed Capital Cost $56,000,000
Annual Operating Cost $ 1,000/Yr
Soda Ash ($54/Ton) 840
Silica Gel Nil
Steam ($1.00/1,000 Ib) 1,540
Water ($0.03/1,000 gal) 420
Power ($0.015/kWh) 2,710
Labor & Supervision (Labor @ $6.00/hr) 185
Payroll & Plant Overhead (140% of Labor) 255
Maintenance (4% of Capital/yr) 2,240
Miscellaneous (3% of Capital/yr) 1,680
Capital Charges (20% of Capital/yr) 11,200
S02 Credit ($110/Ton) (8,460)
Total $12,610
Installed Capital Cost $84/kW
Annual Operating Cost 2.4 Mils/kWh
733
-------
SUMMARY OF OPERATIONS OF THE CHEMICO-BASIC MgO FGD SYSTEM
AT THE PEPCO DICKERSON GENERATING STATION
R. B. Taylor, Manager
P. R. Gambarani, Project Manager
Project Management Department
Chemico Air Pollution Control Company
New York, New York
D. Erdman, Project Engineer
Potomac Electric Power Company
Washington, D. C.
ABSTRACT
The Chemico/Basic MgO FGD prototype regenerative wet scrubbing
system at Potomac Electric Power Company's Dickerson Generating Station
is reviewed. Objectives of the two-year prototype test program are stated
and operational experience is summarized. Attainment of the objectives
in terms of process, performance, equipment operation, and economics is
evaluated. Projected costs and future direction of wet scrubbing at
the Dickerson Station are also discussed.
735
-------
SUMMARY OF OPERATIONS OF THE
CHEMICO-BASED MgO FGD SYSTEM
AT THE PEPCO DICKERSON
GENERATING STATION
INTRODUCTION
In September 1970, the Potcrnac Electric Power Company, (PEPCO) selected the
Chemico-Basic MgO Process for flue gas desulfurization of one of its coal-
fired boilers to demonstrate the feasibility of regenerative wet scrubbing.
The process utilized a two-stage vessel - the first stage for particulate
removal, and the second stage for SO- absorption using MgO as the absorbing
alkali. The particulate stage handles flue gas from either the inlet or the
outlet of an electrostatic precipitator. The purpose of varying the particulate
inlet loading was to determine how much particulate contamination can be toler-
ated in the SO2 removal stage and to determine the overall effect on particulate
removal by the first stage.
Pilot testing was performed at the Dickerson Station Unit No. 3 to evaluate
system operating and performance conditions as well as to evaluate certain
process factors in the scrubber recycle liquid system. Testing was a follow-on
to the Boston Edison/Chemico-Basic/EPA MgO-SO2 removal system on an oil fired
installation at Boston Edison's Mystic Station. Some modifications made to the
Dickerson Unit during its design and construction were the result of experience
gained on the Boston Unit, although many of the system components remained the
same because of the closely concurrent schedules of both installations.
The objectives of the project were to:
1. Achieve 90% or more S02 removal efficiency using MgO as the
absorbing agent,
2. Determine the effect of particulates on S02 absorption and to
evaluate the effect of inlet grain loading on particulate
collection efficiency,
3. Produce anhydrous magnesium sulfite for ultimate conversion
to MgO and re-use as an S02 absorbant,
736
-------
4. Produce commercially salable sulfuric acid made from the
absorbed SC>2,
5. Determine the impact of boiler load swings upon the scrubbing
system performance as related to materials storage and handling,
6. Obtain data relating to system efficiency, energy requirements,
mechanical performance, maintenance, and overhead burdens to
project operating costs for 600 I>5w installations or larger,
7. Optimize design parameters so that capitalization for systems
of more than 600 Mw could be determined.
This paper reviews the prototype system, and examines its performance,
equipment design and operation, process, and resulting economics.
SYSTEM DESCRIPTION
Figure 1 is a cutaway section of the 2-stage scrubber. The scrubber vessel
consists of a variable throat venturi in the first stage for flyash removal
and a fixed throati in the second stage for SO2 absorption.
Flue gas enters the scrubber at 250°F and is saturated and cooled to 120°F in
the first stage where flyash is removed. In the second stage the flue gas is
contacted with atomized MgO slurry for SO2 absorption. It then passes through
a mist eliminator section within the vessel.
Figure 2 is a schematic of the Dicker son Regenerative Wet Scrubbing System.
The system was designed to treat one half of the flue gas of a 190 Mw boiler,
or 295,000 ACFM. It is capable of taking flue gas, both before and after an
electrostatic precipitator.
After leaving the scrubber, the flue gas proceeds through the I.D. fan and a
separate mist eliminator before entering the breeching of the stack.
The liquid streams are distinct and separate, one for each stage. The first
stage utilizes two 100% recycle pumps of alloy 20 construction. A stream is
bled from the discharge of these pumps to two 40' diameter thickeners. The
overflow water from the thickeners is returned to the scrubber; thickener
underflow is diluted and pumped to a pond approximately 2000 feet away.
The second stage has three 50% recycle pumps. The MgSo_ bleed stream is taken
737
-------
from, the second stage recycle pump discharge and brought to a centrifuge.
The magnesium sulfite cake from the centrifuge is further dewatered and
dried in a cocurrent rotary dryer and stored in silos for transportation
to the calcining facility. Centrate from the centrifuge is used to slake
all magnesium oxide feed.
The dryer off-gas is returned to the second stage of the scrubber vessel.
As originally designed, this stream of 450°F gas was to be returned to
the scrubber outlet duct for reheating the flue gas. However, during
initial operation of the Boston Edison Unit it was discovered that this
was a source of MgO loss as the MgO would go up the stack and fall out
onto the surrounding area.
The silo storage capacity is based upon six days of full load operation.
OPERATION
Initial start-up was in September 1973 and the system was operated until
January 1974. Problems during this period ranged from corrosion leaks
resulting from non-specified material in expansion joints, to minor mechani-
cal problems in material handling equipment, to problems in the feeding
and slaking of MgO, to plugging in the MgO mix tank and suction lines to
the MgO make-up pumps.
Maintenance and modifications were performed on the system from January
to April 1974. The major modification was the addition of a pre-mix tank
in the MgO feed system. The system was again started in April. Operations
from April to July 1974 were limited, however, as PEPCO did not have access
to the EPA calcining facility located at the Essex Chemical Company sulfuric
acid plant in Rumford, Rhode Island. EPA had built and operated the calcining
facility as part of the demonstration of the MgO process at Boston Edison.
PEPCO was to use this facility upon completion of the Boston Edison test
program, but it became apparent that the calcining facility would not be
available prior to July. By the end of June 1974, all the MgO at Dickerson
had been used and a silo and three rail cars were full of magnesium sulfite.
Rumford started calcining the PEPCO sulfite in July 1974. It was discovered
that the sulfite from the coal-fired Dickerson Station was different from the
738
-------
sulfite from the oil-fired Mystic Station, necessitating sorne testing
to determine the proper operating parameters for calciner operation. In
August 1974 the Dickerson scrubber system was started and the MgO slaking
problems appeared to be eliminated by the pre-mix tank. The Rumford
calciner was ready and arrangements were made for trucking the material
between Dickerson and Rumford. A decision was made to use a relatively
low inventory of MgO in order to obtain the maximum number of regeneration
cycles possible in the planned operational period scheduled to end on
December 31, 1974. During this time most of the system operated at 75%
of design gas flow.
A problem developed in the bucket elevator carrying magnesium sulfite from
the dryer to the storage silo. At design load, the buckets, when slugged
with material, tended to overload and trip. The slugging was traced back
to the discharge chute from the centrifuge. Material hung up in the chute
fell into the screw conveyor,thereby slugging the dryer. The dryer held
material a little longer before slugging the bucket elevator, and this
limited continuous operation at design gas flow.
Unit No. 3 was taken out of service in February 1975 for an 8 to 12 week
turbine overhaul. At this point, most of the technical data required for
confirming the process had been obtained. During this period the scrubber
system was inspected and maintenance and modifications made. Unit No. 3
remained shut down until July 1975. Modifications made during this period
were the design of the centrifuge outlet hopper to improve flow, and the
installation of larger buckets on the sulfite elevator.
System maintenance included minor repairs to the flake glass lining in the
vessel, replacement of several sections of first-stage rubber-lined pipe,
and replacement of sections of the second stage carbon steel pipe. All of
this work was based on the minimum necessary to place the system in condition
for a three (3) month run.
The FGD system was re-started in August 1975. Although start-up was smooth,
steam was lost to the MgO mix tank resulting in a very wet product from the
centrifuge. Operation continued, but caking in the dryer occurred. This was
believed to be caused by unreacted MgO carryover due to lack of heat in the MgO
739
-------
mix tank. The dryer was then cleaned and" the steam line repaired. The
system was re-started and again a wet product and caking in the dryer was
experienced. The system was shut down in order to determine the cause and
corrective actions to be taken. It was decided to make some adjustments to
the centrifuge to see if it would extract more water. During this period,
PEPCO decided to test only the first-stage, taking gas ahead of the precipitator.
When this testing was completed an attempt was made to start the second stage
but the centrifuge had frozen, probably due to inadequate flushing on shutdown.
FGD operation at Dicker son terminated at this point.
Operation in the particulate removal mode continued. The objective was to
obtain information on sustained operation for 10 to 20 days. It was previously
reported that problems existed with rubber-lined pipe. This piping was replaced,
and during this run, which was terminated by choice after 20 days, no failures
in rubber-lined pipe occurred.
RESULTS
Results of performance, hardware, process, and economics for two years of
operation are summarized below:
Performance
Test data show that first stage particulate removal is very efficient. The
outlet particulate loading from the scrubber is continually below .02 grains/
SCF dry whether the scrubber is accepting flue gas from ahead of or after the
electrostatic precipitator, and at any flue gas volume up to design flow. Inlet
loading to the scrubber ranged from .06 g/SCF dry to 4.5 g/SCF dry. The SC>2
removal system was designed for 90% removal efficiency when burning 3% sulfur
coal, which equates to 1850 PPM inlet SO2 loading. Although actual operation
above 1700 PPM inlet did not occur, there is no doubt that the system is capable
of meeting design. The SO2 removal efficiency was consistently above 90% when-
ever operating at design gas flow and inlet concentrations above 1000 PPM.
Hardware
Generally, the major items of equipment operated well throughout. The scrubber
vessel, fan, thickeners, centrifuge, and dryer all performed well.
740
-------
A great deal was learned about the materials required for handling MgO
slurry. The carbon steel pipe and slurry feed pumps for this service
were not adequate for the corrosive-erosive action of the slurry. In-
dividual feed pump suction and discharge lines were found to be very de-
sirable. The material handling system was an area which required improve-
roent in items such as the bucket elevator, the centrifuge discharge hopper,
and the weigh belt feeders.
In addition to obtaining test data, every effort was made to achieve maximum
operating hours. The major cause of forced outages during this period were
failures in the first and second stage piping. The first stage piping is
rubber-lined and it was difficult after a failure of the lining to determine
the cause. In some cases, the lining has appeared thinner than specified and
not properly bonded to the pipe. However, there have been no failures in any
rubber-lined piping since replacement. The second stage piping is carbon steel
and the problem was a combination of corrosion-erosion.
The materials of construction in the first stage were specified for low pH and
abrasive material service. The MgO stage was essentially carbon steel and did
not allow for the abrasiveness of MgO nor did it provide for excursions into
the very low pE range when difficulty was experienced in slaking and adding
MgO. Inspection of the scrubber vessel, however, revealed it to be in good
condition. There was no scaling or build-up problems. There were some spots
in the vessel where the protective polyester flakeglass lining was worn. These
were, for the most part, random small spots. In the second stage throat there
was erosion of the lining caused by a modification that was made to increase
pressure drop. This modification distorted the venturi flow. Some failure
spots of the lining were caused by pieces of tramp material damaging the lining.
The wet I.D. fan was very clean.
Following are some of the unknowns that may have had an effect on the wet centri-
fuge product:
1. Internal problem with the centrifuge due to wear or buildup of
material.
2. Change in the MgO. This was the first time that a mixture of re-
generated MgO from the Boston-Edison and PEPOO projects was used.
741
-------
3. Appearance of the centrifuge cake was similar to that of the
tri-hydrate but analyses showed it was still the hexa-hydrate.
It should be noted that the dryer is designed with internal chains to allow
for operation with a wet product. These chains were never installed in the
Dicker son dryer. With chains in the dryer, continued operation with the wet
product should be possible.
Process
The process has demonstrated that it can achieve a 90% or better SO2 removal
efficiency. It can produce commercially salable sulphuric acid, and MgO can
be rerycled and used in the absorber.
Process related problems appeared in two areas:
1. MgO slaking and injection were clearly the most troublesome
areas. It is now known that regenerated MgO and the virgin
MgO have widely varying physical properties. The virgin
material is very dusty and presents buildup problems in the
feed chute. It also resists wetting and agglomerates in
large chunks which would block pump suction lines. The
addition of a pre-mix tank and high speed agitator solved
this problem. The regenerated MgO had different physical
properties, practically no dusting problem, and requires an
elevated temperature to maintain its SO2 absorption capacity.
2. The MgS03 product obtained in the pilot plant had a hexa-hydrate
crystal. At Boston Edison, the tri-hydrate crystal was predominant
in the MgSO^ and initially caused material handling problems.
Although the Dickerson Unit produced a product which was approxi-
mately 90% hexa-hydrate MgSC>3, there were times when this percentage
diminished in favor of the tri-hydrate form, and may have contributed
to the material handling problems.
Economics
As mentioned earlier, one purpose of the demonstration project was to develop
cost estimates to project the economic impact of a full-scale system applied to
742
-------
a unit or total plant. The actual costs incurred during the operation of
the Dickerson units are not too meaningful when quoted in absolute dollars,
but the technical knowledge learned can be input into the second generation
design and gives a utility a yardstick to use in evaluating a proposed design.
To give an idea of the costs of operating the prototype scrubber, PEPCO, in
1974, spent $493,000 in operation and maintenance, and $435,000 in operation
and maintenance of the Rumford facility, including trucking material to and
from the site. In addition, Chemico contributed six months of project manage-
ment and three shift advisors at Dickerson, and EPA retained York Research to
do test work.
For forecasting the economics associated with the MgO process, the existing
Dickerson Station will be used. The Dickerson Station consists of three units
each rated at 190 Mw (gross). The planned Unit 4 addition, rated at 850 Mw
(gross) , is located about one-half mile from the existing plant and is scheduled
for operation in the Spring of 1982.
Following are the assumptions pertaining to the design of a MgO desulfurization
system. No decision has been made to install a desulfurization system; this
is for evaluation purposes only.
1. Flue gas from the three existing units will be headered, passed
through a two-stage scrubber system for particulate removal and SO2
absorption, and the scrubbed flue gas discharged to a single new
tall stack.
2. The new Unit 4 will be equipped with high performance hot electro-
static precipitators with a guaranteed outlet dust loading of less
than .02 grains/SCFD. Space also will be left for the future addition
of particulate scrubbers.
3. The proposed design includes a common facility for MgO storage,
MgSO3 storage, and an on-site sulfuric acid plant. This facility
is located midway between the old and new plants.
4. The assumption is-made that the sale of sulfuric acid will cover the
743
-------
fixed charges, the operation, and the maintenance costs associated with
the acid plant.
5. Capital costs of scrubbing systems for the three existing units, the
one new unit, and common facilities (excluding the acid plant) , are
difficult to determine. The estimated costs in 1975 dollars is in the
range of $130,000,000 to $150,000,000 including the cost of a new 700
foot stack. Based on escalation of 9% per year, Units 1, 2, and 3 in-
service January 1980, and Unit 4 in-service January 1982, the cost
escalated to the $170,000,000 to $200,000,000 range. A problem in
determining the costs of a new unit is to decide what should be included
as part of the scrubber system. For instance: Should any part of the
precipitator be charged to the scrubber, how much of the fan cost, duct
work, foundation, land, etc. be charged?
6. Engineering for four units is done concurrently. Part of Unit 4 cannon
facilities such as foundations will be done with Units 1, 2, and 3.
7. All four units will be assumed to operate at a 75% capacity factor
producing a net output of 8.8 x 10 KWHRS per year.
The problem is to translate these assumptions into the cost of producing
electricity. Based on PEPCO's evaluation of the scrubber system, operation
of Units 1, 2, and 3 will require a minimum of two operators per shift plus
proportionate time of the shift supervisor. The Unit 4 scrubbers which are
located one-half mile from the existing units will also require two operators
per shift and supervision. The common facility will require an estimated
three men per shift. Maintenance for the total complex will need an additional
four men per day. Total complement for stack gas desulfurization is, therfore,
estimated at a rninimum of 39 men.
It is estimated that the system would burn in excess of 30 million gallons of
No. 2 fuel oil per year at Dicker son, but No. 6 could be used for the dryer,
calciner, and reheater at some other site. About 2% - 3% of the station's
output will be required for station service. MgO makeup is estimated at 7%.
Figure 3 is a tabulation of the estimated annual expenses.
744
-------
SUGARY
PEPCO operates three base loaded generating stations in Maryland - the
Dickerson Units, 570 £&;; Chalk Point - 1300 Mw; and Morgantown - 1148 Mw.
It should be noted that all three of these stations meet EPA primary air
quality standards for ground level SO2 concentrations when burning coal
containing approximately two(2)% sulfur. A 630 1-far addition is planned for
Chalk Point in 1930, and an 850 nv addition at Dickerson in 1982. To
install stack gas desulfurization systems at all three of the above plants
for the total planned capacities will require a capital investment of close
to $500,000,000. In addition, it will be necessary to install an additional
133 Mw of capacity to replace the loss of capacity due to scrubbers. This
would require an additional 30 to 80 million dollars, depending on the type
of capacity (base or peaking) installed.
Testing has shown that taking half the flue gas from ahead of the existing
electrostatic precipitator, passing it through a particulate scrubber and
mixing it with the other half of the flue gas for reheat, will result in
compliance with particulate emission regulations for the Dickerson Station.
PEPCOis proceeding to modify the Chemico scrubber system to have it operational
on a full-time basis by the end of 1976 as a particulate scrubber. Scrubbers
will be added to Units 1 and 2 for particulate scrubbing only. These will be
two-stage vessels to allow for the future addition of an SC>2 removal loop, if
required. It should be noted that the current operation of the Dickerson
Station, burning up to two(2)%sulfur coal, is in compliance with EPA primary air
quality standards for ground level S02 concentrations.
PEPCO believes that regulatory agencies should work closely together to achieve
the goal of reliable power with an improved environment for the customer and at
the lowest cost to the customer. PEPCO feels that each power plant should be
assessed on its particular environmental impact, and controls should be required
based only on this impact. For instance, at Dickerson, when its capacity is
expanded by the addition of a new 800 Mw unit now scheduled for 1982, it is quite
possible that ground level SO-, concentrations can be met by scrubbing only a por-
tion of the flue gas. PEPCO believes this would result in substantial capital
savings, lower operating costs, and, most important at Dickerson, considerable
savings in No. 2 fuel oil due to stack reheat being obtained from the mixture of
scrubbed and unscrubbed flue gases in the stack.
745
-------
Figure 1. Two - stage venturi scrubber.
746
-------
POTOMAC ELECTRIC POWER CO. PROTOTYPE
PRECIPITATOR/SCRUBBER - ABSORBER
MgO ADDITIVE SYSTEM FOR S02 RECOVERY
SCHEMATIC PROCESS FLOW SHEET
COAL
v VENTURI
J SCRUBBER/ABSORBER
ELECTRO-STATIC
PRECIPITATOR
MIST
ELIMINATOR
FAN
TO l«o STflOE
TO DRY ASH HANDLING SYSTEM
DUST COLLECTOR
R£(iycc£D POND
WATER
CENTRATE
TANK
CRYSTAL
DRYER
CHEMICO
MgO FROM ACID PLANT
MgSCV, TO ACID PLANT
Figure 2. The Dickerson Regenerative Wet Scrubbing System.
-------
ANNUAL COSTS (1975 $)
$ x 10
-6
MILLS/KWHR
Fixed (15%)
Operating Labor
Operating Supplies
MgO (7% Make-Up)
Maintenance
No. 2 Fuel Oil @ 30<:/gal
Contingency (10%)
Electricity
21.00
.60
1.00
1.20
3.00
9.00
1.48
3.75
2.39
.07
.11
.14
.34
1.02
.17
.43
Sub-Total
41.03
4.67
Fixed Charge on Replacement
Capacity 3.78
.43
TOTAL
44.81
5.10
Figure 3. Tabulation of Estimated Annual Expenses
748
-------
MAGNESIUM OXIDE SCRUBBING AT
PHILADELPHIA ELECTRIC'S EDDYSTONE STATION
James A. Gille
Philadelphia Electric Company
Philadelphia, Pennsylvania
ABSTRACT
The particulate and sulfur dioxide removal system at Philadelphia
Electric's Eddystone Station employs a two-phase approach to air pollution
abatement. Phase I consists of complete particulate scrubbing on
Eddystone Unit 1 (314 MW net) and one-third sulfur dioxide scrubbing
to develop an optimized system for Phase II, which will consist of
complete particulate and sulfur dioxide removal systems on Eddystone
Units 1 and 2.
The Phase I sulfur dioxide removal system uses magnesium oxide
in a recovery system to ultimately produce sulfuric acid. An off-site
MgO regeneration facility is included to close the scrubbing loop.
The system start-up and initial operation have uncovered some
problems to be described in this presentation together with some
project delays.
749
-------
MAGNESIUM OXIDE SCRUBBING AT
PHILADELPHIA ELECTRIC 'S EDDYSTONE STATION
Philadelphia Electric Company embarked on a program to install a
particulate and sulfur dioxide scrubber system at Eddystone Station in
mid -1971. This paper reviews briefly the system design, highlights
the problem and working areas on particulate scrubbing., the problem
areas on sulfur dioxide scrubbing and reviews the projected schedule
for the continuation of this developmental system.
SYSTEM DESIGN
The scrubber plant at Eddystone Unit 1 consists of three parallel
scrubbing trains with a particulate scrubber, reheater and scrubber
booster fan in each train. One train also has the SO 2 scrubber in
series with the particulate scrubber (Figure 1). Flexibility for opera-
tion was incorporated in the ductwork through systems that permit by-
passing the S02 scrubber, each train or the entire scrubber plant, since
this system was retro -fitted between the Unit induced draft fans and
stack.
The particulate and S02 scrubbers have separate surge tanks for
slurry collection, pH control and make-up (Figure 2). These surge tanks
were carbon steel with a polyurethane coating, and the SC>2 scrubber is
also carbon steel with the same coating,, The particulate scrubbers are
316 L stainless steel and the piping systems are all rubber-lined.
blowdown from the S02 scrubber is recovered by dewatering in
a series combination thickener - centrifuge, drying in a rotary-kiln
drier and storage in a silo for shipment to the off -site MgO regenera-
tion facility (Figure 3).
The S02 scrubber incorporates two (2) stages of scrubbing Venturis
and two stages of mist elimination (Figure 4) .
Dry powder MgS03 is calcined in a fluid bed reactor in the off-site
MgO regeneration facility (Figure 5). This facility includes an air
preheater, cyclone separator system, venturi scrubber and gas cooling
tower for the separation of the MgO crystals, and the cleaning and cool-
ing of the SO 2 rich gas for transfer to the su If uric acid plant. This
facility is located at the OLIN Corporation sulfuric acid plant in
Paulsboro, N J , across the Delaware River from Eddys tone . The transfer
of dry material, MgO and MgS03, between Eddystone and Paulsboro is by
truck transportation.
750
-------
OPERATING PERIODS
There was a brief operating period on the particulate scrubbers
from mid-November, 1974 into March, 1975, that uncovered a number of
problem areas, which are outlined later in this paper. Due to a number
of extenuating circumstances the scrubbing was not restarted until
July 23, 1975 with one particulate scrubbing train, 1-C, followed by the
other two trains, 1-B and 1 -A, on August 15 and October 2, respectively.
The S02 scrubber, 1-C, was started for the first time also on October 2,
1975, and the MgO regeneration facility first processed MgSO^ from the
Eddystone scrubber on October 28, 1975.
The 1-C particulate scrubber train has had the most operation with
a total of 2831 hours through January 31, 1976, with an availability of
70% since July 23, 1975. Priority for operation was given to this train,
since the one (1) S0£ scrubber is in series with it, when operated.
The 1-B particulate scrubber train has been operated for a total of
1933 hours through January 31, 1976, with an availability of 55% since
August 15, 1975.
The 1-A particulate scrubber train has been operated for a total of
626 hours through January 31, 1976, with an availability of 2470 since
October 2, 1975.
The 1-C SC>2 scrubber loop has been operated for a total of 556 hours
through December 31, 1975, with an availability of 33% since October 2,
1975.
The MgO regeneration facility has been operated to process all
MgS03 material sent to it from Eddystone. This operation was usually on
a batch basis due to the ups and downs at the Eddystone scrubber.
PROBLEM AREAS - PARTICULATE SCRUBBING
The problems referred to previously in the brief operation period of
November, 1974 to March, 1975, centered around the scrubber booster fans
and the flue gas and liquid conditions.
The 1-B scrubber booster fan, when started for the first time, develop-
ed high shaft vibration after about 1 to 1% hours. Since all three (3)
fans are identical, it was imperative to determine and correct the pro-
blem. The problem was that this fan had an excess clearance of the slip
fit between the shaft and the wheel hub that caused high vibration upon
reaching operating temperature. Fortunately, this fan was not on the
train that has the S02 scrubber, and, while it was returned to the factory
for a shrink fit repair, we concentrated on operating the 1-C train with
the S0 scrubber.
751
-------
The next problem during this early operating period was corrosion
of the 1-C particulate scrubber internals. This scrubber is made of
316 L stainless steel with a 2.75% min. molybdenum content. During a
100 hour run we had experienced a slurry pH level as low as 0.5, and a
chlorides level of 2000 + ppm. These conditions, coupled with the
thermal shock of our "hot start" procedure, caused this corrosion, be-
cause, with caustic addition for pH control at 2.5, higher blowdown to
reduce the chlorides level and a "cold start" procedure the corrosion
has been reduced materially.
About this time we had been checking out the 1-C S0£ Scrubber with
water to check out flow rates and spray patterns. As mentioned before,
this scrubber and its surge tank were carbon steel with a polyurethane
coating, and, after this check-out, we found the coating of these two
vessels was blistering and peeling. An immediate inspection of the
particulate surge tanks revealed a similar problem, although not as
extensive. However, the particulate scrubber surge tanks now operated
at 2=5 pH, while the S02 scrubber surge tank would operate at 6.5 pH.
Thus, we proceeded to blast and recoat the three (3) particulate surge
tanks with a "flake glass" polyester coating. This recoat was delayed
by an operating engineers strike, which forced our re-start into July,
1975.
The other problem areas to be highlighted here have occurred during
our more continuous operation from July 23, 1975 to January 31, 1976.
Circulating Pump Capacity
The circulation pumps for spray supply on the particulate scrubbers
are direct-drive, open impeller, rubber-lined impeller and casing with a
capacity of 1500 gpm. However, the capacity checks indicated only 1200
gpm at rated motor current, which did not adversely affect the 1-B and
1-C particulate scrubbers. The other, 1-A, due to its design, needs full
1500 gpm flow to insure good mixing and fully wetted walls. This problem
is being resolved with the pump manufacturer.
By-Pass Damper Drive Units
One critical point in the by-pass system are the three (3) scrubbing
plant by-pass dampers located between the supply and return headers.
These dampers permit complete, partial or zero flue gas flow around the
scrubbing plant and were designed for two (2) second opening to prevent
upsetting the main unit on a scrubber train malfunction. The drive unit
on these dampers is a pneumatic cylinder operating through a lever arm
to obtain 90° rotation in this short time. The lever arm attachment to
the damper shaft end has failed a number of times following a "scram"
or snap-action. Heavier design lever arms with longer stroke pneumatic
cylinders are being installed at this time, while the main unit is shut-
down for an eight (8) week overhaul.
752
-------
Reheat Burner Refractory Failure
The three (3) reheat burners consist of an oil torch mounted in a
separate combustion chamber at right angles to the flue gas duct upstream
of the scrubber booster fans. The combustion chamber is five (5) foot
in diameter, eight (8) feet long and double-brick refractory-lined. This
refractory has failed due to the criticality of torch placement and start-
up problems with the ultra-violet flame scanners. In addition, insulation
had been placed on a small section of the chamber, which did not allow
heat dissipation and subsequent failure of the refractory and steel shell.
We are still working on this failure.
WORKING AREAS - PARTICULATE
There are a number of areas in the parti culate scrubbing plant that
have been working well and are worthy of notice here.
Scrubber Train Isolation Valves
Each train has an inlet and outlet flue gas valve (damper) for the
purpose of isolating the train while on by-pass operation. These valves
were specified for practically zero leakage in order to do maintenance
on the scrubbers, reheaters, duct instrumentation and booster fans.
These valves are a single 130 inch disk that is lever-guided as opposed
to a center shaft butterfly. The lever-guiding mechanism permits a
single surface seat and disk for sealing, and they have performed very
well. We have had no problems in working on the equipment, while the
main unit was in service.
Parti culate Scrubber Performance
Scrubber performance tests were just completed before the present
outage on the main unit, and test results are expected within the next
week. However, preliminary indications are that the ash quantity going
into these scrubbers is about seven (7) times greater than design due
to poorer coal conditions and deterioration of the mechanical collectors
and electrostatic precipitators on the main unit. The stack appearance,
when all three (3) scrubbers are in service, would say they are doing
quite a job on this quantity of ash.
Mist Eliminator Performance
Performance testing in this area was also just completed and the
test results are also awaited,,
Surge Tank Recoating
The "flake-glass" polyester recoating on the three (3) tanks is
holding up very well so far „
753
-------
PROBLEM AREAS - SO2 SCRUBBING
The short time that the S02 scrubber loop has been in service has
uncovered some mechanical problems and some chemical problems. Due to
the short operating time, only those problems that we have a good handle
on will be reported.
Circulating Pump Motor Base
The three (3) circulating pumps on the S02 scrubber are belt-driven,
open impeller, rubber-lined impeller and casing rated at 6500 gpm. The
belts are of the "poly-V" design and tensioning is with a slide mechanism
under the motor base. This slide mechanism is too light-duty for this
service and has failed. A redesign is being incorporated.
MgSOo Carryover
The SOo scrubber has two stages of Venturis and two stages of mist
eliminator So With both stages of scrubbing in service there is an excess
carryover of slurry from the second stage into the first mist eliminator
catch troughs, causing the overflowing of the mist eliminator tank. A
series of deflectors will be installed to prevent this excess carryover.
MgO Slaking
The MgO powder is slaked in the slaking tank equipped with an agita-
tor and heating coils. The MgO was first slaked with the liquor removed
by the centrifuge, but this caused the formation of grits that plugged
the lines and transfer pumps. A pre-slaker arrangement will be tried
when the scrubber returns to service.
PROJECTED SCHEDULE
All of the problem areas reported have contributed to the extension
of our schedule. In addition we have recently ha d to change the direction
of our S02 removal system "closed loop" due to the shutdown of the OLIN
sulfuric acid plant on December 31, 1975. This means we must relocate
our MgO regeneration facility in order to continue developing the system
for additional installations on the P.E0Co. system.
This relocation of the facility is still being worked out, but it
will probably extend our developmental schedule by almost one (1) year
to mid-1977.
The particulate and sulfur dioxide removal system described here on
Eddystone Unit 1 is the first phase of a two-phase project. Following
the successful line-out of this system with more continuous operation,
it will be incorporated into the design for the complete sulfur dioxide
removal on Eddystone Unit 1, and particulate and sulfur dioxide removal
on Eddystone Unit 2 and one of two units at our Cromby Station.
754
-------
FROM NO. 1
BOILER
Cn
On
BY PASS
1 A PART.
SCRUB.
BY PASS
I I
] PUT. [
I S02 ]
I I
RHTR
IDF
1 B PART.
SCRUB.
I ---- 1
' PUT. I
I I
IDP
RHTR
I ____ I
TO NO. 1
STACK
BY PASS
1 C PART.
SCRUB.
B.
P'^^~"l
S02
SCRUB.
4.. . F?HT
^ r\n i
•-
IDF
SCRUBBER PLANT
Figure 1. Eddystone no. 1 SO^ removal - phase I.
-------
STACK GAS
FROM I..D.
FANS
PARTICULATE
SCRUBBER
HUMIDIFIED
GAS STREAM
S02
SCRUBBER
CLEANED GAS TO
REHEAT AND STACK
Mg (OH)2 SLURRY
PARTICULATE
SCRUBBER
SURGE TANK
1
S02 SCRUBBER
SURGE TANK
-*-TO WASTE WATER
TREATMENT SYSTEM
TO Mg
RECOVERY
Figure 2. SO^ scrubber system - SO^ scrubber.
-------
Mg S03-6H20
SLURRY
TO SUCTION OF
S02 SCRUBBERS
THICKENER
TANK
TO MgO
SLAKING
TANK
Mg S03
STORAGE
SILO
TO MgO
REGENERATION
PLANT
Figure 3, SO^ scrubbing system - Mg 803 recovery.
-------
VENTRI ROD
MODULES
TO SCRUBBER
SURGE TANK
TO SCRUBBER
SURGE TANK
STACK GAS
INLET
BACKWASH
/SPRAYS
ppp] - [oc>qpqp] - [°P°9,°
1>S ' % *»»/^ ' ""XN1 t's '
«? ^ «? ^ «?
-ip-[oopopp] [oppppo] [opppqo]-
- — SM' "r-i"'1' ^rA** '
>
CLEAN GAS
OUTLET
FRSH H20
MIST ELIMINATORS
ENTRAPMENT
SEPARATOR
-1 MOTHER LIQUOR
SCRUBBER
SCRUBBING
SLURRY INLET
TO SCRUBBER
SURGE TANK
Figure 4. S0? scrubber unit detail
*- 758
-------
1750°
SOLIDS
FEED
V
FLUIDIZED
BED
REACTOR
(REFRACT-
ORY
LINED)
\\\\\\\\\\\
FUEL OIL
AIR
PREHEATER
K
1000'
AIR
COMBUSTION
AIR BLOWER
1200
S02 RICH GAS
164'
CYCLONE\7
RECOVERY
SYSTEM
TO
RECOVERED
MgO STORAGE
BLOW DOWN
VENTURI
WET
SCRUBBER
.MAKE-UP
^° WATER
Figure 5. MgO regeneration plant.
-------
INTERIM REPORT ON CHIYODA THOROUGHBRED 101
COAL APPLICATION PLANT AT GULF POWER'S SCHOLZ PLANT
Richard B. Dakan
Chiyoda International Corporation
Seattle, Washington
Reed A. Edwards and Randall E. Rush
Southern Services, Inc.
Birmingham, Alabama
ABSTRACT
This paper presents a summary of the operating experience during
1975 of Chiyoda International Corporation's 23 megawatt flue gas
desulfurization plant at Gulf Power Company's Scholz Steam Plant.
Emphasis is placed on system availability, chemical consumptions,
solid and liquid waste disposal, mechanical problems and their resolution,
and planned future testing. Also included is an appendix showing
construction costs and utility and chemicals consumptions for a hypo-
thetical 500 megawatt FGD facility. Testing and evaluation of the
Scholz Plant are under the direction of Southern Services, Inc.
761
-------
INTERIM REPORT ON CIIIYODA THOROUGHBRED 101
COAL APPLICATION PLANT AT GULF POWER COMPANY,1 S SCHOLZ PLANT
I. INTRODUCTION
The Chiyoda THOROUGHBRED 101 (CT-101) FGD system at Gulf
Power Company's Scholz Plant in Sneads, Florida has now
completed its first ten months since initial startup in
February, 1975. As Chiyoda's first FGD plant in the U.S.
and its first to operate on a coal fired boiler, this
CT-101 is both a demonstration and a testing facility.
The technical capabilities of this process have been well
demonstrated by the thirteen commercial FGD systems currently
scrubbing stack gases from oil fired boilers in Japan.
The first of these plants, which range up to 350 mw, came
on line in November, 1972, and the most recent in October,
1975. (Table 1). Initial operations of these plants were
reported previously in the literature. (1,2)
At the Scholz Power Plant emphasis is on testing the per-
formance of CT-101 on scrubbing flue gases generated from
a coal fired boiler, for comparison with the performance
on oil generated flue gases in Japan. The testing and
evaluation is being done under the direction of Southern
Services, Inc. for the Southern Company (an electric
utility holding company including Alabama Power Company,
Georgia Power Company, Gulf Power Company, Mississippi
Power Company, Southern Electric Generating Company,
and Southern Services Inc.). This evaluation is part of
a program to evaluate three prototype "2nd generation"
flue gas desulfurization processes. This report presents
the operating experiences of Chiyoda's facility during
1975.
II. PROCESS DESCRIPTION
The 23 mw capacity of the desulfurization unit corresponds
to approximately one-half of the rated boiler capacity.
The design basis of the CT-101 at the Scholz Plant is
shown in Table 2. At 100% of design basis the plant
consumptions are: 1800 Ib/hr limestone, 60 gpm water
(excluding recyclable cooling water), 50 gal/hr No. 2
fuel oil (reheat to 300°F), 50 Ib/hr catalyst, and 1200 kw.
The process has four basic operations (refer to Figure 1):
1) Prescrubbing
2) Absorbing
3) Crystallizing
4) By-product/waste handling
762
-------
TABLE 1
THE CHIYODA THOROUGHBRED 101 FLUE GAS DESULFURIZATION PLANTS
Owner of Plant
Nippon Mining
Fuji Kosan
Mitsubishi Rayon
Tohoku Oil
Daicel
Amagasaki Cokes
Hokuriku Electric Power
Mitsubishi Chemical Ind.
Mitsubishi Petrochemical
Mitsubishi Petrochemical
Denki Kagaku Kogyo
Hokuriku Electric Power
Toyama Cooperative Power
Location
Mizushima
Kainan
Otake
Sendai
Aboshi
Hyogo
Toyama
Yokkaichi
Yokkaichi
Kawa j iri
Chiba
Fukui
Toyama
Capacity
(scfm) (MW)
20
98
56
8
62
15
466
250
93
435
75
650
466
,800
,000
,000
,800
,000
,500
,000
,000
,000
,000
,000
,000
,000
-
-
30
-
33
-
250
135
50
235
40
350
250
Avail-
ability
95%
97%
97%
100%
96%
97%
100%
94%
99%
100%
100%
100%
100%
Period of Availability
May '
May '
May '
June
Dec .
July
Oct.
July
Dec .
Dec .
June
June
Sept.
73 to Dec. '75
73 to
73 to
'73 to
'73 to
'74 to
'74 to
'74 to
'74 to
'74 to
'75 to
'75 to
'75 to "
-------
ON
-£•
Table 2 SUMMARY OF DESIGN AND OPERATING CONDITIONS •
Conditions Design Operating Range
CT-101 FACILITY AT THE SCIIOLZ PLANT
Typical Values Future Conditions
Gas Rate, scfm
SO2 Loading, ppm
Inlet
Outlet
S02 Removal, %
Particulate Loading,
Inlet
Outlet
Particulate Removal,
o2, %
H20, %
Excess air, %
Fuel
Sulfur content,
HHV, BTU/lb
Ash content, %
Bottom Ash, %
Limestone
Purity, % CaC03
Max. % MgCO3
Min. Particle Size,
% 200 mesh
Catalyst
% Fe+++
53,000
81,000 acfm
@ 350°F
2250
225
90
gr/scf
0.1
0.01
% 90
3.0
9.0
20.0
% 3.0
12, 200
14
15
98
2
90b
20,000-55,000 35,000-45,000 40,000-50,000
800-2000 1000-1300 2000-3500
50-250 90-150
80-95 82-87
1.0
0.02
98
6-10 7-9
1.7-2.5 1.7 & 2.2a
11,700-13,100 12,500
10.6-13.7 12.3
93-95 94
1-1.9 1.3
a. 1.7 during 2/75 through 7/7
c _cc 2-2 during 9/75 through 12/
bb °q '" b. wet sieve analysis
c. dry sieve analysis
7.8-10.7 io.O
-------
Limestone
Silo
Waste
Disposal
Waste
Disposal
Mother
Liquor
Tank
Gypsum
Figure 1 CT - 101 flow diagram
Gulf Power - Scholz Steam Plant
-------
The prescrubber section is responsible for cooling the
incoming flue gas to its saturation temperature and for
scrubbing out fly ash and chlorides. The suspended solids
concentration in the prescrubber recycle liquid is con-
trolled by an intermittent discharge of settled ash from
a thickener.
Sulfur dioxide is removed via counter current scrubbing
with weak sulfuric acid in the fixed bed absorber. The
absorption of SO? by t^O gives sulfurous acid (E^SO-,)
which is catalytically oxidized by ferric ion to H2SC>4
in the oxidizer. (Equations 1, 2).
HO + S02 —-> H2S03 (1)
H2S03 + h02 ^g$+H2S04 (2)
This scrubbing liquid is not saturated in calcium sulfate,
obviating scaling in the absorber/oxidizer or mist elim-
inator. The flue gas then passes through a two rank
chevron type mist eliminator before reheat is added, and
exits through a local test stack. This mist eliminator
has never been washed and pressure drops are consistently
less than 0.2 inch H20.
The concentration of sulfuric acid in the scrubbing liquid
is maintained at about 2.0 wt% by continuous withdrawal
of absorbent to the crystallizer. In the crystallizer,
this absorbent is partially neutralized to about 1.0 wt%
1*2504 by limestone, the reaction product being gypsum
crystals. (Equation 3).
H2S04 + CaCO3 —-> CaSO4 + H2O + C02 (3)
Gypsum is removed from the system by a continuous screw
decanter type centrifuge as an 80-88% solids cake. This
solid by-product is currently being held in a dry pond.
The overflow from the crystallizer and underflow from
the centrifuge pass to a clarifier where small gypsum
crystals settle out and are returned as seed crystals to
the crystallizer. Overflow from the clarifier is essen-
tially free of suspended solids and is returned to the
absorber as mother liquor.
There are two waste streams in the system presently.
One is fly ash bleed from the prescrubber section and the
other is a mother liquor bleed from the crystallization
section for water balance and for control of the chloride
concentration in the absorbent. Maintaining chloride
concentrations in the absorption/crystallization section
below 200 ppm is presently necessary in the CT-101 process
at Scholz to prevent pitting corrosion in the stainless
steel vessels. These two waste streams are combined,
neutralized by limestone, and discharged to a wet pond
for removal of suspended solids before overflow into the
plant ash pond.
766
-------
III. EXPERIENCE
A. Availability
Availability as defined here is the hours of scrubber
operation divided by hours of boiler availability to
the scrubber, times 100. Table 3 shows the periods
of plant operation and the reasons for outages since
initial startup with flue gas in February, 1975. The
first few months shakedown period was characterized
by many small problems resulting in a very low avail-
ability (30% February - May). Many of these problems
were resolved by June, and availability during June
and July was somewhat improved (84%) but still below
expectations. Many of these early outages were perhaps
longer than would normally be the case, because the
maintenance crew was overburdened and working with
new equipment and a new process. Additionally, oc-
casional delays in repair occurred while the exact
nature of a problem was investigated by Chiyoda en-
gineers and manufacturers' representatives. In early
July, a separate maintenance contract was issued to
increase the available maintenance forces for the
FGD systems and, in general, repairs were completed
more quickly. The six-week outage starting in August
was the result of unbalancing of the centrifuges.
After repair of the centirfuges and installation on
September 15, the availability through December, 1975
averaged above 97%.
B. Equipment
During the shutdown of August and September most of
the vessels and rotating equipment were inspected.
These inspections verified that there was no scaling
or plugging in the absorber, oxidizer or mist eliminator
The crystallizer showed some settled solids on the
bottom but no evidence of scaling.
The unbalancing of the centrifuge is primarily caused
by a high non-uniform wear rate on the screw, and
this continues to be a problem. The Stellite No. 1
hard facing on the screw conveyors is wearing much
faster than experiences in Japan predicted, although
one unit has now operated continuously for four months
without becoming unbalanced. It appears that this
relatively long run without failure is due to even
wearing of the conveyor screw, thereby avoiding unbal-
ancing. Previously, one centrifuge became unbalanced
after only six weeks operation, but this may have been
an anomalous case. Hardness testing on one conveyor
showed that the Stellite No. 1 facing hardness was
less than that of the stainless steel conveyor. The
reason for this may be that the technique for applying
the facing is different from that in Japan. As a
767
-------
Table 3 OPERATION HISTORY
in
operati
'75
F
E
B
M
A
R
C
H
A
P
R
T
-L
L
M
A
Y
J
U
N
E
——
J
U
L
Y
A
U
G
'y/
&
Y/
vv
te.
7/
Y/
j ,
' //
^
/ /
///
v/
/X
^
7s
V
v//
//
///
A
on*
Outage
Hours
387.33
342.75
241.00
15.00
500.17
442.23
71.58
34.15
14 .00
1044 .22
Cause of Outages & Remarks**
Feb. 11 Initial catalyst charge
12 Crack in FRP oxidizer vessel, additional
mats glued.
March 11 Modification of Waste Disposal. Waste
neutralization moved to pond from plant
site. The move was necessitated by insuf-
ficient capacity of the disposal pump and
by scaling in the discharge line.
April 3 Prescrubber shell flange leakage; bolts
tightened .
18 FRP line leakage
21 Flue gas blower inlet vane broken.
Replaced corroded S.S. parts in Prescrubber
area with corrosion resistent material.
May 15 Prescrubber FRP lined nozzle developed a
pinhole leak. Leakage on FRP piping of
pump discharge. Checked oxidizer feed
pump bearing, thought to be over heating
(major reason for outage) , instrument air
compressor alignment & oxidizing air in-
strumentation .
July 6 Limestone feeder jammed by 7/8" nut,
inadvertently entered with limestone powder
21 FRP pipe leakage on Prescrubber.
28 One centrifuge unbalanced.
Aug. 3 Both centrifuges unbalanced; shipped to
a machinery in Pensacola which failed to
repair and sent to the manufacturer in
Houston .
768
-------
Table 3 OPERATION HISTORY
(continued)
'75
S
E
P
T
0
C
T
N
0
V
D
E
C
'//,
'/S
///
//
W,
%
y/
1
fy
1
V
y/
'//
//
v/
Y/.
//,
'//
//
/ •
//
47 . 0
0.67
0.17
0.33
1.5
1.3
14.7
Sept. 23 Sump pump misalignment; changed
spare pump
to
Nov. 10 Flue gas blower inlet vane broken;
removed broken vane .
24 Prescrubber FRP lined nozzle pir
welded from outside.
Dec. 8 ditto
18 ditto
20 ditto
iho.
26 Prescrubber nozzle lining repaired
inside .
Total system availability, including shakec5
period, February 11, 1975 through December
owr
31
was 60%.
* Shaded area indicates period of operation.
** Occasionally after extended outages, there
were problems with the balancing of the F.D.
fan when starting up. These problems were
due to corrosion of the carbon steel and the
fan impeller, most likely caused by leakage
of hot flue gas around the isolation damper.
Delays in startup due to rebalancing of the
fan are not listed as the cause of outage,
but are included in the outage hours.
769
-------
result of these findings, new hard facing materials
have been investigated in the event that the per-
formance of the current one cannot be improved.
Depending on the results of these investigations, a
new facing material may be used when re-facing is
next required. It should be noted that, while the
use of centrifuges is not unique to the Scholz facility,
in its large FGD plants Chiyoda usually uses less
troublesome basket centrifuges or vacuum filters.
There have been few pump problems during the year.
Plant operation stopped once in September due to a
failure of the disposal pump which did not have a
standby spare. The pump was corroded due to inadequate
sealing and flushing. Pump problems not causing process
upsets include unbalancing of the impeller in the
absorber/oxidizer recirculation pump, tripping of the
limestone slurry pump due to excessive slurry con-
centration, and breakage and slippage of drive belts.
The unbalancing of the oxidizer pump impeller was due
to a defect in the rubber liner and subsequent cor-
rosion. The belts are being changed to polybelts
in an attempt to improve operating life.
Other equipment problems include occasional jamming
of the limestone feeder by "tramp" metal from the
silo, and leakage from the flanges on the FRP duct
between the prescrubber, absorber, and mist eliminator.
There have been many leakage problems with the FRP
used in the plant, usually attributable to excessive
vibration, inadequate bracing, or poor installation.
C. Instruments
Instrument problems have been few, but irksome. Most
notably, the prescrubber level transmitter and the
limestone and gypsum slurry analyzers have been un-
reliable. The problems with the level transmitter
have been traced to an improper gasket material on
the sealed system, and have now been resolved. The
limestone slurry analyzer was a differential pressure
type intended to provide only a rough indication of
concentration but did not perform properly because
of mis-specified components. It is now working after
modifications. The gypsum slurry analyzer was an
ultrasonic unit which failed during the first three
months of operation due to erosion of the probes and
has not been used since. The slurry analyzer problems
were easily circumvented by manual observation and
measurement, although this action has not allowed
optimal operation.
770
-------
D. Process
1- Sulfur Dioxide Removal
Correlations of S02 removal efficiencies with
process parameters have been thoroughly investi-
gated in Japan during the course of development
and commercial operations, and the optimal operat-
ing conditions are well defined. Due to the long
shakedown period and the fact that the Gulf Power
operators were learning three FGD systems simul-
taneously, operating conditons have been maintained
within loose guidelines rather than at specified
setpoints. As a result, process performance has
occasionally been below design specifications.
During periods of proper operation it has generally
equalled design specifications.
Throughout most of 1975 only relatively low sulfur
coal was available at the Scholz Plant, and typical
inlet SC>2 concentrations were 1000-1300 ppm. S02
removal efficiencies have ranged from 80-95%, depend-
ing on inlet SC>2 loading, absorbent concentration,
and L/G. Performance in December of 1975 showed
SC>2 removal efficiencies of about 82% overall.
However, SC>2 removal efficiency data obtained to
date disagree with Chiyoda observations and this
discrepancy is being investigated. The outlet S02
analyzer in particular has given intermittent problems
and is suspected to be the cause of these differences.
The result of this investigation will be reported
in the future.
2. Particulates and Chlorides Removal
The prescrubber section was designed to remove fly
ash and chlorides from the flue gas before it enters
the absorber. Normal inlet particulate concentra-
tions have been quite low, however, adequate testing
has not been done to verify actual values. Testing
to determine chloride removal has been performed,
but the results are inconclusive. Further tests
will be conducted in the future in an attempt to
determine accurate particulate and chloride removals.
3. Limestone Utilization
Material balance and solids analyses generally
indicate a 98-99% calcium utilization, although some
individual results indicate utilizations as low as
96%. The limestone particle size used at the Scholz
facility is of a larger mesh size than specified
(76%<200 mesh instead of 95%<200 mesh), and it was
771
-------
suspected that this would give a lower than
normal utilization. The fact that utilization
has not decreased may reflect differences in
particle size analysis which is done by dry sieve
at the Scholz Plant, while Japan experiences are
based on a wet sieve analysis. A small amount of
unreacted limestone in the gypsum product is useful
to neutralize residual acid in the 10-15% moisture
carried with the dry gypsum. The limestone used
at the Scholz Plant comes from the Alabaster area
near Birmingham.
By-product Disposal
Liquid Waste. As the plant is currently operating,
it is not a closed loop. The discharge is approx-
imately 20 gpm, and disposal of acidic water poses
a problem. At the present time, this water is
neutralized with limestone, then sent to a lined
settling pond where fly ash, gypsum, unreacted lime-,
stone, and Fe(OH)3 (catalyst from mother liquor
bleed) settle out.
The overflow from the settling pond has been sampled
and analyzed monthly since July, 1975. The analyses,
as shown in Table 4, include trace elements as well
as major constituents. Concentrations of total dis-
solved solids, iron, mercury, and fluoride in excess
of state water quality standards for receiving waters
have been measured in the settling pond overflow.
Selenium concentrations in excess of U.S.P.H.S.
Drinking Water Standards have also been measured
in this overflow. However, the settling pond over-
flow is routed to the Scholz Plant ash pond which
ultimately discharges to the Apalachicola River
where state water quality standards are applicable,
and it is important to realize that there have been
no violations of state water quality standards,
as determined by monthly monitoring of the ash pond
overflow.
At a full-scale installation, it is likely that
treatment and/or reduction of the liquid waste will
be required by both state and federal regulations.
Treatment could be accomplished by evaporation if
the volume of waste is reduced. Serious considera-
tion is currently being given by the EPA to demon-
strating vapor compression evaporation on the Chiyoda
liquid waste at the Scholz Plant. Plans to reduce
water consumption and the liquid waste stream and
approach a closed loop are being implemented as
described below.
772
-------
Table 4 CIC EFFLUENT ANALYSES - JULY-DECEMBER 1975
Parameter (Units!
Liquid Purge
Settling Pond
Overflow
Gypsum Pond
Underdra.) n
State of Florida
Water Quality
Standards (1)
Total Dissolved Solids
(mg/1)
Total Suspended Solids
(mg/1)
pH (Standard Units)
Conductivity
(Micromhos/cm)
Temperature (°F)
2000
6
2.3
3050
47
-4994(2) 2250-2754(2) 500(3)
-393(4) -1-4
-7.5 (2) 6.8-7.2 6-8 .5
-6700 2000-2670 500
-85 55-88
Total Calcium
(mg/1 as Ca)
Total Magnesium
(mg/1 as Mg)
Total Sodium
(mg/1 as Na)
Total Potassium
(mg/1 as K)
Total Hardness
(mg/1 as CaCO3)
Total Phosphorus
(mg/1 as P)
Dissolved Silica
(mg/1 as SiC>2)
330-1100
39-246
28-65
0.85-2.93
1093-3360
<0.. 01-0. 036
12-31
1160-2750 (!
Sulfate (mg/1 as S04 )
Sulfite (mg/1 as S03) <1
Carbonate (mg/1 as CaC03) 0
Bicarbonate (mg/1 as CaC03) 0-38
Hydroxide (mg/1 as CaC03) 0
Chloride (mg/1 as Cl) 51-138
540-644
14-87
12-67
0.47-1.65
1406-1830
<0 .01
1.5-2.3
1370-1650
<1
0
35-44
0
2.3-9.1
250
773
-------
Table 4 CIC EFFLUENT ANALYSES - JULY-DECEMBER, 1975
(continued)
Parameter (Units)
Liquid Purge
Settling Pond Gypsum Pond
Overflow Underdrain
State of Florida
Water Quality
Standards (1)
Carbon Dioxide
(mg/1 as C02)
Total Acidity
(mg/1 as CaCOj)
Color (Standard Units)
Turbidity (NTU)
Total Aluminum
(mg/1 as Al)
Total Arsenic
(mg/1 as As)
Total Cadmium
(mg/1 as Cd)
Total Chromium
(mg/1 as Cr)
Total Copper
(mg/1 as Cu)
25-970
0-1080
0.5-13
4.2-92.0 (2:
2.10-4.80
<0.01
<0.01
<0.01-0.04
4.8-12.0
0
0.5-8
0.18-1.6
<0. 05-0. 24
<0.01
0.06-0.60
(2)
•CO. 01
<0.01
Total Iron (mg/1 as Fe) 0 . 54-109.0 (2) 0.019-0.17
Total Lead (mg/1 as Pb) <0.01 <0.01
Total Manganese
(mg/1 as Mn)
Total Mercury
(mg/1 as Hg)
Total Nickel
(mg/1 as Ni)
Total Selenium
(mg/1 as Se)
0.17-1.08
. 01-0. 32
<0. 0002-0. 0068 (2) <0. 0002-0. 0004
0.02-0.15
<0.01
0.016-0.13 (9) <0.002-0.04
Total Zinc (mg/1 as Zn) 0.26-1.3^ <0.01
Oil and Grease (mg/1) <1 <1
50
0.05
0.05
0.5
0.3
0.05
(71
ND
1.0
15
774
-------
Table 4 CIC EFFLUENT ANALYSES - JULY-DECEMBER, 1975
(continued)
Liquid Purge State of Florida
Settling Pond Gypsum Pond Water Quality
Paremeter (Units) Overflow Underdrain Standards (1)
Nitrate (mg/1 as N) 20-105(10) 4-43
Chemical Oxygen Demand
(mg/1) <1-11
-------
By-product Gypsum. The gypsum produced at the
Scholz Plant consists almost entirely of calcium
sulfate dihydrate (CaS02 • 2H20). It is easily
dewatered by centrifuging, and the solids content
is typically 80-85%. The gypsum has been physi-
cally characterized using three soil tests, grain
size analysis, consolidation, and Atterberg limits.
The grain size analysis indicates that the Chiyoda
gypsum is composed of 28% sand size, 66% silt size,
and 6% clay size material (MIT classification
system). The majority of the gypsum particles
ranged in size from 0.01 mm to 0.1 mm. From con-
solidation tests, it was learned that landfilled
Chiyoda gypsum could support significant loads with-
out appreciable settlement. However, loads subject
to vibration could not be placed on a gypsum landfill
because the gypsum tends to liquify with vibration
if the water content of the gypsum does not remain
low (approximately 15% or less) . A range of per-
meabilities was calculated from the consolidation
tests to be 10~6 - io~5 cm/sec. Atterberg limits
testing revealed that the Chiyoda gypsum is non-
plastic and that it will liquify when it is both
wet and subjected to vibration.
Currently the gypsum is being disposed in a lined
pond equipped with an underdrain. Rain water per-
colated through the gypsum, passes through the
underdrain, and is routed to the ash pond. The
water from the underdrain has been sampled monthly
since July, 1975, and the results of the analyses
are shown in Table 4. In general, with the possible
exception of total dissolved solids, it appears
that this water (leachate) poses no water quality
problems. This is to be expected, because almost
all trace contaminants contained in the flue gas
are removed in the prescrubber. The liquid waste
from the prescrubber is neutralized and routed to
the settling pond as described earlier.
IV. FUTURE PLANS
A. Reliability Testing
A six month performance reliability test will be
completed at the end of March, 1976. The final
testing to be done during this period will include
continuing material balance checks and ability of the
process to follow load fluctuations. Additional tests
will also be conducted on particulate and chloride
removal capabilities of the prescrubber.
776
-------
B. Testing with Various Coals
As the Scholz facility is a proving ground for CT-101
applicability to coal generated flue gas, tests
related to variations in coal are being undertaken.
The present Southern Company test program is scheduled
to end in June. However, during the next one or two
years, the plant will continue operation with different
coals, permitting further testing by Chiyoda of higher
sulfur and possibly higher ash coals.
C. Reduction of Clear Water Consumption
As mentioned earlier, there are two liquid waste
streams within the plant, one from the prescrubber
and the other from the mother liquor for controlling
chloride concentrations and for water balance. Present
plans are to reduce these streams and approach closed-
loop operation. The design of the precooler (saturator]
in the prescrubber calls for a flow rate of clear water
equal to twice the rate of evaporation. The excess
water is intended to prevent scaling in the precooler
and to convey fly ash. Inspection of prescrubber
conditions indicates that all of this excess water is
not necessary to prevent scaling and may be eliminated.
In addition, clear water is being used in preparation
of the limestone slurry, and this water has been re-
sponsible for half of the mother liquor bleed for water
balance; the remaining half comes from pump seal water
which cannot be eliminated.
The objective of the water saving plan is to demon-
strate the reduction in the volume of liquid waste
accomplished by utilizing mother liquor as make-up
water to the limestone slurry tank and as the convey-
ing medium for fly ash bleed from the prescrubber
(refer to Figure 2).
The addition of mother liquor to the limestone slurry
tank will eliminate the use of clear water there,
should obviate bleeding mother liquor to maintain the
water balance, and will retain some ferric ion catalyst
currently being lost in the bleed stream. On the
other hand, sufficient mother liquor must be bled to
the prescrubber to maintain the chloride concentration
in the absorption/crystallization section below 200 ppm,
Figure 3 shows the required mother liquor bleed rate
to the prescrubber as function of chloride content of
the coal and the chloride removal in the prescrubber.
The addition of mother liquor to the prescrubber and
to the limestone slurry has already been successfully
implemented at a chemical company in Japan but will
still be carefully monitored at the Scholz Plant.
777
-------
Prescrubber
Section
00
1
Absorber/Ox id izer
Section
|
Limestone
Slurry
Preparation
Crystallization
Section
Gypsum
To Waste
Disposal
Figure 2 Block flow diagram
indicating water saving plan (dotted lines)
-------
30 .
a
Cn
-P
C
Q)
e
0)
tJ1
CD
-0
QJ
0)
H
PQ
20
10
Figure 3 Mother Liquor Bleed
vs
Chloride in Coal
Parameter: Removal % at Prescrubber
90
93
90
93
Calculation Basis:
1. Coal Burning Rate, 24,000 Ib/hr
2. All Chlorides in Coal enters CT-101
3. .-•—Chloride in ML, 150 ppm
—— Chloride in ML, 200 ppm
Chloride in ML, 300 ppm
0.1 0.2
% Cl~ in Coal
779
-------
Chiyoda's laboratories have also noted that NO^ ion
in solution is an inhibitor to pitting corrosion by
chlorides. Nitrogen oxides are absorbed in the CT-101
up to a saturation limit, and if the concentration
of NC>3 is sufficient, the maximum permissible con-
centration of chloride may increase significantly,
thus permitting further reduction of the bleed rate
to the prescrubber (if the water balance permits).
The final process blowdown from the facility is
expected to be 10 gpm or less.
D. Implementation Schedule
Testing of the water saving plan, which will be
initiated after completion of the six month relia-
bility test, calls first for addition of mother
liquor to the prescrubber and the limestone slurry
preparation system. Limestone slurry will be pre-
pared exclusively from mother liquor and the effect on
slurry preparation carefully monitored for a few months
Clear water will be used for cooling and saturating
the incoming flue gas, and mother liquor will be
added to the recycled scrubbing liquid for removal
and discharge of fly ash and chloride.
A subsequent test will involve stopping all addition
of clear water to the precooler and using only re-
cycled scrubbing liquid for both saturating and
scrubbing incoming gas. In this mode, the waste
discharged from the system will be minimized.
E. By-product Treatment and Utilization
As the gypsum is potentially a useful by-product and
a waste problem, future testing will investigate both
of these possibilities.
If total dissolved solids in the gypsum leachate pose
a problem at a full-scale installation, leaching
could be minimized by lining the disposal area or
fixating the gypsum. The cost of conventional liners
is high, so studies are underway at the Scholz Plant
to test usage of compacted gypsum layer as a liner.
Approximately one foot of gypsum will be placed in
a 20 ' x 50' level area and will be compacted using
conventional compaction equipment. Core samples of
the compacted layer will be taken, and their perme-
ability will be determined in the laboratory. Approx-
imately three more feet of gypsum will be placed on
top of the compacted layer, and the gypsum and gypsum
liner will be allowed to weather. Core samples will
be taken periodically to determine if liner perme-
ability varies with time.
780
-------
Chiyoda gypsum has been laboratory tested by a ^wall-
board, company and found to apparently be suitable
for wallboard manufacture through small scale tests.
A demonstration test using 100 tons of Chiyoda gypsum
will soon be conducted at a wallboard plant in Jack-
sonville, Florida.
Agricultural use of gypsum as a calcium source for
peanuts is fairly well established in the southeastern
portion of the United States. Testing the suitability
of the Chiyoda gypsum for agriculture began in December,
1975 with soil incubation studies at the University of
Florida Agricultural Research Center in Quincy, Florida.
The first month of soil incubation indicates that
application rates up to twenty tons of Chiyoda gypsum
per acre of soil will be feasible. The results of
plant response tests will soon be available because
test plants (peanuts and soybeans) were planted in
soil-gypsum mixtures in early February, 1976.
The suitability of Chiyoda gypsum for cement production
is being tested by a cement company. Preliminary
observations indicate that Chiyoda gypsum can be used
as a cement retarder. Results of these tests will be
reported later.
All Chiyoda gypsum produced in Japan is sold to wall-
board and cement manufacturers.
781
-------
APPENDIX
This appendix, while it does not relate directly to
Chiyoda's Scholz Plant, is included here to provide
a realistic estimate of construction cost and utilities
and chemical consumptions for a 500 MW FGD facility.
The process flow scheme for a 500 MW plant is the same
as at Scholz except there would be two trains and a
catalyst recovery section with the larger facility.
Table 5
CONSTRUCTION COST AND UTILITIES AND CHEMICALS CONSUMPTION
FOR A HYPOTHETICAL 500 MW CAPACITY FGD FACILITY
Design Basis:
Gas Flowrate 1,120,900
Gas Temperature, inlet 310
S02, inlet 2,200
Particulates, inlet 0.1
Cl~ in coal 0.1
S02 removal efficiency 90
Particulate removal efficiency 90
Reheated gas temperature 200
Utility and Chemical Consumptions:
Power 17,000
Make-up water 647
Recoverable cooling water 2,500
Limestone (as 100% CaC03) 35
Catalyst (as 100% Fe2(SO4)3) 100
Fuel oil for reheating 14
By-product Gypsum, including 12%
Moisture 70
Waste Water 262
Construction Cost $38,000,000
scf m
Op
ppmv
grain/scf
wt. %
OF
KW
gpm
gpm
s-ton/hr
Ib/hr
gpm
s-ton/hr
gpm
782
-------
REFERENCES
Akiyoshi Tamaki, "Commercial Application of Dilute
Sulfuric Acid Gypsum (The Chiyoda THOROUGHBRED 1011
Flue Gas Desulfurization Process for Large Power
Plant Boilers", presented at the Sixty-Seventh
Annual Meeting, American Institute of Chemical
Engineers, December 4, 1974.
Masaaki Noguchi, "Status Report on the Chiyoda
THOROUGHBRED 101 Process", presented at United
States Environmental Protection Agency, Flue
Gas Desulfurization Symposium, November 4-7, 1974.
783
-------
ADVANCED PROCESSES SESSION
Chairman: Archie V. Slack
President
SAS Corporation
Sheffield, Alabama
785
-------
STATUS AND ECONOMICS OF THE ATOMICS INTERNATIONAL
AQUEOUS CARBONATE FLUE GAS DESULFURIZATION PROCESS
D. C. Gehri and R. D. Oldenkamp
Atomics International
Canoga Park, California
ABSTRACT
The Aqueous Carbonate Process (ACP) is an advance flue gas desul-
furization process which produces elemental sulfur as its byproduct.
The ACP system can be used for desulfurization of gases containing 200
to 20,000 ppm sulfur dioxide. Removal of SO is accomplished in the
ACP gas-treating subsystem, which is operationally decoupled from the
regeneration subsystem. Either petroleum coke or coal is used as the
reducing agent in the regeneration subsystem.
The decoupled nature of the process and the absence of recycle
complications made it possible to adopt a stepwise approach to process
development. All of the key process steps have been tested on the
pilot plant scale, and the regeneration cycle has been proven by
sequential operation of its individual steps. Extensive system engineering
studies have been performed, and a proposal to design, build, and
operate a 100 Mw ACP demonstration system has been prepared.
An economic evaluation of the ACP system has been performed by AI
with the help of the Stack Gas Emission Studies staff of the TVA. The
results of that evaluation indicate that ACP capital costs are competi-
tive with other sulfur-producing FGD processes, and total annual operating
costs are comparable to those of lime or limestone "throwaway" processes.
787
-------
STATUS AND ECONOMICS OF THE ATOMICS INTERNATIONAL
AQUEOUS CARBONATE FLUE GAS DESULFURIZATION PROCESS
INTRODUCTION
The Aqueous Carbonate Process (ACP) was conceived and developed
as an advanced flue gas desulfurization (FGD) process with features
that would solve or eliminate the problems encountered by first gen-
eration FGD processes. The features incorporated in the ACP system
are summarized below:
1) Sodium carbonate solutions were selected as the desulfuriza-
tion media in order to provide effective sulfur dioxide
removal over wide concentration ranges.
2) The equipment chosen for the ACP gas-cleaning subsystem was
selected on the basis of operational reliability and
minimum maintenance requirements.
3) Potential problems associated with plugging, scaling, or
other scrubber malfunctions were minimized, both by the
nature of the scrubbing media and by the type of equipment
selected.
4) For most applications, flue gas reheat requirements were
eliminated.
5) The ACP gas-cleaning subsystem was designed to be decoupled
from chemical regeneration steps and thereby become the sole
operational interface with the flue gas source.
6) A sulfur-producing regeneration subsystem was provided in
order to minimize the handling and/or disposal problems
associated with the byproducts of the FGD process.
788
-------
7) The ACP regeneration subsystem was designed to use a solid
carbonaceous reducing agent (e.g., petroleum coke or coal)
and thereby eliminate the need for natural gas or refined
petroleum products.
8) Electrical requirements were minimized, and the need for
other external energy sources, such as steam, were virtually
eliminated.
9) Whenever technically feasible and economically justified,
commercially available technology and/or equipment was
selected for ACP use.
The ACP system has evolved in accordance with the above criteria
and rationale. This paper describes how the system works and summarizes
the process development work that has been done. The economics of the
ACP system are also discussed, in terms of an evaluation performed by AI
with the assistance of the TVA.
PROCESS DESCRIPTION
The six major process steps in the ACP are: (1) scrubbing, (2)
product collection, (3) reduction, (4) quenching and filtration,
(5) carbonation, and (6) sulfur production. Figure 1 is a block flow
diagram showing these process steps as they would be assembled for an
existing power plant application. The scrubbing and product collection
steps are coupled by the gas stream that flows through the equipment;
they make up the gas-cleaning subsystem.
The dashed lines in Figure 1 emphasize the fact that the remain-
ing process steps (the regeneration subsystem) are decoupled from the
gas-cleaning subsystem. This decoupling is done by providing adequate
storage or surge capacity for both dry product and regenerated scrub-
bing solution.
789
-------
EXISTING STACK
EXISTING s
POWER — H
PLANT V
EXU
IDF
1
z
o
1-
CO
0
o
CM
(O
SULFUR id
PRODUCTION ^
,.„ ,. ... ii-c Dir^u
"^ *r^4 r^*
J * \ J \^J
^ ^ v_/ hi
5TING fe NEW ID
AN QOT FAN
LU<
1— o
LU
"! 1
CO
U
z
LU
_J
0
"Oj-CRUDDirJC ENTRAINED ^ PRODUCT
o02oCRUCDING SOLIDS * COLLECTION
CO
0
V)
CM
REDUCER OFF-GAS J
7° H20 ' ' g
T ^ ^ i
Na2S-RICH
r ft nnnn ATiniM -^ QUENCH ANn ^. nrn
SOLUTION FILTRATION ^
" CO
O O
[CO 0
(9 CD
izz
! £
UCTION
t H2--MIUM i
1, GAS T
AIR
•+COKE
ELEMENTAL
SULFUR
ASH AND
COKE
76-09-2-9
Figure 1. ACP Block Diagram
-------
Gas-Cleaning Subsystem
The flow diagram for this subsystem is shown in Figure 2. The flue
gas enters the gas-cleaning subsystem after leaving an existing electro-
static precipitator. If no precipitator is available, a bank of cyclones
would be provided, to lower the flue gas fly ash content to 1.0 grains/
scf or less before it enters the subsystem. The flue gas enters the
spray dryer scrubber at the top, and passes downward through a spray of
fine droplets of sodium carbonate solution. These droplets are generated
by a centrifugal atomizer, which drives the droplets radially outward,
in crossflow to the flue gas. The following reactions take place:
S02(g) + Na2C03(aq) - -Na2S03(aq) + C02(g) (1)
Na2S03(aq & s) + 1/2 02 — Na2S04 (aq & s) (2)
S03(g) + Na2C03(s) — - Na2$04 + C02(g) (3)
Reaction 1 accounts for most of the SCL removal; the aqueous sulfite
formed is subsequently dried to solid sulfite powder. Reaction 2
accounts for most of the sodium sulfate formed. Reaction 3 also pro-
duces sulfate and removes SCL; it occurs in the cyclones and electro-
static precipitator as well as in the scrubber. The dry product formed
by these reactions and collected for regeneration is usually about 60%
sulfite, 20% sulfate, and 20% unreacted carbonate.
Typical liquid-to-gas (L/G) ratios for the spray dryer scrubber are
0.3 gal/1000 scf of gas. Because of this low L/G, there is not enough
water injected into the spray dryer to saturate the flue gas. There-
fore, its exit temperature is in the 150 to 200°F range, and the spent
reactant is entrained as dry particles. The bulk of these particles are
removed in the product collection cyclones, and most of the remainder
are removed in the precipitator. Particle emissions to the stack will
be 0.01 grains/scf or less.
791
-------
FLUE GAS
REGENERATED"
CARBONATE
SOLUTION
•
X
NF!
•f1
i , , H
¥
—— — SF
D
CYCLONE
SEPARATOR
AQUEOUS
CARBONATE
SOLUTION
FLUE GAS
BLOWER
SPRAY
DRYER
ELECTROSTATIC
PRECIPITATOR
CLEAN GAS
TO STACK
HANDLING
DRY
PRODUCT
SURGE TANK
TO REGENERATION
SUBSYSTEM
FEEDPUMP
75-M3-21-1
Figure 2. Gas Cleaning Subsystem Process Flow Diagram
-------
The key to the operation of the spray dryer scrubber is the
atomization of the carbonate solution. High-speed centrifugal atomizers
are used to achieve the desired spray distribution. These atomizers are
capable of producing fine droplets from either solutions or slurries, at
minimal liquid pumping pressure. The energy required for atomization is
supplied by a spinning disc, and not by a nozzle constriction.
A desirable feature inherent in centrifugal atomization is the
simple turndown method. As gas flow decreases, the solution flow rate
is decreased in direct proportion. Under turndown conditions, any
contact inefficiency caused by a lower gas velocity in the scrubber is
compensated for by the increased atomization efficiency of the disc at
the lower liquid flow rates. Excellent SCL removal efficiency has been
demonstrated at turndown ratios of 4 to 1 for a single scrubber, and
large installations requiring two or more scrubbers could readily
achieve an 8 to 1 or better turndown without sacrificing pollution
control capability.
Turndown of the product collection equipment is also straight-
forward. For single scrubber systems, isolation dampers are provided on
each bank of 8 cyclones to maintain the optimum gas velocity and collec-
tion efficiency at reduced gas flow. Also, as gas flow decreases, the
precipitator collection efficiency increases because of the longer gas
residence time in it. The turndown ratio of the product collection
equipment is therefore compatible with that of the spray dryer, and
particulate emissions of <0.01 grain/scf can be maintained under all
operating conditions.
The gas-cleaning subsystem also possesses excellent pollution
control capability for other acidic gases, such as HC1, SCL, etc.
particular, the presence of an electrostatic precipitator virtually
eliminates an SCL plume problem.
In
793
-------
Regeneration Subsystem
Reduction. The dry product collected in the gas cleaning subsystem
is stored in a silo or hopper, from which it is conveyed pneumatically
to the reducer. The reducer is a refractory-lined steel vessel, which !
contains a pool of molten sodium carbonate and sodium sulfide.
The feed material (a mixture of dry scrubber product and carbon) is
blown. into the molten salt pool with air, and additional air is injected
to react with a part of the carbon reducing agent. (This agent can be
petroleum coke or the same coal used in the power plant.) The temper-
ature of the reducer melt pool is maintained in the 1700 to 1900 F
range; the dry feed material melts and mixes with the salt pool, and the
carbon reacts with the sulfite and sulfate, reducing both to sodium
sulfide. The following reactions take place:
Na2S03(j0) + |c(s) — -Na2SU) + |c02(g) (4)
Na2S04U) + 2C(s) — Na2SU) + 2C02(g) (5)
C(s) + 02(g) — C02(g) + heat (6)
Reaction 4 is the reduction of sulfite; it probably also occurs to a
large extent through a disproportionate step, since sulfite is unstable
at the reducer temperature:
4Na2S03(j0) — -Na2S(j2) + 3NaS0(j2). (7)
This disproportionate is followed by the reduction of sulfate accord-
ing to Reaction 5.
Reactions 4 and 5 are endothermic, requiring heat input, and addi-
tional heat is required to heat up all of the feed material and to melt
794
-------
the salt feed, as well as to compensate for heat losses through the
vessel walls and heat carried out with the melt product stream and the
off-gas. The heat is generated by the oxidation of a part of the carbon,
as is shown simplistically in Reaction 6. Actually, much of the heat is
generated by the reoxidation of sodium sulfide, which is in turn re-
reduced by the carbon. The mechanism for this is
Na2SU) + 202(g) —-Na2S04(j2) + heat (8)
Na2S04(j0) + 2C(s) — Na2S(j2) + 2C02(g) (5)
The sum of these two reactions is Reaction 6.
The reducer vessel is a vertical cylindrical steel shell with a
lining of high-density, fuse-cast, alpha-alumina bricks (Monofrax A,
from the Carborundum Corp.). This material has been found to be very
durable in the reducer environment; indeed, corrosion rates are less
than 0.5 mil/yr. The reducer is provided with an air compressor for
air-feed injection, spent-reactant and carbon feed systems, off-gas
conditioning equipment, startup preheaters, and a continuous melt drain
spout.
Less than 5% excess carbon (over the reduction and heat generation
requirements) is used to insure 95% reduction of the sulfur-containing
salts. Air is injected in sufficient quantity to oxidize the amount of
carbon required for heat generation. The typical reducer off-gas con-
tains less than 1 vol % of oxygen and up to 35 vol % carbon oxides, with
C02/C0 ratios of 10 to 1 or greater. Heat is recovered and steam is
generated in a waste heat boiler prior to cooling the reducer off-gas for
use in the carbonation step.
The most volatile salt in the reducer melt is NaCl, which will be
the predominant particulate in the reducer off-gas. A gas cooling
tower for the reducer off-gas will remove the NaCl-rich particulates
and a part of the recirculating cooling tower liquid stream will serve
as the chloride purge for the ACP system.
795
-------
Extensive pilot-scale tests of the reduction step have been con-
ducted in the AI Molten Salt Test Facility (MSTF). This facility and
the tests are described later in this paper.
Quenching and Filtration. The quenching and filtration steps have
both been demonstrated in AI bench- and pilot-scale tests, and both are
employed in the pulp and paper industry in conjunction with recovery
boiler operation. Recent work has also shown that the quench step can
be eliminated by cooling the reduced melt, allowing it to solidify, and
then breaking it up into chunks for dissolution. The solidified melt is
quite porous, and readily soluble.
Carbonation. After dissolution and filtration, the clarified
solution (green liquor) is contacted with the carbon-dioxide-rich reducer
off-gas to regenerate sodium carbonate and evolve hydrogen sulfide. The
green liquor is first treated in a precarbonator to form sodium hydro-
sulfide (NaHS), by Reaction 9.
2Na2S + H20 + C02(g)— 2NaHS + Na2C03 (9)
The precarbonated liquor is further reacted with C02 in the carbonator,
and the following reactions occur:
NaHS + C02(g) + H20 — - NaHC03 + H2S(g) (10)
2NaHS + C02(g) + H20 — - Na2C03 + 2HS(g) (n)
The Glaus plant feed gas is taken directly from the carbonator. The
bicarbonate slurry produced in the carbonator is subsequently decomposed
by the reaction:
Na2C03 + H20 + C02(g), (12)
and the C02 is recovered for recycle to the inlet of the carbonator.
796
-------
Sulfur Production. The hydrogen sulfide concentration produced by
the carbonation process is sufficiently high to allow economical conver-
sion to elemental sulfur in a conventional Claus plant. Claus technology
is considered commercially available, and therefore has not been tested
as a part of the ACP development program.
STATUS OF PROCESS DEVELOPMENT
Scrubber Development
Two pilot-scale installations have been used to test and develop
the spray dryer scrubber concept. The first of these installations
consisted of a 5-ft diameter spray dryer and matching cyclone, plus
auxiliary equipment. This unit was purchased, modified, and refurbished
for installation at the coal-fired Mohave Station of Southern California
Edison. The performance of the spray dryer as a scrubber was then
tested using actual flue gas from the Mohave Station.
The second pilot plant scale test installation consisted of the
7-ft diameter spray dryer and associated product collection equipment
that is permanently installed at a field laboratory owned and operated
by Bowen Engineering, Inc., a leading U.S. manufacturer of spray drying
equipment. This pilot-scale dryer is the basic tool that Bowen uses for
design information to scale up each special spray dryer application.
Based on data obtained here, Bowen has designed and built spray dryers
up to 44 ft in diameter.
Mohave Tests (5 ft Dryer). The spray dryer system was installed
at the Mohave Station in less than two weeks, and operated on a one-
shift-per-day basis over a period of seven weeks (in May and June 1972)
with no significant maintenance or operational problems. During the
seven-week test period, about 100 tests were conducted to completely
characterize the performance of the spray dryer. Since steady state can
be achieved with this kind of scrubbing system in less than 1/2 hour,
797
-------
the duration of each test was typically limited to two to four hours.
This allowed ample time to obtain all the required data from each test.
A photograph of the Mohave system is shown in Figure 3. In addi-
tion to the fully instrumented and controlled spray dryer system, the
installation included a 560 gal. feed tank; a calibrated metering pump;
duct heaters to allow variation and control of the inlet flue gas tem-
perature; continuous monitoring instrumentation for S02, N0x, and 02;
and particulate sampling apparatus for product characterization. Gas
flow rates were measured with an in-line flow meter, and tests were
conducted at flue gas flow rates of from 1035 to 1375 scfm.
The main objective of the Mohave test program was to demonstrate
90% S02 removal from a flue gas containing 400 ppm S02- This was
accomplished by atomizing -0.3 gal/1000 scf of a 4 to 5 wt% sodium
carbonate solution into the flue gas with the centrifugal atomizer.
Equally effective S02 removal was also demonstrated at S02 concentra-
tions of up to 1500 ppm by maintaining the L/G constant at the 0.3 value
and increasing the carbonate concentration. A typical plot of Mohave
test results illustrating the S02 removal efficiency as a function of
liquid absorbent feed rate is given in Figure 4.
Extensive data were also obtained at Mohave relative to the col-
lection efficiency of the cyclone. Collection efficiency ranged from 86
to 99.5% depending on the test conditions, varying primarily as a
function of the particulate loading in the gas.
The Mohave installation was shut down in mid-June 1972 while the
data from the test program were analyzed and the report prepared. It
was reactivated in September 1972, for the purpose of additional testing
with inexpensive crude soda ash. Ten tests were conducted using a trona
salt obtained by surface mining of deposits at Searles Lake, California.
The results generally confirmed that any soluble sodium-based alkaline
material could be used for effective S02 removal in a spray dryer.
798
-------
72-MA8-22-18
Figure 3. AI Scrubber Test Installation
-------
£
a
a.
<
DC
LLJ
O
z
O
o
CM
o
CO
H
Z
LU
uu
IMa2CO3 CONCENTRATION = 5.5 wt %
GAS FEED RATE = 1290 scfm
40
30 -
0 0.1 0.2 0.3 0.4
ABSORBENT FEED RATE (gal/min/1000 scfm)
72-025-52-18
Figure 4. SC>2 Removal
vs Reactant Flow Rate
800
-------
Bowen Tests (7 ft Dryer). 'A series of 26 tests was conducted in
May 1973 to expand the S02 concentration data base and to demonstrate
the turndown capabilities of the spray dryer scrubber. Tests were run
at S02 concentrations from 175 ppm to 2120 ppm S02 and with gas flow
rates from 1000 to 4000 scfm. The Mohave test conditions were duplicated
in some of the tests, to verify that the data obtained were equivalent
to those obtained when using real flue gas. The results from these
tests were virtually identical to the corresponding Mohave results.
Another important result related to the gas flow capacity of the
spray dryer and the turn-down characteristics. It was found that excel-
lent S02 removal efficiency could be maintained by adjusting the solution
feed rate in direct proportion to the gas flow rate. With all other
conditions except solution feed rate held constant, gas flow rates were
varied from 1000 to 4000 scfm. The S02 removal efficiency of the spray
dryer remained at 90% or better over this entire range. In fact, SOp
removal tended to increase slightly at the higher flow rates.
The rated capacity of the seven foot dryer is -3000 scfm. Since
the performance was excellent at 4000 scfm, 30% above the rated capacity,
it is postulated that large-scale spray dryers for ACP applications can
be designed to handle a greater gas flow than is typical for other spray
dryer applications. This must be verified, however, in a full-scale
test. The first large-scale ACP system will be designed according to
standard spray dryer practice, but tested at gas flows greater than its
rated capacity.
A series of 40 additional tests was conducted at the Bowen facility
in June 1973, to further expand the data base by extending the S02
concentration range to 7200 ppm. Test results in the 200 to 3000 ppm
range were in agreement with data and extrapolations from previous Bowen
and Mohave tests. Tests run with S02 concentrations of 4000 to 7200 ppm
required inlet gas temperatures of 400 to 500°F in order to insure the
production and collection of a dry product; S02 removal efficiencies of
from 92 to 99% were achieved.
801
-------
The 7200 ppm SOp was the maximum concentration that could be gen-
grated with the available S02 injection apparatus. Without this limita-
tion, the spray dryer could have been operated at higher inlet gas
temperatures and with correspondingly increased S02 concentrations. An
estimate of the dryer's performance, with its nominal maximum inlet
temperature of 1200°F, indicates that about 95% S02 removal would be
obtained with an S02 concentration of at least 20,000 ppm.
Large-scale spray dryers for the ACP system will utilize multiple
atomizers in a single drying chamber. Bowen Engineering recently modi-
fied their pilot dryer to accommodate three centrifugal atomizers.
Tests were run in October 1975 to demonstrate the feasibility of the
multiple atomizer concept. About 1/3 of the incoming gas was distri-
buted to the vane ring above each atomizer. In the first series of
tests, about 1/3 of the total Na2CO- solution was also fed to each
atomizer. As expected, S02 removal efficiency was comparable to that
previously obtained with a single atomizer.
A subsequent series of tests was run with one of the atomizers shut
off, but with gas still flowing through its vane ring. Using the same
total Na2C03 solution as in the first series, but feeding 1/2 of the
solution to each of the two operating atomizers, S02 removal efficiency
remained the same as that observed in the first tests. Thus, the mul-
tiple atomizer concept was not only found to be feasible, but its use
will apparently enhance operational reliability of the ACP gas cleaning
subsystem.
Summary. Pilot tests of the ACP spray dryer scrubber and its
associated cyclone collector have been run with gas flow rates up to
4000 scfm. One of the pilot installations is the standard scale-up tool
used by the spray dryer manufacturer for the design of large-scale (up
to 44 ft diameter) equipment. In fact, although the application is
somewhat different, full-scale spray dryers of the size needed in a
100 Mw ACP system are currently in operation (see Figure 5).
802
-------
Figure 5. Large-Scale Spray Dryer
Installation
73-J17-5-21
SO 7>
-------
Reducer Development
The development of the ACP reduction step has proceeded from bench-
scale tests in a 6 in. diameter alumina vessel to full pilot-scale
testing in the Molten Salt Test Facility (MSTF) built by Atomics Inter-
national' for a variety of molten salt development work.
Bench-Scale Tests. The first bench-scale tests of the basic ACP
reduction concept were conducted early in 1973, using small electrically-
heated reactor vessels. The objectives of the tests were to establish
reaction conditions, determine offgas characteristics, and obtain kinetic
data.
A second bench-scale test series was conducted in early 1974 to
demonstrate the feasibility of simultaneous heat generation and reduc-
tion in the ACP reducer. Solids were fed to the reducer at a rate of
-70 gm/min; 2 to 3 scfm of air was continuously injected; and the
offgas was sampled for CCL, CO, and 0-. The CO^/CO ratio was measured
to be >10 to 1 under optimum conditions. Residual carbon in the melt
was kept at the 0.6 to 2.5 wt% level, and steady-state reduction of
between 85 and 90% was achieved.
Pilot Scale Tests. Three series of pilot scale ACP reducer tests
have been conducted in the AI Molten Salt Test Facility (MSTF). A
photograph of this facility is shown in Figure 6. The refractory-lined
vessel is three feet ID and 12 feet high. Solids and combustion air are
fed into the reaction zone below the melt level through a side-entering
port. The major limitation of the MSTF is its batch-type operation, but
recent modifications will allow future tests to be performed with con-
tinuous withdrawal of melt product. Its capacity is sufficient to
obtain steady-state data at feed rates up to 300 Ib/hr (100 Ib/hr
equivalent to ~1 Mw) in batch-type operation, and at feed rates of 500
Ibs/hr using the new continuous withdrawal capability.
804
-------
40412-6CN
Figure 6. Molten Salt Test Facility
SD5
-------
The first series of tests was conducted in the MSTF to verify
bench-scale results and provide preliminary design data for equipment
studies. About 1670 Ib of salt containing 30 wt% Na2C03, 26 wt% Na2S04,
and 44 wt% Na2S03 was fed at a rate of 125 Ib/hr. The extent of reduc-
tion was found to be a function of the excess coke present in the melt.
As shown in Figure 7, at least 90% reduction was achieved with about 3.5
wt % excess coke. This reduction rate is consistent with the bench-
scale data, which are also plotted for comparison. Approximate C02/C0
ratios of 9 to 1 or greater were observed under the test conditions.
This result is again consistent with the bench-scale results.
A second series of ACP reducer tests was conducted to optimize
reducer operating conditions at the pilot plant scale, improve the gas
analysis techniques to obtain continuous off-gas monitoring of C02 and
CO, and obtain data relevant to the design of larger-scale reducers.
About 2300 Ib of spent ACP reactant salts were fed to the MSTF during
the tests at feed rates up to 150 Ib/hr. Temperature was maintained
between 1800 and 1850 F during the tests by reacting coke with air to
generate heat. Because of the high relative heat losses in the MSTF as
compared to larger vessels, and the correspondingly high air require-
ments, carbon oxide concentrations in the off-gas were only about
25 vol % as compared to the 35 vol % expected in full-scale reducers.
The measured C02/C0 ratios were excellent, with 80 to 1 observed during
some of the tests. A more typical OL/CO ratio for the tests was in the
range of 10 to 15 to 1, well above ACP design requirements.
Twenty melt samples were taken during the test period, and analytical
results showed steady-state reduction of 90 to 95%. This again meets or
exceeds ACP system design requirements. Within the constraints of the
MSTF itself, the test conclusively demonstrated the feasibility of the
basic ACP reduction concept, and provided satisfactory design and
operating data for scale-up purposes. The MSTF lining material, Mono-
frax A, has been observed to be very resistant to corrosion by the melt.
Corroborating test data show corrosion rates of <0.1 mil/year at 900°C
(1650°F), and <0.5 mil/year at 1000°C (1830°F).
806
-------
100
80
60
u
D
O
lit
QC
LU
D
CO
40
20
LEGEND
MSTR TEST
LABORATORY
D
O
4 6
COKE CONTENT OF MELT (wt %)
Figure 7. Sulfate Reduction
vs Coke Content of the Melt
10
4137-4001
807
-------
A third ACP reducer test series was run using a low-sulfur, high-
ash, western coal as the reducing agent. Nearly 3000 Ibs of salt and
about 1800 Ibs of coal were fed to the MSTF during the test program.
Average feed rates during steady-state operation were 370 Ibs/hr of salt
and 180 Ibs/hr of coal. The steady-state reduction rate was 89%, and
the off-gas composition was found to be suitable for subsequent use as a
carbonating gas. This test series provided the same type of data for
coal as the second reducer test series provided for coke. A final test
series is currently in the planning stages. It will utilize a high-
sulfur eastern coal as the reductant, and it will also demonstrate the
new continuous withdrawal capability of the MSTF.
Summary. The essential elements of the basic ACP reducer have
been pilot tested. Design information from the MSTF tests has been
incorporated into a preliminary reducer design study, and control diagrams
for ACP reducer operation have also been prepared. Reducer design
studies are continuing, and a preliminary specification for the ACP
reducer-quench subsystem is in preparation.
Quenching and Filtration Development
Early in the ACP development program, laboratory tests were run to
establish the filtering characteristics of coal ash suspended in solu-
tions of sodium carbonate and sodium sulfide. The most important result
established by the tests was that sodium values could be recovered from
the ash filter cake with a simple washing technique. The original cake
contained sodium equivalent to 20% of that remaining in the filtrate.
Two wash cycles reduced that sodium level to less than 1.0% of the cake
weight. More recently, a rotary drum vacuum filter was used to test
the filtration of coal ash from green liquor prepared from the melt
produced in the third reducer test series.
808
-------
Quenching tests were conducted in AI laboratories during 1974 to
provide basic chemical and physical information. Up to 20 Ib/min of a
70% sodium sulfide - 30% sodium carbonate melt were quenched and dis-
solved in a 14 cu ft tank. The key to trouble-free quenching was found
to be adequate dispersion of the melt into fine droplets, and maintain-
ing the bulk dissolving liquor at or near its boiling point.
Carbonation Development
The development of the ACP carbonation step originated with an
extensive literature survey and analysis of existing carbonation tech-
nology used in the chemical and pulp industries. The essential conclu-
sion of this study was that AI should embark on a development program
specifically aimed at devising a carbonation scheme to take advantage of
the key ACP features.
Accordingly, a small pilot-scale development program was initiated.
The pilot unit was designed, built, and tested during the past two years
A photograph of the equipment is shown in Figure 8. It consists of a
series of glass columns with stainless steel internal hardware. A gas
supply system, a small steam generator, and appropriate pumps and valves
complete the equipment array for the pilot unit. An engineering support
effort was initiated in parallel with the development program to make
the work relevant to the design constraints of a large-scale system. As
a result of this cooperative effort, the original carbonation scheme was
modified, and is now considered to be the best alternative for this ACP
process step.
Extensive tests (>100) were run to evaluate the parameters affect-
ing the carbonation sequence. The data obtained included parametric and
steady-state characterization of performance in terms of the gas-liquid
reactions, temperature effects, and column configurations. One of the
key design requirements of the carbonation step is the efficiency of
H2S stripping in the carborator. The reference design value derived in
809
-------
4137-6
Figure 8. Carbonation Pilot Plant
810
-------
engineering studies was 98.6%. Pilot test results were typically above
99%, and were above 99.9% in many cases.
The most recent tests of this process step utilized the clarified
green liquor that had been generated by sequentially processing the melt
from the third reducer test. Over 200 gallons of clarified green liquor
were carbonated. This test was of major importance because it closed
the process loop, and it clearly established that the use of coal will
not significantly affect the basic ACP process steps.
Commercial Components
The two major components of the ACP system which have not been
tested are the electrostatic precipitator and the Claus plant. These
components are commercially available without specific testing for the
ACP application.
Electrostatic Precipitator. Electrostatic precipitators are used
as the final particulate collectors in the basic ACP system. Specifica-
tions have been written and quotes and guarantees obtained for this
component.
Claus Plant. Commercially available Claus plants will be used to
produce elemental sulfur in the ACP system. The Al-developed carbonation
step produces 35 vol% H2$ gas, a suitable feed for the Claus reaction.
PROJECTED PROCESS ECONOMICS
The capital investment requirements and total annual operating
costs for an ACP system installed on a new 500 Mw, coal-fired power
plant have been estimated, in a preliminary economic evaluation per-
formed by Atomics International with the help of the Stack Gas Emission
Studies Staff of TVA. The basis for the estimates were the same as
those used by TVA in a recent EPA-sponsored study of other advanced flue
gas desulfurization processes. ' The costs are for a plant with a
mid-1975 startup.
811
-------
The total capital investment for the 500 Mw ACP system was esti-
mated at $32,067,000. This compares favorably with another sulfur-
producing process evaluated by TVA. The net total annual operating
costs for the 500 Mw ACP system were estimated at $7,783,200. This
compares favorably with the limestone throwaway process, which had the
lowest operating costs of all the processes evaluated by TVA. Based on
this economic evaluation, which provided a unique opportunity to compare
the ACP system with other advanced flue gas desulfurization processes on
the same basis, the ACP system is among the best of the regenerative
processes currently being developed.
Capital Investment Summary
A summary of the estimate for the capital costs of the 500 Mw ACP
system is given in Table 1. The total direct investment is $19,835,000,
and the total capital investment is $32,067,000 or $64/kw.
The indirect costs for engineering, construction, and contingency
were taken as percentages of direct investment, and they are self-
explanatory. They total 37% of the direct investment and are, there-
fore, a very substantial part of the total investment. The other
indirect costs for startup and interest during construction were taken
as 10% and 8% of the fixed investment. They are also substantial.
Operating Costs
A summary of the estimate for the total annual operating costs of
the 500 Mw ACP system is given in Table 2. The direct costs are esti-
mated at $3,063,000/year. Indirect operating costs, including 14.9% of
the total capital investment, are estimated at $5,672,000/year. The
gross total annual operating costs are $8,736,100/year or 2.5 mills/kwh
for the assumed 7000 hr/yr load factor. This total does not include a
credit for the sulfur produced, although a marketing expense to sell the
sulfur was charged under indirect operating costs.
812
-------
TABLE 1
TOTAL CAPITAL INVESTMENT REQUIREMENTS
500 Mw System
(Mid-1974 Equipment Delivery)
DIRECT COSTS TOTAL
• EQUIPMENT 13,392
• PIPING AND INSULATION 636
• DUCTWORK, CHUTES, AND SUPPORTS 1,628
• CONCRETE FOUNDATIONS 278
• EXCAVATION, SITE PREPARATION, RAILROADS,
• ROADS, AND POND 331
• STRUCTURAL 233
• ELECTRICAL 1,293
• INSTRUMENTS 724
• PAINT AND MISCELLANEOUS 113
• BUILDINGS 244
• LAND 18
• CONSTRUCTION FACILITIES 945
• SUBTOTAL DIRECT INVESTMENT 19,835
INDIRECT COSTS
• ENGINEERING DESIGN AND SUPERVISION 2,182
• CONSTRUCTION FIELD EXPENSE 2,182
• CONTRACTOR FEES 992
• CONTINGENCY 1,984
• SUBTOTAL FIXED INVESTMENT 27,175
• ALLOWANCE FOR STARTUP AND MODIFICATIONS 2,718
• INTEREST DURING CONSTRUCTION 2,174
• TOTAL CAPITAL INVESTMENT 32,067
75-M3-21-24
813
-------
TABLE 2
TOTAL ANNUAL OPERATING COSTS
500 Mw System
(Mid-1975 Startup)
ANNUAL DIRECT COSTS
• DELIVERED RAW MATERIALS
• SODA ASH (tons)
• COKE (tons)
« CATALYST
UNIT
COST
($)
52/ton
15/ton
A SUBTOTAL RAW
TOTAL
ANNUAL
QUANTITIES
420
45,500
MATERIALS
TOTAL
ANNUAL
COST
($)
21,840
682,500
24,000
728,340
•CONVERSION COSTS
•OPERATION LABOR & SUPERVISION
8.00/
man-hr
39,840
0.23/gal
0.02/Mgal
0.08/Mgal
0.30/Mgal
0.01/kwh
1,470,000
1,822,000
145,740
24,780
74,492,000
12.00/hr
9,960
•UTILITIES
A FUEL OIL (NO. 6), (gal)
ACOOLING WATER (Mgal)
APROCESS WATER (Mgal)
ATREATED WATER (Mgal)
AELECTRICITY, (kwh)
MAINTENANCE (LABOR & MATERIAL)
AANALYSES
A SUBTOTAL CONVERSION COSTS ($)
A SUBTOTAL DIRECT COSTS ($)
INDIRECT COSTS
•AVERAGE CAPITAL CHARGE AT 14.9% OF INVESTMENT
•OVERHEAD
• PLANT, 20% OF CONVERSION COSTS
• ADMIN & MARKETING COSTS
A SUBTOTAL INDIRECT COSTS ($)
A TOTAL ANNUAL OPERATING COST ($)
318,720
318,100
36,440
11,670
7,430
744,920
758,500
119,520
2,335,300
3,063,600
4,778,000
467,100
427,400
5,672,500
8,736,100
75-M3-21-23
814
-------
The direct operating costs include raw materials, operating labor
and supervision, utilities, maintenance, and analytical support labor.
Maintenance requirements for the ACP system were estimated as percent-
ages of direct investment for the various process areas. The three ACP
areas which are somewhat unique are particulate scrubbing, SCL scrub-
bing, and reduction. Since the ACP particulate scrubbing area consists
of only cyclones and ducts, maintenance requirements should be very low.
The ACP S0? scrubbing area should also be low, with the electrostatic
precipitators requiring the highest degree of maintenance. These two
maintenance factors were specifically checked by TVA against their own
power plant operating data and found to be valid. The major maintenance
user in the ACP system is reduction, which was assumed to require an 8%
factor. This factor is the maximum assessed against any regeneration
area for the processes studied by TVA.
The capital charge used by TVA is 14.9% of total capital invest-
ment. For the ACP system, with an investment of $32,067,000, this
charge totals $4,778,000 or almost 55% of the total annual operating
costs. Plant overhead was assumed to be 20% of the conversion costs.
Conversion costs include all direct operating costs except those for raw
materials. The final indirect cost is for administrative overhead and
marketing of the process byproduct. An arbitrary total figure of
$427,400 was charged for this indirect cost.
If the administrative and marketing overhead is broken down in
accordance with the TVA ground rules for other processes, it includes
$31,900 for administration (10% of operating labor), and $395,500 for
product marketing. Since the ACP system and other regenerative pro-
cesses incur this expense, it is only fair that a product credit be
given. The ACP system produces 5.445 tons/hr of sulfur which is worth
$952,900/yr assuming a credit of $25/ton. This was the credit used by
TVA in their analyses of the long-term operating costs of flue gas
desulfurization systems. Although not shown on Table 2, the net total
annual operating costs for the ACP system, including the product credit,
are $7,783,200/yr or 2.2 mills/kwh.
815
-------
REFERENCES
(1) G. G. McGlamery, et al., "Detailed Cost Estimates for Advanced
Effluent Desulfurization Processes," EPA-600/2-75-006 (January
1975)
816
714-A.84/abg/mim
-------
ENERGY REQUIREMENTS FOR SHELL FGD PROCESS
F. A. Vicari and J. B. Pohlenz
UOP Process Division
UOP Inc.
Des Plaines, Illinois
817
-------
ENERGY REQUIREMENTS FOR SHELL FGD PROCESS
by
F. A. Vicari and J. B. Pohlenz
UOP Process Division
Des Plaines, Illinois
Introduction
In the normal evaluation and comparison of flue gas
desulfurization systems, it is usual to make an overall
economic analysis including capital as well as operating
costs, and to determine annual charges for the projected
operational load over the real life of the equipment. The
cost estimates prepared by the engineers in the Chemical
Development Division of TVA illustrate this procedure in
impressive detail, d) With this type of study the annual
charges are often computed back to present dollars to yield
a single value, characteristic for each process, as applied
to a given problem and a given set of factors, i.e., utilities,
catalysts, chemicals, labor, maintenance, taxes, finance
procedures, interest, and capital recovery rates. This single
number represents the amount of present dollars which at a
preselected rate of interest would finance the entire
operation from engineering and construction through the final
day of the units life.
Such an analysis has the advantage of permitting a
comparison on a single economic scale of complex processes
of different types, e.g., regenerative vs. throwaway. The
basic limitation of this technique lies in the inability of
the analyst to predict the costs of utilities over the normal
life span of the plant, e.g., 30 years. In contrast with
capital charges which occur in the immediate future and
hence which can be cost-estimated with some accuracy, the
cost of fuels are often affected by unpredictable factors.
However, a single predictable characteristic related to
projected fuel costs is the inevitability of continuously
increasing costs of all fuels.
With the major adjustments in fuel prices that have
occurred in the last several years, it has become common
practice to supplement the economic analysis with a survey
of the energy requirements as an aid to process evaluation.
Such a study of the energy requirements for the Shell
FGD process is the basis for this paper.
Process Flow
The energy requirements for the Shell FGD process '2~7'
expressed as fuel, steam or electrical, refer to the capture
and production of a concentrated stream of S02 free from
818
-------
particulate matter and oxygen. Note that there is, therefore,
no net change in the chemical species containing the sulfur.
For a given boiler size the energy requirements can
vary substantially with the amount of sulfur processed.
This condition is described in Table 1, which shows the heat
rates, flue gas rates, and sulfur rates for a 150 MW boiler
fueled with coal containing 0.32 and 3.89% sulfur. The
processing schemes were formulated to effect 85% desulfurization
on flue gas containing 350 ppm (Case I) and 3500 ppm of S02
(Case II) .
The proposed design in Case I uses three reactors in
parallel to obtain the desired sulfur removal (See Figure 1) .
The flue gas in processed through a copper-on-alumina acceptor.
Since the quantity of sulfur removed is small, the acceptance
time is long (290 min.) , and one additional reactor in regener-
ation can support three in acceptance. Thus, the cycle for
each reactor becomes :
Q Minutes
Acceptance 290
Valve Switch and Purge 3
Regeneration 91
Valve Switch and Purge _ 3_
387
Since each reactor experiences the same cycle, the
average number of completed cycles per day is:
Cycles/Day = !!) (4) =
Acceptance
At the start of acceptance, immediately after
regeneration, purging, and valve switching, the copper in
the regenerated acceptor exists as elemental copper. Upon
contact with the oxygen in the flue gas, the copper reacts
rapidly to form copper oxide (CuO) which in turn reacts with
the oxygen and S02 in the flue gas to form copper sulfate
(CuS04). The conversion of CuO to CuSO. is given by the net
average loading ( a ) , the fraction of trie total copper
present as CuSO,. The reactions occurring during acceptance
are:
Cu + 1/2 02 ^ CuO AH=-37,340 kcal/kg-mole Cu
400°C
CuO + 1/2 O2 + SO2 *- CuS04 AH=-74,600 kcal/kg-mole Cu
400°C
819
-------
The heat released during the acceptance period is (see Table 2) :
/ 0-7 0,1^ , -/ » / -,, ™^v -85,084 kcal/kg-mole Cu
(-37,340) + (a) (-74,600) = ^t~400°C
where a = 0.64.
This exothermic heat is absorbed by the flue gas and
results in a temperature increase. In Case I, for example,
with a net average loading of 0.64, the average hourly
temperature raise is calculated:
Basis: 1 kg-mole Cu
(\ / \ f f\ r\ r- \
') I •=—(Yq-p-) " -63,462
where 29.5 is the molecular weight of the flue gas. The heat
release of 85,084 kcal per kg-mole of copper is absorbed into
this flue gas and results in an average temperature increase of:
= /85,Q84\/ 1
\63,462/\0.25
where 0.25 is the specific heat of the flue gas. If this
sensible heat is recovered in the air preheater and returned
to the furnace, the heat of the oxidation and acceptance
reactions represents:
(85,084) - 0.823
which is equivalent to 0.24% of the input heat rate, i.e.,
a reduction in fuel rate of 0.24%. Note also that the
sensible heat represented by (5.4)/0.24 = 22.5°C is therefore
equivalent to approximately 1% on boiler efficiency or 91
Btu/Kwhr reduction in heat rate.
In Case II, where the amount of sulfur being processed
has increased by a factor of 65.34/6.19, the 85% desulfurization
is achieved with 2 reactors in acceptance supported by one in
regeneration; the loading at the end of acceptance is 0.77,
the heat released per kg-mole of copper is 94,782 kcal/kg-mole,
the average flue gas temperature raise is 49.6°C corresponding
to 2.34% on the input heat rate to the boiler (213 Btu/Kwhr).
The heat release, resulting in an average temperature
raise of approximately 49.6°C, is not uniform over the
acceptance period, since the oxidation wave moves through the
reactor in but a few minutes. The reactor internals are
designed with sufficient heat capacity to keep peak temperatures
well below that at which temperature damage to the acceptor
can occur. The combined heat capacity of the SFGD reactors and
the regenerative air preheater serve to smooth out the temperature
820
-------
of the preheated air and cooled flue gas.
Regeneration
In the regeneration portion of the cycle the copper as
CuS04 and as CuO is reduced to the elemental form. These
reactions are:
CuS04 + H2—*- Cu + S02 + HO AH=-5,560 kcal/kg-mole Cu
2 400°C
CuO + H2 > Cu + H20 AH=-20,880 kcal/kg-mole Cu
400°C
The heat release is therefore:
Basis: 1 kg-mole Cu
(1--64) (-20,880) + (0.64) (-5,560) =
-11,075 kcal/kg-mole Cu
Since the regeneration off-gas goes to a waste heat
boiler this heat of reaction is a recoverable steam credit.
It is informative to observe that the copper cycle is
now completed and that each step in the cycle is exothermic,
yielding up heat to the fluid being processed. Furthermore,
with appropriate design of the system, this heat is recoverable
as a credit against the energy requirements of the process.
The reductant which "fuels" the cycle is the hydrogen used
in the copper regeneration and which is rejected in a completely
oxidized form, i.e., water. Note further that hydrogen is
simply an intermediate product in the stepwise oxidation
of fossil fuels, yielding ultimately the carbon as carbon
dioxide, the hydrogen as water, and sulfur as sulfur dioxide.
Thus hydrogen not only can be produced directly by the sub-
stoichiometric combustion of any fossil fuel, but further, a
unit of hydrogen represents a quantity of thermal energy almost
independent of the fuel source, e.g., methane, light hydro-
carbons, liquid residual fuels, coke, or coal. In this study,
this relation between hydrogen and thermal energy is 4298
kcal/NM . Note this thermal equivalent of hydrogen ds
approximately 140% of the heat of combustion.
The stoichiometric amount of hydrogen required per
mole of S02 removed and released can be expressed in terms
of the acceptor loading, a , as
H_2 _ a +1.12
S02 ~ a
applied to Case I, the stoichiometric hydrogen per cycle is:
821
-------
6.19 290 (2.75) ->-,*, •,
5— —^- =27.4 kg-mole
The SO.-, evolution vs. time curve appears as a block
function; as long as SOo leaves the reactor no unreacted
H2 "slips" through. At the end of the regeneration period,
when the quantity of SC>2 produced falls off sharply, H2
appears in the regeneration off-gas.
At the end of a regeneration, the reactor void volume
contains regeneration gas which is displaced during the
subsequent purge through an accepting reactor where the
hydrogen is catalytically oxidized by CuO to water. The
heat released is absorbed by the flue gas being processed
and is recoverable.
Copper sulfate at 400°C is extremely reactive with
hydrogen; indeed the rate of SC>2 release can be controlled
by the rate of hydrogen addition. Furthermore, there is a
mole change (increase) in the copper sulfate reduction so
that the course of the reaction can be easily monitored (and
controlled) by observing gas flow rates to and from the
reactor. An interesting characteristic of the chemistry
occurs at the beginning of the regeneration when hydrogen
is introduced to the purged reactor. As the first copper
sulfate is reacted at the inlet of the bed, the released
SC>2 is re-accepted. All the CuO is reduced before S02
breaks through and appears in the regeneration off-gas.
The S02 production rate change is equally as abrupt at
the end of the regeneration.
The total amount of H.2 required for one regeneration
for Case I is then:
Stoichiometric H2 27.42 kg-mole
Slip H2 (10%) 2.74
Purge H2 1.1
31.26 kg-mole H2
This amount of H2 must be supplied during the time
allotted for regeneration (91 minutes in Case I), i.e.,
0.343 kg-mole, (20.6 kg-mole)
min. hr.
In some instances the hydrogen may be available on
demand, that is, supplied intermittently from a generator
supplying other process units. More frequently, however,
the hydrogen must be produced continuously and during the
valving and purge portions of the regeneration cycle the
reductant is vented to the boiler displacing its heating
value equivalent as fuel.
The total hydrogen demand is converted to equivalent
822
-------
fuel at the ratio mentioned above, i.e., 4298 Kcal/NM^
hydrogen. The vented and purged reductant is credited at
fuel value, i.e., at its heat of combustion.
After regeneration the reactor is purged with steam
at a rate of one reactor volume/minute, and like the
hydrogen can be treated as an intermittent or continuous
supply depending on local conditions. Frequently the
steam purge requirement of one reactor volume/minute
determines the supply rate. Vented steam carries no credit
value.
Additional Heat Recovery
The SFGD process provides almost complete removal of
803. The small amount of SOj remaining in the processed gas
is a function of the acceptor temperature and sulfur loading,
In most cases the SO-j removal results in a reduction in the
sulfuric acid dew point permitting a reduction in flue gas
temperature and increased thermal efficiency of the boiler.
In Case I, however, the total sulfur content of the flue
gas is so low, the S03 (normally 2-3% of the total) permits
heat recovery to flue gas temperatures of 125°C. For
Case II, however, 863 removal in the SFGD reactors permits
a decrease in acid dew point of approximately 40°C.
Figure 2 gives typical 863 vs. dew point-temperature
data from several sources. Translating to an energy
credit: (See Table 2)
492,000 NM3 \ /0.327 kcal \ (40°C)
hr / \ NM3-°C/
6.435 x 106 kcal/hr
Flue Gas Pressure Drop
The energy requirements discussed in the preceding
development are related directly to the amount of sulfur
removed from the flue gas, and have been expressed
appropriately as steam or fuel debits or credits against
the flue gas desulfurization. Note that these requirements
can be met without derating the station being serviced.
However, the electrical requirements to drive the fans
and to overcome the pressure drop across the reactors,
ducting and valves does represent a consumption of a
portion of the station output. Fortunately, the pressure
drop through the SFGD reactors is low because of the
nature of the internals design wherein the gas flows through
open channels. The pressure drops have been converted to
electrical requirements assuming reversible adiabatic
compression on a perfect gas and 75% driver efficiency.
823
-------
For a hot-dry process wherein this compression work
is applied upstream of the air preheater, the heat of
compression is recoverable. For example, the temperature
raise on the flue gas for Case I resulting from a pressure
increase of 30 cm water is 5.6°C. This temperature rise
is carried through the system to the preheater.
Dampen Flow
In this operation steam is generated by heat exchange
with the hot regeneration off-gas and water is condensed
from the released SC>2 in the quench tower. The flow of
concentrated SC^ from this system is necessarily cyclic
because of the nature of the process. In the design used
in this study the flow is smoothed through a gas compressor/
gas holder to provide a steady feed rate to a sulfur recovery
unit, e.g., Glaus or sulfuric acid. The net water condensate
is boiled through a small stripper; the stripped water can be
used as boiler feed water directly or through the boiler feed
water anion exchanger. Because the quantity is small, if
desired, it may be released to the atmosphere in the flue gas.
The results of the calculations are summarized in Table
3, by case, energy class and service function. Since the
units for all entries are the same, per cent on energy input
to the station, the results can be summed into a single number.
In Case I, with 100 units of thermal energy released in the
furnace, the yield at the generator is 37.5 units of electrical
energy and 0.31 of the 37.5 (0.83%) is required to drive the
fans. Further, an additional 0.29 units as net fuel and
steam supplies the other energy requirements to effect the
desulfurization and concentration. In Case II, the net output
of the station drops from 37.5 to 37.0 and total requirements
increase to 3.5% on input.
S02 Work-up
The work-up of the concentrated S02 to final product,
e.g., liquid S02/ sulfuric acid or elemental sulfur, was
purposely not included in the analysis since the energy
requirements differ for each final product. However, it
is instructive to evaluate one processing procedure on
the same scale of percent of station input, e.g., a
modified Glaus processing So2 to elemental sulfur. But,
now we are faced with a new concept where there is a change
in chemical species of the sulfur from input to output.
The fuel requirements and steam credits for a modified
Glaus are readily determined but note that the final form
of sulfur produced is a fuel and has a fuel credit, whereas
the sulfur entering the Glaus as S02 has no real fuel value.
One manner of crediting the process for the energy
value increase of product over feed is to compute the heat
824
-------
release in the furnace from the oxidation to SC>2 of
the sulfur in the fuel which is recovered. This has been
done and the results tabulated in Table 4.
Summary
A survey of the process energy requirements as a
function of the sulfur content in the fuel and based on
the neat rate to the boiler, gives the development engineer
a measure of the operating costs independent of external
factors. The analysis presents a single number, analogous
to the detailed cost estimate, which represents the operating
costs of the particular process design. Such an analysis
again permits the comparison, on a single scale, of complex
processes.
825
-------
LITERATURE CITED
Detailed Cost Estimates for Advanced Effluent
Desulfurization Processes, Environmental Protection
Technology Series, EPA-600/2-75-006, January, 1975.
Dautzenberg, P.M., Naber, J.E., van Ginneken, A.J.J.
"The Shell Flue Gas Desulfurization Process" AIChE,
Sixty-eighth Annual Meeting, Feb. 28-Mar. 4, 1971.
Paper 31d.
Dautzenberg, F.M., Naber, J.E., van Ginneken, A.J.J.,
"Shell's Flue Gas Desulfurization Process" Chemical
Engineering Progress, 67 , Aug. 1971 pp 86-91.
Conser, R. E., Anderson, R.F., "New Tool Combats
S02 Emissions" Oil and Gas Journal, Oct. 29, 1973.
Naber, J.E., Wesseling, J.A., and Groenendaal, W.,
"New Shell Process Treats Glaus Off-Gas" Chemical
Engineering Progress, 69, Dec. 1973 pp. 29-34
Ploeg, J.E.G., Akagi, E., and Kishi, K.
"Shell's Flue Gas Desulfurization Unit at Showa Yokkaichi
Sekiyu K.K." Petroleum International, Vol 14 No. 4,
April 1974.
Groenendaal, W., Naber, J.B., Pohlenz, J.B., "The
Shell Flue Gas Desulfurization Process - Demonstration
on Oil - and Coal-Fired Boilers" AIChE National
Meeting, March 10-13, 1974.
Pohlenz, J.B., "The Shell Flue Gas Desulfurization Process"
EPA Symposium of Flue Gas Desulfurization, Nov. 1974
826
-------
TABLE 1
BASIS FOR CASE STUDY
MW
Heat Rate, Btu/Kwh
HHV, Btu/lb
Sulfur, Wt-%
Flue Gas Rate, SCFM
Flue Gas Rate, NM3/hr
S02 in Flue Gas, Vol-%
Inlet Sulfur, kg-mole/hr
Desulfurization Efficiency,%
Reactors Accepting/Regenerating
Acceptance Time, Min
Regeneration Time, Min.
Valve Change & Purge, Min.
Average hourly SO?
Rate Accepted, Kg-mole/hr
Net Average Loading At End
of Acceptance Cyele,a
CASE I
150
9100
8500
0.32
290,000
466 ,000
0.035
7.284
85.0
3:1
290.0
91
6
6.19
0.64
CASE II
150
9100
8500
3.89
306 ,000
492,000
0.35
76.80
85.0
2:1
52.73
20
6
65.28
0.77
Input Heat Rate x 10~6 kcal/hr 343.98
343.98
827
-------
TABLE 2
PROCESS UTILITY REQUIREMENTS
AS % OF STATION INPUT
CASE I
CASE II
Acceptance
Heat Release, MMKcal/hr
% of input
Desorption of
Heat Release, MMKcal/hr
% of input
Total H intake, kg-mole/hr
% of input
H2 to vent, kg-mole/hr
% of input
H.2 purged, kg-mole/hr
% of input
Total steam intake, kg-mole/hr
% of input
0.823
0.24
0.107
0.03
20.6
0.58
1.28
0.02
0.68
0.01
87.'12
0.28
8.043
2.34
0.939
0.27
238.5
6.68
54.28
1.08
4.98
0.10
781.16
2.50
Additional Heat Recovery
Temperature advantage,
% of input
40
1.87
System Pressure Drop
AP, cm H20 30.2
Electrical requirements, Kwh/D 30241
% of input 0.31
Temperature raise, AT°C 5.6
% of input 0.25
39.1
41685
0.43
7.2
0.33
Flow Dampening
Electrical requirements,Kwh/D 730
% of input 0.01
Net Steam Produced, kg-mole/hr 7.5
% of input 0.02
8500
0.09
71.5
0.23
828
-------
TABLE 3
SUMMARY
CASE I
E S F
(0.24)
0.28 0.58
(0.02)
(0.01)
(0.03)
CASE II
E S F
(2.34)
2.50 6.68
(I. 08)
CO. 10)
(0.27)
Acceptance
Desorb SO?
(Vent H2)
(Purged H2 )
Heat Release
Additional Heat
Recovery -- (1.87)
Pressure Drop
0.31 (0.25) 0.43 (0.33)
Dampen Flow 0.01 (0.02) 0.09 (0.23)
0.32 0.26 0.03 0.52 2.27 0.69
Net energy requirement
% of input 0.61 3.48
829
-------
TABLE 4
SO? WORK-UP
CASE I CASE II
E S F E S
Fuel Requirements - 0.27 - 2.85
CH4 credit (0.06) (0.53)
H9 credit (0.03) (0.32)
Steam credit (0.09) (0.94)
Fuel credit for
elemental
sulfur produced (0.19) (2.24)
(0.09) (0.01) - (0.94) (0.24)
Net credit overall (0.10) (1.18)
830
-------
TO
STACK
00
OJ
OPEN
BYPASS
FLUE
GAS
400 °C
FIGURE 1
SIMPLIFIED PROCESS FLOW DIAGRAM
CASE 1
REGENERATION GAS
\V/
\V/
\v/
'6—I—6-J I L6--I—^6
Jo 1 I L..o|. o
** M >
BLOWER
TO
SULFUR
RECOVERY
UNIT
-------
CO
C/4
FIGURE 2
DEW POINT AS A FUNCTION OF H2SO4 CONCENTRATION
0.01
0.10
.0
10
100
1000
SO3(H2SO4) IN FLUE GAS, ppm
-------
THE DOWA'S BASIC ALUMINUM SULFATE-GYPSUM FLUE GAS
DESULFURIZATION PROCESS
Yoshikazu Yamamichi and Jun-ichi Nagao
The Dowa Mining Co., Ltd.
Okayama Research Laboratory
Chikkosakae-machi 7-2, Okayama, Japan
ABSTRACT
The Dowa Mining Company recently developed a basic aluminum sulfate-
gypsum process to desulfurize waste gas from smelters, sulfuric acid
plants and steam boilers. In this process, SO is absorbed in a clear
solution of basic aluminum sulfate in an absorbing tower. The resulting
solution is oxidized by air in an oxidizing tower and then neutralized
with finely-divided limestone to precipitate gypsum and to regenerate
the basic aluminum sulfate solution.
After several tests on a small scale, the first commercial plant
(3,500 Nm /H) was constructed in October 1972 to desulfurize waste
gas from a molybdenum sulfide roaster containing 7,500 ppm SO at
The Taenaka Mining Company.
In June 1974, two large units each treating respectively 140,000
Nm /H tail gas from a sulfuric acid plant were started up at the
Okayama Works of Dowa.
Another three plants are now under construction. One is for an
oil burning boiler, another for iron ore sintering and the third for
an converter of antimonical sulfide converter.
833
-------
THE DOHA BASIC ALUMINUM SULFATE-GYPSUM FLUE GAS DESULFURIZATIOSF PROCESS
I. nistcry of the Process Development
Dowa fining Co., Lta, is one of the major oroducers of non-ferrous ir.etals
such es cooler, zinc, lead, etc., in Jaoan a.nd ergages in mining, smelting
and refining, and research in these fields. Also in the field of sulfuric
acid manufacturing, JJowa is the largest oroducer in Japan.
Until 1967 the acid plant tyoe to be constructed were the so-called
"single contact type", so Dowa, hss been investigating and developing method
for eliminating sulfur dioxide from acid slant tail gas over the past twenty
years. For examole, we have been scrubbing the acid plant tail gas with
sodium sulfite solution in our Okay a ma plant for eighteen years and with slaked
lime milk in our Kosaka plant for eight years.
bince development of the more efficient "double contact orocess" for
sulfuric acid manufacturing, it seemed unnecessary to treat the acid plant
tail gas any more, riut in our country, the governmental regulation for
preventing air pollution has become stricter year by year, so we have had to
eliminate more than 90$ of sulfur dioxide even from double-contact acid olant
tail gas.
with suclj background, we have investigated to develop a new desulfuriza-
tion process using basic aluminum sulfate solution as an a.bsorbent in our
Okayama la.bora.tory , the work beginning in 1970.
After many laboratory and pilot t>lant tests, the commercial plant was
built, by request of Taenaka fining Co., a.nd has operated successfully from
Oct. 1972. Gas conditions are 7,530 ppm S0? , 3,500 Nmf/h, and 1DO°C. The
gas comes from a. rotary kiln calciner for MoSp fired by natural gas.
tie then constructed the two trains to eliminate sulfur dioxide from two
double-contact acid plants. This was done a.t the same time the second 1,350
T 'D acid ulant was constructed in our Okayema nlant in 1974.
-Design conditions of each plant are as follows;
Gas volume to be treated:
Gas terr.perature :
Gas comoositions:
Adiabatic gas temperature (ts):
Removal of bOp:
Liquid/gas ratio: (l/iJm )
-biameter of the absorbing tower:
Backing height:
834
140,000 ixfm
80"C
SOj, oOO Dom; S03, 5Qporr.
^2* 5f^; -tf? , remainder
?8"C
more than 95f
2.^ - 3
5-8 m
? m
-------
These olants have operated successfully since start-up in July 1974? with
an average removal efficiency of sulfur dioxide during the twenty months of
99^-
Three more plants are now under construction for various tyoes of waste
gas. At the start of this research and development, we aimed to establish a
new process which would fix S02 in waste gas as gypsum. For the neutralizing
reagent, limestone is more oreferable to lime because the latter costs about
twice as much. Moreover, a simple process and cc-^nact equipment with no
plugging problems are desireble. After much tria± and error, we finally decided
on basic a.luminum sulfate as the absorbent.
I. Descriotion of the Process
As shown in Figure 1, Dowa's "basic aluminum sulf ate/gypsum process"
consists of the following four simple unit processes, namely;
l) Absorbing
Tail gas is sent to the lower part of an absorbing tower and contacts
the basic aluminum sulfate solution countercurrently- Sulfur dioxide contained
in the waste gas is absorbed into the solution by the following chemical re-
action.
Al2(S04)3-Al203 + 3S02 - ^ A12(S04)3-A12(S03)3 ---------- (l)
2) Oxidizing
The resulting flows to an oxidizing tower where air is injected through
soecially designed nozzles. Very fine air bubbles are developed to accelerate
the oxidizing reaction. Sulfite in the solution is oxidized to sulfate by
the following chemical reaction.
A12(S04)3-A12(S03)3 + 3/2 02 - ^ 2A12(S04)3 ------------- (2)
After oxidation, the major part of the solution is recycled to the absorb-
ing tower.
3) Neutralizing
A small part of the oxidized solution is continuously sent to a neutraliz-
ing tank where limestone powder is added to neutralize the solution automatical-
ly until the basicity (see later definition) of the solution reaches an optimum
level.
2A12(S04)3 + 3CaCOj + 6H20 - =»
A12(S04)3-A1203 + 3CaS04-2H20 + 3CO? ---------- (3)
Then the slurry overflows to a thickener to separate the gypsum crystals.
Large crystal size can be attained by recycling a portion of the gypsum slurry
from the thickener to the neutralizing tank.
835
-------
4) Filtering
Underflow from the thickener is pumped to a centrifuge where the gypsum
is filtered and washed with water automatically- The thickener overflow, the
filtrate, and the washed water are recycled to the absorbing tower- Also, as
shown in Figure 1, the flow of all solutions is completely closed, and there
is no effluent drained out from the system.
GAS EXIT
Tower AirCompresor
Demisrer
Packing Zone
WASTE
GAS
INLET
Washing Water
Centrifuge
Absorbing
Tower
Figure 1 Flow Diagram of Dowa's Aluminum Sulfate/Gypsum Process
M, Principle of Absorption
Basic aluminum sulfate solution is a well-known sulfur dioxide absorbent,
and was once used in I.G.I.'s process for making liquid sulfur dioxide. This
solution is easily made from the liquid aluminum sulfate available on the
market by neutralizing with lime powder. Basicity is defined as the degree
of neutralization of aluminum sulfate solution, namely;
A12 (S04 )3
basicity, 0 ^
basicity, 50 ^
i°
basicity, 100
Al(OH)3
(l-x)Al2(S04)3.xAl203 basicity, 100x$
Aluminum sulfate dissolves easily in water to give a solubility of 360 g/l»
equal to 57 g/1 aluminum concentration. We have measured the absorption
curves at various levels of concentration and basicity of basic aluminum
sulfate solution. General speaking higher concentration, higher basicity,
836
-------
and lower temperature give higher S02 absorption, but even at the required
operating conditions of lower than maximum concentration and basicity, solution
has much better absorption capacity than water.
Figure 2 shows SC>2 absorption curves for lower levels; of aluminum concent-
ration. Figure 3 gives curves as a function of basicity and temperature when
1,000 ppm SOp is present in gas phase. Even 5-7 g/1 aluminum solution has ten
times the absorption capacity of water-
Figure 4 indicates the range of aluminum solubility at various basicities.
The basicity should be lower than 40^ in commercial operation to prevent
plugging the packings in the absorbing tower caused by aluminum precipitation
from the solution. Also, Figure 5 shows that minimum aluminum loss in the
gypsum will be obtained at a concentration of 15 to 20 gpl of aluminum in the
solution.
As for the liquid/gas ratio, owing to the good SO? absorbility the ratio
can be relatively small with good S02 recovery. The L/G ranges from 2.5 to
5.0 1/m3, depending on solution temperature and by SO? and 0? content in waste
gas. It will be necessary to use an L/G of 10 or more when flue gas from
boilers, and metallurgical plants is being cleaned; the higher gas temperature
reduces absorotion capacity as compared to sulfuric acid plants. However we
found a way to decrease this ratio by addition of a snail amount of a certain
inorganic catalyst into the absorbent. This catalyst activates the oxygen
contained in the flue gas and promotes the oxidation reaction of SO^ to SO^
in the solution during the absorbing step. For instance, our pilot plant
results indicate that we can use on L/G of 5 "to remove more than 95$ of the
S02 when boiler flue gas containes 5$ oxygen, and 8 when it containes 3f oxygen.
A rapid rate of oxidation is attained in the weakly acidic solution in
the oxidation tower, with a resulting retention time of only a few minutes.
We use a newly developed oxidation tower, in the bottom of which are installed
special bubble-making devices. And at 3 kg/cm2'of pressure is blown into the
tower. Oxygen consumption is 60~JOtf. of that fed.
Limestone is ground to a sizedistribution of over QQff through 200 mesh.
Through every step of this process, the aluminum stays in solution and
this is fully recoverable. However, because of incomplete washing, some loss
is inevitable, but only about 0.5 kg per 1 ton of gypsum. The cost is very
small as compared with the total cost.
IV. Problems Encountered and Solutions
1) Taenaka plant
The first plant in Ta,enaka Mining Co. started in Oct. 1972, treats a
waste gas from MoS? calcination in a natural gas-fired rotary kiln. A TCA
scrubber is used at an L/G of 5? giving 95/^ ^0? removal. The solution
temperature is 40°C, inlet SO? 7,500 ppm, HpO content of gas 4$, oxygen content
10$, gas flow 3,500 Nm3/h, and inlet gas temperature 100°C. Gypsum production
is 100 mt/d. There ha.ve been no problem with plugging.
837
-------
20
o
CO
c
o
O
CD-
CO
10
A. Al 378 ^| Basicity ,9.3%
B. Al 1 1.9 &1 '' 149%
C. AJ 57 3/l * 14.9%
D. Woter
c
/
0
L
/
/
,A
/
i
^°
J>^
^
1 2 2
^r
/-^
o
1 !
Temp, 20°C
o-
3
-D
5456
S02 Content In Gas ( x 10* PPM )
Figure 2 Solubility curves
of S02 in BAS solutions
20
A. AI 17.3% (20«C)
B. AI 10.8 ty (20°O
C. Al 173 9/1 (5O°C)
D. Al IQ8 9/j (50eC)
Basicity (%)
Figure 3 Solubility curves
of S02 at various temoeratures
of the solution
Basicity (%)
Figure 4 Hange of Alrainum
"being soluble at various
basicities of the solution
i?
6
o
^c
0)
4-
c
o
O
0 5 JO C 20 ^5
Al Content in the Solution(^j)
Figure 5 Al loss in gypsum
at various A.1 content in the
solution
838
-------
?) Okayama plant
During the twenty months of operation in our Okayama plant, the main
problem we have encountered was how to measure the basicity of the absorbent.
First, we measured the pH value of the solution; however the pli didn't vary
linearly with basicity. Next we selected electrical conductivity of the
absorbent as a possible measure of basicity. This gave us very good results
at the beginning, but the magnesium ions coming from the lime added for
neutralization gradually accumulated in the soluLon and interfered because of
the much higher conductivity of magnesium ions. At last we found a new method
to measure the basicity of the absorbent.
Solution
from
Main Line
(Extra)
Neutralizing
Tank
-CaC03
to Main
Neutralizing
Tank
Setlling Tank
Storage Tank
Figure 6 Flow-sheet of de-magnesium unit
Although magnesium ion does not disturb S02 absorption, it is necessary
to prevent the accumulation of magnesium ions to produce marketable gyosum.
We solved this problem by adding a small neutralizing unit to our plant. The
additional neutralizing unit consists of a small neutralizing tank and a settl-
ing tank as shown in Figure 6. First, a small part of the absorbent is neutra-
lized with excess limestone and the aluminum in the solution is precipitated
with the gypsum. Since the magnesium remains in solution, the overflow from
the settler can be thrown away to keen the concentration of magnesium ion at
a certain level in the absorbent. Underflow from the settler is sent to the
main neutralizing tank to avoid loss of aluminum.
V. Cost Analysis
From the results of twenty months of operation, the direct operating cost
of our Okayama desulfurization plant is estimated at 20 Ub cents per 1,000 Nm3
of gas treated. The items contained in this estimate are as follows;
839
-------
Plant capacity of 280,000 3m3 /hr and S02 650 ppm
CaC03 ------------------------------ 0.81 t/hr (15 A)
A1(S0) solution (A120 8#
243
Water
15 kg/hr (4 d/kg)
7 t/hr
Electric power ---------------------- .000 K.WH
Operator ___________________________ one person -oer shift
Figure 7 shows a plant construction cost estimate for boiler flue gas
desulfurization in Japan in 1975-
too
10
v
•
e»
(U
1
-O-
0 ZS 3.0
Estimated Plant Cost
(xlO«$)
Figure 7 Boiler flue ^"s desulfurization ulant
cost estimation in Jana.n
VI. Application of our Process
We sumr;iari/;e desulfurization olants usin^ our process as follows:
840
-------
Completion
of construction
Oct. 1972
Jun. 1974
Oct. 1974
under
construction
under
construction
under
construction
Plant
name
Taenaka
Mining Co.
Okayama
Okayama
Naikai Co.
Y
Itfihon
Seiko Co.
Gas
source
M
A
A
B
M
M
Gas condition
Volume
Nm3/Hr
3,500
140,000
140,000
80,000
50,000
30,000
so2
ppm
7,500
6jO
600
1,500
3,000
5 , ooo
°2
#
10
5
5
4
18
19
Temp.
°C
100
80
80
250
250
100
ts
°c
40
28
?8
55
56
34
S02 removal
efficiency
%
95
99
99
95
95
95
Note K: Metallurgical furnace
A: Acid plant
B: Boiler
VI. Advantages of the Process
Compared with other desulfurization processes, Dowa's "basic aluminum
sulfate/gypsum orocess" has the following advantages,
l) £asy operation by a simple process; One stage filtration.
2) Stable plant operation because there is no slurry in the absorbing
and oxidizing steps.
3) Low operating cost because of using limestone as the neutralizing
reagent.
4) Small equipment size because of low liquid/gas ratio.
5) Gypsum with good quality is recovered as by-product.
6) No secondary pollution since the system is closed circuit system.
Acknowledgement
The authors are indebted to the management of the Dowa Mining Co., Ltd.
841
-------
for their kind permission to oresent this paper. We also grateful to manager
S. Koh, former manager T. Otsuka and T. Morikawa and others involved at the
Gkayamn Works; to those involved at the Okayama .Research Laboratory; and to
manager T. Tomoda and others involved in Kowa Engineering Co., Ltd.
Photo: A bird's-eye view of ?80,000 Nm3/Hr (140,000 Nm3/Kx?) plant
842
-------
CITRATE PROCESS FOR FLUE GAS DESULFURIZATION
A STATUS REPORT
W. I. Nissen, D. A. Elkins, and W. A. McKinney
Bureau of Mines
Salt Lake City Metallurgy Research Center
Salt Lake City, Utah
ABSTRACT
The Federal Bureau of Mines citrate process for removing SO
from stack gases contains two basic steps, (1) absorbing S0? in a
solution of sodium citrate, citric acid, and sodium thiosulrate, and
(2) reacting the absorbed S0_ with H0S to precipitate elemental sulfur
and regenerate the citrate solution for recycle. The Bureau's research
and development of the citrate process includes the operation of a
pilot plant to assess feasibility for flue gas desulfurization.
The pilot plant treats 1,QQO cubic feet per minute of a 0.5-percent SO
flue gas from a lead sintering furnace, removes more than 95 percent
of the SO and produces about 600 pounds of sulfur per day. It has
operated for a total of about 4,500 hours and produced more than 50
net tons of sulfur. Plans are being made to demonstrate the citrate
process on a 50- to 100-megawat (MW) base-loaded electrical utility
burning high-sulfur coal. Operating costs for a citrate plant,
including capital charges, installed at a 1,000-MW base-loaded electric
utility burning high-sulfur coal are estimated to be about 3.8 mills
per kilowatt-hour.
843
-------
CITRATE PROCESS FOR FLUE GAS DESULFURIZATION,
A STATUS REPORT
INTRODUCTION
In 1968, the Bureau of Mines Salt Lake City Metallurgy Research Center started
research on flue gas desulfurization with particular emphasis on control of S02
emissions from the nonferrous smelting industry. Pioneering research had
indicated that flue gas desulfurization might be achieved effectively by the
absorption of sulfur dioxide (S02) in a suitable solution, followed by reaction
of the absorbed S02 with gaseous hydrogen sulfide (H2S) to precipitate sulfur and
regenerate the solution for recycle. After a year of screening many possible
reagent combinations of inorganic and organic solutions, it was established that
an aqueous solution of citric acid and sodium citrate was an effective absorbent
for S02. Among the desirable characteristics affecting the choice of citrate
were its' good chemical stability, low vapor pressure, and adequate pH buffering
capacity, and the purity and physical character of the precipitated sulfur.
As first studied in the laboratory, the citrate process comprised (l) absorbing
S02 in citrate solution, (2) reacting the absorbed S02 with H2S, (3) filtering
and melting the precipitated sulfur, and (k} recycling the regenerated citrate
solution to the S02 absorption step. Preparation of the H2S required for the
precipitation step of the citrate process was studied separately. Hydrogen
sulfide was generated by reacting recycle sulfur, natural gas, and steam over
an alumina catalyst.
Subsequently, in 1970 and 1971 > a pilot plant to remove S02 from flue gas from a
copper reverberatory furnace was constructed and operated jointly by the Bureau
of Mines and the Magma Copper Co. This plant, located at the San Manuel smelter
in Arizona, treated 300 standard cubic feet per minute (scfm) of gas containing
1.0 to 1.5 percent S02 and consistently removed 93 to 99 percent of the S02.
Results of the initial laboratory and pilot plant research were reported in 1970
and 1971 (3_, 8).i/
The preliminary Bureau of Mines laboratory and pilot plant research demonstrated
that the citrate process is capable of substantially complete removal of S02
from industrial waste gases. Most of the S02 is converted to sulfur with less
than 1.5 percent converted to sulfate regardless of the S02 and oxygen content
of the feed gas. The citrate process produces an elemental sulfur end product
that can be marketed or stored with a minimum of environmental disturbance.
After the encouraging preliminary results, two other pilot plant investigations
were undertaken to obtain data for engineering evaluation and cost estimates.
One pilot plant independently built and operated by Arthur G. McKee and Co.,
Peabody Engineered Systems, and Pfizer, Inc., at Terre Haute, Ind., is described
in four publications (l, 2, k_, 10). This pilot plant operation treated a stack
gas from a coal-fired steam-generating station simulating a utility application.
After several modifications to arrive at a final equipment configuration, the
pilot plant operated from March 15 to September 1, 197^, logging 2,300 operating
hours. The longest sustained run was 180 hours. The Terre Haute pilot plant
consistently removed 95 to 97 percent of the S02 from flue gas containing 0.1 to
0.2 percent S02. The other pilot plant was constructed by the Federal Bureau of
Mines and operated jointly by the Bureau of Mines and The Bunker Hill Co. at the
lead smelter in Kellogg, Idaho. The design, construction, and operation of the
Bureau's Bunker Hill citrate pilot plant are described in three publications (5-7)
I/Underlined numbers in parentheses refer to the list of references at the end
of this report.
844
-------
The present report describes the further operation of the Bureau's Bunker Hill
citrate pilot plant operation, process development, plans for a 50- to 100-MW-
scale citrate process demonstration plant, and estimated costs for a 1,000-MW
base-loaded electrical utility burning high-sulfur coal based on the pilot plant
operation.
PROCESS DESCRIPTION
As a result of further process developments, the citrate process, as new-
envisioned and shown in figure 1, comprises the following steps:
1. The S02-bearing gas is cooled to between ^5° and 65° C (113° and 1^9° F)
and cleaned of sulfuric acid (H2S04) mist and solid particles.
2. The S02 is absorbed from the cooled and cleaned gas by a solution of
sodium citrate, citric acid, and sodium thiosulfate.
3. Absorbed S02 is reacted with H2S at about 65° C (1^9° F) and atmospheric
pressure to precipitate elemental sulfur and regenerate the solution for
recycle.
h. Sulfate is removed from a slipstream of regenerated recycle citrate
solution by cooling the solution and crystallizing Glauber's salt
(Na2S04'10H20), which is removed by filtration and then washed.
5. Sulfur is separated from the solution by oil flotation and melting.
6. The H2S for step 3j> if not otherwise available, is made by reacting
two-thirds of the recovered sulfur with natural gas and steam.
CHEMISTRY OF THE PROCESS
Absorption of S02 in aqueous solution is pH dependent, increasing at higher pH.
Because dissolution of S02 forms bisulfite (HSOs) ion with resultant decrease in
pH by the reaction
S02 + H20 ?± HSOg + H+, (1)
the absorption of S02 in aqueous solution is self-limiting. However, by incor-
porating a buffering agent in the solution to inhibit pH decrease, high S02
loadings and substantially complete S02 removal from waste gases can be attained.
The principal function of the citrate or other carboxylates that have been
tested is to serve as a buffering agent during S02 absorption.
The chemistry for the production of sulfur (S°) and regeneration of absorbent by
reacting H2S with the S02 in the aqueous solution is complex, but the overall
reaction is as follows:
S0£ + 2H2S -* 3S° + 2H20. (2)
Actually, thiosulfate (S203) and polythionates are found in solution at equilibrium
concentrations after several S02-absorption and H2S-regeneration cycles. Oxidation
of S02 in the aqueous_solution is sharply depressed by the complexing of HSOs from
reaction 1 by the S20s ion, according to the following reaction:
H+ + HSOs + S20s *> (S02-S203)= + H20. (3)
845
-------
GAS CLEANING j S02 ABSORPTION J SULFUR PRECIPITATION I SULFATE I
AND ' S AND REMOVAL 1
SOLUTION REGENERATION | I
AND
COOLING
SULFUR
RECOVERY
H2S GENERATION
CO
Cleaned and
Flue
gas-
" cooled gas
* I
Acid water
and solid
participates
To gas reheater
and stack I
I I
S02
liquor
H2S-C02
Recycle liquor
C02
^
slurry
Sodium
sulfate
y
Sulfur
i \
I
s
j
n '
'i
Sulfur
powder
Molten sulfur
Steam
FIGURE I.- Generalized citrate process flowsheet
D-26I4-SL
-------
To insure satisfactory operation of the system on startup, sodium thiosulfate
(Na2S203) is added to the initial absorbing solution.
Hydrogen sulfide for regenerating the absorbent and precipitating elemental sulfur
can be produced by reacting sulfur with methane (CH4) and steam (H20), as shown in
reaction h:
+- hS° + 2H20 -* C02 + k-E2S.
Other reducing gases, such as hydrogen and carbon monoxide, individually or
combined, can be used in place of methane.
More detailed information on the chemistry of the citrate process is provided in
a Bureau of Mines publication (9) and a paper presented at the American Chemical
Society National Meeting in April 197^- (]+)•
BUNKER HILL CITRATE PILOT PLANT
Details of construction of the Bunker Hill citrate pilot plant are described in
previous Bureau of Mines papers (5-7) • The nominal capacity of the pilot plant
is 1,000 scfm of 0.5 percent S02 gas yielding about 600 pounds of sulfur per day.
A block diagram of the complete pilot plant is shown in figure 2. The plant will
be discussed in three sections:
1. Gas cooling and cleaning required prior to S02 absorption.
2. Citrate absorption system, including S02 absorption, S02 reduction to
regenerate the absorbing solution and precipitate sulfur, and sulfur
recovery.
J>. Generation of the H2S required for the reduction of absorbed S02.
Between February 197^ and November 1975; the citrate pilot plant operated for more
than ij-,500 hours and produced more than 50 net tons of sulfur. The only antici-
pated further operation will be a h- to 6-week campaign in the spring of 1976.
Gas Cooling and Cleaning
A slipstream of the lead smelter sinter plant tail gas, which contains dust, acid
mist, and from 0.3 to about 1 percent S02, was used as pilot plant feed. This
sinter tail gas normally passes through a baghouse and then is discharged to the
atmosphere through Bunker Hill's main stack. To simulate conventional lead
smelter practice, the gas cooling and cleaning section was designed to recover
most of the valuable dust from 1,000 scfm of the tail gas in a baghouse, cool the
gas in a packed wet scrubber, and remove H2S04 mist and traces of particulate
matter with a wet electrostatic precipitator. One of the goals in the pilot
plant operation was to determine the minimum gas cleaning requirement compatible
with the citrate process.
Shakedown runs of the gas cooling and cleaning section of the pilot plant showed
that the slipstream of sinter plant tail gas had an S02 concentration of up to
2 percent, which was higher than anticipated and had to be diluted with air to
avoid exceeding the pilot plant's capacity of 1,000 scfm of 0.5 percent S02 gas.
The gas also required heating to prevent acid mist condensation and corrosion in
the baghouse. After making these changes, the gas cooling and cleaning section
functioned satisfactorily and was operated with the S02 absorption-regeneration
section of the pilot plant.
847
-------
Co
.fc»
Co
300-gal absorbent makeup tank
Sinter plant tail gas
1
1,200-sq-ft boghouse
-*-Flue dust
2.5-ft-diam by 18-ft-high packed scrubber tower
*» Particulate sludge
Electrostatic mist precipitotor
300-gal absorbent feed tank
•*• H2S04 mist
2.5-ft-diom by 30-ft-high packed absorption tower
Absorbent solution
50-gol recycle
liquor tank
Oi!
50-gai recycle
liquor tank
Treated flue gas
-*> to atmosphere
Three 100-gal H2S precipitation reactors
•C05
•Oil
100-gal oil-sulfur contactor
H2S-C02
50-gal sulfur flotation vessel
200-cu-ft sulfur float product storage bin
Steam
l-ft-diam by 4-ft-long
sulfur autoclave settler
2/3 molten
sulfur
•Natural gas
ton/day
generator
1/3 molten sulfur
cast into 100-lb molds
FIGURE 2.-Bunker Hill pilot plant.
-------
Some pilot plant operations were conducted on sinter plant tail gases that were
diluted below 0.5 percent S02; these would be similar in S02 concentration to
coal-fired powerplant flue gases. During these operations the sinter plant tail
gases were diluted to between 0.07 and 0.13 percent S02.
During the operation of the pilot plant, investigations were made of the
effectiveness of and requirements for the gas cooling and cleaning equipment.
The baghouse was used during all of the operation. The effect of bypassing either
the electrostatic mist precipitator or both the electrostatic_mist precipitator
and the packed wet scrubber on S02 absorption and sulfate (SO^) buildup in the
citrate scrubbing solution was studied.
Gas samples from the gas cooling and cleaning section showed that the sinter
plant tail gas contained about 3 grains of dust per cubic foot (7-0 grams per
cubic meter). When the tail gas was diluted, the average dust loading was about
0.6 grain per cubic foot (1.4 grams per cubic meter). Gas sampling results showed
that over 99-6 percent of this dust was removed in the baghouse, leaving less than
0.002 grain of dust per cubic foot (k.6 milligrams per cubic meter). Gas samples
for dust contained in the packed wet scrubber or the electrostatic mist precipi-
tator were not taken. However, these pieces of gas cleaning equipment were assumed
to remove some dust.
Analysis showed that the dilute sinter plant tail gas contained 0.02 to 0.5
milligram S03 per cubic foot instead of the 2 milligrams it was originally esti-
mated to contain. Although this low S03 content is equal to, or less than, the
guaranteed S03 content of the gas at the outlet of the electrostatic mist precipi-
tator, about two-thirds of the S03 from the diluted sinter plant tail gas was
removed by the combination of cleaning equipment. The design intention of the gas
cooling and cleaning section was to use water sprays to precondition the tail gas
and form H2S04 mist. The water sprays were not used owing to the unusual method
of operation, which was to dilute and heat the tail gas. With little H2S04 mist
formed, it was doubtful that the electrostatic mist precipitator removed much S03.
Besides removing some S03, the packed wet scrubber, when used, provided the
necessary gas cooling. During one period, the packed wet scrubber was bypassed
to determine the effect of a hot dry gas on the_S02 absorption. The only
detrimental effect observed was an increased SO^ buildup in the scrubbing solution.
This indicates that most of the S03 removal occurred in the packed wet scrubber,
as described in more detail in the absorption-regeneration section of this paper.
Despite the problems encountered, the gas cooling and cleaning equipment satis-
factorily cleaned and cooled the gas for the S02 removal step of the citrate
process.
Citrate Absorption
Following gas cooling and cleaning, as shown in figure 2, the S02-bearing flue gas
stream passed through a 2.5-foot-diameter by JO-foot-high packed absorption tower
countercurrent to the citrate solution, in which over 95 percent of the S02 was
usually absorbed. The absorption tower contains three 6-foot sections packed with
1-inch polypropylene Intalox—/ saddles and a stainless steel mist eliminator.
From the absorption tower, the citrate solution flowed by gravity to closed,
stirred vessels for reaction with two-thirds ton of H2S per day to form 1 ton per
day of elemental sulfur. From one to three 100-gallon stainless steel reactor
vessels, which were arranged for countercurrent flow of citrate solution and H2S
Reference to specific trade names is made for identification only and does not
imply endorsement by the Bureau of Mines.
849
-------
generator product gas, were used for the S02 reduction reaction-sulfur precipita-
tion step. A slurry containing 1 to 3 percent solids of elemental sulfur in the
regenerated citrate solution overflowed the reactors and passed through a common
header to a stainless steel reactor effluent tank. From this tank, the dilute
slurry was pumped to a 100-gallon conditioner tank where kerosine or other hydro-
carbon oil was added for the sulfur flotation-separation step.
From the conditioner tank the slurry flowed to a specially designed sulfur
flotation skimming device resembling an Esperanza drag classifier. The sulfur
separated from the oil-conditioned slurry by floating to the surface as a 35- to
•1+5-percent-solids product. Some pilot plant operation was conducted using air
flotation in which no kerosine was added to the dilute slurry. Such flotation
produced a wet sulfur product of about 8 percent solids. Both oil and air
flotation satisfactorily separated the sulfur from the bulk of the citrate
solution, leaving clear regenerated citrate for recycle to the absorption tower.
The sulfur was skimmed off the surface of the citrate solution, pulled up an
inclined chute, and discharged to a conical storage bin.
Regenerated citrate solution from the feed well of the sulfur skimmer passed to
a 300-gallon absorber feed tank. Citrate solution from the absorber feed tank
was pumped through parallel backwash clarification filters and a water-cooled
heat exchanger to the absorption tower.
On day shift only, the sulfur float product was withdrawn from the storage bin and
pumped by a Moyno positive-displacement auger-type pump through a single-tube,
steam-jacketed heat exchanger where the sulfur was melted at about 135° C (275° F).
Molten sulfur and citrate solution passed into a closed settler tank at 135° C
under a pressure of 35 psig. Part of the molten sulfur was tapped from the bottom
of the autoclave settler and cast in 100-pound blocks; the balance was pumped to
the H2S generating plant. Citrate solution and the oil used for flotation were
withdrawn from the top of the settler and then passed through a sulfur knockout
pot and a water-cooled heat exchanger into a decanting vessel for separation and
reuse. Citrate solution from this tank drained into the absorber feed tank.
The pilot plant was completely instrumented and controlled from a panel mounted
in a 40-foot-long instrument trailer that was connected to the pilot plant build-
ing and also served as an onsite laboratory.
An incinerator was provided outside the main building to burn H2S vented from the
system or released under emergency or upset conditions. In a commercial plant,
the offgas from the incinerator would be returned to the gas stream entering the
absorber, but in the pilot plant this offgas was vented to the smelter stack.
Operation of the Bunker Hill citrate pilot plant was started on February 15, 197^.
The plant operated for a total of lj-,500 hours through November 1975, and produced
about 50 net tons of bright yellow, high-quality sulfur. During about lj-50 hours
of this operation, a glycolate scrubbing solution was satisfactorily substituted
for the citrate solution.
Because of interruptions resulting from irregular operation of the lead smelter
sinter plant with frequent unavailability of feed gas, the longest continuous
operation was about 265 hours.
850
-------
Laboratory tests preceding pilot plant studies indicated that a pH of about 4.5
was needed for good absorption. During continuous campaigns in late summer 197^
when only one of the three precipitation reactors was being used, the pH of the
recycled citrate solution dropped below 4.5, and S02 absorption efficiency sub-
sequently decreased. As the pH dropped from 4.5 to about 4.0, the S02 removal
efficiency at the design gas flow of 1,000 scfm of 0.5 percent S02 feed gas at
45° C (115° F) dropped from 99 to 85 percent. During this time, the S02 content
of the offgas from the pilot plant increased from about JO parts per million
(ppm) to 750 ppm. Laboratory regeneration tests on samples of the pilot plant
citrate solution showed that a combination of low pH, high polythionate content,
and low thiosulfate content resulted from incomplete regeneration of the
absorbent solution. The principal cause of incomplete regeneration was found
to be insufficient retention time of H2S gas in the single sulfur precipitation
reactor. Apparently, slightly more contact time than that provided in one
reactor was necessary for effective utilization of the H2S. If the H2S-S02
reaction for producing elemental sulfur is not allowed sufficient time, more
polythionates, principally in the form of polythionic acids, are produced, per-
mitting the pH of the solution to drop. Low solution pH causes two problems, the
more critical of which is a drop in the S02 absorption efficiency. This problem
was solved by (l) increasing the contact time of H2S with the loaded liquor by
using two reactors with countercurrent flow of H2S and citrate solution, (2) add-
ing sodium thiosulfate to bring the concentration back up to the desired level,
and (3) adding sodium carbonate to neutralize the sulfuric acid in the plant
solution. These measures were successful in regenerating the pH at 4.5 and
restoring high S02 absorption capability to the solution. In subsequent campaigns,
the indicators of incomplete regeneration (low pH, high polythionates, and low
thiosulfate) were closely monitored. The second problem was thermal decomposition
of thiosulfate in the sulfur-melting system, which accelerates at pH k or below,
thus increasing the sulfate concentration of the solution. In practice, this
would require additional sodium carbonate for neutralization and purging of the
additionally formed sodium sulfate from the system.
In the early stages of plant operation, excess H2S absorbed in the citrate
solution resulted in cloudy recycle solution being recovered from the kerosine
flotation step, apparently due to delayed precipitation of colloidal sulfur- Some
of the absorbed H2S escaped at times from the sulfur skimmer into the plant build-
ing; this was corrected by using the second or third stirred reactor as a delay
tank to allow more contact time with the H2S and bypassing about 5 volume-percent
of the S02-loaded solution from the absorption tower to the reactor effluent tank,
where it reacted with the excess absorbed H2S. These measures have resulted in
consistently clear citrate solution from sulfur flotation for recycling to the
absorption tower and have stopped the escape of H2S from the sulfur flotation
unit.
During precipitation of sulfur with H2S gas, sulfur buildup along the walls of
the stainless steel reactors or on the impellors was minimal when the tip speed
of the impellors was at least 900 feet per minute. However, sulfur plugs were
a chronic problem in the pipelines between reactors and required periodic cleanout.
The kerosine conditioner tank operated well with no sulfur buildup at the design
liquid flow rate when the impellor tip speed was at least 700 feet per minute.
Initially, some trouble was experienced with holdup of floated sulfur in the
freeboard of the tank. This required an occasional cleanout of the J-inch-
diameter overflow line to the skimmer- However, the addition of a second
impellor operating just beneath the liquid surface prevented buildup of the large
lumps of powdery floated sulfur that were blocking the overflow line.
851
-------
Some plugging problems were encountered in the citrate solution lines from the
autoclave settler, apparently because of sulfur being dissolved in the kerosine
flotation reagent and then crystallizing out upon cooling. Such problems were
reduced by installing an additional knockout, heat-exchanger vessel. The automatic
sulfur level control, which controls the molten sulfur flow from the autoclave
settler, gave some trouble, and at times scrubbing solution was transferred through
the molten sulfur pipeline to the H2S generator when the automatic control mal-
functioned. This scrubbing solution caused plugging in the liquid sulfur lines
and pumps. Accurate material balances to determine citrate losses could not be
made whenever such scrubber solution losses occurred.
Table 1 summarizes results obtained under steady-state continuous operation at gas
flow rates between 1,000 and 1,300 scfm and S02 concentrations of 0.07 to 0.55
percent. The absorption temperatures ranged from J>k° to 58° C (93° to 136° F).
Citrate solution flow rate in the absorber was 7-5 and 10 gallons per minute,
citrate concentration was 0.5 M, the sodium-to-citric acid molar ratio in the
citrate solution was 2:1, and the pH of the citrate feed solution to the
absorption tower was about k.k to h.8.
TABLE 1. - Results of Bunker Hill citrate pilot plant
operation, February
Feed
Concen-
tration,
pet S02
0.07
.11
.13
-52
,55
gas
Flow,
scfm
1,000
1,100
1,300
1,000
1,000
Citrate
Flow,
gal/min
7-5
7,5
7.5
10
10
solution
Loading,
g/1 S02
1.9
3.2
l+.l
10.7
11.2
1974-October 1975
Absorption
temper-
ature,
0 F
93
109
136
100
96
Off gas
concen-
tration,
ppm S02
2
5
61
hk
k6
S02 removal
efficiency,
pet
99-7
99.6
95.3
99.1
99-2
The citrate pilot plant was designed to treat 1,000 scfm of 0.5 percent S02 gas at
113° F with 10 gallons per minute of 0.5 M citrate solution. Because of the
unusual mode of operation of the gas cooling and cleaning section of the pilot
plant, the gas temperature at the inlet of the S02 absorption tower ranged from
90° to 100° F. As shown in table 1, the S02 removal efficiency was 99.2 percent
with the offgas containing U6 ppm when treating 1,000 scfm of 0.55 percent S02 gas
at 96° F and using 10 gallons per minute of 0.5 M citrate solution. However,
preliminary test results of the citrate absorption section, which were reported
previously (6), showed that under design conditions (l,000 scfm of 0.5 percent gas
at 113° F and 10 gallons per minute of 0.5 M citrate solution), an S02 removal
efficiency of 99 •& percent and an offgas containing less than 30 ppm S02 were
obtained. For this test, the solution loading of 9.8 grams S02 per liter repre-
sented 55 percent of the maximum equilibrium loading of 0.5 M citrate solution at
the absorption temperature of 113° F. Preliminary test results also showed that
the S02 removal efficiency exceeded 96 percent and the offgas contained less than
200 ppm S02 even when the feed gas temperature was increased to 65° C (1^9° F).
As shown in table 1, excellent results were also obtained when treating gases
containing from 0.07 to 0.13 percent S02, which would be similar in S02 concen-
tration to coal-fired powerplant flue gases. When treating 1,100 scfm of 0.11
percent gas at 109° F and using 7.5 gallons per minute of 0.5 M citrate solution,
an S02 removal efficiency of 99.6 percent and an offgas containing 5 ppm S02 was
obtained. The solution loading of 3.2 grams S02 per liter represents about 50
percent of the maximum equilibrium loading of 0.5 M citrate solution at the
852
-------
absorption temperature of 109° F. Test results also show an S02 removal efficiency
exceeding 95 percent and offgas containing 6l ppm S02 when treating lower strength
gases at higher gas flow rate, even when the gas temperature was increased to 58° C
(lj6° F). The solution loading of 4.1 grams S0£ per liter under these conditions
represents over 80 percent of the maximum equilibrium loading of 0.5 M citrate
solution at the 136° F absorption temperature. At the design capacity of the plant,
a gas flow rate of 1,000 scfm and a solution flow rate of 10 gallons per minute,
the total pressure drop through the absorption tower was 6 inches of water.
In the early stages of Bureau research to develop the buffered S02-H2S process for
removing S02 from stack gases, laboratory tests indicated that other carboxylate
solutions could be substituted for citrate solution as the buffering agent. Citric
acid was initially chosen for development of the process because of its chemical
stability, low vapor pressure, and good pH buffering capacity, and the purity and
physical character of the precipitated sulfur. Other carboxylates tested appeared
to have about the same properties as citric acid and cost less in some cases, but
in the closed system projected, with slight solution or decomposition loss, the
carboxylate price seemed relatively unimportant. However, depending on reagent
losses and relative price, the use of less costly carboxylates might become
worthwhile.
One of the more promising of the other carboxylate systems investigated in the
laboratory was a solution of sodium glycolate, glycolic acid, and sodium thiosul-
fate. Citric acid is nontoxic and is, in fact, used in many food products, whereas
the biological effects of glycolic acid are not well understood. However, glycolic
acid is one-third the molecular weight and less than half the cost of the citric
acid. Because of these possible advantages it was also tried in the Bunker Hill
pilot plant. Table 2 summarizes results obtained under steady-state continuous
operation at gas flow rate of 1,000 scfm and S02 concentrations of 0.2 to 0.5
percent when using a sodium glycolate scrubbing solution. The absorption tempera-
ture averaged about 38° C (100° F). Glycolate solution flow rate in the absorber
was 7 and 10 gallons per minute, glycolate concentration was about 0.75 M, the
sodium-to-glycolic acid molar ratio in the glycolate solution was 0.9:1, and the
pH of the glycolate feed solution to the absorption tower ranged from 4.2 to 4.3-
TABLE 2. - Results of Bunker Hill pilot plant operation
(glycolate), October-November 1975
Feed gas
concen-
tration,
pet S02
0.21
.4i
.42
Flow,
gal/min
10
10
7
Glycolate
Loading,
g/1 S02
4.6
8.5
12.3
solution
Regenerated,
pH
4.3
4.2
4.3
Offgas
concen-
tration,
ppm S02
19
155
187
S02 removal
efficiency,
pet
99.1
96.2
95.6
The test results show that the glycolate solution may be used to effectively remove
S02 from flue gases. When treating 1,000 scfm of 0.4l percent gas at 100° F and
using 10 gallons per minute of 0.75 M glycolate solution, an S02 removal efficiency
of 96.2 percent and an offgas containing 155 ppm S02 were obtained. The solution
loading of 8.5 grams S02 per liter represents about 50 percent of the maximum
equilibrium loading of 0.75 M glycolate solution at the absorption temperature of
100° F. Results also show that the S02 removal efficiency exceeded 95 percent and
the offgas contained less than 200 ppm S02 even when the solution loading was
855
-------
increased to 12.3 grams per liter,, which represents over 66 percent of the maximum
equilibrium loading of an 0.75 M glycolate solution at the absorption temperature
of 100° F. Excellent results were obtained when treating a gas containing 0.21
percent S02, which would be similar in S02 concentration to a coal-fired powerplant
flue gas. In this test, an S02 removal efficiency of 99.1 percent and an offgas
containing 19 ppm S02 were obtained. The solution loading of 4.6 grams S02 per
liter represents about JO percent of the maximum equilibrium loading of 0.75 M
glycolate solution at the absorption temperature of 100° F.
The S02 reduction reaction-sulfur precipitation reactor circuit effectively regen-
erated the citrate or glycolate solutions using the dilute (56 to 79 percent) H2S
gas product of the H2S generator. Test results showed that to obtain 96 percent
utilization of H2S in a dilute gas required an average solution retention time of
about kO minutes and an agitator impellor peripheral speed of 9°° feet per minute.
The flotation circuit operated satisfactorily at the design capacity of 1 ton of
sulfur per day. A powderlike sulfur product of about ho percent solids was
produced by adding 28 pounds of kerosine (4 gallons) per long ton of precipitated
sulfur. At this kerosine addition rate, very little kerosine was recovered. About
75 percent of the kerosine loss was attributed to volatilization from the hot
sulfur slurry in the kerosine conditioner and sulfur skimmer. The other 25 percent
of the kerosine was contained in the sulfur product and sulfur recycled to the H2S
generator. To eliminate the necessity for kerosine addition, some pilot plant
operation was conducted using air flotation. Air flotation satisfactorily separated
the sulfur from the bulk of the scrubbing solution; however, air flotation produced
a wet sulfur product of about 8 percent solids.
The sulfur melting step functioned satisfactorily at the design capacity, 250 to
300 pounds of sulfur per hour, when melting kerosine-floated sulfur. When melting
air-floated sulfur, the sulfur melter-settler did not provide enough residence or
settling time to make a clean separation between the liquid sulfur and the scrub-
bing solution at design capacity.
The sulfur produced by the Bunker Hill pilot plant has been bright yellow and of
better than 99-5 percent purity. Carbon content, when using kerosine flotation,
ranged from 0.2 to 0.3 percent, and when using air flotation, was less than O.oi
percent. Ash content was about 0.03 percent.
During operation of the Bunker Hill pilot plant, the SO^ buildup rate, which occurs
from the absorption of S03, oxidation of S02, and oxidation of S20s, was determined
to be between 1.3 and 1.5 percent of the S02 in the feed gas. This is quite low
considering that the feed_gas, which is predominately air, contained about 20 per-
cent 02. Although the SO^ buildup rate was not affected by bypassing the electro-
static mist precipitator, the higher rate, 1.5 percent, did occur when the gas
cooling and cleaning packed wet scrubber was bypassed. Laboratory studies show
that about one-third of the SO^ buildup occurs by oxidation of S20s in the sulfur
melter. The highest sulfate concentration in the plant was about 75 grams per
liter. However, this figure is not representative of the greater sulfate buildup
expected because of solution losses during pilot plant operation, which required
a makeup of fresh scrubbing solution.
The most important chemical makeup requirements for the S02 absorption-regeneration
section of the pilot plant include citric acid, sodium carbonate, and kerosine.
The average of many citrate loss tests showed that about 19 pounds of citric acid
was lost per net long ton of sulfur produced. A sodium carbonate addition of
854
-------
about 100 pounds per net long ton of sulfur product was required to neutralize
sulfuric acid produced in the process. Although the kerosine consumption in the
pilot plant was about 85 pounds per net long ton of sulfur produced, it may be
possible to recover the 75 percent that is volatilized in the flotation circuit.
EPS Generation
The H2S generator indicated as a single block in figure 1 is depicted in some
detail in the flow diagram, figure 3. Solid lines in this diagram indicate flows
for a two-stage reaction considered to be a proven process. This two-stage
operation is discussed first. The dashed lines indicate provision to bypass
reactor 2 for later experimentation with a single-stage reaction. The design
capacity of the plant was O.h to 1.25 tons of H2S per day.
Molten sulfur produced in the pilot plant was transferred to a sulfur feed tank
from which it was pumped through a filter to remove carbon and ash impurities and
then to a gas-fired superheater. The sulfur was vaporized and superheated to
about 730° C (1,350° F). Natural gas containing about 92 percent methane and 6
percent heavier hydrocarbons heated to 650° C (1,200° F) in another gas-fired
superheater joins with the hot sulfur vapor ahead of reactor 1. The first stage
of the sulfur-methane-steam reaction to produce H2S and CS2 takes place in the
presence of a catalyst in this reactor according to the following equation:
CH4 + ^S _ CS2 + 2H2S. (5)
The H2S-CS2 reactor product gas was first air-cooled to about 315° C (600° F) and
then further cooled in a steam-cooled heat exchanger to about 150° C (300° F).
Excess sulfur was condensed in the reactor product cooler and was removed from the
gas mixture in the first sulfur knockout vessel.
Steam was superheated to ^25° C (800° F) in a gas-fired steam superheater and then
blended with the cooled H2S-CS2 gas prior to entering reactor 2. In this reactor,
the CS2 was hydrolyzed to H2S and C02 in a catalyst bed at 370° C (700° F) accord-
ing to the following reaction:
CS2 + 2H20 - C02 + 2H2S. (6)
Another important reaction that occurs during CS2 hydrolysis forms COS according
to the following reaction:
CS2 + H20 - COS + H2S. (7)
In the laboratory it was determined that reaction 7 was depressed by feeding excess
steam and favored by high reaction temperatures.
A steam-cooled heat exchanger cooled the reactor 2 product gas to 150° C (300° F)
to condense any free sulfur. The condensed sulfur was removed in two additional
knockout tanks. Sulfur collected in all three knockout vessels was periodically
drained to the sulfur feed tank. Product gas from the second reactor containing
H2S, C02, and some water vapor was cooled to about 65° C (1^-9° F) in a water-
cooled heat exchanger and then flowed to the S02 absorption and sulfur recovery
section of the citrate pilot plant.
The major problem experienced during the pretesting of the H2S generator was the
overheating and damage of the sulfur vaporizer piping when attempting to provide
the heat required for efficient reaction in reactor 1. This problem was solved
by shortening the process piping between the superheaters and reactor 1, adding
insulation to process piping and reactor 1, and feeding an excess of sulfur.
855
-------
Molten sulfur from
citrate plant
Sulfur feed tank
Sulfur superheater
Reactor No. I
Air cooler
Steam cooler
Molten
sulfur
Natural gas
Natural gas superheater
Sulfur knockout
tank
Sulfur knockout
tank
Sulfur knockout
tank
Water
Steam superheater
Reactor No. 2
Steam cooler
Water cooler
Flow during single-stage reaction
GC>2 product gas
to citrate plant sulfur
precipitation reactors
FIGURE 3.-Bunker Hill citrate pilot plant-r^S generation section.
D-3I8I-SL
856
-------
After pretesting, the operation of the H2S generator was integrated with the
operation of the S02 absorption-regeneration section. Two serious operating
problems developed during continuous operation. The first, and probably the most
chronic, problem was sulfur plugging of the cold H2S product lines and valves
because sulfur vapor was not effectively removed from the product gas. The second
problem was carbon plugging of the first reactor and the piping carrying super-
heated natural gas and superheated sulfur. This was due to natural gas cracking
to form elemental carbon. The sulfur-plugged valves caused the automatic H2S
product flow control to malfunction. Because of this, the H2S production had to
be controlled by manually changing the natural gas and steam feed rates.
After achieving reliable operation with a two-stage reaction, the bypasses,
indicated in figure 3 removing reactor 2 from the system, allowed the entire
reaction to take place in reactor 1 according to equation k. This operation
proceeded without problems; the only noticeable difference was that the product
H2S contained more COS and CH4.
Based on the initial Bureau of Mines laboratory work on H2S generation, two other
modes of operation were tried. During one-stage operation, the steam superheater
was eliminated and steam was superheated together with the natural gas. This
operation also proceeded without problems. Next, both the steam and natural gas
superheaters were eliminated, and all reactant feed materials were superheated in
the sulfur superheater. Again the plant worked satisfactorily; the only noticeable
difference was that the temperature of the superheater could be reduced about
100° F without affecting the H2S product quality. The advantages of such single-
stage operation would be the elimination of capital equipment.
Between April and November 1975; the H2S generator operated for more than 2,^00
hours and provided the H2S required for S02 reduction in the citrate process.
During the latter 1,600 hours of this time, the single-stage reactor was used.
Table 3 summarizes results of the H2S generator operation under reasonably steady-
state continuous operation. During this time sulfur was fed to the H2S generator
while varying the natural gas and steam flows to change the H2S production rate.
Throughout most of the H2S generator operation, the sulfur flow was about 160
pounds per hour. This flow provided between 150 and 300 percent excess sulfur,
which was recovered and recycled. The large amount of excess sulfur was used to
prevent overheating of the sulfur vaporizer coil and to transfer heat to the first
reactor. During two-stage operation, a 25-percent excess of steam was used to
prevent COS formation. During one-stage operation, where the CS2 hydrolysis
occurs at a higher temperature, a 50-percent excess of steam was necessary to
achieve low COS in the product gas. The first reactor inlet and outlet tempera-
tures both averaged about 566° C (1,050° F). While operating with two stages,
the second reactor inlet and outlet temperatures averaged 221° C (^30° F) and
3^6° C (655° F), respectively. In addition to the principal gas analyses listed
in table 3; the H2S product gas contained about 1 percent W2 and traces of CS2,
CO, and 02.
857
-------
TABLE 3. - H2S generation results of The Bunker Hill
pilot plant operation, April-November 197?
Reactor
stages
2
2
2
1
1
1
Number of
super-
heaters
3
3
3
3
2
]
Natural
gas flow,
Ib/hr
12.2
8.1
6.0
10
9
7
H2S
production,
ton/day
1.15
-75
.55
• 95
.85
.65
H2S product analysis, vol-pct
H2S
79
79
79
78
77
78
C02
19
19
19
18
17
18
COS
0.7
-5
= 3
1.5
2A
1.6
CH4
0.1).
.2
.2
2.0
2.1
1.0
The test results show that the two-stage operation of the pilot plant makes the
most efficient use of natural gas and produces the least amount of the impurity COS.
The decrease in CH^ utilization during one-stage operation, as indicated by the
higher CH4 in the product gas, may be caused by the increased space velocity through
the first reactor. The increased COS content of the product gas formed during one-
stage operation may result because the formation of COS is favored at higher
temperatures. The COS in the H2S product gas would be vented from the sulfur
precipitation reactors, incinerated to S02 and C02, and returned to S02 absorption.
By all methods of operation, the H2S generator produced a gas which analyzed
between 77 and 79 percent H2S and was suitable for the S02 reduction reaction.
The quantities of sulfur, CH4, and steam required for the H2S generator for each
short ton of H2S produced were
1. Sulfur, 1,900 pounds
2. CH4, 5,600 standard cubic feet
3. Steam, 1,000 pounds
These quantities assume that the H2S product gas contains 1 percent COS and 0.5
percent CH4 and that a 50-percent steam excess is necessary for the most efficient
H2S generation reaction.
To test S02 reduction with lower grade H2S, such as would be made by using a CO-H2
mixture from coal gasification in the manufacture of H2S, the generator operation
was modified to yield such a low-grade gas. By operating at a reactor inlet
temperature of only 482° C, unreacted methane was increased, thus diluting the
H2S in the final product gas. This gas in the pilot plant proved suitable for
S02 reduction and contained, in volume-percent on a dry basis, 56 H2S, 15 C02,
24 CH4, 1 COS, 0.5 CS2, 2 N2, and trace CO and 02.
In general, the H2S generator worked reasonably well after initial shakedown
problems were eliminated. With additional refinements it would be able to produce
suitable H2S at a rate to match the normal requirements of the citrate process.
However, corrosion and plugging of lines may be a continuing maintenance problem,
even though probably reduced in a larger scale operation.
PROCESS DEVELOPMENT
Research continues at the Salt Lake City Metallurgy Research Center to develop and
improve the buffered S02-H2S (citrate) process for removing S02 from industrial
858
-------
stack gases. Two areas of this process development research that look encouraging
are (l) H2S generation using high-sulfur coal or coke in place of natural gas as a
source of reductant, and (2) steam stripping of loaded solution as an alternative
to sulfur precipitation by H2S to produce strong S02 gas for feed to a sulfuric
acid plant.
HgS Generation Using Petroleum Coke
Research was initiated to investigate generation of H2S for the buffered S02-H2S
process by reacting a 5-percent-sulfur petroleum coke with steam and citrate-
produced sulfur. The manufacture of H2S from this raw material, typical of high-
sulfur coke produced at oil refineries, would eliminate the dependence of the
citrate process on natural gas. Utilization of this high-sulfur petroleum coke,
which cannot be burned in many areas because of air-pollution standards, would
help in solving some of the stockpiling problems presently developing at many oil
refineries.
Initial research was directed toward a two-stage procedure whereby the coke was
first reacted with steam in a vertical tube furnace at 800° to 9°°° c "to produce
a water gas containing predominately hydrogen (H2) and CO according to the following
reaction:
C + H20 _ H2 + CO. (8)
The water gas was then reacted with vaporized sulfur and more steam in a horizontal
tube reactor where H2S and C02 were produced by the following reaction:
H2 + CO + 2S° + H20 - C02 + 2H2S. (9)
In the best test to date, reaction of the sulfur-bearing coke and steam at 860° C
(1,580° F) resulted in a gas containing, in volume-percent on a dry basis, 52 H2,
k2. CO, k C02, 2 H2S, and a trace of CH4. Reaction of this gas with sulfur vapor
and more steam over an alumina catalyst at 550° C (1,020° F) produced a gas con-
taining, in volume-percent on a dry basis, 68 H2S, 31 C02, and the remainder CO,
COS, and CH4. Such a gas would be suitable for precipitation of elemental sulfur
and regeneration of absorbent liquor in the citrate process.
Steam Stripping
Investigations have been continued to determine the feasibility of utilizing steam
stripping in the buffered S02-H2S process as an alternative to H2S stripping for
removing S02 and regenerating the absorbent liquor. Strong S02 gas would be
recovered as an end product in place of elemental sulfur. Such a procedure might
be applicable where markets exist for S02 or H2S04, or where acid plant tail gas
could be reduced to acceptable levels by the citrate process and the strong S02
gas produced could be recycled to the acid plant feed gas stream.
Continuous tests have been initiated in a bench-scale, integrated absorption-
stripping apparatus to treat gases containing 0.25 and 0.5 percent S02. This
covers the S02 concentration range normally found in acid plant tail gas.
Absorption of S02 is carried out at ^5° to 50° C using a 0.5 M citrate solution
at pH 4. The steam stripping column is operated at 9^° C. Preliminary tests
indicated high steam consumption, but better results are being obtained with heat
losses in the system reduced by added insulation, near-isothermal conditions
maintained' in absorption and stripping columns, and improved utilization of the
steam.
859
-------
In the best test to date, 95 percent of the S02 was removed from feed gas, using
about 12 pounds of steam per pound of S02 recovered. Sodium thiosulfate is added
to the citrate solution to inhibit oxidation of S02 to SOj. To date, S02 oxida-
tion in the system has been about 1.3 percent, which is comparable with the
oxidation rate obtained when H2S is used for regeneration of citrate solutions.
The results of Bureau investigations indicate that steam stripping might be an
alternative to H2S regeneration in the citrate process when extremely high
absorption is not needed and inexpensive steam is available.
DEMONSTRATION PLANTS
Plans are underway for construction of a large-scale plant to demonstrate the
applicability of the citrate process for S02 emission control at powerplants
burning high-sulfur coal. The demonstration plant would operate on a 50- to 100-MW
powerplant, and would be constructed and operated under a cooperative arrangement
and cost-sharing basis between the Bureau of Mines, the Environmental Protection
Agency, and interested industrial firms.
Work on the citrate demonstration plant will be divided into four phases. Phase I,
consisting of process design and definitive cost estimates, should be completed by
the end of 1976. Phase II, which includes detailed engineering design, construction,
and mechanical acceptance of the plants, should be completed by the end of 1979-
Phase III, consisting of startup and performance acceptance testing, would take
place at the conclusion of Phase II and would be followed by Phase IV, which con-
sists of comprehensive emission testing programs to be conducted at the demonstra-
tion plant by independent contractors for 1 year.
COST ESTIMATE FOR CITRATE PLANT AT COAL-BURNING POWERPLANT
Based on results of the Bunker Hill pilot plant operation, the Process Evaluation
Group at the Salt Lake City Metallurgy Research Center estimated capital and
operating costs for the treatment of flue gas from a 1,000-MW coal -burning power-
plant. This would include fly ash removal in an electrostatic precipitator
followed by removal of 95 percent of the contained sulfur. The powerplant was
assumed to operate 7,000 hours (292 days per year) and to burn 8,^00 tons per
day of coal containing 3^ percent S; the gas flow would be 1,730,000 scfm (60° F)
with an S02 content of o'.2k percent. Under these conditions, the yield of sulfur
would be about 2.1k long tons per day or 62, ^>kk tons per year.
Annual operating costs for the plant are itemized in table k. A summary of
operating labor requirements, fixed capital costs, annual operating costs, and unit
production costs for each of the unit operations and for the entire plant is pre-
sented in table 5» The fixed capital costs were estimated by standard chemical
engineering cost estimating procedures for a "study estimate" and are for 1975 or
a Marshall and Swift (M&S ) index of
860
-------
TABLE 4. - Annual operating cost of S02 removal from
1,000-MW powerplant off gas"
Operating cost
Total cost
Direct cost:
Raw materials:
Citric acid
Soda ash
Kerosine
Lime
Utilities:
Electricity
Natural gas
Process water
Cooling water
Steam
Direct labor:
Labor
Supervision
Plant maintenance:
Labor
Supervision
Materials
Payroll overhead
Operating supplies
Total direct cost
Indirect cost (administration
and overhead)
Total capital charges
Total annual operating cost
$593,000
206,000
22,600
399,500
2,3^6,900
52,600
221,300
237,200
35^00
1,127,000
225,400
1,305,200
.,308,400
5,752,200
272,600
2,657,600
406,300
531,500
10,928,600
1,384,700
14,461,700
26,775,000
861
-------
TABLE 5- - Summary of costs for S02 removal from a
1,000-MW powerplant offgas
Electrostatic
precipitator
Gas cooling
S0a absorption
Sulfur precip-
itation
Sulfur recovery
H2S generation
Na2S04 removal
Facilities
Utilities
Number
of
operators
2.0
2.0
U.2
2.1
2.1
^.3
2.3
_
-
Capital cost
$21, 56^,600
11,5^6,700
1^,159,900
1,013,700
3,2^9,500
12,685,700
598,200
6,464,100
7,756,600
Annual
operating
cost
$4,208,500
4,813,600
6,288,^00
2,275,300
2,220,000
6,690,200
279,000
_
-
Unit
production
cost per
long ton
of sulfur
$68
77
101
36
36
107
5
_
-
Total fixed
capital
Working capital
Total
19.0
79,039,000
5,722,000
84,761,000
26
,775,000
430
The operating cost of the citrate plant includes reheating the final tail gas from
50° to 77° C (122° to 170° F) with a natural gas burner at the base of the power-
plant stack, and the neutralization of liquid effluents with lime. Operating losses
are assumed at 19 pounds of citric acid and 12 gallons of kerosine per long ton of
product sulfur. Costs for Na2S04 removal were estimated on the basis of 1.5 percent
of the removed sulfur being converted to sulfate.
Direct costs include materials and utilities at the following unit costs: Citric
acid at $1,000 per ton, soda ash at $60 per ton, kerosine at $0.65 per gallon, lime
at $35 per ton, electric power at $0.01 per k¥-hr; natural gas at $1.00 per
thousand cubic feet, water at $0.17 per thousand gallons for process use and $0.03
per thousand gallons for cooling, and steam at $1.00 per thousand pounds. These
costs also include direct labor at $6.00 per hour, plus supervision at 15 percent
of direct labor; plant maintenance consisting of maintenance labor at 1.5 percent
of fixed capital costs, plus supervision at 20 percent of maintenance labor, and
maintenance material at 116 percent of maintenance labor; payroll overhead at 25
percent of the cost of labor and supervision for operations and maintenance; and
operating supplies at 20 percent of plant maintenance.
Indirect costs are 40 percent of direct labor, plant maintenance, and operating
supplies.
Capital charges, which include amortization, taxes, and insurance, are 17 percent
of the total capital, which is the average annual fixed capital charge for
privately financed steam-electric plants.
862
-------
As shown in table 5, the capital investment for a citrate plant including an
electrostatic precipitator at a 1,000-MW coal-burning powerplant is estimated at
$8^.8 million. Assuming no credit for sulfur,, operating cost for removing 95
percent of the S02 from the stack gas is about $26.8 million. This is equivalent
to $10.93 per short ton of coal or 3-8 mills per kW-hr.
863
-------
REFERENCES
1. Chalmers, F. S. Citrate Process Ideal for Glaus Tailgas Cleanup. Hydrocarbon
Processing, April 197^, pp. 75-77-
2. Chalmers, F. S., L. Korosy, and A. Saleem. The Citrate Process to Convert S02
to Elemental Sulfur. Pres. at Industrial Fuel Conf., Purdue University, West
Lafayette, Ind., Oct. 3, 1973; 6 pp.; available upon request from Arthur G.
McKee & Co., Cleveland, Ohio.
3. George, D. R., L. Crocker, and J. B. Rosenbaum. The Recovery of Elemental
Sulfur From Base Metal Smelters. Min. Eng., v- 22, No. 1, January 1970,
PP. 75-77.
4. Korosy, L., H. L. Gewanter, F. S. Chaljners, and S. Vasan. Chemistry of S02
Absorption and Conversion to Sulfur by the Citrate Process. Pres. at l6?th
ACS Meeting, Los Angeles, Calif., Apr. 5, 197^, 32 pp.; available upon request
from Pfizer, Inc., New York.
5. McKinney, W. A., W. I. Nissen, D. A. Elkins, and J. B. Rosenbaum. Pilot Plant
Testing of the Citrate Process for S02 Emission Control. EPA 650/2-7^-126-b,
Proc., v. II, November 197^, pp. 10^9-106?.
6. McKinney, W. A., ¥. I. Nissen, L. Crocker, and D. A. Martin. Status of the
Citrate Process for S02 Emission Control. Pres. at the 1975 Lignite Symp.,
Grand Forks, N. Dak., May 1^-15, 1975; available upon request from the Salt
Lake City Metallurgy Research Center, Bureau of Mines, Salt Lake City, Utah.
7. McKinney, W. A., W. I. Nissen, and J. B. Rosenbaum. Design and Testing of a
Pilot Plant for S02 Removal From Smelter Gas. Pres. at AIME Annual Meeting,
Dallas, Tex., Feb. 23-28, 197^, AIME Preprint A-7^-85, 12 pp.
8. Rosenbaum, J. B., D. R. George, and L. Crocker. The Citrate Process for
Removing S02 and Recovering Sulfur From Waste Gases. Pres. at AIME
Environmental Quality Conf., Washington, D. C., June 7-9, 1971, 26 pp.;
available upon request from Salt Lake City Metallurgy Research Center, Bureau
of Mines, Salt Lake City, Utah.
9. Rosenbaum, J. B., W. A. McKinney, H. R. Beard, L. Crocker, and W. I. Nissen.
Sulfur Dioxide Emission Control by Hydrogen Sulfide Reaction in Aqueous
Solution--The Citrate System. BuMines RI 777^, 1973, 31 pp.
10. Vasan, S. The Citrex Process for S02 Removal. Chem. Eng. Prog., v. 71,
No. 5, May 1975, pp. 61-65.
864
-------
APCI/IFP RENERATIVE FGD AMMONIA SCRUBBING PROCESS
C. E. Ennis, Vice President, Corporate Planning
Catalytic, Inc.
Philadelphia, Pennsylvania
ABSTRACT
A brief history of the utilization of ammonia for the absorption
of SO is the preliminary to the introduction of specific technology
developed by Air Products and Chemicals, and Institut Francais du
Petrole. Their combined work has produced a fully regenerable second
generation flue gas desulfurization process. Process chemistry and
pilot plant activities are fully explored.
865
-------
APCI/IFP REGENERATIVE FGD AMMONIA SCRUBBING PROCESS
The very rapid rate of chemical absorption of sulfur dioxide by
ammonia has been known and to some extent applied for many years. Broad adoption
of ammonia scrubbing has been inhibited by the production of a sub-micron particu-
late, which results in a characteristic, but environmentally unacceptable "blue
plume." Key advances have been oriented towards fumeless ammonia scrubbing tech-
nology developed by Catalytic's parent organization, Air Products and Chemicals,
Inc.C1)
A combination of this technology with an existing liquid Glaus process
developed by Institut Francais du Petrole (IFF), the French research and process
development organization, offers a true second generation stack gas scrubbing process.
The ammonia absorbent is completely regenerated and recycled and the sulfur is re-
covered in its elemental form.
Initial ammonia S02 absorption dates back to 1883 when a British patent
was issued to Ramsey(2). Further, in the period of 1935 to 1952, Johnstone and co-
workers developed basic data at the University of Illinois. Five major papers
were published as a result, which dealt with absorption of SO? by ammoniacal
solutions and desorption from the absorber eff luent. (•*, +>5,b, /)
Abroad, considerable data were collected by Chertkov and his co-workers
in Russia during the 1950s and are covered by over 40 papers on sulfur oxide re-
covery and ammonia absorption.
One of the oldest commercial applications of S02 ammoniacal absorption
began in about 1936 by the Consolidated Mining and Smelting Company(8) now Cominco
Ltd., located in Trail, British Columbia. The unit removes SOX from smelter waste
gases. The system products are ammonium sulfate and S02. The unit is still
operating today.
In Japan, the Showa Denko Company has operated a 25 MW equivalent ammonia
absorption test unit on stack gases from an oil fired system yielding an ammonium
sulfate byproduct.(°)
Processes similar to the unit still in operation at Cominco are in full
scale operation on sulphuric acid plant tail gases in Czechoslovakia and Romania.
(10,11) These processes produce ammonium sulfate, which may have a limited market.
In this country, a number of sulfite paper processes presently utilize
ammonia scrubbing on a 24-hour per day, seven day week basis without problems.
Systems operate a scheduled 362 days per year and successfully reduce S02 concen-
trations in the range of 8,000 to 20,000 parts per million down to about 150 ppm
without operating or reliability problems.
In order to adapt the acid plant tail gas ammonia scrubbing method to the
power plant stack, TVA carried out pilot plant activities in 1953-54. (12) It is
apparent that demand for the ammonium sulfate byproduct severely limits use of this
system in the United States.
In more recent TVA work, ammonia absorption processes yielding a fertilizer
byproduct known as far back as the 1920s have been investigated at the Colbert Power
Station facilities.
Program results(") indicated that the most severe problem of the ammonia
absorption process was the formation of a dense, persistent plume. Removal of
the plume by electrostatic precipitation was reported by TVA to be about 50% at
superficial velocities of 15 feet per second.
866
-------
Personnel from Southern Research Institute of Birmingham, Alabama,
determined the fume mean particle size to be 0.25 micron with 10' particles/cm-^
in the range of .005 to .5 micron.
Air Products, as a result of extensive studies, postulated a mechanism
for solid particulate plume formation. Laboratory studies verified that the pro-
posed mechanism was indeed the controlling system of plume formation. With the
mechanism confirmed, Air Products' scientists were able to establish a set of
criteria which fully define an envelope of operating conditions which largely or
wholly prevent or eliminate the formation of -an ammonia salt plume. These criteria
are pa tented (-1-' and are the result of a rigorous theoretical and experimental
program carried out by the Research and Development Department at Air Products.
Utilizing these established operating criteria, Air Products personnel
were literally able to turn the plume on and off in a laboratory system by ad-
justing the operating parameters. During the period 16 April to 4 May, 1973, a
test program was conducted at TVA's Colbert Station. These tests, sponsored
jointly by TVA, EPA and Air Products/Catalytic, were conducted as a factorial
series to explore the pluming characteristics of the system over a range of
operating conditions in an insulated scrubbing tower. Run number I 14A, operating
within the Air Products specified envelope of conditions, resulted in zero opacity
as observed by two EPA certified plume readers.
TVA concluded in its report^ ' that "Plume formation while producing a
liquor with high salt concentration (C^/'IZ) can be controlled by proper operation
of the scrubber."
New technology providing for removal of S02 utilizing ammoniacal solu-
tions without associated plume formation is the basis for our entry into this field.
Our licensing agreement with Institut Francais du Petrole allows us to
combine our ammonia scrubbing technology with a process to convert the spent
ammoniacal brine to elemental sulfur(-^ > -^>l^16,17)with essentially full recovery
of ammonia values for recycle.
At present, Catalytic is operating a 2 MW ammonia scrubbing system at
Calvert City, Kentucky. It is fed a gas slip-stream from two 20 MW equivalent coal-
fired industrial boilers firing 3.27, sulfur coal. Startup of the unit was early
February.
Concurrently, IFP is running their recovery process on a 30 MW scale at
the Electricite du France at Champagne. The unit produces a bright yellow sulfur.
I. THE CHEMISTRY
Air Products and Chemicals has developed a special cycle synthesis com-
puter program which enables the user to establish a material and energy balance
consistent with a given set of conditions for the absorption process. The program
also checks these operating conditions against the plume avoidance criteria.
In a typical 3-1/27, sulfur coal case, four stages of S02 absorption are
employed to effect 907, removal. Each stage is fed fresh ammonia makeup while brine
leaving the stage is split between the downcomer and a stage recirculation loop.
Control of the process is effected by automatic pH controlled addition of fresh
ammonia to each stage.
867
-------
The principal absorption reactions are as shown here:
S02 + (NH4)2S03 + H20 — — -> 2NH4HS03 (1)
S03 + (NH4)2S03 ™— ~i» (NH4)2S04 + S02 (2)
02 + 2(NH4)2S03 — fr 2(NH4)2S04 (3)
The formation of ammonium sulfate is inevitable with absorbent in
contact with S03 and oxygen. Sulfates are readily converted in the recovery
process .
Spent brine from the absorber is first processed in a forced circu-
lation evaporator where about seventy percent of the stream is reduced to S02,
H20, and ammonia as shown by these equations:
(NH4)2S03 — > S02 + H20 + 2NH3 (4)
NH4HS03 — — -4» S02 + H20 + NH3 (5)
The balance of the stream which contains the sulfate load plus remaining
sulfite is converted in the sulfate reducer along with a slipstream of elemental
sulfur to ammonia, S02, and water as follows:
(NH4)2S03 - - »$. S02 + H20 + 2NH3
(NH4)2S04 _ fc,NH4HS04 + NH3 (6)
2NH4HS04 + S — — *. 3S02 + H2<3 + 2NH3 (7)
A small amount of SOo exits with the converted stream which passes through
a catalytic bed where it is converted directly to S02. as shown:
S03 + H2 catalyst S02 + H20 (8)
S03 + CO catalyst S02 + C02 (9)
All vapor streams are combined and next pass to the hydrogen sulfide
generator. Here two thirds of the S02 is converted to H2S by the following catalyzed
reactions:
2/3 S02 + 2 CO + 2/3 H20 - fc. 2/3 H2S + 2 C02 (10)
2/3 S02 + 2H2 - *• 2/3 H2S + 4/3 H20 (11)
The reaction proceeds with either CO or H2 as a reducing agent.
The stream containing H2S and S02 enters the liquid Glaus reactor where
elemental sulfur is produced as a result of the following reaction:
2/3 H2S + 1/3 S02 - — > s|> + 2/3 H20 (12)
Finally, the gas stream leaving the liquid Glaus is treated to recover
the ammonia values and concentrate them for return to the absorber. The non-
condensable gas stream is vented to the boiler combustion inlet air or incinerated.
868
-------
II. THE PROCESS
Now that we have reviewed some of the process background and chemistry,
let us take a look at a typical process schematic which is depicted in Figure 1.
Battery limits begin at the exit flange of the electrostatic precipita-
tor. The process begins with a blower of sufficient power to drive the gas through
gas conditioning and SC>2 absorption steps and finally out the stack.
Although we normally think in terms of a forced draft fan in the dry
position for a typical, application, flue gas composition, especially particulate
loading, can affect this approach.
Gas enters a low pressure drop venturi at about 300°F. When the gas is
brought in contact with the recirculating water, it is brought to its adiabatic
saturation temperature of approximately 125°F. Contact time is a fraction of a
second. During the turbulent liquid to gas contact, over 9070 of the chlorides and
about half of the 503 are absorbed in the recirculating liquid in which the pH is
maintained at about two. The composition of the recirculating venturi liquid is
the basis for control of the purge stream from the system. The purge is utilized
to maintain chlorides below 5,000 ppm and/or total dissolved solids below 570. For
a typical 500 megawatt case, it is estimated that purge stream from the humidifier
would be approximately 5 gpm. It would be discharged directly into the utility
plant's fly ash pond where its acidic components would help to neutralize the alka-
line fly ash discharge.
From the flue gas composition and its thermodynamic characteristics,
operating conditions can be set that result in fumeless operation of the absorber.
It is then simply a matter of maintaining these conditions for each stage such that
expected system perturbations do not carry the absorption process into the fuming
regime. An envelope of operating conditions is maintained to assure that expected
variations in temperature and SC^ concentrations do not cause a plume to form.
Presently, a valve tray absorber configuration is utilized in the design.
We consider it a very desirable initial approach due to the extensive technical
data that exist for such designs. A typical four-tray system is expected to induce
about a 12" water pressure drop. Catalytic anticipates extending its technology
into less energy demanding absorber configurations.
The Catalytic/Air Products approach to plume avoidance comprises stage-
wise control of temperature and composition. In its capability as engineer/construc-
tor, Catalytic has designed and erected many chemical plants whose process con-
trol requirements were far more sensitive and demanded greater sophistication
than the S02 absorption step. We view control of the ammonia system as a routine
process operation. Liquid phase pH and density have been related directly to com-
position, providing the determinant relationship for liquid phase control.
Ultimately, spent liquor exits the absorber from the bottom stage where
it is held in storage or sent directly to the 1FP recovery process. It is most
practical to operate the recovery section at a constant liquid load, and for this
reason, interprocess storage becomes desirable.
The recovery section need not be situated in close proximity to the absorp-
tion section. Use of spent brine and regenerated ammonia storage tankage permits
location of the recovery (IFF) process unit on a distant plot connected only by
spent and fresh liquid absorbent lines.
869
-------
APCI-IFP
REGENERATIVE FGD PROCESS
SCHEMATIC
CLEAN WASTE GAS
MAKEUP WATER
oo
^j
o
FLY ASH TO PIT
HEAT ENERGY
V
H-
m
MOLTEN SULFUR
Ammonia Circulation Loop
-------
Spent ammoniacal brine is fed at a constant rate from the storage tankage
through a filter to a forced circulation evaporator where sulfite decomposition
occurs. The process is driven by low pressure steam derived almost completely from
another step in the process.
Sulfates are combined with a slipstream of molten sulfur and reduced in
a submerged combustor. In this unit, a gas is burned with about 95% of the stoichio
metric air requirement. The exhaust gases are brought in direct contact with the
bath of ammonium sulfate, sulfite, and sulfur. The bath is maintained at approxi
mately 700°F.
The exit gas contains H20, S02, NH3 and trace 803 which is reduced to S02
in a catalytic converter.
There are non-volatile impurities which collect in the sulfate reducer.
These are purged on an intermittent as-required basis. The waste stream purged
from the reducer is small in quantity and can be sold as fertilizer. If inappro-
priate, the sulfate waste could be wet with water and sprayed on the power station's
coal pile for combustion with the coal. The ammonia has some heat value and de-
composes at firebox temperatures. The sulfate will reappear as
The S02 rich gas stream is then united with reducing gas and allowed to
combine to form I^S in a special catalytic reduction reactor. The effluent gas
stream contains an approximate 2/1 mole ratio of t^S to S02- This reaction is
highly exothermic with the exit gas temperature reaching about 900°F. The flow
from the catalytic reducer enters a waste heat boiler and leaves at about 320°F,
ready for the liquid Glaus step. The waste heat is removed as low pressure steam
which is utilized for sulfite evaporation.
Gases at about 320°F pass into the liquid Glaus unit. It is here that the
S02 and l^S combine to form elemental sulfur. The molten product is taken off the
bottom of the unit while the ammonia, water vapor, nitorgen, CC^, and trace amounts
of other combustion gases flow out the top. Recovery of the ammonia values is
accomplished by absorption, and distillation enriches the concentration to usable
levels.
From here, the ammonia is recycled to the fresh absorbent storage tank,
thereby closing the loop. Non-condensables are either catalytically incinerated
or discharged to the combustion air inlet of the boiler.
It is proposed that an air blown gasifier be employed to supply heat for
sulfate reduction and low btu reducing gas for the I^S generation step. The CO
values in reducing gas are used with equal ease as the hydrogen values for the I^S
production step prior to sulfur make in the liquid Glaus. When coal is fired
in the gasifier, the hot producer gas would enter a cleanup system to remove tars,
fly ash and carbon fines by condensation, scrubbing, and electrostatic precipitation.
Ammonia and S02 that pass through to the process are not a problem since
both already exist in large quantity and they have no upsetting effect on the
material balance. The process can utilize both the H2 and CO values of the gas.
It is envisioned that a gas holder would be installed between the gasifier and
the process.
Excluding all contingencies and assuming the coal charge to the gasifier
to be considered a raw material, we envision an approximate 2.870 of the power plant
capacity as an electrical requirement and an additional 1.67= as steam for a total
of 4.47o of the power plant output. This includes steam to provide reheat of 50°F.
This represents the steady state energy requirement for the process.
871
-------
III. APPLICATIONS
Figure 2 shows the 4" diameter glass bench scale scrubber in Air
Products' Allentown, Pennsylvania laboratories which was used to develop
data and to confirm results of the theoretical and experimental studies on
which the process is based.
Figure 3 shows our 2 MW pilot facility which is dedicated to the
optimization of the S02 absorption step. The system has been fabricated on
two transportable skids. One contains the blower, cooler, and absorber, while
on the second we have located the interstage pumps and associated tankage, with
the control house package making up the third. We intend to gain hard data re-
quired for engineering design of commercial systems as well as look into alternate
control systems. In the not too distant future, we hope to be in a position to
offer the unit for demonstration purposes to potential customers.
IFP has a 30 MW demonstration of its recovery process in Champagne,
at Electricite de France. Most of the unit operations are vertically stacked
for a minimum land area requirement. The system utilizes a slipstream from a 200
MW oil-fired boiler. The ammonia scrubber uses European technology and does plume;
however, the exhaust is co-mingled with the boiler effluent.
In conclusion, the ammonia process has a long history of reliable appli-
cations, both in the United States and abroad. At present, the opacity emission
regulations have reduced its potential for use without considerable additional
capital and operating cost to eliminate the plume once formed. Utilizing the
technology discussed here today, we think that the ammonia process is a reliable
and viable method for SOo removal with complete regenerative capability yielding
elemental sulfur. The overall process is reliable in operation, avoids the obvious
scaling problems by virtue of using a clear liquor scrub, does not require constant
replacement of absorbent and catalysts, produces a salable sulfur and does not
generate environmentally objectionable by-products. It can use a reducing gas
obtainable from currently available coal gasifiers, using coal from the plant
supply. Finally, it is attractively cost competitive with all second generation
processes and first generation approaches, particularly in large sizes.
872
-------
APCI 4" Diame
Glass Bench
Absorber
-------
Figure 3.
2 MW Ammonia Absorption
Pilot Plant Located in
Calvert City, Kentucky
874
-------
REFERENCES
(1) Marshall L. Spector, P. L. Thibaut Brian, "Removal of Sulphur Oxides from
Stack Gas," U.S. Patent 3,843,789 (1974).
(2) Ramsey, "Use of the NH3 S02 l^O System as a Cyclic Recovery Method,"
British Patent 1,427 (1883).
(3) Johnstone, H.F., "Recovery of SC>2 from Waste Gas: Equilibrium Partial
Pressures over Solutions of the Ammonia-Sulfur Dioxide-Water System"
Ind. Eng. Chem. 27 (5), 587-93 (May 1935).
(4) Johnstone, H.F., Keyes, D.B., "Recovery of SC>2 from Waste Gases: Distil-
lation of a Three-Component System Ammonia-Sulfur Dioxide-Water," Ind. Eng.
Chem. 27, 659-65 (June, 1935).
(5) Johnstone, H.F., Singh, A.D., "Recovery of S02 from Waste Gases: Design
of Scrubbers for Large Quantities of Gases," Ind. Eng. Chem. 29 (3), 286-97
(March, 1937).
(6) Johnstone, H.F., "Recovery of S02 from Waste Gases: Solvent Concentration
on Capacity and Steam Requirements of Ammonium Sulfite-Bisulfite Solutions,"
Ind. Eng. Chem. 29 (12), 1396-98 (December 1937).
(7) Johnstone, H.F., "Recovery of Sulfur Dioxide from Dilute Gases," Pulp Paper
Mag. Con. 53 (4), 105-12 (March 1952).
(8) Lepsoe, R., Kirkpatrigk, W.S., "SC>2 Recovery at Trail, A General Picture of
the Development and Installation of the Sulfur Dioxide Plant of the Con-
solidated Mining and Smelting Company of Canada, Ltd., at Trail, B.C."
Trans. Can. Inst. Mining Met. 40, (Ind. Con. Mining Met. Bull. No. 304),
399-404(1937).
(9) Pilot-Plant Study of an Ammonia Absorption Ammonia Bisulfate Regeneration
Process, Topical Report Phases 1 and 11, EPA-650/2-74-049-a, (June, 1974).
(10) "S02 Recovery from Sulphuric Acid Plant Off-Gases," Sulfur, No. 80, 36-8
(January-February-1969).
(11) Romania, Ministry of Petroleum and Chemical Industry, "Ammonium Sulfate,"
British Patent 1,097,257 (January 3, 1968) 2 pp.
(12) Hein, L. B., Phillips, A.B., Young, R.D., "Recovery of S02 from Coal Combustion
Stack Gases," Problems and Control of Air Pollution, Frederick S. Mallette,
ed., New York, Reinhold, 1955, pp 155-69.
(13) Sulfur Removal and Recovery from Industrial Processes: Advances in Chemistry
Series 139, pp 100-110, Edit. J. B. Pfeiffer.
(14) Barthel et al, Ind. Pet. Europe (431) 49-53 (1972).
(15) Bonnifay et al, Chem. Eng. Progress, Aug. 1972 68, 51-52.
(16) Chemical Eng. Nov. 13, 1972, 72F.
(17) Petroleum and Petrochemical International, Feb., 197j, Vol. 13, No. 2, p. 47.
875
-------
BF DRY ADSORPTION SYSTEM
PART I
FW-BF Gulf Power Demonstration Unit Interim Results
Joseph J. Strum and William F. Bischoff
Foster Wheeler Energy Corporation
Randall E. Rush
Southern Services, Incorporated
PART II
BF-STEAG Demonstration Unit
Operational Experience and Performance
Dr. Karl Knoblauch
Bergbau-Forschung GMBH
Dr. Klaus Goldschmidt
STEAG Aktengesellschaft
ABSTRACT
This process uses activated char to remove SO and particulate
matter from boiler flue gas. The char is thermally regenerated and
the SO rich gas is converted to elemental sulfur.
PART I of the paper presents interim results of the FW-BF Dry
adsorption System installed as a demonstration unit at the Scholz Steam
Plant of Gulf Power Company in Sneads, Florida. Current status,
operating experiences and plans for the future of the process are
reviewed.
877
-------
PART II of this paper presents the operational experience and
performance of the Bergbau-Forschung SO Removal System installed as
a demonstration unit at the Kellerman Power Station of STEAG in Lunen,
West Germany. The results of extensive operation of the adsorption,
regeneration and sulfur reduction sections are presented and the
flexibility of the unit to cycle successfully in response to peaking
operation is described.
878
-------
PART I
GULP POWER DEMONSTRATION UNIT
Interim Results
I. Introduction
A 20 MW demonstration dry adsorption system was put into operation
at the Scholz Steam Plant of Gulf Power Company near Sneads, Florida,
Figures1&2. This system is one of three advanced flue gas desulfuriza-
tion units of 20 MW size installed at the plant as part of a technology
evaluation program "being conducted by Southern Services, Incorporated,
for the Southern Company. The Southern Company is an electric utility
holding company including Alabama Power Company, Georgia Power Company,
Gulf Power Company, Mississippi Power Company, Southern Electric
Generating Company and Southern Services, Incorporated. The general
design conditions for the unit are shown in Table I.
TABLE I
Summary of General Design Conditions
Unit Rating - Adsorption Section 20 MW Nominal
Unit Rating - Regeneration & RESOX* Sections i^O MW Nominal
Coal
HH7 12,400 BTU/#
°/S 3
% Ash 14
Plue Gas
Qjuantity - #/Hr. 278,940
Temperature - P 326
^Trademark of Poster Wheeler Energy Corporation
879
-------
Figure 1: Adsorption Section - Gulf Power Company
880
-------
Figure 2: Regeneration and Resox* Sections - Gulf Power Company
SSI
-------
The technical basis for this system is derived from a license with
Bergbau-Forschung GmbH, the research group for the German Bituminous
coal industry. The Foster Wheeler - Bergbau-Forschung Dry Adsorption
System was originally developed to enable the utility industry to meet
certain requirements of the Clean Air Act; however, the basic system
can and will be utilized to meet the specific requirements of other
industries as well. The basis of design for both the Gulf Power
Demonstration Unit and the STEAG Demonstration Unit is the very extensive
work done by Bergbau-Forschung over a two year period at their pilot
unit at Welheim, Germany. This plant ran for one continuous period of
6,000 hours and the results of the two year testing period were published
in 1970 at the Second International Clean Air Congress in Washington, D.C.
The initial construction of the Gulf Power 20 MW demonstration dry
adsorption system was completed in May, 1975 followed by a three month
commissioning (shakedown) period. The Adsorption and Regeneration sections
initial shakedown run with passage of flue gas began on August 11, 197?
and was continued for a period of 10 days. During the later part of
October, 1975 sections of the FW-BF system were run for five days. The
purpose for this activity was to conduct an operational run on the RESOX*
portion of the system at full operating temperatures and on process off-
gas from the FW-BF front end subsystem. The delay between shakedown of
the RESOX* section and the remainder of the system was caused by a lag
in the construction of the two plus the need for pre-startup modifica-
tions to the RESOX* start-up heater. This paper covers the operating
experiences during these initial shakedown runs, the subsequent
modifications program and pre-startup testing. In early February, 197&
a formal test program was initiated by Southern Services, Incorporated
and Foster Wheeler Energy Corporation. The results of this test program
will be published in the future.
II. FW-BF Dry Adsorption System
A. Process Chemistry
The process chemistry of the FW-BF Dry Adsorption System for the
Adsorption, Regeneration and RESOX* sections are described below.
Adsorption
Boiler flue gases containing S02, NOX, oxygen, water vapor and part-
iculate matter come into contact with the char pellets, which are the
adsorption media. The S02, oxygen, and water vapor are adsorbed onto the
char. The combination of moving char and flowing flue gas give rise to
the following reaction:
S02 + i 02 = H20 —
882
-------
The char activity acts as a catalyst in this reaction. In addition
SCH and nitrous oxides are similarly adsorbed. While it is easily
understood how SO^ is adsorbed, the principles of the adsorption of
nitrous oxides are still under extensive investigation both in Germany
and here in the United States. Particulate matter is collected on the
surface of the char pellets which act as an impingement filter due to
their size and physical arrangement.
Regeneration of the Adsorbent
When the activated char containing sulfuric acid is heated over
1200 F it undergoes the following reaction:
H2SO^ + i C 1200 p> % C02 + H20 + S02
The nitrous oxides adsorbed are reduced according to the reaction:
2 NOX + C 5£55-f C02 + N2
These reactions result in the production of a concentrated and reduced
volume stream of off-gases consisting of S02> C02, H20, and N2. The S02
concentration of this stream is in a range of 25> to UO percent by weight.
Particulate matter is physically separated from the saturated char prior
to regeneration.
EESOX (Reduction of Sulfur Dioxide by Coal)
The concentrated off-gases produced in the Regeneration Section are
introduced in counterflow to a mass flow of crushed anthracite coal. The
reactions occurring in the reactor can be reduced to the simplified
equation:
S02 + C ^ > C02 + S
The sulfur is produced in the form of a gas which is subsequently
condensed. The nitrogen and carbon dioxide constituents of the Regenerator
off-gas pass through the RESOX* reactor without taking part in the reactions.
The reaction described above is not totally selective. For additional in-
formation refer to Steiner, Juntgen, Knoblauch paper presented at l6yth
National Meeting of American Chemical Society.
B. System Description
The process configuration for the FW-B3T Dry Adsorption System at the
Scholz Steam Plant is shown schematically in Figure 3- The system is
installed on Boiler No. 2 a nominal ij.0 MW (maximum l+J .% MW) pulverized
coal-fired boiler which has been retrofitted with a sectionalized, high
efficiency electrostatic precipitator capable of 99-7^ particulate removal.
883
-------
FW-BF SYSTEM
JL
CRUSHED
COAL
FLUE GAS
FROM DUST
COLLECTOR
00
00
CLEANED FLUE GAS
TO STACK
SULFUR
ADSORPTION
REGENERATION
RESOX
ATRADEMARK OF FOSTER WHEELER CORPORATION
Figure 3
776-015
-------
The FW-BF Dry Adsorption System is sized to handle one-half of the
total flue gas from this "boiler. The 20 MW equivalent in flue gas entering
the system is 85,600 ACFM. The flue gas leaving the boiler air pre-heater
is at a maximum of 350 F. The system is nominally designed to meet Florida
S02 reduction codes of 1.2 # S02/MMBTU heat input which equates to a lk*5%
removal efficiency requirement for 3 percent sulfur, 12,200 BTU/# fuel.
The actual performance of the unit during initial operation far exceeded
this requirement. Further testing during the formal test program will
determine the limits of S02 removal as well as those of NOX and particulate
removal. These test program results will be the subject of future papers.
The operation of the W-BF Dry Adsorption System can be conveniently
broken down into three major sections: Adsorption; Regeneration; and
HESOX*. The equipment utilization and operation of each of these sections
is discussed as follows:
1. Adsorption Section
The adsorber is the key to the system. Adsorption is accomplished
by passing the flue gas horizontally through vertical columns of activated
char in the adsorber.
The Gulf Power Demonstration Unit adsorber consists of 2 vertical
stages of char beds designed in a modular fashion. There are eight
6' x 6' beds in the 1st stage and four I;1 x I;1 beds in the 2nd stage.
All beds are approximately i\Q feet high.
The char in the beds is continuously recycled. A conveyor at the
top of the adsorber feeds regenerated char into a holding tank which
has discharge tubes that gravity feed the regenerated char into the
individual char beds. The char moves downward in mass flow adsorbing
S02 and NOX as it travels. The char flow rate is controlled by a
vibromagnetic feeder at the hopper outlet below each char bed. The
saturated char is collected at the discharge of these feeders and sent
by a combination of natural frequency conveyors and bucket elevators
to the Regeneration Section of the system.
The flue gas entering the adsorber is tempered, if necessary, by
the use of a dilution air fan (vane axial type) which maintains an
inlet flue gas temperature of approximately 281; F. The adsorber dis-
charge fans, ore per stage, (centrifugal type induced draft) restore
the pressure drop suffered by the flue gas during its passage through
the adsorber and associated ductwork. At the Scholz Station, the clean
flue gas is directed to a local test stack for emission measurement
purposes.
885
-------
2. Regeneration Section
The purpose of the Regeneration Section is to provide for the con-
tinuous on-site regeneration of char which has been loaded with S02 in
the form of
Regeneration is achieved by contacting the loaded char with hot sand.
Sand is utilized as an inert heat transfer media and as such does not
take part in the reactions occurring within the regenerator. Its sole
function is to supply heat so that the reactions may take place. The
mixture (1200 F) of hot sand and char flow slowly downward through the
regenerator. Their flow is controlled at the discharge of the re-
generator by a char-sand separator-feeder. The char and sand are
physically separated by means of a vibrating screen deck (profile wire
design). The char is spray cooled to 220 P and returned to the adsorber.
The sand is conveyed to a fluidized bed sand heater where heat is added
by direct combustion of #2 fuel oil. Both the char and sand streams are
closed loop operations.
The flue gas produced by the fluidized bed sand heater is used to
preheat the incoming combustion and fluidizing air. After preheating
it is returned to the boiler air preheater flue gas inlet for additional
heat recovery and to inject the S02 produced as the result of combustion
into the main flue gas stream entering the adsorber, thereby assuring
closed loop operation of this gas stream.
3. RESOX* Section
The purpose of the RESOX Section is to provide for the continuous
on-site reduction of sulfur dioxide to elemental sulfur.
The low volume SOp rich off-gas, stream is directed from the regen-
erator to the RESOX reactor which is filled with crushed coal. The S02
stream is reduced to gaseous elemental sulfur and the liberated oxygen
combines with a portion of the coal carbon to from carbon dioxide. The
gases leaving the reactor enter a sulfur condenser where the sulfur is
condensed to molten elemental sulfur. The sulfur is collected and stored
in an insulated tank which is equipped with steam heating coils to make
up for heat losses through the insulation system to maintain the sulfur
in a molten form pending shipment via tank truck. The mass-flow coal
movement through the reactor is controlled by a discharge feeder. The
combination of non-consummed coal and ash is fed into a receiver vessel
for ultimate disposal after cooling. The tail gases leaving the sulfur
condenser consist of C02, H^O, N2 and those remaining "S" values not
converted to elemental sulfur. These gases are recycled to the boiler
via a centrifugal blower where the sulfur values are oxidized to S02
and then re-enter the adsorber allowing complete closed loop operation
of the unit. The design of the adsorption section is such as to account
for these internal re-cycle loops thereby maintaining a resultant system
efficiency consistent with code requirements. In future designs this
oxidizing step will probably be done internally with the unit and thus
eliminate this connection to the boiler.
-------
III. System Operation
A. General Operating Conditions
Daring the period from May, 1975 "to August, 1975 the system was in
a shakedown and/or startup mode of operation. Various pieces of equip-
ment in the Adsorption, Regeneration and RESOX* sections were operated
individually and then in combinations to simulate sub-system operation.
Finally, the sub-systems were integrated into section operations and
the Adsorption and Regeneration sections were operated with passage of
flue gas across the Adsrober on August 11, 1975 said, continued for a
period of 10 days. Table II gives a summary of general operating
conditions during the passage of flue gas.
TABLE II
Summary of General Operating Conditions
Boiler Load 15-^7 MW
Coal
Sulfur 1.14$ _ 3.148%
Heating Value 12,920 BTU/LB. AVG
Flue Gas Entering
S02 (ppmv) 900-2150
Temperature 235-29^ F
RESOX* construction was not complete at this time, thereby requiring
that the Regeneration section to run on bypass. The off-gases produced
during this period were mixed with the cleaned flue gas from the adsorber
and exited the system via the test stack.
The period of operation without flue gas allowed calibration of
instrumentation and equipment. It also allowed a period of char dryout
which was required due to high moisture content.
B. System Performance
1. Adsorption Section
The operating parameters for the two periods of operation of the
Adsorption Section are summarized in Table III.
887
-------
TABLE III
Summary of Adsorption Section Operating Parameters
w/o Flue Gas With Flue Gas
Char Plow Rate (LBS/tffi) 5000 - 11,850 5300 AYG
Gas Residence Time (Sees) X3 AVG
Char Dwell Time (Hrs.) 102 - k3 96 A^G
S02 Removal (%) ?6 AVG
Gas Flow (LBS/HR) — 17U.OOO AVG
Significant and persistent difficulties with char distribution and
adsorber char level indication systems made clear that modifications
would be required in this area. With the passage of flue gas this was
confirmed,, Poor distribution causing imbalance in bed levels led
ultimately to the shutdown of the system.
The imbalance of char levels throughout the system required portions
of the adsorber to be run continuously and others intermittently. This
resulted in stagnation of char in certain areas causing local hot spots
in several hopper bottom zones. This leads to a reduction in removal
efficiency and if not controlled could lead to a dangerous situation.
In order to investigate the reasons for this situation the unit was
shutdown.
However, despite these difficulties, there were several areas of
considerable encouragement. One of these is the fact that actual
pressure drops across the adsorber were considerably lower than had
been calculated. This confirmed, previous measurements on the Lunen
unit. Figure 1; shows the relationship between bed pressure drop and
flue gas flow. This figure is based on actual measurements taken during
the August operational period and subsequent testing conducted in
January, 19?6. Note that the system pressure drop is quite low.
Inasmuch as the adsorber fans represent approximately 80 percent of
total system power consumption it may be concluded that our power
consumption will be between 0.6 and 1.5 percent of steam generator
name plate rating depending on the system design and the mode of
operation. This bodes well for future designs which will be able to
take advantage of the lower pressure drops in terms of both lower
capital and operating costs. Another point of interest is that of
S02 removal. Removal efficiency is a function of several parameters
including S02 ppm at inlet, inlet flue gas temperature, char dwell
888
-------
125'
o
UJ
ffi
UJ
CD
g
CO
Q
(0
CO
o
o:
o
Q.
O
DC
U)
C
D
CO
CO
UJ
100-
75-
50-
25'
50,000
100,000
©
MEASURED ACROSS 6 BED
AT GULF POWER
MEASURED AT LUNEN
(ADJUSTED TO A 6'BED
FROM ACTUAL 5.75* BED)
PREDICTED FOR DESIGN
GAS FLOW AT GULF POWER
NM3/HR
150,000
GAS FLOW
Figure !(.: Bed Pressure Drop versus Flue Gas Flowrate
Gulf Power Company Demonstration Unit
-------
time and, gas residence time. While the period of operation was too
brief to draw final conclusions, certain positive trends can "be seen.
The averages for flue gas flow and char flow were 57 and 33 percent
respectively during the operating period. This gives rise to a gas
residence term equal to 21? percent of design. In addition, while
char flow rate was only 33 percent of design, it was only flowing
61 percent of the time. Even with these diverse operating conditions,
S02 removal efficiency was between 96 and 100 percent consistently.
The preceding indicates that the char and, therefore, its adsorption
capacity, is greater than had previously been anticipated. Of course,
additional periods of,longer term operation are needed to confirm this
point but, if present trends continue an additional reduction in
capital and operating costs based on a reduction in size of the
Adsorption and Regeneration Sections will be possible.
Finally, there is the matter of flue gas temperature across the
adsorber. As previously stated the temperature of flue gas entering
the adsorber was in the range of 235-29U F. Other than a brief period,
and only for effect purposes, the dilution air fan was not put into
service. With an average inlet temperature of 265 F the flue gas outlet
temperature ranged from zero to 30 F higher thus providing additional
buoyancy at the test stack. This temperature increase in the exit flue
gas temperature is a result of the exothermic nature of the adsorption
of S02.
During the period of operation the Adsorption Section, flue gas
treatment, availability was 100 percent and SOp removal was well above
design values.
2. Regeneration Section
The operating parameters of the Regeneration Section are summarized
in Table IV.
TABLE IV
Summary Of Regeneration Section Operating Parameters
Char Flow Rate (LBS/HR) 5300 AVG
Sand Flow Rate (LBS/HR) 180,000 AVG
Sand/Char Ratio (Vol) 13-9:1 AVG
Regeneration Temperature (F) 1200
890
-------
Here again the period of operation without flue gas permitted
calibration of certain instrumentation as well as operational checkout
of equipment at varying loads under actual operating conditions of
temperature and pressure. There occurred a variety of equipment
problems. The most serious result of these was the fact that Regener-
ation could not be carried out continuously. The specific problems can
be divided into four areas: char feed, char/sand separation, hot sand
conveying, and char cooling.
Char feeding into the regenerator had been accomplished by means
of a rotary airlock (star valve). This proved unsatisfactory for
metering purposes and a vibratory feeder was added for flow control.
Steam mixing with char fines (plugging the valve) and the jamming of
the valve due to the presence of foreign material caused most of the
problems in this area.
The design of the char/sand separator was inadequate for the
service intended, resulting in frequent outages to repair or replace
the screen deck and to reweld various cracks that appeared due to
differential thermal expansion between dissimilar metals.
The design of the hot sand bucket elevator called for extremely
tight clearances between moving parts at high temperatures. When
these clearances could not be maintained it resulted in damage to
stationary airlocks and the tripping of the elevator. Once this
occurred, regeneration could not proceed until the damage had been
repaired.
Difficulty in cooling char resulted from eratic char feed into
the regenerator and insensitivity of the control system at low flow
rates. These problems resulted in sporadic operation of the cooler
at less than optimum efficiency.
As may be seen from the preceding discussion none of the
difficulties encountered were related to the process chemistry. To
the contrary, the process aspects worked as expected but mechanical
problems prevented continuous operation.
As a result the Regeneration Section was only available to the
Adsorption Section $6 percent of the time. It should be noted that
it is not an absolute requirement that the Regeneration Section be
available 100 percent of the time the Adsorption Section is in operation;
however, 56 percent availability is unacceptably low.
891
-------
On the positive side of the ledger, it may be said that when
regeneration did occur it was complete. There have been no indications
of corrosion due to lingering acid and the majority of the equipment
met or exceeded design requirements.
3. RESOX* System
The initial run of Adsorption and Regeneration Sections did not
include the RESOX* Section as previously explained. However, two con-
secuative runs were made in October, 1975- The first without Regenerator
off-gas and the second with it. The general operating conditions for
both runs are summarized in Table V below:
TABLE V
Summary of RESOX* Section Operating Parameters
w/o Off-gas With Off-gas
Coal Plow (LBS/HR) 100-250 100-250
Gas Flow (LBS/HR) 1900 750-2000
Bed Temperature (P) AMB-1200 1200-1500
The initial run was made to check out gas and coal distribution and
temperature controls within the reactor. During this period the prev-
iously modified start-up air heater was used to bring the reactor up to
reaction temperatures. The function of this start-up heater is to bring
the reactor up to the proper temperature range in a scarcity of oxygen
so as to obtain proper temperature control in the reactor. Therefore,
the heater, which burns number 2 oil with air, does not use secondary
air for a diluent, but rather steam. Thus the heated flue gas produced
is essentially inert and very low in oxygen content (0.5-1.5%).
Once the reactor was brought up to temperature and all control
functions were tested, the Regeneration Section, operating in a hot
idle mode, was brought on line by the introduction of saturated char
into the regenerator. Due to the limited amount of saturated char
(approximately 13 hours worth at a 33% char flow rate), the RESOX* run
with off-gas was necessarily brief. All of the char previously
saturated had already been regenerated. Only the amounts left in the
various surge tanks was still saturated. In order to compensate for
the low flow of off-gas, the start-up heater was left in service.
This resulted in a diluted off-gas stream entering the RESOX*
reactor. However, the test showed that the RESOX* reaction is control-
lable and responded very nicely to the controls system. The control
system monitors temperature profile in the reactor and then uses control
air to increase temperature or steam to dampen the temperature. By this
method a reaction zone can be maintained at the proper level in the
reactor.
892
-------
Even with the start-up heater in service the flow of gas through the
sulfur condenser was too low for proper operation of the condenser. The
sulfur produced would not condense as a liquid, but rather was subject
to shock cooling resulting in the production of a sulfur mist which passes
through the condenser without being collected.
After the heater modifications the RESOX* Section was operated with
few problems and these were all of a minor nature. Because the
difficulties in the Regeneration Section, discussed above, had not yet
been resolved totally, we encountered a repeat of some of those problem
areas during the BUSOX* runs.
IV. Modifications Program
The preceding section presented the major difficulties encountered
in each part of the system. In this section we will examine the solutions
to these various problems. As pointed out previously, no major process
problems were encountered during the initial runs and it is not anticipated
that any will present themselves during the testing program.
1. Adsorption System
The major problem areas in the Adsorption Section were related to
char distribution and level control. To overcome these difficulties
the char distributing conveyors on top of the adsorber were replaced
with a "spider leg" system. This system accepts char from the regener-
ation section via a bucket elevator and oscillating conveyor and
deposits it in a right regular cylindrical tank atop the adsorber.
From there it is fed by gravity through a series of ten 8 inch pipes
into the eight first stage beds and the four second stage beds (one
pipe per two second stage beds). Each pipe or "spider leg" is
provided with a sight glass to observe flow in that leg. These legs
enter the top of the adsorber so as to be positioned above the
geometric center of each bed. This arrangement obviates the necessity
for individual level controls on each bed and transfers total control
over char level to the adsorber bottom extraction feeders. Running
these feeders at the same feed rate ensures that all portions of the
beds move down at an equal and continuous rate. Overall level
observation is then made via a single nuclear level detector in the
new surge tank. The level detector has a single source with two
pick-ups, one for high level and the other for low level. This new
system has been tested and functions as predicted.
893
-------
In addition, numerous new thermocouples have "been added at various
locations in and around the adsorber to allow better monitoring of
temperature distribution within the adsorber and, in effect to act as
a check on char and gas distribution. These thermocouples have now
provided confirmation that both gas and char are being distributed
evenly assuring maximum removal efficiency.
2. Regeneration Section
The four areas of equipment malfunction in the Regeneration Section
and the corrective action taken are as follows:
a. Char feeding
(l) A vibratory feeder and site glass were added upstream of the
star valve. This provides for a continuous smooth feed rate to the
star valve obviating the requirement for the valve to act as a feeder.
It now only acts to seal or lock. In addition a TV camera was installed
at the sight glass so that changes in the vibratory feeder feed rate can
be monitored in the control room.
(2) To avoid foreign material from jamming the feeder a removable
scalping screen has been added at the inlet of the star valve.
(3) To prevent steam from plugging the star valve the regenerator
operating pressure has been modified to slightly negative pressure as
opposed to slightly positive pressure. In addition a new flyash removal
system was installed. This system was added to the scope of supply too
late to be installed for the earlier shadesown runs. It's inclusion
prevents the carryover of fines to the regenerator. This eliminates the
problem of plugging in the star valve when steam is present.
b. Char/Sand Separation
The modifications made to the char/sand separator include the
following:
(l) A profile wire screen replaced the existing woven cloth screen.
This eliminates screen failure from relative motion between the pan and
the screen caused by the inability to maintain proper screen tensioning
at operating temperatures. The profile wire screen is rigid and does
not require tensioning.
(2) The channels that connect the hot stainless steel pan section and
the relatively colder carbon steel support frame and drive assembly have
been replaced. New connection clips allow structural integrity to be
maintained while allowing for thermal expansion when joining stainless
steel to carbon steel.
894
-------
(3) A revised insulation system to minimize thermal gradients in the
pans and its cover has been applied. The previous insulation system was
adequate for the temperatures involved but, could not withstand the con-
stand vibration. The new system consists of two inches of insulating
blanket compressed on studs to 5>0 percent of its original thickness.
This in turn is covered with expanded metal mesh and aluminum lagging.
The compression is made possible by large washers and nuts placed outside
of the lagging. The result is a rigid insulation system that moves with
the vibrating part without motion relative to it. To date, the insulation
system has performed satisfactorily.
c. Sand Conveying
The original design of the hot sand bucket elevator called for a
system of intricate airlocks that separated the 1200 P sand being con-
veyed in the buckets from the chain which was to operate at a maximum
temperature of 800 F. This was accomplished by the separation of chain
and bucket with a labyrinthine passageway and the injection of cooling
air in the area of the chain. Actual operation proved that the clearances
involved in the airlocks were unacceptably close causing interferences
between the moving chain and buckets and the stationary airlocks and
resulting in damage to the buckets and the airlock. It was further
observed that no cooling air was required to maintain the chain temperature
below 800 P. It was therefore decided to eliminate a portion of the
airlock labyrinth thus eliminating the source of interference. Operation
after this modification has been perfect and no cooling air has been used.
d. Char Cooling
The insensitivity of the control system has been alleviated by the
changout of the water control valve with one of greater linearity making
the valve response more even over the control range i.e., finer, more
even transition in valve opening versus signal input. In addition, the
use of steam temperature as an inferential measurement of exit char
temperature as opposed to direct char temperature measurement has been
installed. These two improvements coupled with a more even char flow
have resulted in adequate and controllable char cooling.
3. RESOX* Section
The only modifications made in the KESOX* Section in the period
following initial operation were the installation of added alarm points,
and the wring-out of existing control loops in the boiler feed water and
steam circuits.
895
-------
Due to certain wiring errors two circuits could not be placed in the
automatic mode and required manual operation. This caused wide variations
in both condenser water level and condenser steam pressure. These
deficiencies have been corrected and the system now operates on automatic
controls with good stability.
Y. Calendar of Operations
The modifications program outlined in the preceding section commenced
in late August 1975 and continued through January 1976 with the exception
of the five day RESOX* operation in October 1975- January and part of
February were devoted to a systematic testing of the modifications. This
testing program is outlined in Table YI.
With the completion of the pre-startup testing program the unit was
declared ready for the introduction of flue gas and all materials handling
loops as well as the RES OX* loops were put into continuous operation at
operating temperatures and pressures for a three-day period prior to the
introduction of flue gas to ensure stability and continuity of operation.
With the re-introduction of flue gas the operations at Scholz Steam
Plant entered a new phase, system testing.
Because of the long delay in obtaining successful operation, the
Scholz program is nearing the end of the timetable and budget provided
by the Southern Company. Consequently a test program has been organized
to work within these constraints. Any future operation of the system
will depend heavily on the success of this program. The schedule is
outlined in Table VII.
TABLE VI
Modifications Testing Program - January 1976
Test # Mame Test Objectives Results
1 Air Test To calibrate all gas side flow All test
measuring devices, establish objectives
pressure drop vs. gas flow were met.
characteristics, establish gas
distribution
2 Hot Sand To test modifications to All test
Circulation components in the hot sand objectives
loop, and to provide data for were met.
component calibration
896
-------
TABLE VI (Continued)
Test #
Name
3 Hot Sand/
Char
Circulation
Char
Circulation
Hot Sand/
Char
Circulation
EESOX Steam
BFW Circuit
Test
Test Objectives
To test the revisions made to
the char cooler and provide
data for future control loop
modifications. Test was to
use "old" char/sand separator.
To test spider legs distribu-
tion system and calibrate
various pieces of char handling
equipment. Also, to provide
testing of newly installed
fines removal system.
To test new char/sand
separator. Added Require-
ments of test #3 previously
To test the ability of these
circuits in the automatic mode.
Results
Test was cancelled
due to late start
and interference
with other operating/
modification schedules
test re-scheduled.
All test objectives
were met.
All test objectives
were met.
All test objectives
were met.
TABLE Til
Test #
1
2
3
k
Name
Startup
Var I
Var II
Var III
Long Term Testing
Program
Flue Gas Condition
Fuel Sulfur (%]
3-k
3-U
3-h
3-k
) Load
Constant
Constant
Flue.
Constant
Ash
No
No
No
Yes
Duration
[(. wks
1+ wks
il wks
U wks
897
-------
The objectives of the above testing program are to determine:
a. S02 removal performance
b. NOX removal performance
c. Flyash removal performance
d. Operating and Maintenance costs
e. System reliability
Inasmuch as this paper is being written in advance of the formal
testing program, due to Symposium deadlines, the outcome of the tests
will be the subject of future reports. At the Symposium we hope to be
able to share with you some of the early data taken in February.
VI. Conclusions
Based on the foregoing it is fair to conclude the following points:
A. Early start-up operations were slower and less successful than
desired.
B. Brief operations in August and October of 1975 did not show any
process deficiencies. To the contrary they showed the process to be
viable. However, longer term observations of process variables must be
conducted to verify the early results.
C. All of the difficulties encountered to date have been mechanical
and initial testing indicates that these have been corrected.
D. Early test results have indicated the capability for reduction
in both capital and operating cost factors due to lower than predicted
pressure drop and higher than expected char loading performance.
E. System reliability has not been demonstrated as yet but, with
the modifications testing successfully completed the unit will be run
integrated for an extended period to demonstrate this point.
898
-------
PART II
BF-STEAG DEMONSTRATION TOUT
OPERATIONAL EXPERIENCE AND PERFORMANCE
I. Introduction
With the financial support of the Federal Ministry of the Interior,
Bergbau-Forschung is presently operating a demonstration plant for the
desulfurization of flue gases. The plant accepts 150,000 M3n/hr. (88,275
CFM) of flue gas in the form of a slipstream from a 350 MW coal fired
boiler owned and operated by STEAG at Kellermann Power Station in Lunen,
Federal Republic of Germany.
The flue gas is desulfurized, without cooling, in a temperature range
of 110 to 160 C (230-320 F). The captured S02 is converted to elemental
sulfur.
This prototype demonstration unit is the largest dry desulfurization
plant operating thus far in the Federal Republic of Germany.
The only effluents emanating from the unit are utilizable and/or
disposable sulfur, cooling water, dry flyash and, small quantities of
char fines. The flyash is that which is captured in the adsorber, which
is located downstream of a high efficiency electrostatic precipitator.
II. Development of the Process
The long known principle of the process is the adsorption of S02
on carbonaceous adsorbents and its conversion to sulfuric acid by means
of combining with oxygen and water vapor also present in the flue gas.
The sulfuric acid is enriched adsorptively in the pore system of the
carbonaceous adsorbent. This adsorbent material, called char, loaded
with sulfuric acid is thermally regenerated. In this portion of the
process the acid is decomposed to S02» C02» and H~20 with the consumption
of carbon. The S02 rich gas thus produced is then further processed as
desired.
¥hen Bergbau-Forschung started their work in 19^5 for the development
of a process for the adsorptive desulfurization of flue gas, the first
problem was to develop a suitable carbonaceous adsorbent. The activated
char which was developed is characterized by a high abrasion resistance,
a particle size that guarantees low pressure drop in the adsorber (9 mm 0),
a very low reactivity with oxygen, and a high S02 adsorptive capacity.
899
-------
During the development of the process, the thermal regeneration concept
was studied in particular to obtain conditions that would not result in a
loss of activity of the activated char but, would result in an increase in
the level of activity up to some higher equilibrium value. Regeneration is
accomplished in the temperature range of 500 to 6^0 C (932-1202-F). The
heat is supplied by sand which acts as an inert heat transfer media. Thus
a regeneration temperature with a high heating-up velocity can be achieved.
This results in a regeneration reactor which is small in volume and has
short char residence times. In addition side reactions, e.g., with oxygen,
which would consume additional char, can be eliminated. It is a special
and important advantage that the regenerator off-gas is obtained in an
undiluted condition and has, thus, the highest possible concentration.
Therefore, the S02 in the flue-gas low concentrations (0.1 to 0.2% by
volume), can be concentrated up to more than 30 percent by volume. This
S02 rich gas can then be further processed to liquified S02» sulfuric
acid, or elemental sulfur.
III. Technical Description of the Demonstration Plant
The demonstration plant in Lunen is comprised of three main components,
viz., the desulfurization reactor, the regeneration plant, and a modified
Glaus plant for the processing of S02 rich gas to elemental sulfur. Figure 5
shows the desulfurization reactor (adsorber) designed in the form of a
travelling bed reactor with traverse gas flow. The beds of activated char
are placed in a circular silo. The flue gas from the boiler flow into
the center of the reactor and then radially outward through the beds of
activated char. The beds are each 1.75 m (5-7U feet) thick and the motive
force for gas flow is provided by an induced draft fan on the flue gas
outlet of the reactor. The char moves slowly from top to bottom between
sets of fixed louvers. Freshly regenerated char is continuously added to
the top of the reactor and loaded char is continuously drawn off at the
bottom of the reactor and conveyed to the regeneration plant.
The degree of desulfurization can be varied by means of the residence
time of the char in the adsorber thereby allowing the desulfurization
plant to follow boiler load.
The flyash collected in the adsorber along with char fines are
separated by screening prior to regeneration.. Loaded char is conveyed
to the regeneration plant via bucket elevators.
The regeneration plant is shown in Figure 6 and is also designed as
a travelling bed reactor. The loaded char is directly mixed with hot
sand at a temperature of between 6^0 and 720 C (1202-1328 F) which
results in a mix temperature of 500 to 650 C (932-1202 F) depending
on the temperature of the incoming sand. This mixture moves through the
reactor in 10 to 20 minutes and is then discharged onto a vibrating
screen deck to separate the two solid components. The char is then
cooled and returned to the adsorber for another cycle of adsoprtion.
The heat required to restore the sand to its proper temperature is
provided by means of combustion gases which also act to pneumatically
convey the sand to the top of the regenerator.
900
-------
activated coke
from regeneration
from the power plant
to the stack
'f&?$)ff blower
to regeneration
Figure 5: Adsorption Section - Lunen
Figure 6: Regeneration Section
Lunen
901
from adsorber
SOj-rich gas
regenerator
vibration screen
to adsorber
>flue gas
combustion
chamberf
4 gas
-------
The S02 rich gas emits from the upper part of the regenerator
reactor and is drawn off to the Claus unit. In this plant, shown
schematically in Figure 7» a certain percentage of the S02 rich gas
is reduced to H2S with town gas in a special reduction chamber. This H2S
is then combined with the remainder of the S02 in a three stage contact
plant with the resultant production of elemental sulfur. The sulfur is
collected as a liquid after condensation and is stored in a heated tank.
The Claus unit is operated at an efficiency of between 90 and 9k percent.
The remaining unconverted "S" values are incinerated back to S02 and
returned to the inlet of the adsorber.
In order to get a better concept of the entire unit Figure 8 shows
the integration of the three plant sections with respect to the power plant.
As can be seen from the left side of the figure, flue gas is withdrawn
downstream of the electrostatic precipitator (1), flows through the adsorber
(2), and is returned to the main flue gas duct prior to entry into the
power plant stack. It is also possible to take flue gas upstream of the
electrostatic precipitator through a duot not shown in the figure. This
allows for the study of the simultaneous removal of flyash and S02. The
concentration of inlet flyash can be as much as 5 g/m3n. In the center
of Figure 8, the regeneration plant is shown. Of interest to the energy
balance of the unit are a heat exchanger (18) for the utilization of waste
heat from the sand heating and conveying loop and a second heat exchanger
for the partial recovery of waste heat liberated during the cooling of
the char. The Claus plant is shown on the right side of the figure where
S02 rich gas is drawn off from the regenerator (11) to the reduction
burner (20). The waste gases from the tail gas incinerator (28) are
taken back to the adsorber (2).
An aerial view of the unit is shown in Figure 9« I"t shows a portion
of the 350 MW Zellermann Power Station of STEAG, as well as the entire
desulfurization unit. The flue gas enters the circular adsorber at the
top and leave near the bottom via an induced draft fan and from there to
the power plant stack. The regeneration plant is contained in the
rectangular tower. The S02 rich gas leaves the regeneration plant via
the vertical pipe seen on the side of that structure. The Claus plant
(partially hidden by coal conveyor) is in front of the adsorption and
regeneration section. Figure 10 shows a general view of the unit. The
entire unit is centrally controlled and monitored from the control room
shown in Figure 11. Inasmuch as this is a demonstration unit, a multitude
of data is recorded and/or monitored which will not be required in a
commercial unit.
IV. Commissioning and Initial Results
The adsorption and regeneration sections of the unit were designed and
constructed by Deutsche Babcock and Vilcox AG of Oberhausen in accordance
with the basic engineering provided by Bergbau-Forschung. The modified
Claus plant was designed and constructed by Davy Powergas. The order for
incorporation of the entire plant into the power station was placed with
STEAG. After completion of the entire plant, commissioning of
individual loops was started early in the summer of 1971).. Difficulties
were experienced with vendor supplied equipment such as bucket elevators
902
-------
SO2~rich gas gas cooler to adsorber,
after
burning
chamber
reduction |
chamber
condensation
stage I,
blower
trn-*
sulphur-pump
sulphur tank
sulphur
Figure 7: Modified Glaus Unit
Lunen
Figure 8: Flowsheet of prototype plant in Lunen
905
-------
Figure 9* Aerial view of
Kellermann Power Station
Liinen with desulfurizati
unit in foreground
Figure 10: View of BF unit
at Lunen
904
-------
Figure 11: Desulfurization Plant Control Room - Liinen
Figure 12: BF unit - Lunen - Operating Hours
905
-------
and fans. These had to be modified and/or changed before the operation
was satisfactory.
During the first trial run of the overall desulfurization plant,
difficulties arose after the adsorber was taken off the line due to
the fact that the adsorber shut-off dampers were not sufficiently tight
to isolate the adsorber from the power plant stack draft. Due to a vacuum
of 300 mm water glass gage (13.8 inches of water) caused by stack draft,
uncontrolled quantities of air were drawn through the adsorber char beds.
This resulted in the heating up of the char beds due to the adsorption
of oxygen. By installing double shut-off dampers at the inlet and the
outlet of the adsorber with a vacuum break between the dampers the
problem was solved. In the interim the safe operation of the plant was
achieved several times.
The modification work extended to the end of 197U- By early 1975
it was possible for the adsorption and regeneration plants to be put
into operation.
Figure 12 is a bar graph that shows the operating hours of various
sections of the unit as well as operating hours for the boiler, for
the period of January 1, 1975 through December 31, 1975.
During this period the adsorption section was•operated with flue
gas for 1,U50 hours while coupled to both the regeneration section and
the power plant. Further the adsorber was started up and shutdown 15
times in connection with the power plant. This cycling operating is due
to the fact that the power plant is operated as a peaking unit. The
regeneration plant was run for a longer period of 2,850 hours. The
reason for this was optimization work required before integrating all
three sections of the plant. The modified Glaus unit (reduction plant)
was operated for l\.QO hours including 320 hours on model gas.
Figure 13, shows some results for the period from February 28, 1975
to May 15> 19755 during which period flue gas was taken from the power
plant to the desulfurization plant for a total of 1,000 hours. In the
upper portion of the figure the inlet and outlet S02 concentrations are
given. In the central portion desulfurization efficiency is plotted
along with the char flowrate. The through-put of flue gas is shown at
the bottom. During this 1,000 hour period of trouble free operation
the through-put of flue gas varied between 100,000 M3n/hr. and
170,000 M3n/hr. (58,850 CFM and 100,OU5 CM) and the char flowrate
varied from 3.5 to 7.5 M3/hr (l2i| to 265 ft3/hr). This corresponds
to a char dwell time in the adsorber of from 1^0 to 90 hours. As a
result of these operating parameters the degree of desulfurization
varied between 65 and 99 percent. The measured values in this chart
were plotted as daily mean values.
Figure 1l±, on the other hand shows the effect of short term
variations of S02 inlet concentrations on the degree of desulfurization
and adsorber outlet flue gas temperature. These results were obtained
while holding flue gas flow constant at 150,000 M3n/hr. (88,275 CFM)
and char dwell time constant at 50 hours. During this period inlet SOo
906
-------
(O
Ul
(E t.w
^ 1-8
°* '"6
z 1.4
i 1.2
3 i.o
n- o a
•
iv
9
7
5
3
MMMM
•^M
4
/
^^
v
\
>
/
.
x^
/
I^^H
»
88 170 252 355 460 564 668 772 876 980
TIME (MRS.)
Figure 13: Various operating parameters vs. time for a 1000 hour
test run at Lunen
907
-------
1500
E
Q.
Q.
O'
l/l
1000
500
22.1..75 23.t-.75
I
degree of dfsulphuriza t i
Pigure 11).: Effect of short term variations of inlet S02 concentration
on the degree of desulfurization and outlet gas temperature
908
-------
concentrations varied between 850 and 1300 ppm with outlet concentration
of 100 and 1^.00 ppm respectively. Hence a degree of desulfurization
varied between 77 and 99 percent. The inlet flue gas temperatures were
between 118 C and 135 0 (2iji| F and 275 F) depending on the load on the
power plant. Due to the reaction heat, the outlet flue gas temperature
varied between 128 C and 1^2 C (262 F and 288 F)„ In this case the char
dwell time was kept constant. Without stating the dwell time the degree
of desulfurization can be improved and adapted to any desired value indepen-
dent of the load fluctuations in the power plant.
Figure 15, shows the relationship between pressure drop across the
entire adsorber and flue gas flow rate. With a flue gas flowrate of
110,000 M3n/hr. (61;, 735 CFM) the pressure drop is 85 mmWG (3.3 inches of
water) and with a flowrate of 170,000 M3n/hr. (100,0^5 CFM) the pressure
drop is 160 mmWG (6.29 inches of water).
As the coal fired power plants of the Federal Republic of Germany
cover only peak loads, they must change their loading frequently in the
course of the day and, as a rule, shutdown entirely during the weekends.
Therefore, high demands are placed on the flue gas desulfurization plant
with regard to flexibility. The test results obtained thus far with the
Bergbau-Forschung process show that the process can be adapted to
variations of loading without difficulty and that it can be started up
and shutdown simultaneously with the power plant.
In the event the desulfurization process must be interrupted for a
longer period, the activated char can be regenerated and cooled down
to 50 C to 60 C (122 F to 114.0 F). In the case of short term interrup-
tions in the supply of flue gas, e.g. during weekends, it has shown
to be advantageous not to cool down but, rather to add only inert gas
to the adsorber. The frequent starting up and shutting down to the
demonstration plant in Lunen proved that only small quantities of
nitrogen are necessary for this. This type of operation has the advantage
that the downstream reduction plant need not be shutdown, but can be
continued in operation at reduced capacity. To operate in this mode
the activated char loop has to be circulated at a reduced rate via the
regeneration plant. ¥hen flue gas flow is restored, the circulating
activated char can be kept at reduced flow rate for several days with the
same degree of desulfurization as that prior to the interruption in the
supply of flue gas.
V. Conclusion
As can be seen from the preceding, the results of this demonstration
plant clearly establish not only the viability of the process, but also
the high degree of desulfurization and low pressure drop capabilities.
They further show the flexibility of the unit in a difficult power plant
application. During operation of the demonstration plant at Lunen in 1976
we will continue to demonstrate this dry adsorption system and hope to be
able to share the results of new testing programs with you at a future
meeting.
909
-------
p [mmWs 1
160-
150
110- -
130
120
no
700
90
80
z
110000
13OOOO
150000
170000
Nnfjh
Figure 1^: System pressure drop vs. flue gas flow rate
910
-------
ACKNOWLEDGEMENT S
The authors wish to express their thanks to the United States Environ-
mental Protection Agency for their invitation to present these results to
the Symposium. Further, we wish to note the spirit of mutual cooperation
between Gulf Power Company, Southern Services, Incorporated, Bergbau-Forschung,
GmbH and Foster Wheeler Energy Corporation. Without this cooperation this
report would not have been possible.
911
-------
THE CONSOL FGD PROCESS
Robert T. Struck, Everett Gorin, and William E. Clark
Research Division
Conoco Coal Development Company
Library, Pennsylvania 15129
ABSTRACT
This paper describes a new regenerative process for the removal
of 95% or more of sulfur dioxide from power plant stack gases. Scrubbing
is conducted by circulation of a concentrated aqueous solution of
potassium thiosulfate through a pump-around loop containing a packed
bed scrubber for SO removal and an external reaction drum.
A solution containing a mixture of hydrosulfides, polysulfides,
carbonates and bicarbonates is injected into the circulating salt
solution just ahead of the reaction drum. The hydrosulfides and poly-
sulfides are quantitatively removed from the circulating salt solution
before entering the scrubber by reaction with sulfites formed in the
scrubber to produce thiosulfate.
The carbonates in the feed solution pass through the reaction
drum unconverted., but react in the scrubber with SO to produce the
aforementioned sulfites. The net product from the scrubbing process
is thus potassium thiosulfate.
The scrubfrer feed solution is produced by noncatalytic reduction
of the thiosulfate with a CO-rich gas. H S is simultaneously produced
which is converted to elemental sulfur in a Claus plant.
A very low by-product sulfate yield of about 1% of the SO
absorbed and very high SO removal efficiency as high as 99% in a simple
single-stage scrubber were achieved. The scrubbing solution is a
clear salt solution which provides for excellent operability in the
scrubber which is free of problems due to scaling and deposition
of solids.
The scrubbing process was demonstrated on both a 1000 ACFM and
a 30,000 ACFM pilot plant. The latter plant operated with flue gas
drawn from a coal-fired power plant of Philadelphia Electric Company.
Regeneration of the spent reagent was demonstrated in a separate pilot
operation. Finally, closed-loop operation was demonstrated with
continuous circulation of reagent between the regeneration and the 1000
ACFM scrubbing pilot plants.
913
-------
An economic analysis showed the process to be competitive with
other regenerable processes.
914
-------
THE CONSOL FGD PROCESS
INTRODUCTION
The original aim of this project was further development of
the aqueous potassium formate process reported by Yavorsky, et al in
1970 . The process chemistry, as a result of concurrent laboratory
research, was modified and simplified early in the course of the study.
The new scrubbing process was demonstrated both in a 1000 ACFM pilot
plant at the Research Division of Consolidation Coal Company in Library,
Pennsylvania, and in a 30,000 ACFM pilot plant located at the Cromby
Power Plant of Philadelphia Electric Company. A regeneration pilot
plant was also built and operated at Library. The two units at Library
were operated together in closed-loop, integrated fashion in the final
phase of the program.
PROCESS DESCRIPTION AND DEVELOPMENT HISTORY
Figure 1 is a schematic flow sheet of the scrubbing system and
corresponds to both the Library and the Cromby pilot plants. A venturi
scrubber was used in each unit to quench and humidify the hot flue gas.
It served also to remove the SO , HC1 and any residual particulate matter
from the flue gas.
The scrubber itself is a packed tower with a pump-around circuit
through a reaction drum. Fresh feed, spent scrubber product and make-up
potash are fed and withdrawn from the scrubber system as shown. The solid
lines indicate the feed and withdrawal points as practiced in the pilot
plant operations, the dotted lines indicate the preferred points for a
commercial system.
The key reaction in the Consol process is the one that takes
place in the reaction drum. Experiments confirmed previous literature
data by Hansen and Werres in 1933 that sulfite reduction to thiosulfate
may be effected by use of KHS as a reducing agent, i.e., by the reaction:
KHS + 2 KHSO = 3/2 K S O + 3/2 HO (1)
J £, £ ~) £
Similar reactions occur with polysulfides and sulfite.
The feed to the scrubber system is an aqueous solution containing
KHS, K S , KHCO , K CO and KOOCH produced in the regenerator by reduction
of the tniosulfate scrub product with CO. The polysulfides and KOOCH are
relatively minor components. The polysulfides are consumed in the reaction
drum while the KOOCH is substantially inert under the conditions of the
process. The feed enters the reaction drum where it is mixed and reacts
with the circulating scrub solution. Conditions are imposed such that the
KHS and K S are completely consumed in the reaction drum via reaction (1)
and the corresponding reaction for K S . A residence time of less than
1.5 min is sufficient at a pH of 7. Complete consumption of KHS is de-
sirable since, if it is allowed to enter the scrubber, some evolution of
H S into the stack gas occurs.
915
-------
The actual removal of SO in the scrubber occurs by virtue of
reactions with carbonates in the feed solution as, for example,
K CO + HO + 2 SO = 2 KHS03 + CO2 (2)
The overall reaction in the scrubbing system is now the sum of
reactions (1) and (2) , and may be written as reaction (3) for the KHS and
KHS + K CO. + 2 SO = 3/2 K S 0 + 1/2 HO + CO (3)
zo / z z o z ^
Similar equations apply for polysulfides and bicarbonate.
Control of the scrubbing process is effected in two ways. The
first is by controlling the rate of the fresh scrubber feed to maintain
the pH in the scrubber effluent at about 7. This insures nearly complete
SO removal as well as quantitative removal of the KHS and K S from the
feed reaction drum via reaction (1). The second is to control the com-
position of the scrubber feed to maintain the proper proportions of KHS
and K S with reference to the content of K?CO and KHCO . The stoichio-
metric relationships may be expressed by the equation:
2 S° + 3 S~
B K (4)
o =
where S , S and K refer to the mols of zero-valent sulfur, sulfur with
a valance of -2, and total mols of K in the feed solution associated with
the compounds KHS, K S , KHCO and K CO .
£, X -j £3
The above ratio R is termed the "acceptability ratio." When R
equals 1, reaction (3) would proceed quantitatively to produce pure thio-
sulfate. When R is greater than 1, a surplus of sulfur is present in the
feed solution and H S may be evolved into the stack. In practice R should
be close to but below 1 to maintain an excess of sulfite to help drive
reaction (1) to completion.
Lower values of R are undesirable as this permits the concen-
trations of sulfite and bisulfite ion in the circulating scrub solution
to increase. The result is a decrease in the absorption of SO and an
increase in sulfate formation as a result of the oxidation reaction.-
2 KHS03 + 02 = K2S04 + H2S04 (5)
The sulfate yield in the process is, however, quite low and is easily
kept to about 1 mol % or less of the SO absorbed. This is a consequence
of the low sulfite concentration maintained in the process, and the fact
that concentrated solutions of thiosulfate strongly inhibit the oxidation
reaction (5). The low sulfate yield in the revised process is one of the
principal advantages. The K SO also is readily saleable.
Because of the high solubility of K S O in water, it is pos-
sible to operate the scrubber system with a highly concentrated salt
916
-------
solution. Clear solutions are handled throughout the process except at
the point of sulfate removal. Particularly important is the fact that no
deposition of solids or scaling takes place in the scrubber.
The regeneration system in the Consol process is illustrated
schematically in Figure 2. The basic reaction is the noncatalytic re-
duction of thiosulfate with a carbon monoxide-rich gas. The reaction
may be written as follows:
3 K2S203 + 12 CO + 7 H20 = 2 KHS + 4 KHCO3 + 4 H2S + 8 CO (6)
The optimum temperature for the reaction from the practical point of view
is in the neighborhood of 450°F. Total pressures in the range of 500-1000
psig are suitable. Typical operating conditions for the various process
blocks are shown in Figure 2.
Hydrogen, in the absence of a catalyst, is inert under these
conditions. Hydrogen may be utilized, if desired, by introduction of a
catalyst containing metal sulfides of Group VI and/or Group VIII supported
on active carbon as reported in 1973 by Urban and Alpert and by Urban
The use of a catalyst was tested, but was rejected to maintain simplicity
of the system.
Some polysulfides are formed during regeneration. They probably
result from reduction of K S o., with H S:
K2S2°3 + 3 H2S = K2S5 + 3 H2° (7)
The ratio of KHS to KHCO in the reduction product is dictated
by establishment of equilibrium in the reaction:
KHS + C02 + H20 = KHC03 + H2S (8)
The acceptability ratio thus may be controlled by the amount of CO_ in
the reducing gas. Operation of the pilot regenerator was normally carried
out so that the acceptability ratio of the reactor product was above 1.0.
The final adjustment of acceptability ratio to the desired value of about
0.95 was achieved in the flash drum (cf. Figure 2). Here the acceptability
ratio is reduced by the decomposition reaction,
KHS + KHC03 = K CO + H S (9)
Simultaneously CO is released by the decomposition reaction,
2 KHCO = K CO + CO + HO (10)
J /,, 3 £ £
The release of CO? does not affect the acceptability ratio, but it does
convert slightly soluble KHCO into more soluble K CO .
The off-gas from the regenerator contains unreacted CO, H ,
CO and H S. The H S is selectively removed, after cooling and conden-
sation of the steam, by scrubbing with a polar organic solvent. Several
proprietary processes are available for this service. The absorbed
917
-------
H S-CO gases are released from the solvent in its regeneration. These
are combined with the H2S-CO gas released in the flash drum and sent to
a Glaus plant where the H S is converted to sulfur. The Glaus tail gases
are incinerated with a portion of the scrubbed regenerator offgas to pro-
vide hot flue gases for direct reheating of the scrubbed stack gas.
The naior portion of the cleaned regenerator offgas is recycled
through an oxygen-fired, partial-oxidation unit. No steam is added to
the unit and hydrogen in the recycle gas is converted to CO by reverse
water-gas shift:
H + CO = CO + H20 (ID
EXPERIMENTAL
The entire Library scrubbing unit, i.e., venturi, scrubber and
reaction drum, was constructed of type 316 stainless steel. The bottom
section of the scrubbing tower served as a cyclone separator to remove
the venturi water. A chevron-type mist eliminator was used above the
separator to prevent water droplets from entering the packed section.
The scrubbing tower was 20 inches ID and contained a 42 inch
depth of No. 1 Intalox packing in the form of polypropylene saddles. A
distributor was used to supply an even flow of the recycle liquid over
the top of the packing. A York mesh-type demister was used above the
packing.
Hot flue gas was supplied to the Library unit from a packaged
boiler fired with No. 2 fuel oil. A forced-draft fan was used to drive
the flue gas through the scrubbing unit. Sulfur dioxide was metered
into the gas to achieve concentrations similar to that in the coal-fired
units. Provisions were also made to feed flue dust into the gas prior
to the venturi.
The Cromby pilot plant was constructed in a similar manner.
The flue gases were taken from the output of a 160 MW, coal-fired boiler
as a side stream and a separate forced-draft fan was used to drive the
gases through the unit. The venturi was constructed of type 316 stain-
less steel.
The scrubbing tower was 8 ft ID and, in contrast to the Library
unit, was constructed of rubber-lined carbon steel. The lower section
again was used as a cyclonic separator above which was a chevron-type
mist eliminator. The packed section was 5 ft high. The packing initially
used was No. 3 Intalox polypropylene saddles. A weir-trough liquid dis-
tributor was used, but distribution between the two weirs proved to be
unequal. After Run 2, the packing was changed to No. 2 Intalox plastic
saddles, and an improved liquid distributor was installed to ensure equal
flow to the weirs. The initial configuration also used a York mesh,
style 931, teflon mist eliminator at the top of the tower. The reaction
drum and all the piping for the pump-around circuits were constructed of
rubber-lined carbon steel.
918
-------
The regeneration pilot unit is shown schematically in Figure 3.
The regeneration vessel itself was a mechanically-agitated, gas sparged
unit equipped with a Hastelloy liner for protection against stress cor-
rosion cracking due to chloride ion contaminants in the feed. Hot liquid
piping also consisted of Hastelloy tubing. All other tubing and equip-
ment such as the flash drum were constructed of 316 stainless steel."
The agitation was effected by a six-bladed turbine impeller,
8.5 inches OD, driven through a mechanical seal by a variable-speed
drive, generally at 330 rpm. The rate of absorption of CO was found to
vary with the 1.8 power of the rpm, which is similar to the value of 2.0
found by Yoshida et al in 1960
The liquid pool in the regenerator was 15 inches ID x 15 inches
high. Level was controlled by an overflow weir through which the regen-
eration product flowed into the surge pot below the regenerator vessel.
Integrated operation with the scrubber and regenerator was
carried out by continuous circulation of salt between the two units.
The regenerator and scrubber feed tanks provided surge capacity. The
closed-loop operation was conducted continuously for a period of ten
days. This corresponds to 12 turnovers of the inventory.
DISCUSSION OF RESULTS
Scrubbing
At Cromby, a series of tests with water scrubbing was run
initially to check the mechanical components and the operation of the
venturi. The venturi pressure drop was 150% of the design value even
with the variable plug fully extended. All tests were therefore run
with higher than desired pressure drop. With a flue gas input of 16,000
SCFM, the venturi operated with a pressure drop of 11 to 13 inches of
water. The inlet dust loading was 0.22 grains/SCF and the outlet 0.01
grains/SCF for 95% dust removal. Eighty percent of the removed particu-
lates were within the range of 3 to 10 microns. No measurable particu-
lates were removed in the scrubbing loop.
The results of the SO scrubbing tests are given in Table 1.
Tables 2 and 3 give the composition of the feed and product solutions
and gas, respectively.
The first test, CS-2, was carried out with the larger size,
No. 3 saddles. It was a continuous ten-day run. No increase in pressure
drop over the packing was observed during this run. The York demister
showed a very slight increase in pressure drop for the first seven days
and thereafter it went up sharply due to collection of fly ash. It was
necessary after this to periodically back-flush the demister with water
every 22 hours for a period of 45 to 60 seconds. The back flushing
temporarily restored the pressure drop to its original value.
919
-------
Tests showed that at the 20,000 ACFM gas rate used in this run,
less than 0.01% of the potassium in the fresh feed solution would exit
the stack without a mist eliminator.
Average absorption of SO was relatively low in this run, i.e.,
88%. It was found, however, that proper control could be maintained,
and the percent removal of SO held constant during wide swings of as
much as a factor of 4 in the SO concentration of the entering flue gas.
Inspection of the packing after the run showed it to be clear with only
an insignificant film of salt on the packing.
For subsequent runs the No. 3 Intalox packing was replaced by
No. 2 packing. A new liquid distributor and pumps of higher capacity
also were installed. The mist eliminator at the top of the scrubber
was replaced by the Heil blade mist eliminator from the top of the cyclone
separator. Runs CS-3A, 3B and 4 were carried out in sequence for a period
of continuous operation of 6.5 days. During this period, no increase in
pressure drop over the packing was observed. The mist eliminator re-
mained clean and no appreciable foaming or carryover of scrubbing solution
was noted. The only difference between Runs CS-3A and 3B was the higher
liquid circulation rate employed in the latter run.
The use of the improved distributor and the smaller packing in-
creased SO absorption from 88.0 to 96.2%. Increasing the liquid circula-
tion rate in Run CS-3B raised the absorption of SO to 98%. The use of
the higher gas rate increased the SO absorption to 99%, but only at the
expense of a considerable increase in pressure drop.
The sulfate formation was small throughout all of these runs and
averaged about 1% of the SO absorbed. No appreciable NO absorption could
be detected within the precision of measurement. Mass transfer coeffi-
cients for SO absorption were in the same range reported by Johnstone and
Singh in 1937 ( '.
The rubber lining in all the vessels was in good shape after this
operation. Corrosion information was also collected via strips and spools
in the system. Based on these short-term results (518 hours), the choice
of materials would be among rubber-lined carbon steel, type 316 stainless
steel, and fiberglass reinforced pipe.
Closed-Loop Operations
The closed-loop run was conducted over a steady operating period
of ten days. During this period, the pressure drop through scrubber and
demister remained quite constant. A mesh-type demister was used in the
Library scrubber and, in contrast to the experience at Cromby, no plugging
problem was encountered. It is concluded from this and other evidence
that the plugging of the Cromby demister was due primarily to the entrap-
ment of fly ash.
The sulfate yield averaged slightly less than 1 mol % of the SO
absorbed. Only a fraction of the sulfate was removed by the line filters 2
920
-------
in the scrubber product line (cf. Figure 1) or the regenerated feed
system (cf. Figure 2), The major portion of the sulfate settled out in
the regenerated product tank. It is at this point in the system that
the sulfate is least soluble. Recent work has demonstrated continuous
rapid removal of this sulfate after the flash drum via low-pressure
filtration.
The onlv materials added or removed during the run were water
and a small amount of K CO,, added to make up for the K^SO. removed.
23 24
In a commercial process, no dilution water need be added except to
humidify the qas, since the condensate liquor would be returned to
the product salts (cf. Figure 2).
The initial inventory was turned over twelve times in the
course of this run. No foreign materials accumulated in the product.
The compositions of all the feed streams, their rates, temperatures
and pressures are given in Table 4. The stream numbers correspond to
those given in Figures 1 and 3. The figures given in Table 4 are those
midway through the operation, i.e., after five days. Sulfur absorption
was high throughout this run and averaged about 99%.
Potassium formate is one of the components of the product but
is not in reality a foreign material. The amount present is controlled,
other conditions being constant, by the percent conversion of the thio-
sulfate. The higher the conversion, the higher is the formate content
of the product. The formate content of the regenerator product shown
in Table 4 was, for example, 4 wt. % and the thiosulfate conversion was
95 wt. %. Five days later, at the end of the run, the formate concen-
tration was reduced to 1.7 wt. % when the thiosulfate conversion was de-
creased to 88 wt. %.
PROCESS ECONOMICS
A cost estimate based on an engineering study for a specific
site is shown in Table 5. The scrubbing section is designed to handle
the full output from a 200 MW unit with a heat rate of 9,000 Btu/KWH
and 35% excess air in the stack gas to be scrubbed. The regeneration
section provides the capacity to handle coal with up to 3.5 wt. % sulfur
(at 12,000 Btu/lb) when the heat rate is 9,000 Btu/KWH and the average
load factor of the utility is 80%. The overall sulfur removal is 93%
after the incinerated Glaus tail gas has been added to the scrubbed
stack gas.
The costs do not include charges for Venturis and fly ash re-
moval. Retrofit costs are included. The costs were developed for 1977
startup at a time when the rate of inflation was quite high. The invest-
ment estimates correspond essentially to a Chemical Engineering Plant
Cost Index of 200. The operating costs also reflect estimates for 1977
startup. Sufficient details have been given in Table 5 to adjust the in-
vestment and operating costs to other bases as desired.
921
-------
The engineering study also may be used to estimate the cost at
a new 1000 MW coal-fired power station. On the same basis, i.e., 1977
startup and firing 3.5% sulfur coal, the estimated investment and operating
costs are 70 $/KW and 4.1 mills/KWH, respectively.
ACKNOWLEDGMENT
Appreciation is expressed to the utility companies who' provided
a large part of the funding for this work. These are: Allegheny Power,
Consumers Power, Detroit Edison, Niagara-Mohawk, Ontario Hydro and Phila-
delphia Electric. Appreciation is also expressed to C. Itoh & Company,
Inc., for their financial contribution to the work reported here.
Appreciation is also expressed to the Consolidation Coal Company
Management for their support of this work and their permission to publish
the results.
A large number of engineers and chemists participated in this
project and the success of the project was largely due to their efforts.
Specifically noteworthy were the contributions by M. D. Kulik, E. B.
Klunder, P- J. Dudt, W. E. McKinstry, and N. J. Mazzocco. The coopera-
tive spirit of George Kotnick, J. A. Gille, E. G. Boyer, Jr., and M. W.
Moore of the Philadelphia Electric Company during the work at Cromby
Station was deeply appreciated.
REFERENCES
1. Yavorsky, P.M., Mazzocco, N. J., Rutledge, G.D., and Gorin, Everett,
Environmental Science & Technology, 4_, 9, 757-765 (September 1970).
2. Hansen, C.J., and Werres, H., Chemiker-Zeitung, 57, 25-27 (1933).
3. Urban, P., and Alpert, P.J., U. S. Patent 3,725,303, April 3, 1973.
4. Urban, P., U. S. Patent 3,737,515, June 5, 1973.
5. Yoshida, F., Akio, S.I. , and Muria, Y., Industrial and Engineering
Chemistry, 52, 5, 435-438 (May 1960).
6. Johnstone, H.F., and Singh, A.D., Industrial and Engineering Chemistry,
29, 3, 286-297 (March 1937).
922
-------
Table I, SUMMARY OF CONDITIONS AND RESULTS
CROMBY SCRUBBING TESTS
Run Number
Packing Size
Flue gas entering venturi, ACFM
Flue gas leaving scrubber, ACFM
Liquid recirculation rate, gpm
Length of run, hours
Superficial gas velocity in
scrubber, ft/sec.
Temperature, °F
Flue gas entering
Gas from venturi
Gas from stack
Liquid to packing
Liquid from packing
Pressure Drop, Inches of Water
Over venturi
Over packing
Total
CS-2
No. 3 .
20,900
18,200
205
240
6. 0
300
128
130
146
142
11
0.5
22
CS-3A
21, 100
18,200
190
36
6.0
310
136
138
150
148
11
1.2
15
CS-3B
VT-, 0
JNO . ^
21,300
18,400
250
24
6. 1
310
136
138
151
149
12
1.3
16
CS-4
31,900
28,500
255
94
9.
317
132
140
157
140
25
5.
38
S
,5
, 0
pH of scrubber effluent 7.5 7.4 7.6 6.9
S02 absorption, % 88.0 96.2 98.0 99.1
Mol % of SO2 absorbed producing
sulfate 0.9 1.3 1.4 1.0
Calculated value, K a,
Ib mol/hr ft3 atm 21 33 40 74
923
-------
Table 2, COMPOSITION OF FEED AND PRODUCT
SOLUTIONS FROM CROMBY PILOT OPERATION
Run Number
CS-2
CS-3A
CS-3B
CS-4
Average Feed Composition, Wt.%
KHS
KoCOo
Acceptability Ratio
Product Composition, Wt.%
K2S2°3
KHS 03
K2S04
K2C03
Fly Ash
15.5
31.9
0.95
54.20
1.34
0.66
0.2
-
15.3
32.3
0.94
54.60
1.20
0.92
0.2
—
15.4
32.5
0.93
55.60
1.33
1.00
-
—
15.4
32.3
0.94
32.50
1.54
0.42
0.1
0.008
Total salts (by drying), Wt.% 57.6 57.8 59.5
Distribution of Sulfur,
Mol % of absorbed SO2
K2S2O3 96.2 96.2 95.9
KHSOs 2.8 2.5 2.7
K2S04 0.96 1.33 1.40
35.3
93.7
5.3
0.99
Table 3, FLUE GAS ANALYSES
CROMBY PILOT OPERATION
Run Number
Dry Gas Composition, Vol.%
C02
02
N2
Argon
S02, ppm
Wet Gas
Vol.% H20
CS-2
CS-4
Intake
14. 1
4. 1
80.5
1.0
2975
Stack
14.2
4.2
80. 6
1.0
337
Intake
14.5
4.3
80.2
1.0
2010
Stack
14.5
4.3
80.2
1.0
18
6.7
14.6
924
-------
Table 4, OPERATING CONDITIONS AND STREAM
COMPOSITIONS, CLOSED LOOP OPERATION
Scrubbing Section
Gas Stream No.
Flow, SCFH
Temperature, °F
Pressure, psig
Gas velocity, ft/sec
Composition, Vol.%
CO
H2
co2
°2
H2S
S02, ppm
N2 + Argon
H20
% S02 Removal
Liquid Stream No. 7
Flow, gal/hr 648
Flow, Ib/hr
Temperature, °F 144
pH 7.4
Composition, Wt.%
H20 49.7
K2S203 41.5
KOOCH 5 . 7
KHS03 0. 9
KHS 0. 0
K2S04 0.8
K2C03 1.4
KHCOs 0. 0
K2S2 0.0
1
45,000
350
0.7
-
0. 0
-
12.0
3.3
0. 0
2200
74.5
10. 0
(
8
648
-
140
6.8
-
-
-
-
-
-
-
-
-
3
47,600
131
0
7.3
0.0
-
11.3
3. 1
0.0
15
69. 1
16.5
QQ 5 >
9 10 11
10.6 -
126 - 123
149 70 144
7.4
57.8 - 49.7
1.8 - 41.5
4.6 - 5.7
0.0 - 0.9
7.7 - 0.0
0.4 - 0.8
9.2 47 1.4
15.5 - 0.0
3.0 - 0.0
Regeneration Section
Acceptability Ratio
Absorbed Sulfur
Converted to
Sulfate, mol %
0.97 -
16
430
403
595
792
0.3
19.9
0. 0
0.0
0.0
0.6
0. 0
17
704
220
595
7. 4
0.2
39.3
0. 0
14. 1
0.0
4. 7
34.3
18
148
219
26
0.0
8.2
8.2
0.0
8.8
0. 0
24.2
58.6
13 14
40 105
70 150
100
51.5
39.7
6.0
0.8
0.0
0.8
1.2
0. 0
0. 0
15
28
100
19
105
57.7
2.2
4.0
0.0
6.9
0.9
10. 1
14.5
3.7
0.97
0.95
925
-------
Table 5, ESTIMATED COSTS FOR 200 MW PROTOTYPE, 1977 START-UP
Investment Summary
Scrubbing and reheat
K2SO4 Recovery
Regeneration, including oxygen plant, CO generation, acid gas
separation and Glaus plant
Support facilities, including storage and boiler feed water,
cooling water, process water; and waste water systems
Total
MM$
7. 1
0.5
14.3
2.8
24.7
Operating Costs
Operating Labor, incl.
Benefits
Supervision, incl. Benefits
Ind. O.K. & Admin., Incl.
Benefits
Electricity
High-Sulfur Fuel Oil
KOH
Glaus Catalyst
Misc. other Chemicals
Maint. & Misc. Supplies
Insurance
Int. on Working Capital
Sub-Total
Capital Charges
Sub-Total
Credits:
Sulfur
K2S04
Sub-Total, Credits
Annua 1
Quantity
45,990 hrs
s 8,760 hrs
(55% of above)
22,118 MWH
721,000 MM Btu
898 Tons
13.9 Tons
-
(5% of $24.7 MM)
Rate
$10. 13/hr
$10.93/hr
25$/MWH
3$/MM Btu
355.60$/ton
200$/ton
-
(0. 175% of $24.7 MM)
(10% of $1.43 MM)
(16% of $24.7 MM)
16,013 LT
982 Tons
25$/LT
42.50$/ton
Annual Costs,
MM$/Yr
0.466
0.096
0.309
0.553
2. 163
0.319
0. 003
0. 025
1.235
0.043
0. 143
5.355
3.952
9.307
-0.400
-0. 042
Mills
KWH
3.82
2.82
6.64
-0. 442
-0.32
Net Total
8.865
6.32
926
-------
FLUE GAS
MAKE-UP
WATER
TO FLY ASH
VENTURI
POND 8r-
NEUTRALIZATION
ROM
or
K2C03
FEED
TO STACK GAS
REHEATER
3:
SCRUBBER
X
REACTION
DRUM
FILTER
CAKE
HFILTER
TO
REGENERATOR
i
SCRUBBER FEED
FROM REGENERATOR
Figure I Schematic flow diagram, scrubbing section.
-------
SPENT SOLUTION
FROM SCRUBBER
CO DEPLETED
l~'
t
"8
REGENERATOR
450° F
750 PSIG
REGENERATED
""SOL'N. OUT
SELECTIVE
H2S
REMOVAL
RECYCLE GAS
CONDENSATE
0-RCHVNTHESISGAS
?s
CO!
xl
PARTIAL
OXIDATION
2500°F
850 PSIG
02 IN
HI-SULFUR
FUEL OIL
-.H^-l
CO 2
CONDENSATC
CLAUS
TAIL^
"GAS ""
r
AIR
INCINERATOR
_HOT_
TO STACK
REHEAT
K2S04
OUT
SULFUR
OUT
Figure 2 Block diagram, commercial regeneration process.
-------
WATER
(£5
CD
CO
Z>
a:
o
o
en
LJ
CL
en
REGENERATOR
FEED
r
0
n
-a
REGENERATOR
GAS
P| SURGE
U POT
TO
AFTER-BURNER
FLASH
DRUM
REGENERATED
REAGENT
TANK
REGENERATED
REAGENT
TO
SCRUBBING
Figure 3 Schematic flow diagram, pilot regeneration.
-------
UNPRESENTED PAPERS
931
-------
INFORMATION TRANSFER PROGRAM
Timothy W. Devitt and Thomas C. Ponder, Jr
PEDCo-Environmental Specialists, Inc.
Suite 13, Atkinson Square
Cincinnati, Ohio 45246
933
-------
INFORMATION TRANSFER PROGRAM
1.0 INTRODUCTION
In September 1975, the Process Technology Branch of the
Industrial Environmental Research Laboratory/RTF, U.S.
Environmental Protection Agency, contracted with PEDCo-
Environmental Specialists, Inc. and Radian Corporation to
develop an Information Transfer Program related to air
pollution control technology. This program would establish
a method for effectively disseminating data developed, at
least in part, by IERL/RTP research programs. The program
would be sufficiently comprehensive to address control of
potential water and solid waste impacts resulting from use
of air pollution control technologies.
Since the Process Technology Branch is responsible for
research in the area of flue gas desulfurization (FGD), this
program was initiated with FGD related information transfer
products. If this program is successful, it may be expanded
to cover other pollutants, control techniques and indus-
tries. Potential pollutants to be addressed include par-
ticulates, nitrogen oxides, hydrocarbons, and carbon mon-
oxide. Other sulfur dioxide control strategies include
This project has been funded at least in part with Federal
funds from the Environmental Protection Agency under Con-
tracts Number 68-02-1321 and 68-02-1319. For additional
information on this project the following individuals
should be contacted: Mr. Timothy W. Devitt, Project Direc-
tor, Mr. Thomas C. Ponder, Project Manager, PEDCo-Environ-
mental Specialists. Mr. R. Murray Wells, Program Manager,
Mr. James Dickerman, Project Director, Radian Corporation.
Mr. Wade H. Ponder, Project Officer, Mr. Richard Stern,
Branch Chief, Process Technology Branch - IERL - U.S. EPA.
934
-------
clean fuels (physically and chemically cleaned coal), proc-
ess modifications (e.g., fluidized bed combustion), and fuel
conversion (coal gasification and liquefaction).
2.0 PROGRAM" APPROACH
To assess the need for the Information Transfer Pro-
gram, the following three tasks were undertaken, using FGD
as the evaluation vehicle.
Assess the FGD Data Base -.Radian Corporation examined
the data at IERL/RTP and other sources to determine the
extent of information available for the transfer program.
Prepare Examples of Program Outputs - Both Radian and
PEDCo prepared examples of potential program outputs. An
audio-visual presentation was prepared to explain the pro-
gram to potential users. Radian was responsible for the
summary reports, newsletters, direct contact meetings and
symposia outputs, while PEDCo-Environmental was responsible
for example products related to FGD system data books and
FGD system evaluation procedures.
Assess the Information Needs of Prospective Users -
When the proposed program was first formulated, the fol-
lowing groups were identified as potential users:
0 EPA Regional and state/local environmental agen-
cies
0 Electric utilities
0 Environmental consulting firms
0 Architecturaland engineering firms
0 Selected industries with potential S02 control
needs, such as metal processing, paper and allied
products, petroleum refining and related products,
sulfuric acid production, and coal mining.
935
-------
An audio-visual program was presented at the ten EPA
Regional Offices to determine, first, their overall interest
in the program and then specifically, their recommendations
for, ITP outputs. Presently, the Information Transfer Pro-
gram is being explained to representative electric util-
ities, again to determine whether there is sufficient in-
terest and need, and to define the content and format of ITP
products. A final phase of this program involves presenta-
tion to selected industrial users.
3.0 POTENTIAL ITP PRODUCTS
Various potential "products" of the ITP were identified
for the various "user" categories described above. These
products are briefly described below.
Summary or Capsule Reports - Reports will be written to
cover specific aspects of an SO control strategy. Capsule
^S.
reports would address first and second generation FGD
processes. Each report would include:
0 Process description
0 Design considerations
0 Status of development
0 Raw material and utility requirements
Costs
0 Land requirements
0 Environmental considerations
Newsletters - Newsletters would present information not
contained in current commercial or governmental newsletters,
such as:
936
-------
0 Status reports on control technology. This would
include new developments in technology or improved
methods for utilizing existing technology.
0 A series of abstracts covering recently published
reports on selected subject areas.
0 A list of knowledgeable contacts in the field of
sulfur dioxide control technology.
Symposia and Direct Contact Meetings - Two levels of
meetings are being considered.
0 National and regional symposia - This FGD sympo-
sium is a good example of an effective information
transfer symposia related to a specific subject
area. Other symposia could cover such areas as
coal cleaning, low sulfur coal, coal liquefaction,
and coal gasification, and be conducted on a
regional or national basis.
0 Direct contact meetings - Meetings would be held
in workshop format. This would provide close
interaction between individuals requiring in-
formation on SOX control and experts who can
provide such information.
Data Books - Data books would be prepared for such
developed FGD processes such as lime and limestone. Data
limitations prevent the preparation of detailed data books
for other processes. The purposes of the data book are:
0 Provide comprehensive, up to date information on
existing FGD systems in a format suitable for use
by design engineers and those evaluating proposed
FGD system designs.
0 Provide data to aid in the selection and evalua-
tion of FGD systems on a site specific basis.
0 Provide a format for periodic updating of process
information.
937
-------
The data books would be technically oriented, address-
ing a user audience involved in the design, operation, and
evaluation of FGD systems.
A summary of the proposed outline for the data books
follows:
1. Process Description - A detailed process analysis
will be presented. Flow sheets and material and
energy balances would be presented for "typical"
installations.
2. Process Data - Data required to evaluate a process
will be identified. Data from operating systems
will be presented along with the results of tests
conducted by pump manufacturers, etc.
3. Process Design - Design data will address such
areas as velocities, L/G ratios, efficiencies,
pressure drops, stoichiometric ratios, and solids
content for various system operating modes.
4. Project Design - Installation schedules will be
provided for single and multiple module FGD
systems. Information on space requirements and
retrofit layouts will be provided.
5. Process Operation - Manpower requirements, in-
cluding skill levels, for the maintenance and
operation of an FGD system will be identified.
Information on potential FGD system operating
problems will be presented.
6. Process Cost Data - Calculational procedures will
be presented to permit the user to estimate the
cost of a proposed FGD system or to analyze costs
submitted by FGD system vendors. The user will
then be able to obtain relative cost estimates for
different design and operating conditions.
7. Case Histories of Process Applications - Case
histories will be presented for selected FGD
installations. Problems and solutions encountered
at these FGD installations will be described.
938
-------
FGD Evaluation Procedures - Development of the fol-
lowing procedures is being considered to aid in the evalua-
tion of FGD systems:
0 Procedure for FGD system reliability evaluation.
0 Procedure for an interprocess comparison/selection
of an FGD system on a site specific basis.
0 Procedure for FGD system cost/reliability evalua-
tion.
Several versions of each evaluation procedure will be
prepared to accommodate multiple users and different desired
levels of complexity. The procedures are described below:
0 Reliability Evaluation Procedure - This model will
aid in evaluating the reliability of flue gas
desulfurization and particulate control systems.
This simple procedure will permit estimation of
the percentage of down-time for the control sys-
tem. Since reliability data are not always avail-
able, the procedure will contain data that can be
substituted for actual system data as a first
level approximation. The procedure can be per-
formed manually in two or three hours.
0 Interprocess Comparison/Selection Procedure - This
procedure will aid in the selection and comparison
of FGD systems. The procedure will determine the
annualized operating and capital costs for dif-
ferent FGD processes at specific sites. Site
specific constraints such as retrofit difficulty,
sludge disposal, by-product marketability, fuel
penalties, and process complexity would be fac-
tored into these procedures. The procedure will
not only aid in determining the lowest capital
cost system for a specific site, but the lowest
cost system over the life of the installation.
0 Cost/Reliability Evaluation Procedure - This
procedure will aid in the evaluation of various
intraprocess alternatives. Cost/benefit analysis
will be performed to determine the impact of spare
components on total system cost and reliability.
This procedure will be fairly sophisticated,
939
-------
require good FGD component reliability data and be
computerized to be properly utilized. Figures 1
through 5 present system design scenarios illus-
trating the impact of component sparing and in-
creasing component reliability on process opera-
bility and cost.
Figure 1 presents a two module scrubber on a 300 MW
boiler with a single spare absorber recirculation pump for
both modules (system reliability is 64.1 percent). Figure 2
shows the addition of a second s.pare recirculation pump;
reliability has improved to 65.6 percent. Figure 3 illu-
strates the sparing of additional pumps in the system and
the subsequent improvement in reliability. Figures 4 and 5
illustrate two methods for obtaining the same system reli-
ability. In Figure 4, the reliability of the component has
been improved by design changes; fans, absorbers, and
reheater reliability can be improved by changes such as use
of forced draft fans, simpler absorber internals, and moving
reheaters out of the flue gas stream). Although these
changes raise costs, they are still more economical than
adding a spare absorber as shown in Figure 5. Table 1
summarizes the reliability, capital costs, annual costs, and
expected costs, and untreated gas percentages for Figures 1
through 5.
4.0 BENEFITS FROM THE INFORMATION TRANSFER PROGRAM
Potential benefits of the various ITP products are
listed by user group.
Regional and Local Environmental Agencies
0 Provides effective training tools.
0 Presents technical and cost data on FGD systems in
a consistent format.
940
-------
0.94* 0.96 0.94
Tl Tel [T
SYSTEM RELIABILITY: 64.12
0.98 0.97
SYSTEM
COMPONENTS:
1. ABSORBER
2, SUPPLY PUMP
3. RECIRCULATION PUMPS
4, SLURRY PUMP
5. TANK WITH AGITATORS
6. FLUE GAS REHEATERS
7. FLUE GAS BOOSTER FANS
* COMPONENT RELIABILITY
Figure 1. FGC system reliability evaluation.
OPTION B
0.98 0,94
njil* 0,96 0.94
— Qj — [U — HZ) —
— (U— I
(L98
HEh
0,94
— [3)—
0,94
-a-
0.94
'—CD-1
0.98 0.94
i -i r 1
0,94 0,96 0,94
STEM RELIABILITY: 65.6%
— I5J— i
i — -i
0,98
UTV^
rlJH
0.94
— m —
0,94
— Q}—
0,94
/_TTl_J
—
0,98 0,97
—IH — LiK
SYSTEM COMPONENTS:
1, ABSORBER
2, SUPPLY PUMP
3. RECIRCULATION PUMPS
4. SLURRY PUMP
5, TANK WITH AGITATORS
6. FLUE GAS REHEATERS
7. FLUE GAS BOOSTER FANS
"COMPONENT RELIABILITY
'HYPOTHETICAL DATA
Figure 2. FGC system reliability evaluation.
941
-------
0.94 * 0.96 0.94
0.94 0.96 0.94
U]—[T
SYSTEM RELIABILITY: 69.0*
OPTION C
0.98 Oi9i)
HsH
0.98
/
n_ij—
0,94
0.94
4xH
0.98 ^
-m-
0.98
/
H3_n
0.94
0.94
0.9^
0.93
0.98
H1EH
SYSTEM COMPON1
1 ABSORBER
2 SUPPLY DUMr
' RFCIRCULAT
4 SLMRRY PUMF
5 TANK '-'ITC '
Ti FLUE RAS RE
/ rLUF RAS BC
0.97
0.97
-{jLr—
'NTS:
•)
ON PUMPS
i
\GITATORS
HFATFRS
10STER FAHS
COMPONENT RELIABILITY
Figure 3. FGC system reliability evaluation,
OPTION D
QJ8* 0_,i3 OJ8
1
0,98 0,98 0,98
Tj—[¥]—TT}
SYSTEM RELIABILITY: 83.OX
0.98
0.98
0,94
0.94
/
0.94
hCH
o.->;
-CH
0,9c
SYSTEM COMPONENTS:
.1 . ABSORBER
2. SUPPLY PUMn
3. RECIRCULATION r"UMPS
4. SLURRY PUMP
5. TANK WITH AGITATORS
6, CLUE GAS REHEATERS
7. FLUE GAS BOOSTER. CANS
COMPONENT RELIABILITY
* HYPOTHETICAL DATA.
Figure 4. FGC system reliability evaluation,
942
-------
OPTION E
(L98 0_.94
0.94* 0.96 0.94
m i FI m
UJ LiJ i — \
— | 5J—
0,98
/
— Qj—
0.94
— [Tl —
n",94
0.94
I 7 1
— L2J — " — LLJ —
0.98 0.94
0.94 0.96 0.94
(~7~| rr~| pn
UJ IkJ L/J
i — CD — i
0,98
HUH
1
i — rn — i
0.94
— IT] —
0,94
0.94
HI}—1
0.98 0,94
p-TTl— _m_
0.94 0,96 0.94
1 -x 1 16 1 1 71
L2J
0,98
1
ULJ
0.94
0,Q4
— 1 3 1 —
0.94
/__JT1_.
I ^ i i j \
0.98 0.94
0,9*1 0.96 0.94
UJ LLJ Lli
SYSTEM RELIABILITY: 82.27,
i — H — i
0,98
— fsj—
j
— UJ — i
0.94
On/i
, JM
0.94
— tzi — '
0.98
H3H
c YCTC
O I O 1 C
1. AP
^. SL
3. RE
4, SL
5. TA
C r- i
6. Fl
7. GL
0.97
SYSTEM COMPONENTS:
ABSORBER
... SUPPLY PUMP
3. RECIRCULATION PUMPS
4, SLURRY PUMP
5, TANK WITH AGITATORS
FLUE CAS REHEATERS
GLUE GAS BOOSTER FANS
'COMPONENT RELIABILITY
'HYPOTHETICAL DATA.
Figure 5. FGC system reliability evaluation.
Table 1. ESTIMATED CAPITAL AND OPERATING COSTS
CONTROL
OPTION
A
B
C
D
E
GAS FLOW
RATE
ACFM
900,000
900,000
900,000
900,000
1,200,000
CAPITAL
COST,
S/KW
71.68
72,62
73,19
76,17
93,18
ANNUAL
COST,
MILLS/KWH
4.97
5,02
5,04
5,19
6.22
ANTICIPATED
SYSTEM
RELIABILITY,
64,1
65,6
69,0
83,0
82,2
EXPECTED
COST,
MILLS/KWH
8,19
8.23
7,35
6.28
7,55
PERCENT
OF TIME
FLUF GAS
NOT TREATED
]0.75
10.70
7,7
3.6
4,4
ASSUMPTIONS
1. LOST OF REPLACEMENT POWER IS 30 MILLS/KWH,
2. REPLACEMENT POWER is USED 100% OF THE TIME TO REPLACE
UP TO 707o OF CAPACITY, WHEN ACFM < 900,000,
943
-------
0 Can be used to estimate the potential operating
reliability of FGD systems.
0 Presents data for the evaluation of control system
permit applications.
0 Provides a forum to discuss air pollution control
requirements.
Utilities and Industries
0 Permits systematic evaluation of FGD process
alternatives.
0 Aids in the selection of the optimum FGD process.
0 Provides a basis for comparing requirements of FGD
systems.
0 Provide background documents on new processes.
0 Permits the transfer of experience gained by
solving operating problems at other installations.
Architects/Engineers and Environmental Consultants
0 Aids in intraprocess FGD system selection.
0 Identifies FGD process problem areas and methods
to improve process reliability.
0 Aids in preparing equipment and system specifica-
tions.
0 Provides a forum to update and transfer process
data.
944
-------
CATALYTIC/WESTVACO DESULFURIZATION PROCESS
PROTOTYPE DEMONSTRATION PROGRAM
Catalytic, Inc.
Centre Square West
Philadelphia, Pennsylvania 19102
ABSTRACT
A dry, fluidized-bed process using activated carbon for recovery
of SO from flue gas has been developed and patented by Westvaco
Corporation. The process features several improvements over other
carbon-based processes, such as minimizing carbon loss normally
attributable to chemical consumption and physical attrition, improving
heat transfer characteristics while reducing equipment size by employing
fluidized bed reactors, and generating sulfur as the sole by-product.
The technology was successfully demonstrated in an extended 20,000
CFH integral pilot run on a slipstream of an oil-fired boiler. Over
90% SO removal and virtually 100% SO removal was observed over a span
of some 300 hours on-stream. Bright-yellow, 99.9% pure elemental
sulfur was by-produced. After the test program, Westvaco selected
Catalytic, Inc. as its engineering partner with direct responsibility
for all facets of continuing commercialization: marketing, engineering,
installation and, when appropriate, operation of the process.
Since the technology has never been demonstrated in a coal-fired
situation, or when using producer gas as the chemical reductant,
Catalytic included these aspects in preparing an engineering evaluation,
design, cost estimate and schedule for a 30 MW (60,000 SCFM) prototype
demonstration program. Catalytic's evaluation supported Westvaco's
initial appraisal of process viability. Capital expenditure for the
prototype installation was estimated to be $5.9 million. An operating
budget of $1.2 million was projected for a nine-month test program.
All phases of the demonstration including design, installation, operation
and evaluation can be completed within a 2.5 to 3 year period.
Conceptual economics for a 500-MW installation (available on request)
indicated competitiveness with other regenerable FGD processes under
development. Coupled with documented technical viability, these economics
strongly suggest that the Catalytic/Westvaco process merits demonstration
on the prototype scale.
945
-------
CATALYTIC/WESTVACO DESULFURIZATION PROCESS
PROTOTYPE DEMONSTRATION PROGRAM
INTRODUCTION
"Activated carbon" is the generic term for specially processed carbonaceous
substances which possess high internal surface areas and are commonly
employed as adsorbents in air and water purification systems. Less
exploited but equally well understood is the ability of certain grades of
active carbon to adsorb and to catalyze the oxidation of sulfur dioxide
to 503. The sulfur trioxide thus formed is subsequently hydrolyzed to
sulfuric acid which remains loosely bound to the carbon until a regener-
ative operation is implemented. Historically, many approaches featuring
activated carbon for desulfurization have been explored. Although effective-
ness for SO-^- removal was usually found to be acceptable, most of the early
processes had fundamental problems, some of which were:
1) Carbon loss due to burn-off (chemical degradation) or abrasion
(attrition).
2) Carbon ignition hazard due to "hot-spots" in adsorber.
3) Sulfurous by-products, such as dilute acid, often undesirable
in the U.S.
Westvaco, a major U.S. corporation and a leader in the pulp and paper
industry, recognized the potential in exploring the desulfurization market
for active carbon, which is a major product of their specialty chemicals
group. In a conceptual development phase, Westvaco attacked the carbon loss
problem by effectively eliminating chemical burn-off, thereby allowing a
high quality (abrasion-resistant carbon) to be used. Consequently, adsorbent
make-up requirements were stabilized at an acceptably low level. The hazard
of potential spontaneous ignition of the adsorbent bed was eliminated, not
only because the carbon selected for use has a high ignition temperature,
but also because the fluid-bed method of gas/solid contact offers excellent
heat transfer properties (eliminating hot-spots) and promotes thorough
mixing of the carbon granules. Regarding by-products, Westvaco developed
a unique sulfur-producing regenerative technology. Merchant-grade sulfur is
potentially the most valuable and certainly the most convenient of all possible
end products from a desulfurization process.
The culmination of Westvaco's preliminary laboratory and pilot work was a
closed-loop, fully regenerable process comprising only three-major steps —
each one taking place in a fluidized bed. The integral pilot plant (in
Charleston, S.C.) treated a 20,000 CFH slipstream from an oil-fired boiler.
The pilot unit was operated over a period of years to obtain fundamental
data and to establish long-term operational stability. Westvaco experienced
consistently satisfactory performance in removing greater than 90% of the
incumbent SOX over a period of more than twenty cycles of the carbon (greater
947
-------
than 300 hours on stream). A comprehensive review of Westvaco's development
program and pilot activities has previously been presented.*
Recognizing the need to convert its technological expertise into a reliably
optimized process, Westvaco sought a licensee with proven competence in
the engineering and construction industry, experience in clean energy
technology, and established contacts in the electric utility industry.
Consequently, in May, 1975 Catalytic, Inc., was granted the exclusive right
to continue the technical development and assume the marketing responsibility
for Westvaco's activated carbon desulfurization system.**
As the first phase of its involvement, Catalytic developed a preliminary
engineering study and cost estimate for an unsited 30-megawatt (60,000 CFM)
prototype scrubber to desulfurize a slipstream from a coal-fired utility
boiler. Catalytic and Westvaco had agreed that the extensive pilot and
bench scale programs previously implemented by Westvaco had already fulfilled
their ultimate purposes. Further advancement of the technology would require
a demonstration of significant scale in commercially available apparatus.
In selecting the size of the prototype, overall cost must be weighed against
scale-up potential and "believability". Although a 30-megawatt basis was
chosen for Catalytic's study, th_e.. optimum capacity might be in the range
of 10 to 40 MW.
The results of Catalytic's engineering study form the theme of this presentation.
In effect, it is actually the prospectus for a proposed demonstration program
featuring the Catalytic/Westvaco technology. Although this paper describes
a potential demonstration in an electric utility station, it should be emphasized
that the subject process is equally applicable in non-utility situations; e.g.,
smelters, industrial boilers, oil refineries (CO boilers), sulfuric acid plants,
and others. Data presented here are easily convertible in terms familiar to
various industrial plant operators.
Catalytic, Inc., before accepting the major financial responsibility for
development and promotion of this process, had thoroughly reviewed its
technical aspects and carefully assessed its potential marketability. As
a result, Catalytic became convinced that it constitutes a significant
*Ball, F.J., et al., Westvaco Activated Carbon Process for SOX Recovery as
Elemental Sulfur, EPA Atlanta, Ga. conference (Nov. 1974). (Copies
available on request to Catalytic, Inc.)
**Catalytic, Inc. a wholly-owned subsidiary of Air Products and Chemicals, Inc.,
is a major U.S. engineer/constructor/plant maintenance contractor. Catalytic'
and Air Products maintain active research and development program in gas
desulfurization, coal liquefaction (SRC) and coal gasification. Further
details upon request.
948
-------
advancement in the state-of-the-art in stack gas desulfurization. The
process clearly embodies a number of advanced features and appears to be
highly competitive in capital and operating economics. Some of its out-
standing features are:
1) Process is dry — no aqueous waste streams are handled.
Corrosion problems are minimized.
2) Process operates at normal stack gas temperature — reheat is
not required. Steam plumes are not formed.
3) Raw-material make-up is minimal. Carbon loss by chemical or
physical means is effectively minimized.
4) Process is conceptually simple — only three major unit steps
are required between the raw flue gas and pure sulfur.
5) 503 is effectively removed together with SOT.
6) The process is modular and separable.
7) Elemental sulfur is the only by-product. No objectionable
substances are discharged from the system. Special treatment
of waste is unnecessary.
8)
When a coal gasifier is integrated into the process scheme
(as this paper illustrates), the system's only energy
inputs are electricity, coal and low-pressure steam.
The Catalytic/Westvaco process satisfies the generally accepted criteria
for an advanced, second-generation technology because it is purge-free,
sulfur-producing, modular, and not dependent upon natural gas, LPG or
petroleum distillates. It offers great versatility in that it can
desulfurize gaseous emissions over a wide SOX concentration range -- from
several percent to practically zero. Intrinsically simple to control, the
process responds rapidly and positively to changes in inlet gas composition.
Turndown of fifty percent can be accomplished in a single scrubber; greater
turndown can be achieved with multiple modules.
This paper presents the preliminary engineering design and associated costs
for a 30-megavatt equivalent FGD system installation and test program.
Its intent is to guide the reader through the steps necessary to establish
viable commercial technology using Westvaco's experimental results as the
starting point. Catalytic's underlying objective is to explain why the
subject process merits additional investment of time and resources.
949
-------
PROCESS SUMMARY AND SCHEMATIC
The Catalytic/Westvaco system Is designed to convert sulfur oxides
scrubbed from flue gas into merchant-grade sulfur. Activated carbon
functions as the recirculating adsorbent and catalyst to promote the
following principal reactions:
Adsorption:
so
1/2 0
SO. (adsorbed) +
Regeneration:
H2S04
3H2S
3H9 + 4S (adsorbed) =
S03 (adsorbed) . . . (1)
S03 (adsorbed) ... (2)
HS0 (weakly adsorbed) ... (3)
AS (adsorbed) + 4H20 ...(4)
3H2S + S (vapor) ... (5)
Where particulate-laden gas is to be treated, an electrostatic
precipitator is normally recommended to reduce stack opacity. The
Westvaco adsorption system is relatively insensitive to incumbent fly
ash, but it is ineffective for particulate removal.
The adsorber is a multi-stage fluidized bed which operates in the
temperature range of utility exhaust gas (about 300°F) . SO^ is removed
on the bottom stage via reaction (1), and the gas is cooled slightly by
partial humidification to enahance S02 removal in the upper stages via
reactions (2) and (3). Clean gas is exhausted at temperatures above
200°F under conditions which normally do not lead to the formation of
steam plumes. Since SOX is not condensed in the gas phase, corrosion
is not expected to be a problem.
The sulfur recovery loop, in which carbon is regenerated, is distinct
and separate from the gas scrubber. Carbon transfer is the only link
between the scrubbing and regeneration sections.
Spent carbon is restored to its original effectiveness through two
fluidized bed unit operations in which reactions (4) and (5) occur.
Reaction (5) requires an external source of reducing gas, such as hydro-
gen.
Although hydrogen can be obtained by burning natural gas, LPG or petroleum
distillates, the only fuel likely to be abundant in utility stations is
coal. Therefore, a coal gasifier is featured as the source of reducing
gas and process heat in this prototype installation.
950
-------
A schematic flow diagram of the Catalytic/Westvaco process is presented
on Page 6. This sketch illustrates the three major process units in
which carbon is continuously saturated with SO,, and subsequently
regenerated. The gas producer will be a present-generation coal gasifi-
cation unit generating a low-Etu gas. A shift converter and clean-up
train will form the interface between the gas producer and the main
process.
951 '
-------
CATALYTIC/VESTVACO FGD PROCESS SCHEMATIC
makfe- „_'--,.,
up '*•-""
carbon
flue
gas
active
carbon
loop \
V
X
5
i
i
i
i
'
i
*
SULFUR
OXIDE
SORSER
1
j^__clean
flue
gas
SO + %0 4- H 0 = H SO
2 22 24
I £1
i
;
i
i
r-
"t.
,
1
SULFURIC
ACID
CONVERTER
v?
SULFUR
STRIPPER/
H2S
GENERATOR
?1
y
^rfV^
^^teKwaprawnfijHW1*^*"^
coa
952
H SO + 3H S = 4S + 4H 0
242 2
^ elemental
sulfur
product
S = S
(sorbed) (vapor)
H + S, , = H S
2 (s) 2
Producer Gas
Ojj
GAS
V
ash
-------
TECHNICAL APPROACH - PROTOTYPE UNIT
Project Scope
The technical feasibility of the Catalytic/Westvaco process has already
been established in integral pilot plant tests. In order to obtain funda-
mental information for a detailed process scale-up and economic optimization,
a demonstration program is proposed featuring the erection and testing of
a 30-megawatt prototype (60,000 CFM measured at 60°F and 1 atmosphere).
The scope of the proposed program is to assess process performance and
to obtain engineering data in the following key areas:
1. Interface analysis
a) turndown, surges, upsets and outage in boiler
b) reliability and availability
c) fuel feed variations
d) safety
2. Control of process chemistry
a) response to upsets and variations (controllability)
b) long term stability
3. Performance of activated carbon
a) mechanical
b) chemical
c) poisoning and contamination
4. Performance of large scale fluid bed vessels
a) solids distribution
b) gas distribution
c) thermal gradients
d) ease of operation
e) materials of construction
In attaining these objectives the scope of this program includes:
1. Definition of boiler operating characteristics
2. Preparation of detailed prototype testing schedule
3. Preparation of prototype design specifications
4. Detailed engineering design and bid evaluation
5. Construction
6. Start-up
7. Operation
8. Data evaluation and process technical and economic assessment
953
-------
The prototype plant will contain the following major processing capabilities'
1. SOX removal from a fossil fuel boiler flue gas
2. Sulfur production from recovered SOX
3. Thermal stripping of sulfur product and in-situ production of
hydrogen sulfide
4. Condensation and recovery of the elemental sulfur "product
5. Production of chemical reducing gas (hydrogen) by gasification
of coal and shift conversion of carbon monoxide
As an adjunct to the prototype program, the scope of work will also
include, as necessary, additional testing in the original pilot (and
other) equipment to refine the prototype design with a view toward
optimization.
General Design Basis
The basis for the design and cost of the proposed demonstration facility
is outlined below. It is considered to be indicative of a "typical" coal-
fired utility installation and, of course, must be re-evaluated for any
specific application. This basis in no way suggests a limitation on
the range of effectiveness of the Catalytic/Westvaco Process, but merely
provides meaningful criteria for the development of technical and economic
details.
Host Boiler Slipstream Characteristics:
a. Capacity:
b. Type of Fuel:
c. Sulfur Content:
30 MW equivalent (60,000 SCFM)
Coal (Bituminous)
3.5%
Process Conditions
Sorber
1) Contact Mode:
2) Carbon Feed Rate:
3) SOX Inlet Rate:
4) S02 Removal:
SO^ Removal:
5) Temperature:
Sulfur Production
1) Contact Mode:
2) Carbon Feed Rate:
3) Sulfur Formation Rate
4) Temperature:
Gas/Solid Fluid Bed — Stagewise
11,025 Ib/hr.
1,664 Ib/hr. (about 2,800 ppm)
95% (min.) from inlet stack gas
100%
150-300°F
Fluidized Bed — Stagewise
11,025 Ib/hr. (mean)
2,337 Ib/hr.
250-325°F
954
-------
c. Sulfur Stripping/t^S Formation
1) Contact Mode:
2) Carbon Feed Rate:
3) Chemical Design Rates:
4) Temperature:
d. Sulfur Recovery
1) Type:
2) Duty:
3) Efficiency:
4) Temperature:
e. Gasifier
1) Type:
2) Feed:
3) Product Gas:
3. Active Carbon Characteristics
Fluidized Bed — Stagewise
11,025 Ib/hr. (mean)
Sulfur Stripping 700 Ib/hr.
H2S Formation 330 SCFM
750-1000°F
Condenser, Shell & Tube
700 Ib/hr. sulfur recovery
99.5% recovery of Inlet Sulfur
250-300°F
Coal Feed with Steam/Air Blast
Bituminous Coal 3,000 Ib/hr.
60 SCF Gas per Ib. coal
(max)
1) Size:
2) Density:
3) SO2 Number:
4) Attrition Number:
12x40 mesh nominal;
40 Ib/cubic foot.
75 (minimum)
30 (maximum)
1.5 mm mean particle siz>
4. Instrumentation
Adequate instrumentation is included to control and to monitor
temperature, pressure, gas and carbon flow. Additional instru-
mentation is also included in the design to assess alternative
methods of process control under the varying modes of utility
boiler operation.
Detailed Process Description
The Catalytic/Westvaco prototype unit is described in this narrative and
illustrated in the process flow diagram on Page 10. This diagram has
been simplified for the benefit of the casual reader. For those desiring
more detailed information, Engineering Flowsheets A-201 through A-204 are
included in the Appendix, accompanied by a tabular material balance. Stream
numbers and equipment designations employed throughout this paper are keyed
to the engineering flowsheets for easy reference. (Some secondary equipment
described in the narrative is not shown in the simplified flow diagram).
Flue gas leaving an electrostatic precipitator at approximately 300 F
passes through a forced draft blower (C-101) into S02 adsorber (FB-101).
The adsorber is 23 feet in diameter and contains five stages of fluidized
activated carbon. Sulfur trioxide is removed from the hot gases in the
bottom stage of the adsorber. Water sprayed into the second stage fluid
bed cools the gas to 175°F prior to. completion of 95% S02 removal in this
and in the remaining three stages. Conversion of sulfur oxides to sulfuric
955
-------
to
stack —fiS—}
CYCLONE
to boiler
>
3H2S + H2S04 = 4S + 4H20
ACID
CARBON
STORAGE
(V-102)
SULFUR GENERATOR/
ACID CONVERTER
(FB-102)
BLOWER
(C-101)
ADSORBER
(FB-101)
WATER
QUENCH
(V-1.13
STEAM HEAT EXCHANGER
SULFUR
CONDENSER
SULFUR
FILTER
REGEN.
CARBON
STORAGE
(V-101)
3H2 + 4S = 3H2S + S
CARBON
COOLER
(FB-105)
G
o
en o
Sulfur
Prod.
-------
acid, an exothermic process, reheats the flue gas to approximately 215°F.
A cy lone dust collector (F-101) removes entrained carbon dust from the
flue gas prior to its exit through the stack. Recycled activated carbon
continuously passes through the adsorber by gravity at a rate of 11,025
Ibs./hr. When discharged, it is loaded with 22.0 Ibs. ^SO^ per 100 pounds
of carbon. The H2S04 is produced by chemisorption , oxidation and hydrolysis
of the SOx in the flue gas.
Carbon leaving the scrubber is diverted to acid carbon storage silo V-102,
from which it is fed to the regeneration system at a steady rate.
Sulfuric acid-laden carbon from the silo is fed to sulfur generator acid
converter (FB-102) by a bucket elevator (M-107) and a feeder (W-103) . In
sulfur generator (FB-102), a 5'-3" diameter, 10 stage fluid bed, the
sorbed sulfuric acid is converted to elerrental sulfur by reaction with
hydrogen sulfide at 300°F. The activated carbon becomes progressively loaded
with elemental sulfur as it flows by gravity through the sulfur generator
and exits with a loading of 20 Ibs. sulfur per hundred pounds carbon. The
sulfuric acid not converted to sulfur in this reactor will be converted in
the top section of the H^S generator/sulfur stripper /preheater (FB-104) .
The elemental sulfur product is thermally stripped, and some sulfur is immediate-
ly converted to H2S in H2S generator/sulfur stripper (FB-104) , which is a
two-section unit. The 6'-7" diameter bottom section contains a total of
seven stages - - three for H2S formation and four for sulfur stripping. The
carbon is preheated in the top two stages using reducing gas from gas heater
H-106 at 1300°F, containing about 17% H2. The carbon flows from the pre-
heater section at 750°F, and bypasses the three l^S formation stages to
the four sulfur removal stages (at 1040°F) . Any sulfuric acid remaining
in the carbon after leaving the sulfur generator- will react with the reducing
gas to form sulfur and water. The off-gas from the carbon preheater passes
to a quench tank (V-113) and is cycled at 200°F by a compressor (C-lll) to the
gas heater (H-106) for reheating. Additional 1300°F reducing gas from the gas
heater (H-106) goes directly to the bottom of the ^S generator/sulfur
stripper where it contacts the carbon countercurrently to thermally strip the
sulfur. The hydrogen and sulfur are subsequently converted to H2S (necessary
for sulfuric acid reduction in the sulfur generator) . The gas leaves the
H2S generator/sulfur stripper at 1040°F and passes through cyclone (F-109)
to a shell and tube sulfur condenser to recover the elemental sulfur. The
liquid sulfur at 260-280°F is filtered of dust before storage. The cooled
gas at 260°F is effectively free of sulfur- The regeneration gas is supplied by
a coal gasifier (X-101) capable of gasifying anthracite, coke, charcoal, or
bituminous coal. The gas from X-101 passes through a shift converter (R-101)
before being used in the carbon preheater-H2S generator/sulfur stripper.
Regenerated carbon from the ^S generator at 1040°F is cooled to 300°F in
carbon cooler (FB-105) . Cooling is by evaporation of water sprayed over the
single bed of fluidized carbon in the 6'-0" diameter unit. Inert gas,
used as the fluidizing gas, passes through a cyclone (F-104) , to a water
condenser (E-103) . The gas is recycled to the carbon cooler and condensed
water is collected in a tank (V-104) for recycle.
957
-------
The regenerated and cooled carbon is then returned to storage silo (V-101)
for supply to the adsorber. Make-up carbon is added to this silo inter-
mittently as needed.
Technical Alternatives
The foregoing design is sufficiently versatile to permit "certain modifications
in the interest of expanding the demonstration program or, conversely,
reducing its overall cost. Several potential alternatives are described
below:
Gasifier - The prototype plant contains a coal fired gas producer as
a source of reducing.gas (hydrogen). This route was chosen since
most coal-fired utilities would prefer to use coal in this service
rather than fuel oil, natural gas or liquified petroleum gas.
Considerable savings in cost can be achieved by using an oil-fired
gasifier in the demonstration effort. For a prototype plant in
certain locations, fuel oil gasifiers nay be an acceptable alternative.
Gas Cooling and Tar Removal - A spray water quench system was
designed for the prototype plant in order to prevent plugging of
tubular type coolers. By this means the cooling water can be
utilized directly as part of the steam needed for shift conversion.
Shift Converter - A by-pass has been shown around the shift converter
to allow for bypassing all or part of the producer gas. It is
believed that activated carbon may "catalyze" the conversion of CO
and H20 in the production of tb • It may be possible to eliminate
the shift converter in a commercial plant. This possibility will
be checked during the test run of the prototype plant.
Flue Gas Inlet - Provision has been made in the design to bypass the
electrostatic precipitator and enter the scrubber directly with flue
gas. The design also allows for ducting scrubbed gas through the
precipitator as the last step before entering the stack. (It may
not be possible to conduct this test at some demonstration sites due
to space limitations.) The purpose of this multiple path duct system
is to test the long-range effects of flyash on the carbon system
and to find the most efficient means of eliminating traces of carbon
dust from the stack.
Equipment Arrangement
A clear area of approximately 100 x 130 feet will provide ample operating
space for the proposed Catalytic/VJestvaco prototype plant. Ideally, the
site selected should be near the power boiler gas exit duct to facilitate
tie-in. It should be unobstructed overhead since some of the carbon
elevators will reach a height of approximately 100 feet.
958
-------
It was assumed in developing this program that electric power, steam,
domestic water, boiler feed water and instrument air will be supplied by
the host utility at the battery limits of the demonstration plant.
A potential Equipment Arrangement is given in Catalytic Drawing A-101,
which follows this page.
959
-------
L A N \/\ ErW
960
-M-H
-------
-------
DEMONSTRATION PROGRAM
General Objectives
The overall intent of the proposed program is to demonstrate the Catalytic/
Westvaco Process on a coal-fired boiler slipstream. Prior to actual
engineering of the test unit, a comprehensive operating program and overall
process specifications for the unit will be developed. These will be
used by Catalytic as a basis for detailed design of the system. Prior to
start-up, Catalytic will train technicians and operators to perform "hands-
on" duties during the start-up, test and demonstration periods. Data
reduction.evaluation and process assessment will be performed by Catalytic
with appropriate assistance, by Westvaco.
As an adjunct to the prototype design, tests will be made to evaluate
proposed control modes, operating ranges and design features of the
prototype unit. Input: from these evaluations will be used to modify and to
optimize the process design.
Program I.lcTieuts and Schedule
The operating schedule for the prototype plant calls for a two-month
period of start-up, a six-month test program and an optional six-month
demonstration run. The scheduled activities are presented in chart form
on Page 16.
The start-up period will be used to place all units individually and jointly
into operation, checking their operability over specified temperature,
flow and pressure ranges, and to make any required adjustments or modifications
At the conclusion of the start-up period, the plant will be ready for an
extended continuous test run accepting flue gas and circulating carbon through
the adsorption and regeneration equipment on an integrated basis.
The initial operating period will be devoted primarily to establishing the
operating characteristics of the process. A material and energy balance
will be obtained and compared with calculated values. The S02 removal
capability per carbon cycle, as a function of the number of cycles, will
be det.erii.ined. The process control characteristics, particularly stability
and turndown capability, will be verified. These data will be analyzed
as they are obtained. At the conclusion of the initial operating period,
the data will be reviewed and any changes needed in the program for demon-
stration operation will be made.
The schedule chart presented on Page 16 contains a six month
demonstration run (Item 6-d) which has not been included in the operatin^
budget presented on Pages 18 - 20. It is expected that the budgeted test
program (Item 6-c) will provide the fundamental data required to assess the
viability of the process. An additional run may be desirable to establish
industry-wide confidence in the technology.
962
-------
WESTVACO S02 REMOVAL
PROTOTYPE PHO;;RAM SCHEDULE
f5S3S!^ffl[^SS2B^W2J®ES5Sa«E3EM3Bia3SVEi!S^^
!
1 1. PREPAEE DETAILED TEST PROGRA.M.
I
2. PREPARE PROCESS DESIGN
P SPECIFICATIONS.
1 3. DETAILED EN, R. DESIGN AND BIDS .-
4. PROCUREMENT AND CONSTRUCTION .
5. DEFINITION OF BOILER OPR.
CHARACTERISTICS
6. OPERATION
a. OPERATOR TRAINING ... . .
1
1
1 b. START-UP
! c. TEST PROGRAM
3
I r\ T)V f'lOM^ TK* A TTOTvl RTfM ^
7. DATA EVALUATION AND PROCESS
ASSESSMENT
?
9
MONTHS I
03 6 9 11 15 10 41 14- 11 .3033 \
ESS
!!3£2S!
StrtSSCSKI
fajzzxzsF—iA
EC.
-tKii^rasi
»««
S^il-i^sUSbs
tS£3S3^
-- - -
•^^^5
-
f Onf
V VJJ L
Hfy^KUfBT^fTf
-
1 •
I
1
|
|
i
!
i
i
!
1
!
s
I
i
"'""
^»H«J
cr\
* Deuionslra Lioa Run has tioL been considered in tlie economic asstissmciiit ot Lhc cosLs to impieivient trjis prop.ram.
-------
DEMONSTRATION PROJECT COSTS
Estimating Bases
The estimate presented in this section reflects all the costs associated
with designing, constructing and operating the subject prototype/
demonstration module of the Catalytic/Westvaco Flue Gas Desulfurization
Process. The estimate as presented is for an undetermined but "typical"
site. The equipment costs are based specifically upon the following
attached drawings: A-101 (Equipment Arrangement) and A-202 through A-204
(Engineering Flow Diagrams).
The project completion schedule to which the estimate is tied is predicated
upon starting engineering on 1 May 1976, beginning construction on 1 January
1977 and completing construction on 30 October 1977. The estimate is
based on December, 1975 dollars. Appropriate escalation to carry through
the construction period is included in the calculations.
The estimate is an order-of-magnitude type, based upon major equipment
costs. Information for the major equipment items was obtained by several
methods. In the case of the fluid bed towers (FB-101, 102, 104 and 105),
two prospective vendors were each asked to quote on designs and fabrica-
tions. On items of mechanical equipment (pumps, compressors, blowers
and miscellaneous conveyors), vendors with whom Catalytic has continuous
contact submitted estimates based upon Catalytic equipment data sheets.
Items such as heat exchangers and vessels were estimated in-house using
recent quotations and experience. The gasifier package, X-101, is based
upon the Wellman-Galusha Agitated design, 6.5 feet in diameter. Vendor
pricing was obtained for approximately 90% of the major equipment costs.
Estimates for other materials (piping, instrumentation, electrical, etc.)
were calculated by one of two methods. Whenever possible, engineering
estimates were made based upon preliminary takeoffs. This method was
used for piping, instrumentation, electrical, concrete, steel, insulation,
painting, chemicals and catalyst, and ductwork. Where such a technique
was not feasible, allowances based upon experience were, made (as in the
case of sewers, site development and fire protection). The buildings
were, priced on a per-square-foot basis.
As previously stated, the capital cost estimate is for an undetermined
site. An hourly craft labor rate of $10.00, including fringes, was used
in calculating labor costs, and a payroll burden of 12% was assumed.
Productivity was assumed to be comparable to Gulf Coast standards. When
a specific site is selected, these factors plus items such as sales or
use taxes must be re-evaluated.
Catalytic's Home Office Service includes detailed Engineering and follow-
up support throughout construction. Costs for these tasks were prepared
as follows: Each engineering Division Chief Engineer used his experience
and judgment to estimate the manpower requirements necessary to fulfill
his discipline's responsibilities efficiently and on schedule.
964
-------
Installed Cost
The installed cost for a 30-negawatt demonstration module of the
Catalytic/Westvaco FGD Process has been estimated at $5.85 million.
This cost reflects process design capability for continuous SO,, removal
in excess of 95% when raw flue gas issues from a coal-fired boiler
fitted with an electrostatic precipitator. The inlet SOX concentration
corresponds to a 3.5% sulfur coal.
A breakdown of the installed cost estimate is presented below. These
figures take into account the nature of the proposed installation as
a demonstration plant, containing considerable reserve capacity and
fully adaptable for a rigorous experimental program. Therefore, it
would be inaccurate to scale the numbers in order to estimate the costs
of a commercial installation. Catalytic will provide, on request,
realistic costs for large-scale commercial svstems.
PRELIMINARY _IXSTALLED COST ESTIMATE
Material, subcontracts, shop labor $4,050,000
Total field charges and payroll burden 666,000
Total Hone Office charges 600,000
Contingencies and escalation 336,000
Overhead and Fee 198,000
Installed Cost $5,850,000
Operating Costs
A summary of operating costs for the proposed demonstration program
is presented below. Of the $1.24 million project budget, 46% is allocated
for utilities and chemicals, with 21% for operating labor. Maintenance
is estimated at 4.5% of the installed cost (on an annual basis) and
plant overhead' is 20% of the sum of items (1), (2) and (3) below. The
operating labor force includes 31?: operators per shift (average of 4 shifts
per day), a Plant Superintendent, Lab Technician, Secretary/Clerk, Develop-
mental Process Engineer and Demonstration Program Manager. A summary
of utilities and chemicals requirements are given on Page 20.
Since the proposed program is primarily for testing and demonstration,
the attached operating costs have been developed on a conservative basis,
including significant reserves of utilities and chemicals and an expanded
operating labor roll. Catalytic will supply, on request, operating data
applicable to large-scale commercial installations.
965
-------
SUMMARY OF OPERATING COSTS
Direct Costs
(1) Operating Labor, Payroll Burden and Training $ 262,000
(2) Maintenance - Material, Labor and Supplies 150,000
(3) Utilities and Chemicals 566,000
(Not including initial charges)
Indirect Costs
(4) Plant Overhead (Approximate) 195,600
$1,173,700
Travel 20,000
Laboratory Analyses 20,000
Computer Usage 5,000
Miscellaneous 20,400
Total Operating Program* $1,239,000
* Includes: Operator Training (1 month)
Start-up (2 months)
Program (6 months)
Not Included: Demonstration Run (Optional) (6 months)
966
-------
Utilities and Chemicals
Basis: 6 month program 4- 2
Total - 8 months at
Item
Electricity
Coal
Steani
Dcr"ir, .
Wa t e r
Instrument
Ail-
Cooling
Water
Activated
Carbon
Catalyst
Rate of
Usage
1800 KWh/Hr.
1.5T/Hr.
2900 tf/hr.
38.4 GPM
100 cfm
340 GPM
20?:-'/hr.
month start-up:
80% utilization
Rate of
Cost
0.025/Kwh
$25/T
$3.00/M Ib.
$3.00/M Gal.
0.05/M scf
0.10/M Gal.
$0.50/Jb.
= 4675 hours
Total-
Usage
8,415,000 KWh
7,000 T
13,550,000 Ib.
10,800,000 Gal.
28,000,000 scf
95,300,000 Gal.
93,500 Ibs.
Total
Cost
$210,400
175,000
41,000
32,000
14,000
9,000
46,700
5,500
Carbon & Catalyst loss due to Cleaning,
Maintenance and spills during startup
and tests - 90,000 Ibs.
Total Utilities and Chemicals
Based on x:675 hours
45,000
$ 566,000
Allocation: Start-up (2 mo.) = $1/0,000
Program (6 mo.) - $396,000
967
-------
CONCLUSIONS AND RECOMMENDATIONS
Among various techniques for controlling sulfur oxides emission, post-
combustion gas scrubbing is expected to predominate during the 1980's.
Regenerable FGD processes producing useful effluents, such as sulfur,
should be in greatest demand. The Catalytic/Westvaco system shows great
promise of competing successfully both technically and economically in
this market. Full-scale commercialization of the technology is premature
at this time, however, since additional data for scale-up and optimization
purposes is still needed.
Clearly, a prototype demonstration facility — large enough to supply
reliable information yet small enough to be "affordable" — is the next
logical step. This prototype plant should be equipped to provide meaning-
ful test results for Catalytic while serving as a representative and
believable demonstration for the Utilities and for Industry. To this end,
Catalytic welcomes cements and suggestions relating to the process or
to the proposed demonstration program. In the same light, we will provide
any information at our disposal to facilitate the analysis or evaluation
of the technology. Catalytic's his_tory is one of service to the chemical
process industry. The organization has built a firm reputation for
"making things work". The proposed demonstration project is viewed in
this spirit.
ACKNOWLEDGEMENT
Much of the Westvaco's experimental data was developed under a contract
partially funded by the U.S. Environmental Protection Agency, with Mr.
Leon Stankus acting as project officer. EPA neither approves nor
disapproves of the Catalytic/Westvaco FGD process.
968
-------
APPENDIX
The Appendix contains Catalytic1s Engineering Flowsheets A-201 through
A-204, which are followed by a tabular Material Balance. This information
supplements the Detailed Process Description presented on Pages 9 through
12 of this report. Item designations and stream numbers presented on
the Drawings are keyed to the Material Balance and to the preceding
•narrative.
969
-------
C«3 Ult tiltt
>C U.ME '*'t
era uu. u«.«t
f\J UCU t*L>E
[7]
ff f Bn*ni* i F
Q Sl'SJ"
0 W3TWOT i
1X3 MIK.E Mi-t
GO «-'•*
p CttTVl »«.*£
-tT- F« ,«a».)
f.e. F.C (Fin a»»»l
_j^ r£>^rim
-CL* "VJ££>r*
_^1^ (utT2i2nsr **LA7f
Tmn-«iT *LMII »*irt
FUM «rWI VJtTt
txi ut IUL «
CM H*L OJltB
CX MIT i*3;ur««i win
A
IT piEwfcitiM
-Q- n«i ejn
©
^7
^TTi
O«ir»3«r
(UKJL waits)
Y
MEIWC MIKMBT
•
—((3 V-<«-<>«©
KCTi^i '»-*•
» L*uri*«.
970
'
/\
<&
Al
•• i *
• " " ' " nC-
I
5?S
1AI j
1/3 i
l/^
3
I
nil. L^"r
Tgrry-.--?
Al
Ai~
r i « i
4 1
-------
\, f f ^ ^ ^K|( uj|jii
-'"'""""'
O
« i Hi Iffn lJ.*fl 9f rK |
CAT>LY'*C ISC
-j A-201
-------
*?.H^5L - ~t±:-_=i.* ^?Jt i T
'VL'X.a ^y^^ cr-.;y, -^ j I „
j[ -KT^_ T
|FHO«tl«~1 !*">* V-" A= f> ,—, I
972
£^_
-------
T^ZI^zL*" 2crf
-------
i^u!K-T%~- ?-—
r^J?-7^!'
-v ty'J ±2:
,S, __ /ITf
^T" T ~^«>
974
ZS^_
A_L
j£L
-------
R£_v;.tC *-^Ki^.cN ^n>F T -'>i*r»"re_H__ 5t-t^f u »-T\>q i11^ '-*1*-
io.r_«. s c^qt
-^?»-
^
T rj / : ' >
~ V T T'
/ft** 'y?s
\^
^"
H-,«.
1
975
STATdS AS 3f 10-2-75
C*^»V^J tOt »^O^Si
-203
-------
976
-------
-------
SOLIDS STREAMS - #/Hr .
ALL GAS STREAMS IN LB . MOLES /HR.
!
t£>
00
i
i
COMPONENT
1
S02 M.W. 64
H2 2.016
K?S 34
NH3 17
C02 44
CO 28
JN2 28
02 32.
CH4 16
H20 18
C2H6 30
C 12
Resld S 32
H2S04 98
S 32
TOTAL #/»r. or lb/mo]es
hr.
Temperature °F
Pressure PSIA
SCFM-GPM
ACFM
Density Ibs/cu.ft.
<7>
26.0
1174.0
4.7
6727.4
414.9
954.4
' 9301.4
300°
14.7
59,994
86,036
0.061
-
1.3
1174.0
4.6
6727.4
394.2
1392
9693.5
216
14.7
62,523
79,746
0.058
Q
10.5
.5
11.0
216
40
G)
i
11,035.5
441.5
11,477
40
u^
266
428
11,025
441
2,426
14,586
300
40
Q
8751
8751
160
17.5
61
6 ,-"
52.15
62.87
5.86
118.11
5.46
229.69
.99
475.13
260
18.0
3065
4164
0.049
, — 3
;•
3.71
0.16
62.87
5.86
118.11
5.46
322.93
.99
520.09
300
ATM
3355
4811
0.043
,,
SCFM = 70°F
-------
( j ALL SOLIDS STREAMS - #/HR.
_ '">ALL GAS STREAMS IN LB. MOLES/HR.
COMPONENT I
S02 M.W. 64 j
H2 2.016 |
! H,,S 34 I
t z !
NH3 17
\ C02 44 |
CO 28 1
| N2 28 j
02 32 j
i"" " ' '
i H20 18 j
C 12
Res Id S 32
H2S04 98
S 32
I C-;Il6 30
I
1 TOTAL #/IIr. or Lb/molo
i'nr .
P Temperature °F
-i
1 Pressure PSIA
I SCFM-GPM
a
| ACFM
| Dt.Misi ty Ibs/cu . f t .
i , , <
(IIB)
35.36 !
.47
31.99
3.02
60.86
2.81
112.36
0.51
I
';/ 247.38
1217
17.0
1596
5050
0.0177
"0 |
325
64
11,025
441
364
2,218
i 14,437
[~ 300
40
(7o)
,
\
325
1
1
11,025
441
2,337
14,128
751
40
,;iu- | ,'u •
f — '
70.01 jj 24.22
.92 I 0.47
J
62.08 jj 31.99
5.86 | 3.02
«
118.11 \ 60.86
I
|
!
5.46 ! 2.81
210.91 130.75
|
I
... |
0.99 ! 0.51
J
474.34 254.63
"t
i
1217 f 500
18.8 | 15.2
3059 1 1642
9679 I 2974
0.0177 ) 0.0313
13 ""'
103.37
1.39
94 .07
8.88
178.97
8.27
323.27
1.50
721.72
1217
18.8
4655
14729
0.0177
'" 14 ">
81.15
.92
62.08
5.86
118.11
5.^6
152.87
0.99
427.44
823'
21.8
2757
6674
0.0231
' , " •'
33.62
.92
14.55
53.39 1
118.11 1
5.46
86.06
0.99
313.10
I 534
24.7
2020
3788
0.0366
1
-------
,) ALL SOLIDS STREAMS - #/Hr.
;- ALL GAS STREAMS IN LB. MOLES /HR.
to
no
O
COMPONENT
S02 M.W. 64
H2 2.016
H2S 34
Nil 3 17
C02 44
CO 28
N2 28
02 32
H2 2
CII4 16
H20 18
C2H6 30
C 12
Resid S 32
H2S04 98
S 240
TOTAL #/Hr. or Ib/moli
hr.
Temperature °F
Pressure PSIA
SCFM-GPM
ACFM
Density Ibs/cu.ft.
^ 16 >
;
1
52.15
62.87
5.86
118.11
5.46
229.69
0.99
2.91
s/ 478.04
1040
15.1
3083
8725
0.0248
©
I
11,015.5
441
11,456.5
1040
40
c- IS^P
1
i
0.31 I
121.17
11,29
227.64
10.52
177.39
1.91
I
550.23
160
16.1
3549
4152
0.062
r_~~*~~j
V19> |
0.31 |
j
I
121.17 )
11.29 !
227.64
10.52
275.22
1.91
648.06
300
15.5
4180
5994
0.0479
~@ |
i
ij
t
|
1
!
!
11,015.5
441
11,456.5
300
40
-' — \ """N
< 21) &( 22
v J V .-/
1761
1761
160
3.52
61
./\
<^23j^:
24.22
0.47 |
!
31.99 1
3.02
60.86
2.81
170.40
0.51
294.28
200
15.0
1898
2364
0.045
<^24 ">
j
33.62
0.92
g
14.55
53.39
118.11
5.46
200.40
0.99
427.44
600
22.9
2757
5514
0.028
-------
THE SPRING-NOBEL HOECHST PROCESS FOR
SULFUR DIOXIDE RECOVERY FROM STACK GASES
W. H. Stark, H. A. Syme and J. C. H. Chu
Spring Chemicals Limited
Toronto, Canada
ABSTRACT
Nobel Hoechst Chimie of France and Spring Chemicals Limited of
Canada have developed a regenerative process for flue gas desulfurization.
The process uses an organic absorbent which is highly selective, possesses
excellent stability and has a high capacity for absorption in the lower
temperature ranges. It desorbs readily at higher temperatures producing
a stream of SO at a concentration in excess of 95% for liquefaction
or conversion to sulfur or sulfuric acid.
A description of the process and flow diagram are presented. Capital
and operating cost estimates for this process have been made based on
the procedures used in EPA's January 1975 Detailed Cost Estimates for
Advanced Effluent Desulfurization Processes. The resulting data are
compared with the data given in the EPA report for alternative processes.
Direct operating costs are also given for a representative range of SO
concentrations in the gas to process for other applications.
The Spring-Nobel Hoechst Process has application to flue gas
concentrations of SO ranging from power plant concentrations up to
those levels which permit direct feed to sulfuric acid plants or conversion
to elemental sulfur. It is being offered for recovery of SO in stack
gases from smelters, Glaus plants and sulfuric acid plants among others
and is ready for trials on power plant flue gas desulfurization.
981
-------
THE SPRING-NOBEL HOECHST PROCESS FOR
SULFUR DIOXIDE RECOVERY FROM STACK GASES
INTRODUCTION
Nobel Hoechst Chimie of France and Spring Chemicals Limited of Canada have
developed a regenerative process for flue gas desulfurization. An aqueous
solution of glyoxylic acid is used to absorb S02- The absorbent is highly
selective, possesses excellent stability and has a high capacity for absorption
in the lower temperature ranges. It desorbs readily at higher temperatures
producing a stream of SO at a concentration in excess of 95% for liquefaction
or conversion to sulfur or sulfuric acid.
The absorbent is produced by Nobel Hoechst Chimie of Puteaux, France who have
developed the basic process data and obtained patents on the use of glyoxylic
acid as an S0? absorbent. Spring Chemicals Limited have continued the process
development and engineering including economic studies for a wide range of
plant sizes and S0? concentrations in the gas to process. Spring, the exclusive
world-wide licensee, is offering the process for flue gas desulfurization.
DESCRIPTION OF THE PROCESS
Figure 1 is a flow diagram of a typical application of the process. Gases enter
the system through a direct contact cooler to bring the gases to the absorption
temperature of approximately 35 C (95 F). Cooled gases enter the base of the
tower where water is separated from the gases. For large applications such as
coal burning power boilers where fly ash must be considered, the water flows
to a sludge separation system and then to a pond or tower for cooling and
returns to the direct contact cooler. For smaller applications and cleaner
gases an indirect water cooled heat exchanger is a suitable alternative to the
pond or tower. In either case the purge and water make up flows control the
water balance of the direct contact cooling system.
The cooled gases pass upward through the required number of absorption stages of
the tower where they contact a counter current flow of an aqueous solution of
glyoxylic acid introduced to the top stage. The absorbent leaves the bottom
982
-------
<£>
CO
Gas Out
1
Gas In
^ c
^ ,
JJ
Water
Purge
Absorption
Tower
Sulfur
Dioxide
Direct
Contact
Cooler
Condensate
FIGURE 1
THE SPRING-NOBEL HOECHST PROCESS
-------
absorption stage rich in SO . Circulation pumps for each stage provide the
proper ratio of liquid to gas flow. The absorption conditions are maintained
within the range necessary to meet the S0? emission level which has been
prescribed for the exit gases. These gases are reheated if required before
discharge to atmosphere.
Absorbent rich in SO from the bottom absorption stage is pumped to the upper
section of the stripping column which is heated with indirect steam. Sulfur
dioxide is removed as the absorbent passes down the column. The sulfur dioxide
product exits the condenser at the top of the column at a concentration of about
955? SO and 55? HO, dependent on plant cooling water temperatures.
The stripped absorbent from the column is pumped to the absorption tower top
stage for reuse. A heat exchanger transfers sensible heat from the stripped
absorbent to the absorbent entering the stripping column, reducing the heat
input required at the column. A small heat exchanger on the absorbent feed
line to the top stage provides final cooling which includes compensation for
the heat of absorption to maintain the design operating temperature in the
absorber. A tank incorporated into the system acts as a surge tank. A storage
tank for absorbent make up is also included.
THE ABSORBENT
Three very important characteristics of this absorbent are the high degree of
selectivity, the excellent stability under process conditions and the fact that
there has been no evidence of sulfuric acid formation with this absorbent.
Glyoxylic acid is marketed world-wide for use in the pharmaceutical and chemical
industries. It is an aldehydic acid. The formula of the monohydrate is
•"*" OH
HOOC-CH . The product has a number of uses in the chemical industry for
^" r\tj
manufacture of Pharmaceuticals, cosmetics, plastics, rubber, textile, paper,
dye and coatings, as well as others. It is marketed as a 505? aqueous solution.
984
-------
PROCESS ECONOMICS
Cost estimates have been made for a number of different applications of this
process. The design bases for these estimates have covered a wide range of
feed gas rate and SO concentrations. The capital and operating costs for the
gas cooling and absorption system are a function of the gas to process, its
temperature, flow rate and SO content to and from the absorber. On the other
hand, the desorption system costs are related to the quantity of S0? to be
processed and to the S0_ concentrations in the gas to and from the process.
Accordingly a specific case has been selected for the preparation of comparative
data on power plant flue gas desulfurization; the second case reflects the
effect of SOp concentration on operating costs for other applications.
Power Plants
The U.S. Environmental Protection Agency has made and published an in-depth
evaluation of Flue Gas Desulfurization Processes. The published data are
detailed and very well presented, permitting a good understanding of the design
basis and costing methods used. Spring has selected the Base Case in the
reference, a 500 mgw new boiler burning coal containing 3.5% sulfur as a basis
for developing comparative capital and operating costs for the Spring-Nobel
Hoechst Process. In developing these data, Spring has adhered to the design
parameters and the scope of work as defined in the Base Case.
In the reference Base Case neoprene lined carbon steel was used for the contact
cooler and absorber. For comparative costing Spring has followed the practice
employing neoprene lined carbon steel shells and 316L stainless steel internals
for the cooling and absorption equipment and all 316L stainless steel desorbers.
At present the well established material of construction for the process
equipment is 316L stainless steel. Its use for all material in contact with
process streams would increase the investment for the Spring-Nobel Hoechst Process
in this study by about $2,000,000.
The capital estimates for cooling, absorption and desorption equipment prepared
by Spring for this comparison are order of magnitude costs which fall within the
range of +25 to -15% accuracy established in the reference Base Case study.
Operating costs prepared by Spring are expected performance values.
985
-------
For ease of comparison, these data and the reference EPA data have been summarized
in two tables. The first table presents capital and operating costs for the
complete facility through to an H0SO. product as defined in the EPA reference.
2 4
The second table provides a more direct comparison of the regenerative processes
by excluding the costs downstream of SO desorption; the costs of the Lime Slurry
and- Limestone Processes through to disposable sludge are shown again for
information.
Table 1 summarizes the capital investment and operating costs for the Spring-Nobel
Hoechst Process including the sulfuric acid plant at the EPA report costs and
compares these combined costs to the costs for the Magnesium Slurry and
Catalytic Oxidization Processes producing H0SO , the Sodium Solution Process
^ 4
which produces elemental sulfur and the Limestone and Lime Slurry Processes
which yield a disposable sludge. The capital investment of $26,4/70,000 for the
Spring-Nobel Hoechst Process compares favourably to the other processes as do
the operating costs of 2.283 mills per kwh compared to 2.20 to 2.31 for the
disposable sludge processes and to 2.63 to 3.31 mills per kwh for the other
regenerative processes.
Table 2 summarizes the capital investment and operating costs for the process
through to the intermediate product S0? compared to the same point in the
Magnesium Slurry Process and the Sodium Solution Process where SO is produced
as an intermediate process stream and to the Limestone and Lime Slurry Processes
which produce a disposable sludge. The capital investment of $20,060,000 for the
Spring-Nobel Hoechst Process compares favourably with the other processes. The
operating costs including fixed charges totalling $6,199,000 per year, or 1.767
mills per kwh, are substantially lower than the operating charges for flue gas
desulfurization for the other processes which range from $7,703>000 to
$9,580,000 per year.
Other Stack Gas Desulfurization Applications
The process has application to stack gases from smelters, Glaus Plants, sulfuric
acid plants and other industrial operations. The engineering evaluation of a
number of these applications provides data which illustrates the effect of S09
concentrations in the feed gas on the direct operating costs with this process.
986
-------
TABLE 1
FLUE GAS DESULFURIZATION SYSTEMS FOR POWER BOILERS
THROUGH TO H.SO., SULFUR OR SLUDGE1
2 4
PROCESS
Spring-Nobel Hoechst
Limestone Slurry
Lime Slurry
Magnesium Slurry
Sodium Solution
I
]Catalytic Oxidation
2
DOLLARS/TON
COAL BURNED
6.09
5.87
6.17
7.02
8.84
6.76
DOLLARS/TON
SULFUR REMOVED
221.78
214.68
225.81
255.43
323.34
247.32
I UUOiO
DOLLARS/TON
100$ H SO.
1 ' 2 H"
72.43
-
83.43
354.79
80.75
CENTS/106
BTU INPUT
25.38
24.45
25.72
29.24
36.83
28.17
MILLS/KWH
2.283
2.200
2.310
2.630
3.310
2.540
CAPITAL
DOLLARS INVESTED
PER KV/
52.94
50.33
44.84
52.81
60.98
85.47
END
PRODUCT
H2S°4
Sludge
Sludge
H0SO . ,
*L 4
Sulfur
H0SO,
t- 4
BASIS:
(80%)
U.S. Environmental Protection Agency data from Report EPA-600/2-75-006
500 mw new coal-fired power unit, 3-5$ S in fuel; 90% removal;
15.8 tons/hr 100% H SO •
No credit taken for recovered product.
Stack gas reheat to 79°C (175°F).
Midwest plant location represents project beginning mid 1972,
ending mid 1975, average capital cost basis mid 1974.
Minimum in process storage, only pumps spared.
Closed loop water utilization for gas cooling stage.
Investment for disposal of fly ash excluded.
Construction labour shortages with accompanying overtime pay incentive not included.
Operating costs are projected 1975 values.
2. Spring estimates
-------
PROCESS
p
Spring-Nobel Hoechst
Magnesium Slurry
Sodium Solution
Limestone Slurry
Lime Slurry
(JD
00
oo
TABLE 2
FLUE GAS DESULFURIZATION SYSTEMS FOR POWER BOILERS
THROUGH TO SCL OR SLUDGE1
COSTS
$000 ' S
20,060
18,974
23, 590
25,163
22,422
OPERATING
COSTS $000 's
6,199
7,528
9,580
7,703
8,102
DOLLARS/TON
COAL BURNED
4.71
5.73
7.30
5.87
6.17
DOLLARS/TON
SULFUR REMOVED
171.71
208.77
267.00
214.68
225.81
UWO1O -r —
CENTS/10
BTU INPUT
19.61
23.90
30.41
24.45
25.72
MILLS/KWH
1.767
2.149
2.733
2.200
2.310
O-tt-T _L_L.H_U
DOLLARS
INVESTED/KW
40.12
37.95
47.18
50.23
44.84
END
PRODUCT
so2
S°2
S°2
Sludge
Sludge
BASIS:
1. U.S. Environmental Protection Agency data from Report EPA-600/2-75-006
500 mw new coal-fired power unit, 3.5% in fuel; 90% removal; 10.3 tons/hr SO .
No credit taken for recovered product.
Stack gas reheat to 79°C(l75°F)
Midwest plant location represents project beginning mid 1972, ending
mid 1975, average capital cost basis mid 1974.
Minimum in process storage, only pumps spared.
Closed loop water utilization for gas cooling stage.
Investment for disposal of fly ash excluded.
Construction labour shortages with accompanying over time pay incentives
not considered.
Operating costs are projected 1975 values.
2. Spring Estimates
-------
The design basis selected is 73,000 acfm of gas to process at 400 F at the SO
concentrations shown and an SO emission of less than 200 ppm by volume, dry
basis. The operating costs include the induced draft gas fan power requirement,
gas cooling and all operations through delivery of product SO . Fixed charges
and capital related costs are excluded.
S0_ in gas to process,
percent by volume, dry basis
S02 Recovered, percent
SOp Recovered, Tons per day
Power @ $0.009 per kwh
Steam @ $2.00 per M Ib
Cooling Water @ $0.02 per M gal
Absorbent Process Losses
Cost per Short Ton of SO Recovered
0.2
90
10
$ 7.34
16.50
0.10
1.10
1.0
98
50
$2.13
6.52
0.12
0.55
2.0
99.8
100
$1.36
5.32
0.13
0.55
TOTAL
$25.04
1.32
$7.31
The process system is well instrumented to ensure stable operation with minimal
labor requirements. It is essentially free of purge streams other than the
discharge of the particulate scrubbed from the feed gas. The high concentration
of the SO product permits efficient conversion to liquid SO , elemental sulfur
or sulfuric acid.
Reference 1 EPA-600/2-75-006, January 1975 publication of the Office of Research
and Development U.S. Environmental Protection Agency, entitled
"Detailed Cost Estimates for Advanced Effluent Desulfurization
Processes". gg9
-------
STARTUP OF AMERICAN AIR FILTER'S
SULFUR DIOXIDE REMOVAL SYSTEM AT THE
KENTUCKY UTILITIES COMPANY'S GREEN RIVER STATION
Albert H. Berst, Manager, Scrubber Projects Engineering
American Air Filter Company, Inc.
Louisville, Kentucky 40201
Jack Reisinger, Plant Superintendent
Green River Station
Kentucky Utilities Company
Central City, Kentucky
991
-------
I. INTRODUCTION
The sulfur dioxide and particulate collection system for generating units
#1 and 2, and boilers #1, 2, and 3, (Figure 1), at the Kentucky Utilities
Green River Station consists of a tail end lime scrubbing system. A turnkey
contract for this project was awarded to American Air Filter in June, 1973,
and startup commenced September, 1975. Generating capacity of the combined
units is 64 MW, with fuel being primarily high sulfur Western Kentucky coal.
The system consists of one scrubber module to handle a maximum of 360,000 ACFM
at 300°F. The blow through scrubber module, see photo, contains an adjustable
throat flooded approach venturi for flyash removal, a mobile bed contactor
for SO removal, and centrifugal demister. Pebble lime storage, slaking
facilities, and pumping system provide reagent slurry which is pumped to the
three-compartment reactant tank. Bleed slurry from the tank is pumped to a
settling pond, and clear water from the pond is to be returned to the system
for closed loop operation. This report depicts events through October 9, 1975.
II. SYSTEM DESCRIPTION
The two generating units are used for peak loads; therefore, the boilers
supplying steam to the generators are normally run only five days per week,
with one or more of the boilers often at reduced capacity. Wide fluctuations
in gas flow to the scrubbing system, as well as system shutdown every Friday
afternoon and startup every Monday morning, had to be considered in the
design.
The flue gas is drawn from the existing breeching through a guillotine type
isolation damper and duct system to the scrubber fan. From the outlet of
the scrubber fan, the gases flow through a venturi scrubber which was provided
primarily to remove the particulate escaping the existing mechanical collectors
The gases then flow upward through the mobile bed contactor, where the slurry
reacts with the SO.. The contactor bed consists of ten separate compartmental-
ized sections with overhead slurry sprays and below-bed dampers to accommodate
gas volume turndown requirements.
The slurry/recycle system, Figure 1, consists of a reactant tank with recycle
pumps supplying the contactor and venturi, a lime slurry slaking and feed
system, and a bleed system discharging to a large pond with return water from
the pond to be used as primary make-up. Spent slurry exits the scrubber into
the return section of the reactant tank, where make-up water and fresh reagent
are added. The second and third compartments of the tank provide for comple-
tion of the chemical reaction. Agitators in all three compartments maintain
992
-------
solid suspension. Rubber-lined recycle pumps convey the screened water from
the third (feed) compartment back to the venturi scrubber and mobile bed
contactor.
The lime slaking system consists of a 500-ton capacity pebble lime storage
bin which is filled by pneumatically unloading lime received by rail. The
lime storage bin has a vibrating bottom and screw feeder, which feed the
pebble lime into an agitated slaking tank. From the slaker, slurry is dis-
charged into a large agitated hold. Reactant pumps feed the reactant as
approximately 20% slurry to the return section of the reactant tank in res-
ponse to pH sensing instrumentation.
The bleed pumps remove reaction products and collected flyash from the
reactant tank to maintain an approximate 10% slurry within the tanks by
pumping slurry to the pond where the solids settle, and clear water is
returned from the pond by pumps to the return section of the reactant tank
to maintain the proper liquid level.
Treated water (cleaned river water) for make-up is brought into the system
through the pump gland seals, the slaking operation and through the demister
wash nozzles.
Extensive instrumentation is provided to match gas flow from the boilers by
a pressure sensing loop in the ductwork to the main scrubber fan, to control
the number of bed dampers that are open, to control the system pH, to control
the venturi pressure drop and to monitor other important functions.
It will not be the purpose of this paper to discuss the design considerations
of the system nor the chemical aspects of the startup, as these considera-
tions will be covered in a separate paper. It will also not be the object
of this paper to discuss the methods of training and operator participation
as this is also to be covered in another paper. The primary purpose of this
paper is to discuss the philosophy of the startup and some of the problems
encountered.
III. PHILOSOPHY AND PLANNING FOR STARTUP
Experience had shown that careful preparation must be made for startup to
prevent the chemistry episodes experienced with other SO- systems. Hence,
startup was to be divided into four phases. The first phase would consist
of mechanical and electrical debugging, which would include running the
agitators and pumps.
993
-------
The second phase would include operation of the unit on air and water only.
During this operation, various gas and liquid flows, operation of dampers,
and spray patterns would be observed and calibration accomplished. During
the latter part of the air and water test, the seed crystals would be added
so that a slurry could be pumped through the system.
The third phase would be a mechanical proving phase in which the unit would
be run on air and water under the control of the operators as a normal
system. The purpose of this phase was to catch any early failure in mechani-
cal equipment so that the final flue gas run would not be complicated by such
problems.
The fourth phase would consist of operation handling flue gas on a round-the-
clock basis for a sufficient length of time to verify that the system would
operate satisfactorily before commissioning and final turnover.
To perform the work under these various phases, manpower was to be supplied
by AAF, Kentucky Utilities, and their mechanical and electrical contractors.
In addition, once the flue gas phase had begun, an analytical team would be
employed around the clock from AAF's Research and 'Development Group to give
advance warning of potential chemical scaling.
IV. AIR AND WATER TESTS
Following general electrical and mechanical debugging, startup for the air
and water phase began in August. Upon startup of the main fan, stack
oscillation occurred. This was not a consistent problem, but appeared to
vary in relation to air flow through the system. AAF, utilizing its Pulsco
Division, determined that a standir^g wave was forming within the system
that exactly matched the demister and stack natural frequency, causing the
oscillation. An anti-spin device was designed and installed within the
upper portion of the demister, solving this problem. Also, during this
test, minor modifications were made to the cyclonic demister to achieve
the required performance. By positioning deflectors and drip rings at care-
fully analyzed points, mist carryover was eliminated. Concurrently, a care-
ful analysis was made of the mobile bed contactor. Engineering personnel
observed spray patterns from within the unit to verify sphere movement
through a range of gas flows, and performed other tests. Nozzle locations
were optimized, and further nozzle study work was begun.
While this work was being done, a number of the control loops were confirmed
or modified. Other minor work, such as installation of additional pipe
supports, and access modifications was performed.
994
-------
Just prior to the completion of the work, the lime bins were charged with
pebble lime and the slaker was operated. The slaking equipment operated
satisfactorily, although there were initial minor problems with the lime
feed system. Bridging would occur across the slake tank, which would then
result in cave-in and splashing. The lime feed screw conveyor was simply
adjusted to the proper speed, and this problem was eliminated. The mix
hold tank was filled approximately half full of 20% lime slurry in prepara-
tion for the flue gas run.
V. MECHANICAL RUN
After making the abovementioned adjustments, the system was run continuously
for approximately two weeks to uncover any additional mechanical difficul-
ties. Mixer malfunctions occurred, including a motor failure and faulty
drive train assembly. Near the end of the two-week run, gypsum seed crystals
were added to the reactant tanks, and the system was run for a few days under
this condition.
VI. FLUE GAS RUN
The flue gas run commenced the morning of September 13, 1975. Startup
occurred and subsequent operation was at 50% capacity as one of the genera-
ting units was off line for major repairs. Slurry samples were withdrawn
from specified points hourly to monitor pH and per cent solids and to verify
sulfate saturation level within the system. The inlet SO analyzer plugged
with flyash within a few hours, and has since required frequent calibration
and cleaning.
During the first day's operation, bunkers were filled with 1.6 to 2.0%
sulfur coal having ash content ranging between 15% and 25%. This low SO
concentration and low evaporation rate due to 50% gas flow resulted in a
minor water imbalance as water entering the system through pump glands
and demister wash was greater than the amount lost through bleed plus
evaporation. Solids content in the recycle system increased very slowly
because of the low sulfur content and resulting low solids precipitator
rate. The unit was shut down for a routine inspection after several hours'
operation, no problems were observed, and the system was put back on line.
With one generator down for repairs, it was necessary to begin operation
of the I.D. fan for the third boiler at nominal flow, as flue gas leaked
from the third boiler breech connection. This resulted in high excess air
and low SO inlet concentration. During this fan startup, the system auto-
matically went on bypass proving the duct pressure control loop. Taking
995
-------
advantage of this bypass, the scrubber fan was shut down and an inspection
was made. Additionally, a shield was installed around the SO- analyzer
inlet probe to prevent it from plugging with flyash. Upon re-start, the
SCL analyzer indicated approximately 1,200 ppm entering the system, consistent
with the level expected from the combination of low sulfur coal and dilution
air from the third boiler. Operation during this period was primarily under
a manual mode; pH readings were inaccurate due to poor slurry circulation at
the probes. The sensing probes had been placed within still wells within the
reactant tanks, and although the agitators provided more than adequate move-
ment within the tanks, the still wells employed for probe protection didn't
allow adequate slurry circulation and blockage occurred. The probes were
then removed from the still wells and installed elsewhere in the tank which
provided^for adequate circulation and with ultrasonic cleaning resulted in
reliable operation. Ultrasonic probe cleaners were further modified at the
manufacturer's request to provide better performance. A pH override control
was modified to increase its sensitivity to scrubber exit slurry pH, resulting
in improved control response.
SO removal efficiency decreased during the week, with inspection revealing
that nozzles were plugged with debris that had not been adequately flushed
from the system prior to startup. Nozzles were periodically removed, flushed,
and re-installed during the flue gas run.
AAF was aware of the scaling that has occurred in other systems and took
pains to carefully inspect this system during periodic shutdowns to identify
the formation of scale.
Scaling has not been a problem, although some film of scale has been noted.
This has tended to disappear with continued operation.
The SO- removal efficiency of the system was generally in the mid-nineties
range, with a varying inlet concentration of SO- as the sulfur in the coal
changed.
SUMMARY
The initial phases of the startup of this system have followed a carefully
planned procedure. Some mechanical problems have been experienced, but
solutions have been found as the problems occurred.
The wide range of sulfur content in the coal burned as well as the high
equivalent excess air in the flue gas tested the system's ability to respond
to changes.
The S02 removal capability and the absence of measurable scale in the system
are most encouraging in this startup.
996
-------
UNIT #1
BOILER #1
Mech.
Clctrs.
I.D.
Fans
Mobile
Bed
Contactor
UNIT #2
BOILER #2
BOILER #3
Exist. Stack
Booster
Fan
Venturi
From
.Pond
M.U.
Reactant Addition
i n
TT TT
Recycle
Bleed to Pond
997
I )
-£—y FIGURE 1
Hold Tank
-------
-------
TCA SPHERE DEVELOPMENT AND EVALUATION
Paul Sorenson, Principal Eng., Technical Services § Development
Dr. Nurhan E. Takvoryan, Mgr. of Scrubber Operations
Raymond J. Jaworowski, Mgr., Technical Services § Development
Air Correction Division
UOP, Inc.
Darien, Connecticutt
ABSTRACT
The requirements for fluidized-bed packing (spheres) have become
more stringent as the application of TCA scrubbers has expanded from
use in industrial processes to flue gas desulfurization and particulate
removal. To meet these requirements a continuing program of sphere
improvement has been maintained by the Air Correction Division of UOP.
The development of representative screening and evaluation tests is
critical to this program since the service life of the spheres is to
be a minimum of 1-2 years. It is the intent of this report to describe
both the sphere program and the various testing and evaluation procedures
used.
Economic and operation specifications have indicated that thermo-
plastic and elastomeric materials should be emphasized. Presently
foam rubber spheres have been developed that exhibit extraordinary
resistance to service conditions and thus seem to be the ultimate
solution of the sphere question. The comparison of laboratory and
field experiences demonstrate the continual improvement of spheres
for various TCA applications.
999
-------
TCA SPHERE DEVELOPMENT AND EVALUATION
THE PROBLEM
The Turbulent Contact Absorber (TCA) utilizes a fluidized bed
of low density spheres, Figure 1, to provide a contact surface
between the liquid scrubbing medium and the gases to be scrub-
bed. The first standard to be used was a blow molded polypropyl-
ene (PPO) sphere, Figure 2. As long as the TCA was being used
for chemical absorption with a minimal amount of particulate the
PPO spheres performed satisfactorily- As the market for the TCA
developed, a high percentage of the applications was for flue gas
scrubbing. The presence of fly ash and the use of lime/limestone
slurries in these applications produces a high abrasive environ-
ment which significantly reduces sphere service life. Field test
reports showed service life to be from 2000-4000 hours whereas
acceptable minimum service life for power plant application would
be at least one year or over 8000 hours.
In 1972, the Air Correction Division (ACD) initiated a comprehensive
development program to analyze the sphere life problem and to develop a
sphere of the required quality. Analysis of sphere life indicated
that abrasion, impact, and oxidative degradation were pertinent
factors in PPO sphere failure. The first step in improving service
life was to find a material which could be processed to form the
required sphere and would withstand the severe impact and abrasion
without degradation. The second and more difficult step was the develop-
ment of a test procedure to justify the material selection and fabrica-
tion technique. This accelerated test should help to predict sphere
service life without waiting a year for field data. Only with continual
progress on the second step could the results of the first be fully
realized. In completing evaluation, it is necessary to show that the
test procedures are relevant to field service. It will be shown how
improvement in technique, and confidence in test techniques has lead
to a closed cell foam rubber sphere which meets the requirements for
scrubber service.
THE PRODUCT
The sphere under study is 3.8 cm (1.5 inches) in diameter and weighs
approximately 5.5 grams. During the development program the design
weight has steadily increased from 4.5 grams to the present 6.5 grams.
This change has been made by the requirements of scrubber performance.
Except for correcting data for different weights, the changes in
1000
-------
weight have not influenced the program. For the best combination of
integral strength and production economy, blow molding was chosen.
In the blow molding process, a tube of appropriate size is extruded
and located between mold halves. One end of the extrusion
is clamped shut and the tube is inflated from the other end. The
molds close forcing the material into a spherical shell and cutting
away all excess material. Five spheres can be blown from each ex-
trusion. During the blowing and molding process a new extrusion
is started so that the cycle may be repeated following the stripping
of the spheres and waste from the mold. While the process is simple
the entire cycle requires the closest control. The quality of the
weld seam alone is highly dependent on the strict control of process
parameters.
DEVELOPMENT AND EVALUATION
Embrittlement
The earliest serious problem with TCA spheres was embrittlement.
The PPO spheres would crack away from the poles towards the weld
seam after several hundred hours. Failure rendered the spheres
useless long before the anticipated life to the sphere due to
abrasion. The mechanism of failure was established as brittle
fracture due to impact. This embrittlement was caused by oxida-
tion, accelerated by elevated temperatures. The mechanism was
considered inherent to the PPO polymer, so a comprehensive material
survey was initiated to find an acceptable substitute and if
possible improve the service life of TCA spheres.
The survey resulted in a substitution of high density polythe-
lene (HOPE) for PPO and the development of thermoplastic rubber
(TPR) spheres. The embrittlement and loss of impact strength
was determined with the Gardner Impact Test. The same test on HDPE
indicated an obvious solution to the problem. This was further
verified in field tests and actual service. HDPE spheres have
served to the full expected life in abrasive systems with no
evidence of brittle fracture. Brittle fracture was the first and
simplest problem to overcome.
Abrasion
Of the three factors contributing to sphere failure abrasion has
been the most difficult to evaluate. There are many types of
abrasion tests; some are designed to simulate specific field
conditions and others to measure abrasion resistance in a labora-
tory. At the outset of the program it was realized that methods
must be developed to permit specification of material for abrasion
1001
-------
resistance and to project service life for sample spheres.
Abrasion studies developed along the following areas: first,
a method to screen sphere materials; second, an accelerated
lab test to evaluate sample spheres and finally, service life
data to verify accelerated tests.
A wet abrasive high velocity slurry spray was used to screen a
large variety of materials. Forty-five different thermoplastic and
elastomeric materials were tested in this manner. The results
were erratic. However, by assigning fixed values to the control
samples it was feasible to assign a relative abrasion resistance
ranking. The difficulty of testing and the limited value of the
results discouraged use of this technique.
The first attempt at accelerated sphere testing was made with a
modified Taylor Ro-Top shaking machine (Figure 3). The specially
designed shaker pans were loaded with 30 spheres and a 15% slurry
silicon carbide grinding powder. The vibrating action in the
closed container with an abrasive slurry simulated sphere to sphere,
sphere to slurry and sphere to structure contact. Tests were run
during the course of each working day until significant weight
loss was recorded. Actual running time varied from 100 to 400 hours,
Tests with original polypropylene spheres showed a 17% weight loss
per 100 hours of shaker testing.
When compared to results from the pilot scrubber described below,
the PPO spheres exhibited a tenfold acceleration for several
different tests. Weight losses of 17% per 100 hours in the shaker
were equal to 17% per 1000 hours in the pilot scrubber. The weight
loss experienced by the HOPE spheres in the shaker was 5% per 100
hours as compared to 11% per 1000 hours for the pilot test. Thus
it was realized that there was no simple correlation between the
shaker and field test results.
At this point results of the wet abrasive spray tests indicated
that TPR would offer a tenfold improvement over the abrasion re-
sistance of either PPO or HOPE. This would almost negate the
accelaration factor offered by the shaker test. The shaker test
was therefore no longer useful.
Concurrently with the beginning of the above program a 3 stage
pilot TCA scrubber became available for sphere evaluation. The
unit used ambient air and a counter current flow of a 10% slurry
of fly ash that was continually recirculated. The pilot data
mentioned above was generated with this system. The TPR sphere
tests were run at scrubber operating temperatures. The duct work
1002
-------
was modified and insulated so that the gases would recirculate.
Heating elements were installed such that the spheres were in
contact with 49°C recycled liquor. It was felt that operation at
this temperature would give a reasonable simulation of actual field
conditions.
Tests were started immediately to obtain abrasion data for both
TPR and HDPE for 7000 service hours. Figure 4 shows the results
of these tests. A realistic means of studying abrasion was there-
fore available, although not on an accelerated basis. Samples were
taken from the pilot scrubber after 950 and 5500 hours, subjected
to microscopic examination and compared to virgin samples. Included
were spheres made of Ethylene-Ethyl Acrylate (EEA) which, while not
acceptable, were tested to extend the range of correlation. Both
optical and scanning electron microscopy were used. A mechanism
for abrasive wear was then projected. On hard surfaces such as
HDPE and PPO the particulate is partially imbedded on the surface
of the sphere. Subsequent action tears the particulate out along
with some material. This process continues with gradual loss of
weight. In addition, the protruding particulate acts as an abra-
sive against other spheres. The TPR spheres on the other hand are
elastomeric and therefore have a softer surface. Particulate be-
comes completely imbedded in the surface. The sphere does not,
therefore, have an abrasive surface when contacting other spheres,
nor is material lost due to particulate breaking away from the
surface. The curves in Figure 4 strengthen this theory. The HDPE
spheres are subjected to a nearly constant rate of weight loss.
TPR spheres lose some weight initially but between 500 and 2000
hours they appear to gain weight and thereafter the overall weight
loss is negligible. It is projected that the weight gain is
caused by the assimilation of the particulates imbedded in the
surface.
The understanding of a mechanism for sphere abrasion led to
accelerated tests. Three methods were tested: sand blasting,
pin and disc abrasion and planetary ball mill. In all three
methods the mechanism of abrasion was similar to that experienced
in the scrubber. In the sand blasting and pin and disc tests the
quantitative results could not be easily correlated with scrubber
data.
The planetary ball mill approach was more successful. Following
the test, the samples are examined microscopically to verify the
mechanism of abrasion. Tests on four samples for which scrubber
data was available showed a direct correlation, Figure 5. This
test has now become an important tool in sphere development.
The present understanding of the mechanism of abrasion has taken
1003
-------
three years to develop. With the constant cross-reference of
test results it has been possible to develop test techniques
which can be used to evaluate the ability of a sphere to with-
stand abrasion.
Impact and Seam Strength
The effect of impact on sphere service life has been misleading-
The obvious modes of failure in PPO were abrasion and brittle
fracture. While impact accelerates brittle fracture, eliminating
oxidation embrittlement by the choice of HOPE and TPR as sphere
materials solved this problem. In the first field tests of TPR
a serious problem with weld line failure was encountered. This
led to a comprehensive study of the effects of impact on blow
molded spheres. The results of the study have resulted in successive
improvements in the quality of TCA spheres.
Initial results in TPR from the pilot scrubber reported 0.4% of
the samples failed near the seam. The seam failure was attributed
to manufacturing difficulties which should be corrected before
production began. The first production lot was sent out to field
test based on the pilot experience. Within several hundred hours
of operation, large percentages of spheres failed by splitting in
half along the weld line. Examination of spheres showed that static
loading or single impact would not cause failure. Since TPR is
elastomeric, the surface is flexible. Repeated flexure along the
weld seam could cause failure. Subsequent investigations showed
that material around the weld seam was significantly weaker than
the rest of the shell.
Work to strenghten the seam was started immediately by the blow
molders with the cooperation of the resin supplier. At the same
time a program to quantify seam strength was initiated. The first
tests were to repeatedly compress spheres to approximately one
half of their diameter. The number of cycles until failure would
then indicate the seam resistance to failure. In the test fixtures,
spheres were free to rotate so that the axis of compression was
random. Post test examination revealed that failure always started
on the seam at location shown as "E" in Figure 2.
Results with this approach showed that the cycle life was also
dependent on the initial sphere weight. Sphere weight is a direct
relation to wall thickness and thus shell strength. Therefore, in
an attempt to equalize flexural stress between spheres of different
weights it was decided to adjust the amount of compression to be
equal to the compression caused by a force of 25 pounds. With this
1004
-------
method, samples from the field test batch averaged 10,560 cycles
to failure with approximately 20% of the samples below 1,000 cycles.
The results gave a qualitative relation to the field test i.e.,
high initial rate of failure with a long service life for the re-
maining. One further refinement was added. The temperature at
which the tests were conducted was raised from ambient to 49°C.
Improvements in seam strength were verified using the above tech-
nique. Figure 6 shows an example of the progress that was made.
The improvements were a direct result of changes in the manufactur-
ing techniques and processing parameters.
At about the same time the tests were being run, pilot scrubber
results and field tests showed that HDPE was also subject to seam
failure. Since HDPE is a rigid shell, the techniques of the
flexure test were not applicable. It was thought that the mode
of failure for both types of spheres was repeated impact.
An alternate test method was required. A machine called the "Frag
Tester" was adaptable to sphere testing. The Frag Tester is a
machine designed to measure the tear strength of paper. With a
few modifications, the machine was adapted to measure the seam
strength of spheres. In this machine a weighted arm is raised
to a preset height and dropped on the sphere under test. The
action is repeated until there is a break in the seam. The number
of drops to failure is an indication of seam strength.
Unlike the flexure test, the impacts are repeated at a single
point. It was necessary to establish the point for consistent
testing. The flexure tests showed the "E" point of failure for
TPR. This was verified with the Frag Tester, by testing at 12
locations around the seam. It was determined that the lowest
cycles to failure occurred at the "E" point. The focus for im-
pact testing was established at that location. Figure 7 shows
the correlation between Frag and Flex Test for various production
batches of TPR spheres. Each point represents the average cycles
to failure of 20 samples, 10 for flexure testing and 10 for Frag
testing.
With the help of the Frag tester and the cooperation of the blow
molders it was determined that seam strength is a direct result of
process parameters. In both HDPE and TPR, the molding cycle and
process temperatures are critical. In order to insure acceptable
quality, Frag type tests have been installed at the blow molders
and have become a regular part of their quality control program. In
addition, the molders are required to submit samples to ACD for
quality audits.
1005
-------
Dimpling
The phenomena referred to as "dimpling" is common to all flexible
or elastomeric spheres. It appeared in the first TPR field test.
At that time it was attributed along with splitting, to incorrect
material processing. After some careful thought and related tests,
the simple explanation became evident: most plastics are porous
to gases. Following the thermodynamic history of a sphere will
illustrate how the porosity results in dimpling.
After the sphere has been molded and cooled to ambient temperature
the gas inside the sphere is approximately at 21°G with a pressure
of about 760mmHg and has a relative humidity (R.H.) of say 50%. In a
scrubber the spheres are operating at approximately 54°C and 99%
R.H. The air inside the sphere has increased in pressure to 863mmHg
and the R.H. has dropped to 9%. The porosity of the material per-
mits air to migrate out and water vapor to be absorbed. This may
take 100,500 or even 1000 hours. In any case, the migration will
continue until equilibrium is reached i.e., the inside gases are
at 54°C, 760mmHg and 99% R.H. At this condition, the partial
pressure of water vapor is llVmmHg and the air is at 645mmHg. When
the scrubber is shut down for any reason the first effect is to re-
act to temperature.
The air will loose pressure according to the ratio of absolute
temperatures. The water vapor will condense with a further re-
duction in pressure. The resulting partial pressure of the air
would be 580mmHg and the partial pressure of water vapor is only
18.6mmHg. There is then approximately 160mmHg negative pressure
in the sphere. If the shell is not rigid enough to withstand
this differential, the sphere can collapse. The mechanism occurrs
in all spheres, rigid shells will not collapse and while the more
flexible TPR do not generally collapse, whenever subjected to
outside loading they will collapse or "dimple".
To overcome dimpling it would seem logical to either impart abrasion
resistance to a rigid shell or impart rigidity to an abrasion re-
sistant shell. Attempts to find an abrasive resistant rigid shell
have been fruitless. Blending HDPE or PPO resins with TPR to in-
crease rigidity have always been accompanied with losses in
abrasion resistance.
1006
-------
Field Reports
Of necessity the validity of laboratory procedures has rested
heavily on comparison with results obtained from the pilot
scrubber. It has been possible to gather some field data for
comparison with the pilot scrubber. Figure 8 shows the latest
available data from scrubbers operating on line. The correlation
between field and pilot scrubbers for particulate removal is ex-
cellent and is the basis for accepting all current test procedures.
As might be expected, the wear rates for lime/limestone slurries
are higher in both HOPE and TPR.
Seam splitting in the field has been harder to correlate. Because
of the tedious labor involved in counting failures, field reports
have been a simple statement as negligible, minor or unacceptable.
It has been possible to determine that spheres which last more
than 100 Frag cycles will not split in scrubber service. The difficulty
in reducing seam failure is inherent in that seam strength can only
be monitored by destructive testing. With the use of appropriate
statistical techniques the rate of failure has been held to less
than 0.5 percent. Unfortunately the results of seam failure even
at 0.5% are particularily troublesome when accumulated in the scrubber
sump.
The occurrence of dimpling in the field is an accepted fact. The
percentage of dimpled spheres increases with each shut down. On the
basis o.f present reports, scrubber operators have accepted dimpling
of TPR spheres in lieu of weight loss and splitting in HOPE. ACD
however has not accepted dimpling and has continued to seek a solution
to the problem.
CLOSED CELL FOAM RUBBER SPHERES
The entire sphere development program led ACD to a design paradox.
In using an elastomeric material which does not abrade, a flexible
shell results which dimples. If the shell is made rigid, the
abrasion resistance is significantly reduced. The cause of both
problems is in the basic function of the TCA and therefore cannot
be altered. Because of the low density of the spheres,0.16-0.24gm/cc,
homogenous materials are not applicable. The next approach then
was to use some form of a foamed material.
1007
-------
The first step was to select a material which could be foamed,
would have a solid skin and have the various other properties
which are required for service in a scrubber. The material would
have to be elastomeric so as to be abrasion resistant, at least
as well as TPR. Two catagories were considered, the urethanes and
rubber. Generally urethanes have not been considered as suitable
materials for TCA service due to their acid sensitivity. Although
recent developments are overcoming this problem, it was decided
to delay investigations of urethanes in favor of rubber.
The selection of a suitable rubber compound was simplified Vw
the existing list of compounds already screened as serviceable
for scrubbers. With the help of manufacturers representatives,
it was possible to select a rubber compound which would have the
required specifications and could be processed to obtain the sphere
of the right density- The material chosen was a modified polyvinyl
chloride-nitrile rubber. It is both acid and alkali resistant,
abrasion resistant and flame retardant. It is readily blown to
form a closed cell foam of the required density and has a solid
skin. Feasibility was shown with spheres formed in a single
cavity transfer mold. Various formulations were tried until an
appropriate stock receipe was obtained. The receipe was trans-
ferred to a qualified rubber processor for development. The first
samples were received from the processer in December, 1974, and the
process of evaluation began. Before the end of 1975, almost 2000
spheres had been received and subjected to the various testing
programs. By June, 1975 test results were so encouraging that
an order was placed for production tooling. By the end of 1975
production had started and the first placement of foam rubber
spheres for field test was made.
Although the material was specified for adequate chemical re-
sistence sample spheres were subjected to routine chemical tests
as verification. Similarily, tests were run for flame retardance.
The spheres were more than adequate on all points. Although there
is no seam in this type of sphere, the samples were subjected to
the Frag test to note their reaction to impact. These preliminary
tests provided confidence that the spheres would withstand the
environment in the TCA. A representative sample was placed in the
pilot scrubber for simulation of service life as it was necessary
to determine if there was a mode of failure which had not been
anticipated.
The results of the pilot scrubber were encouraging. in over
4000 hours of operation there were no incidents of sphere failure.
1008
-------
No significant change was noted in shape and skin texture. No
splits, cuts or other flaws were noted. There was definite evidence
of water absorption. Weights increased gradually until equilibrium
was reached at approximately 125% of the original weight. Water
absorption was verified by placing the spheres in a warm dry atmos-
phere for several hours. In that time, spheres would return to
within 1/2% of their original weight. Figure 9 is a graph of the
change in weight in comparison with HDPE and TPR.
The results of the accelerated wear tests were fortunately indepen-
dent of absorption. The tests are run in one hour increments for
a total of 5 hours. At each increment the samples are dried and
weighed. Since it takes over 1000 hours to reach absorption equilibrium,
the effect as compared to the abrasion loss is negligible. The drying
also minimizes any effect. The weight actually lost in 5 hours is
0.135% of original weight as compared to 0.16% for TPR. On this
basis, the estimated time to wear through the skin of the foamed
ball is 12,000 hours.
Since no failures were established in the pilot scrubber, the
accelerated wear tests were run past the 5 hours until surface
flaws occurred which penetrated the skin of the sphere. At this
point numerous craters and splits were produced which exposed the
internal cells. It was of particuliar interest to determine if
the exposure of the inner cells affected tha rate of absorption.
The scarred samples were soaked in water with several new samples
as a control. Almost identical weight gains were observed, thus
when the skin is abraded or otherwise damaged, no appreciable
change in the action of the mobile bed is expected.
In the earlier discussion of dimpling, it was shown that the
mechanism is a function of gases trapped in a shell of elastomeric
material. In the foam rubber sphere, the single gas cavity has been
replaced by millions of tiny cells separated by a lattice of rubber.
The mechanism of dimpling also occurs in these cells but does not
result in an appreciable change in shape or usefullness. In the
newly made condition each of the cells is pressurized. This places
the skin in tension holding the diameter. During service the
pressure is relieved through diffusion and the gases in the cells
reach an equilibrium in both pressure and humidity at operating
temperatures. When the spheres are returned to ambient temperature,
the internal pressure becomes less than ambient and the skin goes
from a state of tension to one of compression. Since the sphere
is made of an elastomeric material this change will result in a
dimensional change. The pneomena is witnessed when cutting samples
in half. When split, a new sphere will bulge out in the center
1009
-------
such that the two halves can not be reassembled surface to surface.
In splitting a sphere which has been in service long enough for
equilibrium to be reached, the cut surfaces do not bulge and may
be placed together in surface to surface contact. Thus it has
been shown that the mechanism of dimpling has occurred but because
of the change in structure, the results cause no change in the
effectiveness of the spheres as a mobile bed.
CONCLUSION
In the continuing effort to improve the quality and service life
of TCA spheres, ACD has been able to reduce the severe environment
of the scrubber to laboratory test procedures and life tests. Realizing
that a commercial TCA installation requires anywhere from 750,000
to 1,500,000 spheres for operation, the importance for sound laboratory
procedures is easily understood. Fortunately, through the cooperatioA
of several customers and test facilities, it has been possible to
correlate the results of the laboratory with field data. Over the
course of the program the solution to each problem has been closely
followed by the challenge to extend the service life still further.
No sooner had HDPE been substituted for PPO and the brittle fracture
problem solved, then the challenge to extend the life due to
abrasion from 4000 hours to 8000 hours was made. In the development
of the TPR sphere the promise of 8000 hours life was a reality.
This was followed by the problems in manufacture which resulted
in weld seam failure. With the cooperation of the blow molders
this problem was finally overcome after lengthy and difficult
experimental programs. With the solution of the seam problem
came the realization that dimpling in TPR spheres was a bothersome
problem that would not be long tolerated. Finally as a result
of all of the technology acquired over the duration of the program,
an approach was conceived and developed which adequately meets the
rigid specifications for TCA scrubbing spheres. The foam rubber
sphere has been and will continue to be evaluated and tested
against standards which were not envisioned when the first PPO
spheres were ordered for TCA service.
1010
-------
CLEAN GAS OUTLET
DEMISTER
-—LIQUOR INLET
MOBILE PACKING
STAGES
LIQUOR OUTLET
FIGURE 1 TURBULENT CONTACT ABSORBER
POLE
WELD SEAM
FIGURE 2 MOBILE PACKING SPHERE
ion
-------
FIGURE 3 TAYLOR RO-TOP SHAKER
1012
-------
X
O
uu
£
I
Z
O
Qi
O
25%
20%
O 15%
co
to
10%
X
® 5%
1000 2000
3000 4000
TIME - HOURS
5000 6000 7000
FIGURE 4 WEAR RATE PILOT SCRUBBER
EEA o
OO
CK
=3
O
X
O
O
O
CM
OO
OO
O
X
O
10 r
TPR
BLENDC
HOPE,
0.1 0.2 0.3 0.4 0.5 0.6 0.7
% WEIGHT LOSS-5 HOURS
PLANETARY BALL MILL
1013
FIGURE 5 WEIGHT LOSS CORRELATION
0.8
-------
IUU
UJ
i—
| 80
1
5 60
^^
®
g 40
£^
X
uj 20
u_
-
-
(
)o
1
, ,
<
i
i <
!
»
{ i
i i i i i i i i i i i i
12-72 8-73 12-73 12-73M 12-73 1-74 2-74
A B B B C C B
SAMPLE - VENDOR CODE,DATE
FIGURE 6 SEAM STRENGTH FLEX TESTS
MAXIMUM
AVERAGE
MINIMUM
O
x
1/1
UJ
__i
U
10
9
8
7
6
5
3
2.5
1.5
10 1.5 2 2.5 3 456789 100
FRAG CYCLES
1014
1.5
FIGURE 7 IMPACT TEST CORRELATION
-------
30
'HDPE PILOT SCRUBBER
O
CfL
O
to
to
O
I
O
25
20
15
10
x
/O
O x
O /
x
/Q
O- HDPE-PARTICULATE
O-HOPE-LIMESTONE
Q-TPR PARTICULATE
A-TPR LIMESTONE
TPR- PILOT SCRUBBER
1000 2000 3000 4000 5000 6000
SERVICE LIFE - HOURS
FIGURES FIELD SERVICE - ABRASION
1000
2000
3000
4000
TIME- HOURS
1015
FIGURE 9 WEIGHT CHANGE IN PILOT SCRUBBER
-------
TECHNICAL REPORT DATA
(Please read Jntlruclions on the reverse before completing}
l. REPORT NO.
EPA-600/2-76-136b
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Proceedings: Symposium on Flue Gas Desulfurization-
New Orleans, March 1976; Volume n
5. REPORT DATE
May 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S) ,
R. D. Stern, Chairman; W. H. Ponder and
R. C. Christman (TRW, Inc.), Vice-chairmen
8. PERFORMING ORGANIZATION REPORT
9. PERFORMING OR9ANIZATION NAME AND ADDRESS
Miscellaneous
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
In-house
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings; 3/8-11/76
14. SPONSORING AGENCY CODE
EPA-ORD
is. SUPPLEMENTARY NOTES jERTj-RTP project officers for these proceedings are R. D. Stern
and W.H. Ponder, Mail Drop 61, Ext 2915.
16. ABSTRACT
The proceedings document the presentations made during the symposium,
which dealt with the status of flue gas desulfurization technology in the United States
and abroad. Subjects considered included: regenerable, non-regenerable, and
advanced processes; process costs; and by-product disposal, utilization, and
marketing. The purpose of the symposium was to provide developers, vendors, users
and those concerned with regulatory guidelines with a current review of progress
made in applying processes for the reduction of sulfur dioxide emissions at the full-
and semi-commercial scale.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED-TERMS_
c. COSATJ Field/Group
Air Pollution
Flue Gases
Desulfurization
ulfur Dioxide
ulfur Oxides
!ost Effectiveness
Byproducts
Disposal
Utilization
Marketing
Air Pollution Control
Stationary Sources
13B
2 IB
07A,07D
07B
14A
3. DISTRIBUTION-STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
463
2O. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
1016
------- |