Proceedings: Symposium on Flue Gas
Desulfurization, Houston, October 1980. Volume 1
Research Triangle Inst.
Research Triangle Park, NC
Prepared for
Industrial Environmental Research Lab
Research Triangle Park, NC
Apr 81
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
NTTIS
-------
EPA-600/9-81-G19c,
April 1981
Proceedings: Symposium on
Flue Gas Desulfurszation -
Houston, October 1980;
Volume 1
Franklin A. Ayer, Compiler
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
Contract No. 68-02-3170
Task No. 33
EPA Project Officer: Julian W. Jones
Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
Research Triangle Park, IMC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
EPA- 600/9- 81-019a
4. TITLE AND SUBTITLE Proceedings : Symposium on Flue Gas
Desulfurization— Houston, October 1980; Volume 1
7. AUTHOR(S)
Franklin A. Ayer, Compiler
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
3. RECIPIENT'S ACCESSION NO. ;
PB81 243156
6. REPORT DATE J
April 1981
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO. •,
10. PROGRAM ELEMENT NO.
1NE828
11. CONTRACT/GRANT NO.
68-02-3170, Task 33 j
1
13. TYPE OF REPORT AND PERIOD COVERED '
Proceedings; 10/80 \
14. SPONSORING AGENCY CODE \
EPA/600/13 }
15. SUPPLEMENTARY NOTES IERL.RTP project officer is Julian W. Jones , Mail Drop 61, *
919/541-^2489. EPA-600/7-79-167a and -167b are the proceedings of the previous
symposium on flue eas desulfurization.
is. ABSTRACT
two-volume proceedings document presentations at EPA's Sixth Sym-
posium on Flue Gas Desulfurization (FGD), October 28-31, 1980, in Houston, Texas.
Presentations covered such subjects as approaches for control of acid rain, the
Nation's energy future, economics of FGD, legislative/regulatory developments,
FGD research/development trends , FGD system operating experience , FGD
byproduct disposal/utilization, developments in dry FGD, and industrial boiler
applications .
17.
KEY WORDS AND DOCUMENT ANALYSIS 1
a. DESCRIPTORS
Pollution
Flue Gases
Desulfurization
Acidification
Climatology
Energy
Economics
Legislation
Regulations
Byproducts
Waste Disposal
Boilers
13. DISTRIBUTION STATEMENT
Release to Public
b.lOENTIFIERS/OPEN ENDED TERMS
Pollution Control
Stationary Sources
Acid Rain
19. SECURITY CLASS (This Report}
Unclassified
20. SECURITY CLASS (This page}
Unclassified
c. COSATI Field/Croup
13 B 05C
21B D5D i
r07A,07D
07B,07C
04 B
14G ISA
21. NO. OF PAGES
22. PRICE '
EPA Form 2220-1 (9-73)
-------
PREFACE
These proceedings for the symposium on "Flue Gas Desuifurization"
constitute the final report submitted to the Industrial Environmental
Research Laboratory, U.S. Environmental Protection Agency
(IERL-EPA), Research Triangle Park, NC. The symposium was con-
ducted at the Shamrock Hilton Hotel in Houston, TX, October 28-31.
1980.
This symposium was designed to provide a forum for the exchange of
information, including recent technological and regulatory develop-
ments, on the application of FGD to utility and industrial boilers. The
program included a Keynote Address on the approaches for control of
acid rain, forecasts of energy and environmental technologies and
economics for the 1980's, and sessions on the impact of recent legislation
and regulations, research and development plans, utility applications,
by-product utilization, dry scrubbing and industrial applications. Par-
ticipants represented electric utilities, equipment and process suppliers,
state environmental agencies, coal and petroleum suppliers, EPA and
other Federal agencies.
Michael A. Maxwell, Chief, Emissions/Effluent Technology Branch,
Utilities and Industrial Power Division, IERL-EPA, Research Triangle
Park, NC, was General Chairman, and Julian W. Jones, a Senior
Chemical Engineer in the same branch was Project Officer and Co-
Chairman.
Franklin A. Ayer, Manager, Technology and Resource Management
Department, Center for Technology Applications, Research Triangle
Institute, Research Triangle Park, NC, was symposium coordinator and
compiler of the proceedings
-------
TABLE OF CONTENTS
VOLUME I
Session I: OPENING SESSION .......................................... 1
Michael A. Maxwell, Chairman
Keynote Address: Approaches for Control of Acid Rain ......................... 3
Stephen J. Gage
The Nation's Energy Future— With Focus on Synfuels .......................... 27
Frank T. Princiotta
FGD Economics in 1980 ............................................. 49
G. G. McGlamery,* W. E. O'Brien,
C. D. Stephenson, and J. D. Veitch
SO2 and NOX Abatement for Coal-Fired Boilers in Japan ........................ 85
Jumpei Ando
Session 2: IMPACT OF RECENT LEGISLATION/REGULATIONS ................... 111
Walter C. Barber, Chairman
Session 3: FGD RESEARCH AND DEVELOPMENT PLANS ...................... 113
Julian W. Jones, Chairman
Recent Trends in Utility Flue Gas Desulfurization ............................. 115
M. P. Smith, M. T. Melia,
B. A. Laseke, Jr.,* and Norman Kaplan
The Department of Energy's Flue Gas Desulfurization
Research and Development Program .................................... 173
Edward C. Trexler
EPRI Research Results in FGD: 1979-1980 ................................. 183
S. M. Dalton,* C. E. Dene,
R. G. Rhudy, and D. A. Stewart
Session 4: UTILITY APPLICATIONS ................... - ................ 231
H. William Elder, Chairman
Test Results of Adipic Acid-Enhanced Limestone
Scrubbing at the EPA Shawnee Test Facility— Third Report .................. 233
D, A. Bwrb,ank,* S. C. Wang,
I. ft Meftw, wrt 4- £• Wife™*
. . .287
S. B. Jackson
Presented by William L. Wells, TVA
DOWA Process Tests, Shawnee Test Facility ............................... 311
S. B. Jackson, C. E. Dene, and D. B. Smith
Presented by William L. Wells, TVA
F.G.D. Experiences, Southwest Unit 1 ........ , .......................... 327
N. Dale Hicks* and O. W. Hargrove
* Denotes speaker
-------
Results of the Chiyoda Thoroughbred-121
Prototype Evaluation 347
Thomas M. Morasky,* David P. Burford,
and 0. W. Hargrove
Forced Oxidation of Limestone Scrubber Sludge at TVA's
Widows Creek Unit 8 Steam Plant - • 3?1
C. L. Massey, N. D. Moore,
G. T. Munson, R. A. Runyan,* and W. L. Wells
La Cygne Station Unit No. 1 Wet Scrubber
Operating Experience
Richard A. Spring
One Button Operation Start-up of the Alabama Electric
Cooperative FGD System 415
Royce Hutcheson* and Carlton Johnson
Operation and Maintenance Experience of the World's
Largest Spray Tower S02 Scrubbers 433
Robert A. Hewitt* and A. Saleem
Dual Alkali Demonstration Project Interim Report 453
R. P. Van Ness,* Norman Kaplan, and D. A. Watson
Operating Experience with the FMC Double Alkali Process 473
Thomas H. Durkin, James A. Van Meter,*
and L. Karl Legatski
Status Report on the Wellman-Lord/Allied Chemical
Flue Gas Desulfurization Plant at Northern Indiana Public
Service Company's Dean H. Mitchell Station 497
E. L. Mann* and R. C, Adams
Magnesium FGD at TVA: Pilot and Full-Scale Designs 543
E. G. Marcus, T. L. Wright, and W. L. Wells
Presented by Landon W. Fox, TVA
VOLUME II
Session 5: BY-PRODUCT UTILIZATION 559
Jerome Rossoff, Chairman
introduction .... 561
Jerome Rossoff
Characterization and Environmental Monitoring of
Fufi-Scate Utility Waste Disposal; A Status Report 557
Chakra J. Santhanam* and Julian W. Jones
Evaluation of Potential Impacts to the Utility Sector
for Compliance with RCRA . . 603
Val E. Weaver
EPRt FGD Sludge Disposal Demonstration and Site
Monitoring Projects 625
Dean M. Golden
-------
Potential Effects on Groundwater of Fly Ash and FGD
Waste Disposal in Lignite Surface Mine Pits In North Dakota 657
Gerald H.Groenewold,* John A. Cherry, Oscar E. Manz,
Harvey A. Gullicks, David J. Hassett, and Bernd W. Rehm
Environmental Compatibility and Engineering Feasibility
for Utilization of FGD Waste in Artificial Fishing Reef Construction 695
P.MJ. Woodhead, J. H. Parker, and I. W. Duedall*
Government Procurement of Cement and Concrete
Containing Fly Ash 701
Penelope Hansen* and John Heffelfinger
Session 6: DRY SCRUBBING 711
Theodore G. Brna, Chairman
Spray Dryer FGD: Technical Review and Economic
Assessment 713
T. A. Burnett, K. D. Anderson, and R. L. Torstrick
Presented by Gerald G. McGlamery, TVA
Spray Dryer FGD Capital and Operating Cost Estimates
for a Northeastern Utility 731
Marvin Drabkin* and Ernest Robison
Current Status of Dry Flue Gas Desulfurization Systems 761
M. E. Kelly* and J. C. Dickerman
Dry SO2 Scrubbing Pilot Test Results 777
Nicholas J. Stevens
SO2 Removal by Dry FGD 801
Edward L. Parsons, Jr.," Lloyd F. Hemenway,
O. Teglhus Kragh, Theodore G. Brna, and Ronald L. Ostop
Dry Scrubber Demonstration Plant—Operating Results 853
T. B. Hurst* and G. T. Bielawski
Session 7: INDUSTRIAL APPLICATIONS 86;
J. David Mobley, Chairman
Applicability of FGD Systems to Industrial Boilers 863
James C. Dickerman
Sulfur Dioxide Emission Data for an Industrial Boiler
New Source Performance Standard ...... 887
Charles B. Sedman
Applicability of FGD Systems to Oilfield Steam
Generators and Sodium Waste Disposal Options 927
A. N. Patkar* and S. P. Kbthari
Performance Evaluation of an Industrial Spray Dryer
for SO2 Control './:., 943
Theodore G. Brna,* Stephen J. Lutz, and James A. Kezerle
-------
Evaluation of Emissions and Control Technology for fi5
Industrial Stoker Boilers ,*
Robert D. Giammar,* Russell H. Barnes, lfs
David R. Hopper, Paul R. Webb, and Albert E. Weljar
987
Unpresented Papers
989
Flakt's Dry FGD Technology: Capabilities and Experience . . • • • • •
Stefan Ahman, Tom Lillestolen, and James Farrington, Jr.
Perspectives on the Development of Dry Scrubbing—
_,. _ __
The Coyote Story
R.O.M. Grutle and D. C. Gehri
The Riverside Station Dry Scrubbing System
Gary W. Gunther, James A. Meyler, and ^ndJCeisJHanser
Evaluation of Gypsum Waste Disposal by Stacking 1031
Thomas M. Morasky, Thomas S. Ingra,
Lamar Larrimore and John 6. Garlanger
Dry Activated Char Process for Simultaneous SO2 and
NOX Removal from Flue Gases • 1067
Ekkehard Richter and Karl Knoblauch
KOBELCO Flue Gas Desulfurization Process 1081
Kobe Steel, Ltd.
APPENDIX: Attendees 1099
VI
-------
Session I: OPENING SESSION
Michael A. Maxwell, Chairman
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina
-------
KEYNOTE ADDRESS
Approaches for Control of Acid Rain
Stephen J. Gage
Assistant Administrator
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C.
Preceding page blank
-------
The current situation in the Persian Gulf has once again brought
home the stark reality of the fragile balance of our industrialized
interdependent society. Once again we learn that our national economy
can be tipped up or down by events thousands of miles away from our
shores. National security and foreign policy deliberations must again
focus on the question, "What are the likely impacts on our oil imports
of a broadened war in the Mid-East?"
We have come to the point where we must find alternatives to foreign
oil and we have recognized that we have our own massive coal resources -
a wealth of "black gold" — among the greatest known reserves existing
anywhere in the world. We have recognized that we must move away from our
dependence on foreign oil to greater reliance on domestic coal. President
Carter and the Congress have mandated this conversion to coal as part of
our overall National Energy Plan. We are beginning to move from a pre-
dominantly oil-based energy supply structure to one emphasizing domestic
coal, oil shale, unconventional natural gas and heavy oil. And we are
also encouraging — and succeeding in — a vigorous energy conservation
program.
What this means, of course, is that we are going to be mining and
burning more of the "dirtier" fuels. And that means there could be a
growing air pollution problem. Coal mining in the U.S. is projected to
increase from the current 700 million tons annually to 1.4 billion tons
in 1990 and 1.9 billion tons in 2000. Conventional combustion will
-------
continue to be the primary method of utilizing this coal well Into
the twenty-first century -- despite the growth of a major coal-based
synthetic fuel Industry.
The challenge we face, therefore, 1s to maintain our air quality
as the production of pollutants from burning fossil fuels rapidly
expands. Because of the Increased use of fossil fuels and the necessary
cost of pollution abatement, there will be increasing pressure in the
future to Improve environmental control technologies, to make them more
cost-effective and — equally Important — to achieve widespread
acceptance and operational utilization of these control systems by the
utilities and industrial facilities. This 6th FGD Symposium is testimony
to a continuing effort by both government and industry to meet these
challenges.
The Congress has also provided impetus for the development and
application of upgraded control technologies, like FGD. The 1977
Amendments to the Clean A1r Act underscored the importance of control
technologies through the requirement for Best Available Control Technology
1n areas where the air is clean....and the requirement for Lowest
ffefelevable Emission Rate in "non-attainment" areas where the air is
already dirty.
The recently issued New Source Performance Standards for utility
boilers and the forthcoming development of NSPS for industrial boilers
are typical examples of recent environmental protection efforts that
-------
will drive the continued research and development of environmental
control technologies.
I think 1t is likely that Federal legislative action In the
future will not significantly weaken current environmental programs.
I believe, rather, that in the face of the pressures to relax environ-
mental controls to allow more rapid expansion of our domestic fuels
utilization, the public and Congress will continue the trend toward
careful consideration of environmental impacts of future energy
development. While we have made progress in improving air quality
throughout the country over the last decade, the struggle is far from
over. The recent smog episode in southern California is a grim reminder
that some parts of the nation are still threatened with severe air
pollution under poor meteorological conditions.
We have made great strides in developing and demonstrating highly
efficient, reliable flue gas desulfurization technologies. While
there are Improved coal cleaning and new combustion technologies that
are in the developmental stage, and some even at the demonstration and
pilot test stages, FGD systems are currently the only viable sulfur
control technology capable of genera] application over the next ten years,
It has been estimated that by 1990, electrical utilities will have
invested between $10 and $20 billion for construction and operation of
FGD units.
-------
I see from the program that Gerald McGlamery of TVA is going to
discuss the economics of FGD systems a little later this morning.
I'm sure that those of you here representing the utilities will be
especially interested in what he has to say about TVA's latest cost
studies and experience. From our own studies in this area, we believe
that there 1s a good dollars and cents case for converting from oil
to coal — and that includes taking into consideration the use of FGD
control equipment. Let me cite a few figures. To produce one million
BTU's of heat, the cost of oil is $5.18, based on a price of $30 per
I
barrel. To produce the same one million BTU's of heat, the cost of
coal is $1.30, based on a price of $30 per ton. A power plant could
save five cents per kilowatt-hour by making the conversion and using
the best available scrubber, one with a 90 percent efficiency in
reducing sulfur oxide emissions. This translates to a savings of
$14 million per year for the average size electric generating plant
being built today.
Where less stringent scrubber controls are required, savings
could Increase. According to conservative EPA projections for burning
high-sulfur coals, a savings of 1/5 of a cent per kilowatt-hour
would be realized by a utility that retires even a modern oil plant,
writes off the Investment, and replaces it with a new coal-fired
facility outfitted with the best scrubber available.
-------
In the United States, Japan and the Federal Republic of Germany,
operating FGD systems using wet processes, such as lime or limestone
scrubbers, continue to show improvement. Most of these processes are
currently capable of removing well over ninety percent of the sulfur
oxides in the flue gas. Here, in the U.S., lime and limestone scrubbers
have been applied to coal with a wide range of sulfur content, and they
have reliably removed the sulfur oxides from burning coals with one to
four percent sulfur content. Many of these U.S. high sulfur coal FGD
installations have operational reliabilities of over 90 percent. FGD
installations on low sulfur coal have operational reliabilities of
over 95 percent which is similar to the Japanese experience with
low sulfur coals.
One example of a key program in nonregenerable systems is the
lime/limestone prototype test facility at TVA's Shawnee Steam Plant.
You'll be hearing about the latest results from that operation during
tomorrow morning's session. The results of this particular program
are important because over 90 percent of the U.S. coal-fired electric
generating capacity presently committed to FGD systems involves the use
of similar lime/limestone processes. The Shawnee program has been
directed toward obtaining answers to some of industry's concerns about
long-term reliability of the process, the large quantities of waste
sludge generated by the scrubber, and the high capital and operating
-------
costs Involved. I believe major technological Improvements and cost
reductions are possible and will be realized, as we learn from programs
such as this one.
FGD systems are now performing reliably and effectively both here
arid abroad. As I mentioned, in Japan, during the past decade FGD systems
have been Installed on a widespread basis. They have operated reliably
and have had outstanding success in improving the air quality. Dr. Ando
will speak on this subject in detail, but I'd like to cite a few statistics
to demonstrate how these systems have proven themselves in Japan. There
is no reason why they should not be just as effective here in the U.S.
Approximately 75% of the utility power generated in Japan is fossil-
fired steam-electric. The balance is hydroelectric and nuclear powered.
Of the fossil-fired capacity, 85% is oil-fired (most of the oil imported)
and only 3% 1s coal-fired — so you can see that their problem with foreign
oil dependency 1s much worse than ours. But they have reduced sulfur
oxide emissions from burning both oil and coal by 50% between 1970 and 1975,
and this has been due in great part to the use of FGD systems. They now
have ambient S02 standards that are among the most stringent in the worlc -•
about half the yearly average emission level that we allow.
Although Japan and the U.S. have both emerged as world leaders in
developing and applying FGD systems, Japan has generally moved ahead more
rapidly, because of its more serious commitment to solving its pollution
-------
problems. As of the beginning of last year, Japanese utilities had FGD
systems installed, under construction, and planned for about 16% of
their fossil-fired steam generating capacity....75% of it already
installed and operating.
In the U.S., on the other hand, only about 3% of the total fossil-
fired utility capacity is presently under FGD operational control.
There are plans or systems under construction, however, for another 12%
of the total fossil-fired capacity. At last count, 73 FGD units were
in operation, with 127 units in design or under construction. When all
of these units are operational, over 25% of the current total U.S.
coal-fired capacity will be equipped with FGD. Because of this growing
use of FGD, the total amount of sulfur oxides emitted to the atmosphere
is expected to remain constant or even decrease slightly by the year 2000
Even though we have made great strides in controlling sulfur oxides*
we still have a long way to go to ensure that our expanded use of coal
will not degrade the quality of our environment. EPA has been pursuing
an aggressive air emissions program to control sulfur oxides, nitrogen
oxides, and particulates — all released from the burning of coal. And
all contributors to a growing problem of acid deposition, more commonly
referred to as acid rain. I am concerned that acid rain may become one
of the most significant environmental problems of the coming decade. It
already poses an environmental threat to our aquatic resources and
possibly to our forest and agricultural resources as well -- a threat
10
-------
that could Intensify with the full-scale development of our fossil fuel
resources. We must therefore continue to work toward controlling the
emission of not only sulfur dioxides, but also nitrogen oxides and
partlculates, before they get a chance to get out Into the atmosphere
create add rain problems.
Far from being a "gentle rain from heaven," add rain can cause
extensive ecological damage. In New York's Adirondack Mountains, for
example, an area that was once a sport fisherman's paradise, acid rain
has killed all of the fish 1n half of the high-attitude lakes. We
cannot even guess at this time the extent of the damage in North American
lakes, but we strongly suspect that tens of thousands of lakes are
threatened, with millions of dollars in recreation benefits and commercial
fishing at stake. Acid rain may also be playing a part in the decline
1n forest growth observed in both the Northeastern United States and
southern Sweden. Experimental studies have shown that acid rain may
damage foliage, Interfere with the germination of seeds and the rooting
of seedlings, affect the availability of nitrogen in the soil, decrease
soil respiration, and deplete its nutrients. The destruction of stone
monuments and statuary throughout the world, including the 2500 year-old
Parthenon in Athens, Greece, has been accelerated by add rain.
Acid rain may even indirectly present humans with a health hazard.
If drinking water reservoirs become contaminated with acids, increases
11
-------
in heavy metal concentrations may exceed public health limits. In
New York State, for example, water from the Hinckley Reservoir has
acidified to such an extent that when the water comes in contact with
household plumbing systems, lead from soldered joints passes into the
water. These concentrations exceed the maximum levels recommended by
the New York State Department of Health.
Acid rain was once thought to be primarily an S02 problem, but
we've since learned that the phenomenon is more complicated than that.
Nitrogen oxides as well as sulfur oxides can be transformed into
potent acids when they combine with water vapor molecules in the atmos-
phere. The result is rain that may be — as we have found in some
parts of the country -- as acidic as lemon juice. Normal rainwater has
a pH of about 5.7; newly hatched fish, which are most sensitive to low
pH, are in serious trouble in water when its pH goes below 5.0. The
average pH of the rain east of the Mississippi today is 4.4, which is
almost 20 times as acidic as normal.
In the United States, the rain is most acidic in the heavily
industrialized Northeast, but the most rapid increase in acid rain seems
to be occurring in the Southeast. This parallels the expansion of South-
eastern urban and industrial activities that result in sulfur and nitrogen
emissions. Here, the trend is more apparent than in the Northeast,
because the atmosphere is more rapidly deteriorating, and fewer acidic
ions are required to cause a pH change. Most of the West has thus far
12
-------
escaped the add rain scourge* but Colorado, the Los Angeles Basin, the
San Francisco Bay Area, Spokane, Tucson, and Portland are known exceptions.
In much of the West, the alkaline nature of the soils and lakes acts to
neutralize add rain, so the effects may not be as pronounced there. But
even 1n the West, ominous signs of vegetation damage have appeared.
The Adirondack fish disaster, which occurred in an area of thin soils
and fragile, closely watched ecologies, may be only a dramatic early
warning of the damage that acid rain may someday cause on a much larger
scale. Were 1t not for the buffering ability of the soil in other sections
of the East Coast, the rains of the 1970's could have killed off most of
the region's freshwater fish.
Clearly, we are not talking about something that sprang from the
overactive imagination of a zealous environmentalist. Acid rain is a
phenomenon that demands careful attention.
What can be done to prevent the rains of the 1980's from becoming
increasingly more destructive? The most urgent task that EPA faces is
to get to the bottom of what causes acid rain. Until the perplexing
mechanisms by which acid rain is formed are better understood, attempts
to control it may miss the mark, resulting in a less than optimum use of
costly investments for control.
It 1_s_ known that, after sulfur and nitrogen oxides are discharged
Into the atmosphere, they are oxidized into sulfates and nitrates, which
then react with moisturp in the air to become acids. There are several
13
-------
complicated pathways or mechanisms by which this oxidation can occur-
Which path is actually taken depends on a number of factors, including
the concentration of heavy metals, the intensity of sunlight, the
temperature, the humidity, the amount of ammonia present, and the
particulate and photochemical smog levels.
In the eastern United States, sulfuric acid is the major component
of acid rain, comprising as much as 65 to 70% of the rain's acidity,
while nitric acid supplies only 25 to 30 percent. In the West, the
acids in acid rain are generally half nitric acid and half sulfuric
acid, although in some western urban areas, as much as 80% of a rain's
acidity can be comprised of nitric acid. Other acids can also contribute
to the acid rain problem. Hydrochloric acid, for example, may be emitted
directly from coal-fired power plants and is frequently found relatively
short distances downwind from such sources.
Acids may be deposited on earth not only by rain or snow, but also
through an atmospheric process called "dry deposition." This is the
process by which particles such as fly ash, as well es 862 and NOX, are
deposited onto surfaces. While these particles or gases are normally
not in the acidic state before deposition, it is believed that they are
converted into acids after contacting water in the form of rain, dew, fog,
or mist after deposition. The precise mechanisms by which dry deposition
takes place, and its effects on soils, forests, crops, and buildings, are
not adequately understood. Much research is being initiated to clarify
14
-------
the contribution of dry deposition to the overall add deposition
problem.
Another aspect of add rain that demands further study, and which
makes regulation of add rain a particularly tricky undertakinq, is
long-range transport. This phenomenon was first recognized in the
early 1970's. At that time, studies on the adverse effects of S02 and
sulfates on human health led to a stringent ambient air quality
standard for S02 as well as technological control of S02 emissions.
The associated control efforts forced the utilization of low sulfur
fossil fuels and scrubbers, and resulted 1n lower sulfur dioxide emissions.
Unexpectedly, however, reductions 1n urban S02 levels did not result in
proportional decreases in urban sulfates.
Several theories were offered to explain this development. One
explanation, the transformation-transport theory, was that reductions in
urban S02 emissions were offset by increases in rural S02 emissions from
new power plants located outside cities. S02 emissions from these power
plants, the theory held, had been transformed into sulfates and transported
over long distances to urban areas.
A project that was recently completed by EPA's Office of Research
and Development on sulfur transformation and transport seems to bear this
theory out. It found that sulfate aerosols could be transported hundreds
of kilometers from the initial S02 source. This validation of the trans-
formation-transport theory reinforces evidence indicating that the
15
-------
acidity of lakes in New York's Adirondack Mountains, for example, may
be caused by adds carried by winds from power plants as far away as
the Midwest.
Under certain conditions, it appears that sulfate and nitrate
compounds can stay aloft long enough to cross continents* oceans, and
international boundaries. This creates a situation in which the acid
rain in one country is caused by the emissions of another, but the
recipient of this damaging rain receives little or no benefit from the
source initiating the pollution. In a few short days, local problems
can become international 1n scope. This aspect of acid rain has caused
us problems with our northern neighbor; Canada receives two to four
times the amount of SOX that the U.S. gets from Canada, and the NOX
exchange 1s 11 times greater from the United States to Canada. Recent
negotiations between the two countries have been aimed at confronting
this problem. These talks are expected to evolve into a bilateral
transboundary air pollution agreement. And, through agencies like the
United Nations Economic Commission for Europe, the acid rain issue
vis-a-vis other countries may also be faced.
EPA is not alone in its efforts to uncover the causes of and the
solutions to the acid rain dilemma. Many government agencies as well as
private industry are participating in these efforts. In recognition of
the seriousness of the acid rain threat, the President, in his Second
Environmental Message, called for,a minimum of $10 million per year to
16
-------
be spent over the next ten years on a comprehensive add rain research
program. He also established an Acid Rain Coordinating Committee
consisting of seven Federal agencies to plan and coordinate a Federal
Interagency program. The Committee Is co-chaired by representatives
from the Department of Agriculture and EPA, and more recently* the
National Oceanic and Atmospheric Administration. As one of the co-
chairmen of the Federal Committee, I am pleased to note that the
federal agencies are now spending over $15 million for acid rain
research under the AEGIS of a cooperative research plan.
In addition to generating Information on add rain that can be
used to develop air quality control strategies and options, EPA has
another fundamental task: to qommunicate to Congress and the public
the effects of add rain, with particular attention paid to the ecologic
and economic consequences of continued high levels of add precipitation.
One tool to accomplish this communications function will be the
development of an "add deposition document," which David Hawkins, EPA's
Assistant Administrator for A1r, Noise and Radiation, and I are mapping
out. This document will be an attempt to quantify and quality, in a
preliminary way, the entire range of pollutants involved in acid rain
creation — sulfur, partlculates, nitrogen oxides, hydrochloric acid,
hydrocarbons and heavy metals.
The document will not be a "criteria document" in the sense that
1t will be used to develop ambient air .standards; rather, It will put
17
-------
the scientific evidence on add rain before the public so that It can be
discussed 1n an open forum, as well as serve as a focal point for future
add rain research. This document, we hope, will be an Important step
toward fostering public debate about how we as a country will meet the
add rain challenge.
We do know, at this time, that some of the methods currently being
used to minimize the local effects of S02 and NOX around large sources
are actually aggravating the acid rain problem. One method long favored
by power companies is the use of tall emission stacks. The rationale
behind tall stacks is that the emitted sulfur dioxide will be carried
away from the local community by winds. Unfortunately, the tall stacks
also keep the sulfur dioxide airborne longer, thus making sulfate
formation more likely.
As the mist that conceals the secrets of acid rain formation and
transport is gradually lifted, we will know better what control methods
will actually stop acid rain at its source, rather than passing the
problem on to someone else. At present, however, it appears that the
only practical approach lies 1n reducing SOX and NOX emissions. Many
innovative schemes have been suggested. There are studies underway to
estimate the costs of various ways to reduce emissions of these pollutants
and to compare these costs against acid rain damage costs, which are
only now beginning to be understood.
18
-------
For SOX control, FGD will probably remain our chief weapon through
at least 1985. As you will hear throughout this conference, this
technology can be applied to a variety of sources without imposing an
unreasonable financial burden. The use of low-sulfur coal is another
piece of the arsenal in the war against SOX emissions, along with the
array of technologies, both under development and on the commercial
market, designed to remove sulfur from fuel before it is burned. These
technologies include coal cleaning, coal gasification, and desulfurization
of liquid fuels. Then, there are also the combustion modification methods
that allow removal of sulfur during burning, such as fluidized-bed
combustion.
But, as we have seen, SOX constitutes only a piece of the acid rain
puzzle. NOX emissions can play an equally large role. And while we
have found ways to hold the lid on SOX emissions, we've only recently
begun to get a handle on NOX control. In fact, as coal use rises, we
expect that NOX emissions could increase by thirty to forty percent by
the year 2000, unless more effective control methods are developed and
quickly put to work by industry. At present, half the current NOX
emissions come from stationary sources; but by 2000, due to the trend
toward greater combustion of coal, stationary sources may be responsible
for up to 75 percent. Of the emissions from stationary sources, over
half are contributed by utility and large industrial boilers alone.
These large boilers now.emit an estimated 6 million tons of NOX every
year.
19
-------
The solutions that are so effective for SOX control aren't much
help when it comes to NOX control. Physical coal cleaning, which can
be used on some coal to reduce sulfur and ash content, has no effect on
coal's nitrogen content, because the nitrogen is chemically bound to
the coal. "Denitrogenation" — that is, chemically removing nitrogen
from coal — is prohibitively expensive at present, and at any rate does
not address the problem of thermal NOX, which is formed by molecular
reaction in super-heated combustion air. Flue gas treatment for NOX
control has been used with a fair amount of success in Japan on oil-
fired boilers, but there are major financial and technical hurdles to
applying that technology to coal-fired units. Even the coming age of
synthetic liquid fuels made from coal, which may consume 120 million tons
of coal in 1990 and 300 million tons in 2000, offers little hope for NOX
control — in fact, the concentration of fuel nitrogen may be increased
when coal is converted to a liquid.
However, there is a promising answer that is both cost-effective and
energy-efficient. By modifying the conditions under which combustion
takes place, an existing coal-fired power plant can reduce its NOX emissions
by 40 to 50 percent. When applied to new burner designs, combustion
modification may reduce NOX emissions by another two-thirds, yielding a
total NOX control of up to 85 percent. And, because combustion modifica-
tion involves changes only in burner design, the cost is quite small —
less than one-h&lf of one percent of the boiler cost. Further, because
20
-------
we are ensuring that the new burners are as efficient as the older designs,
the operating cost 1s nearly zero. EPA 1s aggressively developing low-NOx
burner designs.
Ideally, one technology would simultaneously control both of acid
rain's major components. This, in fact, is the idea behind a particularly
exciting new control technology, which may be retrofitted to many existing
coal-fired boilers with only minor modifications: the limestone injection/
multi-stage burner, or LIMB for short. The LIMB may be able to remove
50 to 70 percent of sulfur oxides at the same time that it reduces NOX by
50 to 80 percent. And 1t can accomplish this at a cost for $03 control
equipment of only $30 to $40 per kilowatt, as opposed to the average of
$150 per kilowatt that wet scrubbing requires.
Although the LIMB has only reached the bench/pilot scale stage of
development here 1n the U.S., Germany 1s currently operating a 60 megawatt
electric boiler using the technology, so we know that it works on a
larger scale.
The Idea of combining limestone injection for $03 control with a
low NOX burner 1s not a new one. In 1967, UOP, building on earlier
limestone Injection experiments by Combustion Engineering, injected
limestone Into an arch-fired burner, which is a naturally low NOX burner.
emissions were reduced by 50 percent at a stoichiometric ratio of 1:3.
The 60 megawatt prototype limestone injection, boiler in Germany,
which I mentioned earlier, has been operating for one year. It fires
v'
West German lignite, and utilizes flue gas recirculation to minimize
21
-------
peak temperature and NOX formation. At present, 1t is achieving 50 to
90 percent S02 removal at stoichiometric ratios of 2.5 to 5.0. Retrofit
capital costs for this technology are only $3.00 per kilowatt.
EPA has proposed a five-year research, development, and demonstration
program that will bring the LIMB technology up to commercial scale. In
the first year, EPA will characterize reactions and furnace conditions;
evaluate impacts on furnace operation; and test the technology with a
wide range of coal types and calcium-based sorbents. Next will come a
year of field evaluation, in which EPA goals will be to demonstrate
sulfur removal efficiency, optimize performance variables* determine if
there are any adverse boiler side effects such as slagging, plugging and
corrosion, and obtain design and cost data. Both wall-fired and tangentially-
fired units will undergo testing. Another year will be spent installing
the LIMB technology on full-sized boilers, which will then be subjected
to two years of performance optimization and long-term evaluation. The
development effort will be co-sponsored by EPA and the Department of Energy.
The total tab for the LIMB program will amount to $16.5 million, which
will be a bargain if LIMB fulfills its initial promise.
Industry as well as government must play a crucial role in the
development of methods to control acid rain. EPA has the resources to
provide the fundamental research and the testing of new control technologies,
but we must rely on industry to provide the host sites that allow tech-
nologies to be tested under real-life conditions. And, we must depend
22
-------
heavily upon the commercial expertise and engineering experience of
boiler manufacturers if a technology is to progress beyond the demon-
stration stage.
Now there's always an element of risk for the private sector
when it invests in new equipment and new technologies. Control processes
that look promising on the drawing board or durinq small-scale experiments
don't always pan out when they are put into practical use. But we at
EPA truly believe that with the kind of cooperation between government
and industry we have enjoyed up to now, and with continued joint effort,
we can solve the acid rain control challenges we face.
With a better understanding of what causes acid rain and with the
necessary control technology under development, we will be able to
begin making strides in the regulatory arena....to pull in the "reins,"
if you will forgive me, on acid rain. As the Clean Air Act stands now,
there are no regulatory requirements concerning acid rain per sj&. As
most of you are aware, this Act comes up for revision next year, and EPA
is consulting with other Federal agencies on the possibility of changes
that would better address the acid rain issue.
The Clean Air Act is currently structured around a presumption
that air pollution can be related to a particular source or a well-defined
group of sources. But, in the case of acid rain, there is no clear-cut
relationship between specific emissions and the acid rain. In other
even though the types of emissions that lead to acid rain are known, it
23
-------
1s currently not possible to accurately trace individual emissions that
cause acid rain back to their origins. And, while the Clean Air Act
has been amended to address the problem of interstate pollution, any
given state is only able to enforce its emission limitations against
sources within its own boundaries. A state can petition the EPA
Administrator if it feels that another state is preventing it from
attaining a national standard or otherwise causing a deterioration in
that state's air quality, but then EPA is faced with the problem of
how to demonstrate that one or several out-of-state sources are
responsible for impermissible air quality violations. Such a demon-
stration would be hard, if not impossible, to make, especially if a
number of sources from several states or nations were involved.
One regulatory option that EPA is reviewing is the development
of national ambient air quality standards for nitrates or sulfates,
two precursors of acid rain. However, it Is not clear whether there
is sufficient data on which to base such a standard. Even if the
data were available, the standard-setting process is a lengthy one.
It would probably be five to ten years before any emission reduction
could be achieved. Other near-term options include: better monitoring
of S02 emissions to improve enforcement of existing standards; the
establishment of federal regulatory requirements for review of interstate
Impacts of State Implementation Plan provisions; or the establishment
of new source performance standards for pollutants for which EPA has
not set ambient standards, such as total sulfur.
24
-------
A longer term option might Involve the Conaress setting
regional SOg and NOX emission reduction goals — say 5 to 10 percent
per year — goals which would be administered on a multi-state basis
and would allow the utilities and Industries to meet the goals on a
system-wide basis using the most cost-effective combination of
approaches — coal washing, combustion modification, load shifting
to cleaner plants, fuel shifting, and early plant retirements, to
name a few.
Whatever path we choose, however, we must be mindful of the
need to consider the regulatory burden imposed on the utility or
Industry and the ratepayer or consumer. In addition, we must fully
support the national energy policy of expanded coal use, and be
sensitive to the fact that the economy cannot regain its vital growth
without the atd of a vigorous industrial base. These are "mighty
tall" orders, as they say, for the Government and the industrial sector.
But then few people really believe that anything worth doing in this
country is going to be easy. Why should reconciling environmental ar.c
energy goals, a priori, be any easier than, say, reconcilings energy
goals and national security, or inflation and,unemployment objectives.
There are no easy answers, only a nation of differing but robust people
trying to work out their future.
25
-------
THE NATION'S ENERGY FUTURE - WITH FOCUS ON SYNFUELS
Frank T. Princiotta
Director, Energy Processes Division
United States Environmental Protection Agency
Washington, D. C.
ABSTRACT
Projections indicate that coal, nuclear energy and oil shale wil"i
become increasingly important as we adjust for static domestic oil
and gas production and minimization oil importation. Environmental
problems can be quite severe for each of these fuel cycles. A massive
synthetic fuel industry based on coal, oil shale and biomass, is
emerging with monumental potential for environmental damage. The
Environmental Protection Agency (EPA) has designed a regulatory program
aimed at mitigating environmental damage while allowing for birth and
nurturing of this critical industry.
Preceding page blank
27
-------
THE NATION'S ENERGY FUTURE - WITH FOCUS ON SYNFUELS
OUR ENERGY FUTURE
America is making progress in minimizing dependence of imported oil.
During the first five months of 1980, gasoline consumption decreased 8.1
percent — compared with the same period last year - and crude oil imports
decreased to 7.8 million barrels per day — the lowest level in four years.
Petroleum stockpiles are at capacity levels nationwide due to a very real,
conscientious effort to conserve energy in all areas: electricity, home
heat and transportation fuel.
Of even greater significance is passage of the Energy Security Act,
signed by President Carter in June of 1980. This bill will promote conser-
vation, increase production of coal and oil, and help harness the power of
the sun, wind and rivers and most importantly spawn a major synthetic fuel
industry based on coal, oil shale and biomass. All of these measures can
serve as effective remedies against further reliance on costly and uncertain
supplies of foreign oil.
To achieve the necessary growth in domestic energy resource development
to meet our future production goals, a substantial increase in extraction,
processing, transport and use of domestic fossil fuels must take place. EPA
has recently made projections attempting to predict our nation's energy future
using the Strategic Environmental Assessment System (SEAS) model and an EPA
sponsored study projected synfuel production. These projections suggest that
coal, oil shale and nuclear energy will allow for the nation's economic growth
despite the leveling off of domestic petroleum and natural gas and without
increasing oil imports (Figure 1-4). For example, the amount of coal mined
in this country must expand from the current 700 million tons annually to
1.1 billion tons in 1990 to 1.6 billion tons in 2000. The production of
synthetic liquid fuel and gas from coal is expected to consume 80 million
tons by 1990 and 350 million tons in 2000. We can also expect that the 1980's
will see the oil shale industry emerge as a significant supplier of fuel,
producing up to 300,000 barrels per day by 1990 and 2.2 million barrels per
day by 2000.
Such projections indicate a trend away from traditional and less environ-
mentally damaging energy sources, toward potentially more damaging fossil fuel
sources such as coal (particularly from western surface mines), oil or gas
from the Outer Continental Shelf, and western oil shale. The trend also
points to the increasing use of nuclear energy to generate electricity and
indicates an increasing interest and use of solar and geothermal energy.
These major shifts toward increased use of less clean fuels can pose a
significant threat to human health and the environment. Potential negative
impacts are likely to result from the extraction, processing and utilization
phases of each major fuel (Figure 5). For example, increases in coal and oil
shale mining can create erosion and subsequent surface water siltation problem?
28
-------
domestic fossil energy
resource requirements
40-
ro
CJ
"o
W
cr
W = western mining
ES = eastern strip mining
EU = eastern underground
OF = offshore—lower 48 states
OS = onshore—lower 48 states
A = alaskan
NOTE: Projected Oil Imports
1975: 17 QUADS
1990: 14 QUADS
2000: 10 QUADS
17.8
1975/1990/POOO
1975/1990/2000
oil
1975/1990/2000
gas
1975/1990/2000
oil
-------
FIGURE 2
U. S. non-fossil energy
resource requirements
CO
o
>.
®
(Q
3
cr
1.8
10.7
I
1975/1990/2000
uranium
3.2 3.2
S = solar
G = geothermal
B = biomass
1.5
S
B
IWM1II f%
S
1975/1990/2000
hydroelectric
1975/1990/2000
other
-------
components of
domestic coal
25
20-
o>
o>
(0
TJ
nj
3-
cr
10 —
0
20.6
16.3
9.1
CM
CM
CM
1975/1990/2000
electric
utilities
CK
4.6 5.7
4.7
CK
CM
CK
CM = combustion
CK = coking
LI = liquefaction
GA = gasification
8.0
-|-jCM
LI
1.7
0
LI
GA
1975/1990/2000
industrial
1975/1990/2000
conversion
-------
FIGURE k
alternative fuels
production: 1985/1990/2000
thousands
of bpdoe
1200-
1000 -
800 -
600 H
200 -
high
low
1985/1990/2000
high BTU gas
high
low
high
Tow
high
low
high
high
• low
1985/1990/2000
low/medium*
BTU gas
1995/1990/2000
Indirect
1985/1990/2000
direct
1985/1990/2000
oil from
oil shale
1985/1990/2000
oil from
tar sands
1985/1990/2000
5,060 I
hlghf
2,965}
low
low
1985/1990/2000
from coal gasification
elhanol from total production
blomass &
Industrial wastes
liquids from coal liquefaction
Source: Hauler, Ballly & Company Alternative Fuels Monitors: Coal Gasification and Indirect UquefaclHtn: Oil from Shale.
-------
lifted U.S. domestic fuel flow: 199O
PRIMARY FUELS
EXTRACTION
c" ' '
CONVERSION/
TREATMENT
PROCESS
SECONDARY FUELS
UTILITY
ELECTRIC POWER
TRANSMISSION/
TRANSPORT
UTILIZATION DEVICES
NUCLEAR
POWER
PLANTS
ELECTRICAL
DEVICES
ELECTRICAL
POWER LINES
DIVERTED TO
PETROCHEMICAL
?rv INDUSTRY
^
COAL POWER
PLANTS
LIQUID/GASEOUS
DEVICES
®®®J7Ffift
WATER I IK? LAHD
&•
-------
groundwater quantity and quality are also likely to be affected. Processing
coal and oil shale to synthetic liquids and gases may yield toxic emissions
and large quantities of solid wastes; and despite current regulations, an
increase in coal combustion will result in increased production of nitrogen
oxides, sulfur oxides and solid wastes (Figure 6). The environmental and
safety uncertainties surrounding the use of nuclear energy have been well
publicized.
Many of the adverse impacts on health and environmental quality, however,
can be controlled or avoided: Most mined land can be reclaimed; particulate
matter and the oxides of nitrogen and sulfur can be scrubbed from flue gas;
acid precipitation and its effects on agricultural and forest production can
be reduced.
EPA has an impressive array of legislative tools available to control
air, water and land pollution from energy and industrial sources (Table 1).
The agency will face the monumental challenge of utilizing these mandates
to achieve maximum benefit of minimum cost.
Controlling these pollutants increases the monetary costs of energy, but
failure to control them lowers the productivity of our natural resources,
degrades the quality of our environment, and imperils the health of our
'population.
Focus on Synthetic Fuels
As the projections suggest our energy future should be characterized
by a massive synthetic fuel industry by the year 2000. Although oil shale
plants will be limited to a relatively limited area (Figure 7) coal gasifi-
cation and liquefaction plants could be constructed anywhere large quantities
of coal are located (Figure 8). Ethanol plants will be initially sited in
corn and wheat farming areas (Figure 9) but could eventually proliferate as
other crops and agricultural wastes become feasible as feedstocks (Figure 10).
oo Synfuel Environmental Issues
Synthetic fuels processes are receiving our most serious
attention because synfuel development activity is clearly
intensifying, because of our concern over the unknown nature
of the pollutants which may be generated, and because of EPA's
recognition that the enormous capital outlays involved in
building these facilities during the next decade dictates the
earliest possible and most stable possible environmental
regulations for this new industry. It is expected that
pollutants coming from coal conversion and shale oil production
will be more diverse in composition than those produced by
direct fossil fuel combustion. The burning of fossil fuels
in conventional processes involves complete oxidation (or
attempts threat) whereas synthetic fuels are produced under
34
-------
FIGURE 6
growth In emissions/wastes
from stationary sources
to
millions
3001 of tons
250'
200 i
150
50
net emissions (after treatment)
1975 1990 2000
•TOTAL SUSPENDED
PARTICULARS
1975 1990 2000
*SULFUR OXIDES
1975 1990 2000
•NITROGEN
OXIDES
1975 1990 2000
UTILITY COAL ASH/
SLUDGES
1975 1990 2000
O!L SHALE
WASTES
"energy
-------
TABLE 1
AIR, WATER AND SOLID WASTE ENVIRONMENTAL
LAWS IMPACTING FOSSIL ENERGY FACILITIES
RELEVANT AUTHORITY
IMPACT
Clean Air Act Amendments of 1977
• Set New Source Performance Standards
(NSPS) for energy industries (Section 111).
• Set National Emission Standards for Hazard-
ous Air Pollutants (NESHAP) for selected
industries (Section 112).
• Implement Prevention of Significant Deteri-
oration (PSD) Program (Section 160).
• Achieve Ambient Air Quality Standards (Sec-
tion 109).
• Set Lowest Achievable Emission Rates
(LAER) (Section 171).
* NSPS set for fossil utility boilers; industrial
boiler NSPS being developed; oii shale, coal
gasification, and liquefaction in planning stage.
• NESHAP requirements for synthetic fuels
industry being evaluated as process plans become
firm.
*
• PSD permits required for all New Sources
(coal-fired boilers and synthetic fuels plants) to
prevent increases in paniculate and SO, levels in
areas having good air quality.
• Require utilization of appropriate control
technology to reduce emissions to levels required
to meet State Implementation Plan (SIP) goals.
« Require level of pollution control technology
greater than that which would normally be
required by SIP for plant siting in non-attaiment
areas.
Federal Water Pollution Control Act Amendments of 1977
• Set discharge limits based on best conventional
technology for energy industries (Section 306).
• Set discharge limits based on best available
technology for toxic pollutants (Section 307).
• Issue and enforce discharge permits to achieve
above limits and to meet water quality standards
(Section 402).
• Effluent guidelines for steam-electric industry
issued, industrial boilers must meet guidelines for
specific industry; effluent guidelines being planned
for oil shale and coal gasification and liquefac-
tion facilities.
• For designated toxic pollutants best available
control technology will be required, and will have
greatest impact on the design of synfuel plants.
• Permits for electric utility plants and other
industries being issued based or, effluent guide-
lines, permits for synthetic fuels plants will be
issued on basis of besi information available until
guidelines are issued.
Safe Drinking Water Acl of 1974
• Review projects for possible danger to under-
ground drinking water supplies (Section 1424).
• All projects receiving federal assistance will be
reviewed for processes impact on groundwater
quality as it may impact drinking water.
Resource Conservation and Recovery Act of 1976
• Set criteria for defining hazardous waste
(Section 3001).
• Define accepiable disposal practices for
hazardous wastes (section 3008). '
• Set guidelines for non-hazardous waste
disposal (Section 4004).
• Proposed procedures for determining if wastes
are hazardous have been issued.
e Utility wastes and spent oil shale classified as
"special" wastes; if hazardous, they must meet
monitoring requirements but not disposal
requirements; best economically attainable
disposable technology will be defined.
a Disposaj guidelines for non-hazardous utility
waste will be completed in 1981, other energy
wastes subject to state guidelines.
36
-------
FIGURE 7
some oil shale development sites
OJ
oil shale deposits of the Green River formation
p!anr:fcd wod potential oil sha!e projects
Wyoming
SOURCE: Office •
cyy Asses;;r.-?ni
-------
FIGURE 8
to
CD
sites for coal-derived
alternative fuels plants
potential for coat-derived alternative
fuels development
high potential for coal-derived alternative
fue*& development
SOURCE: U.S. Bu'oao of. Mines
-------
FIGURE 9
potential sites for large-scale
ethanol production by 1985
co
uo
[JUJ corn
fcd~! wheat
SOURCF. Hagler, Bailiy R C-oi.ipany based on information from O-
of Technology Assessment, U.S Department of Energy. US Depi'
of AgrlcuHure
-------
FIGURE 10
potential sites for large-scale
ethanol production by 2000
sweet sorghum
agricultural & food wastes
corn & agricultural & food wastes
sorghum & agricultural & foc:d wastes
com & sorghum & agricultural & food wast
SOURCE: Hagler, Bailly & Company, bdsed on Information horn Office
of Technology Assessment, U.S. Department of Energy. U.S. Department
of Agriculture
-------
reducing conditions using less air than is required for complete
combustion. The result is that a wide variety of high molecular
weight organics, reduced sulfur compounds, and other potentially
toxic compounds are formed, presenting a different array of
pollutants than have been dealt with in the past.
We believe the air pollution problems may be particularly
serious. The synthetic fuel industry is expected to produce
a wide range of air emissions with potentially adverse environ-
mental effects if not adequately controlled. Oil shale retorting,
for example, will emit nitrogen oxides, sulfur oxides, reduced
sulfur species, ammonia, various volatile and partially oxidized
organics and, of course, particulate matter. The Prevention of
Significant Deterioration increments available may well pose
serious problems. The air pollution problems associated with coal
gasification and liquefaction are similar in many ways to those for
oil shale. These processes can generate significant quantities of
particulates, sulfur compounds, trace metals, high molecular
weight hydrocarbons and nitrogen oxides, etc. The sulfur species
may be particularly troublesome.
Water-related environmental problems from synfuel production
may be just as complex. The oil shale industry will need copious
water supplies for cooling compaction of spent shale, and for
revegetation of surface mined areas. Coal mining and coal conver-
sion will also have substantial water requirements for process
uses and revegetation. Supply of water for these activities will
be particularly crucial at some sites in the arid western part of
the country where oil shale retorting and some mine-mouth coal
conversion will occur. At other sites, mine dewatering and retort-
produced water from shale oil production will produce excess water.
Among the water pollution problems of concern, spent shale, if not
properly handled, could create serious water quality problems from
the leaching of soluble contaminants into nearby ground or surface
water. With underground, modified "insitu" operations being
considered for oil shale, and possibly for coal, the opportunity for
groundwater contamination is even more likely than for surface
operations. Here again, the problem is particularly serious in
the western part of the country where groundwater is a vital resource.
From all types of synthetic fuel operations, raw process water
discharges will be highly contaminated by toxic materials (most likely
including carcinogens, mutagens, etc.) which would represent major
threats to both surface and groundwaters if not properly controlled.
It is expected that synfuel facilities will utilize process water
recycling to a great extent but this may not totally solve the water
pollution problems at all locations.
There are a variety of synfuel-related solid waste problems as
well. Both oil shale mining and coal mining produce enormous amovats
of solid waste. Many of the mining problems are similar to those
41
-------
encountered with conventional coal mining and can be solved
similarly. Surface reclamation techniques for strip mined
areas are particularly successful at least where an adequate
water budget exists. The solid residues of oil shale retorting
and coal conversion are, however, another problem. Shale oil
production, for example, produces spent shale that is greater
in volume than the shale originally removed from the ground;
coal conversion technologies, both gasification and direct lique-
faction, will produce vast quantities of ash. Each of these
wastes will most likely contain a wide variety of potentially
harmful components and will have to be properly managed. Some
special wastes from synfuel plants such as spent catalyst from
coal conversion may be classified as "hazardous" under the
Resource Conservation and Recovery Act.
There is also concern about the possible toxicity of liquid
synthetic fuels themselves, both from the handling and usage
standpoints, including concern for both industrial employees and
the general public. Coal-derived liquid fuels, particularly those
produced by direct liquefaction, are of the most concern. These
liquid fuels are not of the same composition as ordinary crude oil
products. They are higher in nitrogen content, yielding higher
NOX levels upon combustion and they tend to contain more substances
which are potentially mutagenic or carcinogenic so that public
exposure to them through normal usage might represent a significant
health problem. More data are needed., however, on both conventional
petroleum products and synthetic fuels in this regard.
oo Pollution Control Guidance Documents - Part Of The Agency's Regulatory
Strategy
Regulating new, presently non-existent energy industries, of
course, presents different problems from regulating long-standing
segments of United States industry. The differences are of such
an extent that a unique regulatory approach is demanded. The
differences arise primarily from the facts that the new energy
industries are, for the most part, not yet commercialized in the
United States, have potentially different effluents and emissions
from those from existing pollution sources and are being developed
on a telescoped.time frame under a governmentally-mandated response
to "the energy crisis."
Because of these circumstances, the general approach we are
taking is to issue, as preregulatory multi-media guidance, a series
of Pollution Control Guidance Documents, PCGDs—one for each of the
major energy technologies. The focal point of each PCGD is to be a
set of available control alternatives for each environmental discharge
(again, for all media) along with associated performance expectations
42
-------
and the basis for the alternatives presented. The intent is to
present guidance for plants of typical size and for each signifi-
cantly different feedstock likely to be used. PCGDs will not have
the legally binding authority of regulations but each will be
reviewed extensively both within and outside of EPA. These documents
will provide useful and realistic guidance to permit writers within
EPA and the States and to the energy industry itself during its
formative stages. As the energy industry develops, permits for
individual installations are being issued based on best engineering
judgment and, as the various PCGDs become available, permits will be
prepared in light of the information the PCGDs contain. Then, as the
energy 'industries mature and as .large-scale control technology data
become available, EPA will invoke its legally-binding regulatory
procedures, but in a coordinated, multimedia fashion; in the water
quality area, for example, this would mean the issuance of effluent
guidelines and establishment of appropriate water quality standards,
including consideration of related air quality and hazardous waste
requirements.
oo Processes To Be Covered
Although the major objective of a PCGD is to recommend pollution
control options, it will contain a great deal of background information
on the energy processes themselves and on process streams and pollutant
concentrations, and will, on the basis of a series of "case studies,"
offer specific technology-based control guidance for various kinds of
energy processes. Processes to be included will cover those that are
expected to be built for demonstration or commercial application first.
(Table 2 shows planned process coverage for the four PCGD's currently
being written). It is intended that discussion of product (e.g., low
Btu coal gas) uses also will be included if use is integral with the
manufacturing process. The process descriptions will detail the key-
features of each process and their pollution potential. If various
process modifications are likely to be used, the changes in process
configuration will be covered and expected changes in pollutant
releases will be indicated. Pollutant releases that vary non-linearly
with plant size or flow rates will also be identified and quantified
to the extent possible.
The environmental control alternatives to be considered will
include both end-of-pipe treatment techniques and process changes.
Candidate control alternatives will be identified from existing
United States and foreign bench-, pilot- and commercial-scale
facilities or from different United States or foreign processes
that have similar discharges. Performance and design will be
included as will information on capital, operating an annualized
costs. Energy usage for control alternatives will also be included.
Finally, techniques for monitoring control performance will be
identified. The source of all data will be clearly referenced to
allow referral to original sources; uncertainties in the data will be
indicated.
43
-------
TABLE 2
TECHNOLOGIES FOR WHICH PCGDs ARE CURRENTLY PLANNED
Indirect Liquefaction
Lurgi
Texaco Gasifiers
Koppers Totzek (K-T)
Fisher Tropsch
Mobil-M Conversion Systems
Methanol
Oil Shale
Occidental
Rio Blanco
Lurgi
Paraho
Union
Colony
Lou Btu Gasifiers
-single bed, atmospheric, entrained
gasifiers with and without sulfur control
Medium/High Btu Gasifiers
Lurgi
K-T
Texaso
Others to be decided
44
-------
oo Penult Processing
Various action's have been taken which are aimed at expediting
permits on energy facilities issued by EPA. We have set up our own
Permits Coordination Group to carefully track permits on all energy
installations, including the important synfuels ones. The Group
will identify potential processing problems early and enable
appropriate remedial action to be taken almost immediately. We have
designated a single person in each of our Regional Offices to serve
as a special point of contact for new energy facilities. These
individuals have responsibility for assuring that timely review of
permits for new energy facilities takes place, that industrial permit
applicants are well informed as to when EPA will make decisions.
Industry, especially the small and medium-sized firms, has responded
very positively to this concept.
We now set target dates for permit processing based on the
requirements of individual permit applications. The complexity of
individual cases varies considerably and by tailoring the review
schedule to each individual case, a much shorter average turn-a-round-
time can be achieved than if a general schedule sufficient for all
applications is used. For surface water discharge permits involved
with surface mining of coal, a memorandum of understanding is being
developed with the Department of the Interior's Office of Surface
Mining (OSM) . With this arrangement, OSM could issue a single permit
under an agreement with EPA that OSM's comprehensive review procedure
would also meet EPA's legislative requirements.
EPA has already issued several air pollution control permits for
oil shale development. This early group of permits includes the Colony
Development Operation of Exxon and TOSCO Corporations, the first
commercial-scale shale retorting facility for which a permit has beer.
granted in the United States. EPA's permit will eventually allow Co_o.Ty
to expand and produce 46,000 barrels per day of low sulfur distillates
and other by-products. The permit will also allow Colony to construct
and operate: (1) a 66,000-ton/day underground oil shale mine, (2) a
surface oil shale retorting facility and (3) extensive support facilities
including a 194-mile pipeline and a loading terminal. PSD permits have
also been issued for the non-commercial-scale projects of Union Oil,
the C-b tract (Occidental and Tenneco), and Rio Blanco (Gulf and Standard
of Indiana). Another synthetic fuels facility which has received
a PSD permit is the Great Plains Gasification Associates Coal gasification
plant in North Dakota. This commercial facility will produce 125
million standard cubic feet per day of high Btu synthetic fuel gas.
Finally, a recent development in regulatory procedures to expedite
permitting is the Consolidated Permit Program (4). The new consolidated
permit regulations combine the requirements for the following five
programs covered under four different Federal environmental laws:
45
-------
o the National Pollutant Discharge Elimination System (NPDES)
program of the Clean Water Act;
o the Prevention of Significant Deterioration (PSD) program of
the Clean Air Act (but only where EPA itself is the permitting
authority and only to specify permit procedures);
Drinking Water Act (SDWA);
o the Hazardous Waste Management program under the Resource
Conservation and Recovery Act (RCRA);and
o the Dredge and Fill (Section 404) program under the Clean Water Act
The consolidated permit regulations and associated application forms
provide a framework for simultaneously processing multiple EPA permit
applications for the same facility. Standard information can be
provided on a single form along with information required for specific
permiting activities. Also, where appropriate, EPA has the ability
to consolidate draft permits, public notices, public hearings and
administrative records for all permitting activities for the facility
or activity. These procedures should not only expedite the permitting
process but also provide an opportunity for better comprehensive
assessment of multimedia environmental control. The results should be
more consistent and more efficient control requirements.
THE RESEARCH PROGRAM
EPA1s energy and environmental research program is based on the
belief that increased domestic energy production need not come at the
cost of a deteriorating environment and threats to public health and
welfare. The Federal Interagency Energy/Environment Research and
Development Program was established to provide the information necessary
to develop a scientific rationale for policies that strike a balance
between ample domestic energy production, reasonable cost and
environmental quality. This interagency effort is divided into two
major research programs: health and environmental effects, and control
technology.
The health and environmental effects program is designed to
identify energy related pollutants in the environment, the mechanisms
by which they move through the environment and their resulting effects
on human, animal and plant populations.
The control technology program provides information on the types
and quantities of pollutants released by energy supply activities and
develops, or stimulates the development of, control options where
necessary. A major thrust of research in the control technology program
is the generation of technical and cost information on which reasonable
environmental standards can be based.
EPA's research program emphasizes the generation of data
necessary to support the establishment arid implementation of technology-
based environmental guidelines. This information will be used to assist,
and ultimately minimize, environmental damage resulting from a broad
46
-------
of energy fuels and processes. Those systems judged to have the
greatest potential for near-tearm negative impact will receive study priority.
Over the next five years, the focus of the research program will be
on the current and projected coal fuel and oil shale cycles. Over the next
fifteen years, coal and oil shale production and use are expected to grow
faster than any other fuel source, and they both demonstrate the potential
for creating major environmental problems throughout the fuel cycle. In
addition, coal is expected to be the dominant fuel employed for electricity
production and will be used increasingly as a feed stock for synthetic
liquids and gases.
47
-------
FGD ECONOMICS IN 1980
By
G. G. McGlamery, W. E. O'Brien, C. D. Stephenson, and J. D. Veitch
Division of Energy Demonstrations and Technology
Office of Power
Tennessee Valley Authority
Muscle Shoals, Alabama
ABSTRACT
Presented in this paper is a review of recent results from EPA-
sponsored flue gas desulfurization and byproduct/waste disposal economic
evaluations prepared by TVA. Included are a summary of comparative capital
investments and annual revenue requirements from a three-phase effort to
evaluate the leading FGD processes, and similar results from three phases
of sludge disposal studies. Data from a 1985 projection of FGD byproduct
sulfur/sulfuric acid marketing potential are given.
A new series of FGD process evaluations is also previewed including
a set of updated evaluation premises which will be utilized in the early
1980's. Examples of the effects of the revised premises on limestone
scrubbing economics are shown. Finally, results are provided from a recent
evaluation of limestone scrubbing in a spray tower using adipic acid,
forced oxidation, and gypsum disposal by stacking.
Preceding page biank
49
-------
FGD ECONOMICS IN 1980
INTRODUCTION
Through the publication of numerous studies sponsored by EPA and
other organizations, a great deal of understanding and a broadened
perspective of FGD economics have been developed during the past decade.
As we enter the 1980's, interest in FGD economics continues as strong as
it was 10 years ago. Changes in technology, environmental regulations,
economic conditions, and design philosophies all affect the projection
of FGD economics to such an extent that constant reassessment is necessary.
Neither the pace nor the effects of these changes can be expected to
diminish soon.
The interagency EPA-TVA program to evaluate FGD economics that began
in 1967 is now well into its second decade of activity. Projects to
evaluate the economics of leading nonrecovery and recovery FGD processes,
waste disposal processes, coal-cleaning systems, and byproduct marketing
studies have all been a part of this program. Results from much of this
work have been reported at earlier symposiums.
During 1980, additional results have been derived from the continuing
program. This paper summarizes most of the recent published data and
work in progress. First, a summary of results from three reports on
comparative FGD process economics is presented. Second, a summary of
information from three published reports on sludge disposal economics is
given. All six of these reports utilize the same time frame (1977-1980)
and design and economic premises. Reported next are the data from a
1985 projection of FGD byproduct sulfur/sulfuric acid marketing.
A new series of FGD process evaluations was begun in 1980 using an
updated set of design, regulatory, and economic premises more typical of
conditions to be faced in the early 1980's. Evaluation projects using a
costing time frame of 1981-1984 are previewed on dry scrubbing processes,
limestone process alternatives, gypsum-producing processes and ash
disposal systems. The new premises are also described, as is a stepwise
conversion of limestone scrubbing economics from the old premises to the
new premises.
In the final portion of the paper, results are projected for an
advanced limestone scrubbing process using a spray tower, adipic acid
additive, forced oxidation, and gypsum stacking. This particular evalua-
tion is for a limestone system expected to come into common usage in the
future if scheduled large-scale process development is successful and
environmental acceptability is proven.
50
-------
Because the results presented herein are from a variety of studies
using different premises, special caution should be exercised in utilizing
the results. Particular attention should be paid to the different
designs evaluated for the limestone scrubbing process.
FGD ECONOMIC STUDIES
In 1977 TVA began a series of FGD economic studies designed for the
twofold purpose of updating previously evaluated processes and integrating
evolving technologies into the EPA-TVA FGD economic studies. Three
reports (1,2,3), two of which have been published, covering seven FGD
systems and two processes for producing sulfur from FGD S02, have been
prepared. The limestone and lime scrubbing processes were updated from
an earlier report, as were the magnesia and Wellman-Lord scrubbing
processes (4). A generic double-alkali process was included to represent
this important type of nonrecovery FGD process. The citrate process and
the Rockwell International aqueous carbonate process (ACP) were included
as emerging sulfur-producing processes. The ACP represents two areas of
new FGD technology, spray dryer FGD and the use of coal as a reducing
agent to produce sulfur. The latter technology was also represented in
this series of studies by the Foster-Wheeler Energy Corporation Resox®
process and the Allied Chemical coal/S02 reduction process, both of
which utilize coal to produce sulfur from S02- Schematic flow diagrams
of all the processes evaluated in this series are shown in Figure 1.
These processes represent a range of development from established
technology (the limestone and lime), through demonstration and recent
commercialization (the double-alkali, citrate, magnesia, and Wellman-
Lord scrubbing processes), to less-developed processes (the ACP and the
Resox® and Allied coal reduction processes).
The same premises, based on a 500-MW power plant burning 3.5%
sulfur coal, meeting the 1.2 Ib S02/MBtu NSPS, and using mid-1979 capital
costs and mid-1980 annual revenue requirements, were used throughout.
As in other EPA-TVA economic studies, these base-case conditions were
systematically varied to evaluate different fuel, power plant, and FGD
conditions. In all, over 100 case variations of 9 basic FGD processes
were evaluated. In addition, in recognition of the growing importance
of energy in design considerations, a ground-to-ground energy evaluation
was made for some of the processes.
Prbcess Descriptions
The limestone, lime, and double-alkali processes produce a waste
slurry that is disposed of in a pond. In the limestone process the flue
gas is scrubbed with a slurry of ground limestone, forming calcium
sulfur salts that are discarded by pumping a purge stream to a disposal
pond. The lime process is similar except that a slurry of lime, is used
as the scrubbing medium. .In the double-alkali process a solution of
51
-------
LIMESTONE S LIME PROCESSES
SCRUBBER
FLUE.
GAS
ESP
,,
1
, , f"\ ? *• TO STA
RFHFAT ~J
i iMr^Tnwr
0, iiDRY ^— LIMLo ' UNL
* PREP. UK L'ML
i >
1 L-t) s-^l
POND
DOUBLE-ALKALI PROCESS
SCRUBBER
FLUE
GAS
ESP
w
\
REHEAT
f~'
^ ABSORB.
"• REGEN ""
A I
T 1
1 <\T-
-^— ~
Ja2C03
, SLURRY
PREP
•• IU 3 IA
m LIME
OR
1
POND
MAGNESIA PROCESS
PRE-
FLUE _^«
GAS ^t
ESP
w yv >
s
CRUBBER SCRUBBER
4
REHEAT
SLURRY
PREP.
•TO STACK
MgO MAKEUP
CHLORIDES
1
DRYER
i
*MgO
CALCINER
so2
ACID "1
PLANT j
WELLMAN-LORD PROCESS
FLUE
GAS
ESP
PRE-
SCRUBBER SCRUBBER
_NQ2S03
CHLORIDES
[SULFATE
I PURGE
EVAR
REGEN.
•TO STACK
•Na2CO,
TO END PLVJ
(ACID FLAMT, RL'
OR AL- 'iLLi C\
Figure 1. FGD process flow diagrams.
52
-------
CITRATE PROCESS
SCRUBBER
TO STACK
Na2S04
GAS
CHLORIDES
AQUEOUS CARBONATE PROCESS
Na2C03 SOLUTION
CYCLONE SPRAY DRYER 1
rLUt *
GAS ESP
V L J
Y ^^^ .
FLY ASH
W!
RESOX UNIT
R
ANTHRACITE — *
S02 — *
_£_\ y-^-TO STACK
CO/
T
rRE;
P
C/l
•— *01
\L r^pf
1
1 L
RE- — '
•F GAS ..CHLORIDES
ACCESS, a SOLIDS
*C02
fc CARBON.
DECOMP.
H2S
CLAUS
/
Na2C03
— *• SULFUR
*
SOLIDS
E ACTOR
•
»
— *• CONDE
fcTAIL GAS
.N. TQ FGD
SULFUR
Y
CHAR
ALLIED UNIT
COAL-
L»J I m
REACTOR r\ )
Figure 1 (continued)
f
S02
SOLIDS
CLAUSTL^TAIL GAS
UNIT | TO FGD
SULFUR
53
-------
sodium sulfite is the scrubbing medium. The spent solution is regenerated
by adding lime, producing calcium sulfur salts that are discarded^in a
disposal pond. A slurry of magnesium oxide is used as the scrubbing
medium in the magnesia process. The spent slurry is dewjatered, dried,
and thermally decomposed to regenerate the magnesium oxide and produce
SO-> which is converted to sulfuric acid in a conventional acid plant.
Z, i
The citrate process is a wet scrubbing process usinfe a sodium
citrate solution as the absorbent. The absorbent is regenerated and the
SOX compounds reduced to elemental sulfur by liquid-phase reduction
using H2S. The H2S is produced by reducing some of the product sulfur
using natural gas.
In the Wellman-Lord process a solution of Na2S03 is the scrubbing
medium. Reaction with SOX produces NaHSC>3 which is heated to evolve SC>2
and regenerate Na2SC>3. Other sodium compounds, primarily Na2S04, form
and must be removed. Unlike the magnesia process, which produces a
dilute S02 off-gas, the Wellman-Lord process produces an SC^-rich off-
gas more suitable for direct reduction to sulfur. In these studies it
is evaluated with a sulfuric acid end plant and with the Resox" and
Allied coal reduction processes.
The Resox® process consists of a vertical reactor through which
rice-sized anthracite flows by gravity at a controlled rate. The S02~
rich off-gas is mixed with controlled amounts of water and air, heated,
and passed through the reactor. In a complex series of reactions some
anthracite is oxidized to maintain the reaction temperature and most of
the SC>2 is reduced to sulfur. A noncaking coal such as anthracite is
necessary. Careful control of residence time, temperature, and SC^rl^O
ratio is necessary to limit the thermodynamic tendency of the sulfur to
go to H2S. Sulfur is condensed from the emerging gas and the remainder
is burned to convert the sulfur compounds to SC>2 and returned to the FGD
system.
The Allied process uses a slightly pressurized fluidized-bed reactor
containing a mixture of ground power plant coal and silica sand through
which the SC>2 off-gas, mixed with a small quantity of air, is passed.
Most of the SC>2 is reduced to sulfur but appreciable H2S is also produced.
The off-gas is passed through a particulate collector, a liquid sulfur
scrubber to condense the sulfur, and a Glaus unit to oxidize the'l^S to
sulfur before the residue is incinerated and returned to the FGD system.
The process also includes coal drying and grinding facilities and sulfur
cooling and filtration facilities..
The ACP consists of spray dryer absorbers using a soda ash solution
followed by ESP's to collect the sulfur salt particulate matter and
residual fly ash not removed in upstream cyclones. The particulate
matter is mixed with ground power plant coal and injected into refractory-
lined reactors. Air is injected to maintain a reaction temperature of
1500°F, at which the sodium salts are molten. Most of the sulfur is
reduced to the sulfide. The reactor off-gas is scrubbed to remove
54
-------
chlorides and ash and used as a CC>2 source. The melt overflows to a
quench/dissolving tank. The dissolved melt is treated with process
to form NaHS and then with process C02 to produce H^S and NaHC03, which
is further reacted with C02 off-gas to produce Na2C03- The H2S is
processed to sulfur in a conventional Glaus unit.
Economic Results
The base-case costs for each of the nine processes are shown in
Table 1. Except for the AGP, the costs are product-related, falling
into separate groupings of waste-, acid-, and sulfur-producing processes
in both capital investment and first-year revenue requirements. The
differences in cost between the waste-producing and acid-producing
processes are essentially the costs for absorbent regeneration; ponding
costs and acid plant costs do not differ greatly and raw material costs
do not differ sufficiently to produce large cost differences. The
higher costs for sulfur-producing processes are the result of the added
costs for reduction of sulfur oxides. Here coal reduction holds a
strong advantage over other fossil reducing agents. In the citrate
process, 16% of the annual revenue requirements (1.06 mills/kWh of 6.44
mills/kWh) are for natural gas to produce H2S. In contrast, reducing
coal costs range from 9% (Resox®) to 4% (Allied).
TABLE 1. FGD PROCESS ECONOMIC COMPARISONS
Mid-1980 first-year
Mid-1979 capital revenue requirement,
investment, $/kW mills/kWh
Waste-Producing Processes
Limestone 98 4.02
Lime 90 4.25
Double alkali 101 4.19
Sulfuric Acid Processes
Magnesia . 132 5.08
Wellman-Lord/sulfuric acid 131 5.11
Sulfur Processes
ACP 119 4.81
Wellman-Lord/Resox 138 6.03
Wellman-Lord/Allied 141 5.94
Citrate 143 6.44
55
-------
The anomalous capital investment of the ACP results from a credit
for the unnecessary separate fly ash ESP's and from the intrinsic chloride
purge from the reducer off-gas quench. If no ESP credit is given (as in
an existing plant with ESP's in place) its capital investment becomes
137 $/kW. Similarly, if no chloride removal is necessary in the wet
processes, these process costs are reduced about 10 $/kW. Under these
conditions, the ACP becomes the highest in capital investment. Specific
power plant conditions are thus important in the comparative capital
investments of the regeneration processes. In first-year revenue require-
ments the lower costs for the ACP are less site specific. It has low
raw material costs and low utility costs that prevail regardless of
specific fuel and power plant conditions.
Ground-to-Ground Energy Assessment
As a part of this series of FGD studies, a ground-to-ground energy
assessment of the limestone, lime, and magnesia processes was made.
This consisted not only of the FGD energy requirements but the energy
consumed in mining, processing, and transportation of the raw materials,
the disposal of wastes, and an energy credit for the sulfuric acid
produced. The assessment represents, in a sense, the energy removed
from a hypothetical energy reservoir because of the operation of the FGD
systems. A credit is given for the sulfuric acid because it replaces
acid that would be produced from sulfur, and thus the energy that would
have been consumed in mining and transporting the sulfur and producing
the acid. The results are shown in Table 2 and Figure 2.
TABLE 2. FGD GROUND-TO-GROUND ENERGY REQUIREMENTS
Btu/lb sulfur removed
Function
Mining
Absorbent processing
Transportation
FGD
Waste disposal
Total
Byproduct credit
Net total
Btu/kWh
% gross power unit output
Limestone
438
176
14,042
22
14,678
_
14,678
291
3.2
Lime
356
6,198
143
13,165
15
19,877
_
19,877
395
4.4
Magnesia
25
161
33
26,387
26,658
(5,491)
21,115
420
4.7
56
-------
30-
20-
$
o
i
r
<0
i
0-
10-
ABSORBENT
I I
BYPRODUCT
CREDIT
Limestone Lime Magnesia
Figure 2. Ground-to-ground energy requirements for
limestone, lime, and magnesia scrubbing
processes.
57
-------
The ground-to-ground energy comparison shows considerably different
relationships than comparison of FGD energy requirements alone. The FGD
energy requirements of the magnesia process (typical of regeneration
processes) are about twice those of the limestone and lime processes.
The.absorbent energy requirements are low for the magnesia process
because only makeup magnesia is used. In contrast, the lime process,
which has the lowest FGD energy requirements, has much higher energy
requirements when the energy for calcining lime is included. With the
byproduct credit included, the magnesia process is not appreciably more
energy intensive than the lime process.
Energy requirements cannot, of course, be directly related to FGD
costs. Energy consumed in absorbent production and transportation, for
example, is seen only indirectly, as it affects raw material costs. In
addition, the form of the energy may have an important effect on costs.
The magnesia process uses fuel oil for over one-third of its energy
requirements whereas almost all of the limestone and lime energy
requirements are met with coal. The significance of these differences
on costs is dicussed further in the byproduct marketing portion of this
paper.
FGD WASTE DISPOSAL ECONOMICS
Also during the past three years, TVA has conducted a series of
evaluations for EPA on the economics of disposal processes for flue gas
cleaning wastes. The first three studies (5,6,7) deal with the disposal
of fly ash and scrubber wastes from limestone/lime FGD systems. In all,
seven disposal methods were evaluated covering a range of existing or
potential disposal options of the late 1970's. All of the evaluations
were based on the same premises, using as the basis a 500-MW power plant
burning a 3.5% sulfur eastern coal and scrubbing with a limestone slurry
to meet the then-existing 1.2 Ib S/MBtu NSPS. In addition, over 175
case variations representing various power plant, fuel, waste treatment,
transportation, and disposal site conditions were evaluated. Schematic
flow diagrams of the processes are shown in Figure 3.
Except for the gypsum process, the scrubber waste consists of a 15%
solids slurry with a sulfur species composition of 85% CaS03'l/2H20 and
15% CaSQ4'2H20. Fly ash is included in the slurry except for the
sludge - fly ash blending and Dravo landfill processes. In dewatering,
30% solids from the thickener and 60% solids from the filter is used.
For the gypsum process essentially all the sulfur is CaS04'2H20 and the
filtered solids is 80%.
Process Descriptions
The untreated ponding case assumes that the effluent is pumped
directly to an earthen-diked pond. The Dravo, IUCS, and Chemfix processes
are all commercial fixation processes using somewhat different approaches
to treat dewatered FGD sludge. All depend on additives that produce
58
-------
UNTREATED PONDING
DRAVO PONDING
THICKENER
CALCILOX
»-
1
LIME
MIXER
IUCS
THICKENER
FILTER
MIXER
LANDFILL
CHEMFIX
THICKENER
CEMENT
FILTER
1
r
SILICATE
MIXER I +
SLUDGE- FLYASH BLENDING ^
THICKENER
FILTER
&.
MIXER
LANDFILL
GYPSUM
AIR
OXIDATION
THICKENER
FILTER
"
LANDFILL
MINE DISPOSAL
~~j MINE j~
DRAVO LANDFILL
LANDFILL
Figure 3. Process flow diagrams.
59
-------
cementitious chemical reactions. The types and quantities of the
additives and the degree of dewatering can be controlled to produce a
soillike material over a curing period of hours or months. The Dravo
process uses their product Calcilox,® a processed blast furnace slag,
sometimes with lime or fly ash, or both. Depending on the degree of
sludge dewatering and materials added, the treated material is pumped to
permanent or temporary pond storage or it is hauled to disposal after a
curing period. The IUCS process uses lime and fly ash blended with
dewatered sludge to produce a soillike solid. The Chemfix process uses
Portland cement and sodium silicate blended with dewatered sludge to
produce a soillike solid. The process is said to provide an encapsula-
tion that reduces leaching. For comparison, a sludge - fly ash blending
process without purchased additives, is included. The gypsum process
differs in that air oxidation equipment is added to the scrubber loop,
permitting production of the more easily dewatered and denser CaS04«2H20.
It is assumed this material can be dewatered and handled as a solid
without stabilization or fixation with additives. Finally, a process
using the sludge - fly ash blending process with disposal in a surface
mine is evaluated.
Economic Results
Cost breakdowns of the base cases by processing areas were made, as
shown in Table 3, to facilitate identification of cost elements and
comparison of different disposal processes. The sludge - fly ash blend-
ing process, the mine disposal process, and the Dravo landfill process
require inclusion of ESP costs for comparison with the other processes.
In those cases in which fly ash is collected separately the cost of
ESP units and their operation is a major component of the waste disposal
costs. In comparison, simultaneous fly ash removal results in relatively
modest increases in thickening and filtration costs. Separate collection
of fly ash is, of course, possible with all of the processes evaluated
and would require similar costs for all processes. In comparison of
landfill disposal practices having separate fly ash collection, cost
differences would largely be reduced to the raw material portion of the
cost breakdown.
For the processes using purchased fixatives, raw materials are an
important element of both capital investment and first-year revenue
requirements. Fly ash handling is also a relatively expensive element.
The advantage of a single fixative is illustrated by comparison of raw
material costs for processes that use two additives with processes that
use one. Thickening is the largest capital investment cost element,
excluding ESP costs, for all of the nonponding processes. It is also a
large cost element in annual revenue requirements. Filtration is also
a large cost element, though considerably less so than thickening.
Dewatering costs for the gypsum process are lower than the other simul-
taneous fly ash - FGD waste filtration processes because of the predicted
superior filtration characteristics of the high-sulfate sludge. Mixing
costs are a minor part of both capital investment and annual revenue
requirements.
60
-------
TABLE 3. MODULAR COSTS BY PROCESSING AREA FOR EIGHT DISPOSAL PROCESSES
Capital investment by processing
Other Raw materials
Ponding
Dravo ponding
IUCS
Chemf ix
Sludge-fly ash
Gypsum
Mine disposal
Dravo landfill
blending 19.
4.
19.
19.
2a
6b
2a
2a
9.0
4.2
8.5
4.4
4.4 .
6.2
Thickening Filtration
8.4
8.5
9.1
6.3
5.2
6.2
6.0
First-year revenue
Ponding
Dravo ponding
IUCS
Chetnfix
Sludge-fly aeh
Gypsum
Mine disposal
Dravo landfill
blending 0.
0.
0.
0.
56S
29d
56C
56C
0.91
0.44
0.97
0.22
0.22
0.57
0.24
0.29
0.29
0.24
0.29
0.25
0.22
4
4
2
3
2
2
.1
.8
.5
.0
.5
.2
requirements
0
0
0
0
0
0
.18
.19
.11
.16
.11
.10
area, $/kW
Mixing Storage Disposal
1
0
1
1
0
0
0
.4
.5
.1
.6
.9
.9
.8 1.1
by processing area
0
0
0
0
0
0
0
.14
.03
.06
.06
.05
.05
.05 0.03
33.0
30.3
3.5
3.1
3.1
2.6
2.0
3.8
Total
34.4
48.2
21.4
27.1
36.4
15.4
35.3
39.4
, mills/kWh
0.80
0.74
0.54
0.49
0.45
0.44
0.36
0.47
0.94
1.91
1.51
2.00
1.64
1.18
1.54
2.00
$/ton
dry waste
8.1
15.3
12.6
15.9
9.3
7.9
8.2
11.9
Basis: 500-MW
power plant,
removal in scrubber
127,
500-hour life
where cost is not
, 7,000 hr/yr revenue requirement basis; 3.
shown. Limestone
scrubber,
5% S, 16%
1.5 stoichiotnetry, 15%
ash coal; fly ash
solids
waste to
disposal system.
a. $9,614,000
b. $2,3u3,000
c. $1,975,000
d. $1,005,000
ESP cost for
separate fly ash
collection.
air-oxidation modifications.
ESP operatJi.^
costs.
air-oxidation ope.
rating c^sts
-------
Transportation and disposal site costs illustrate fundamental
differences between ponding and landfill disposal methods. Capital
investment for ponding transportation and disposal site costs is an
order of magnitude greater than the capital investment for landfill
transportation and disposal site operations. Capital investment for
transport lines is also an important element in ponding. Among the
landfill and mine disposal processes, transportation and disposal site
costs are a relatively minor element of total capital investment.
First-year revenue requirements for ponding transportation and
disposal site costs are also higher than those for landfill and mine
disposal although the differences are less pronounced. About two-thirds
of the annual revenue requirement direct costs for ponding transportation
and disposal site operations consist of pond maintenance. Transportation
of the waste is a relatively minor cost element. In contrast, about
four-fifths of the annual revenue requirements direct cost for landfill
and mine disposal transportation and disposal site operations is for
labor and supervision, much of it for loading and hauling.
In overall comparison of the processes evaluated, the most important
capital investment cost elements are separate fly ash collection, raw
material handling, thickening, and pond construction. Large cost elements
in first-year revenue requirements are separate fly ash collection, raw
material purchase and handling, and disposal.
The most important variations from the base-case conditions affecting
costs are power plant size, coal sulfur and ash content, and transportation
distance to the disposal site, as shown in Figure 4. Coal sulfur content
affects costs both through the volume of waste to be processed and
disposed of and, for processes using fixatives, the quantity of fixative
required. Costs for the disposal processes increase at different rates
with increasing sulfur content, depending on the relative influence of
these factors. Fixation processes increase in cost more rapidly than
the processes that do not use purchased fixatives. Distance to the
disposal site illustrates an important difference between the ponding
and landfill processes. Ponding investment costs increase dramatically
as the distance increases to 5 and 10 miles, in contrast, transportation
costs for landfill processes decrease more slowly with distance. The
relatively small cost advantages of mine disposal are lost in higher
transportation costs if the comparison is made between a landfill onsite
and a mine over a few miles from the power plant. From a purely econom:.c-
viewpoint, mine disposal requires very close proximity of power plant
and mine for its economic advantages to be realized.
BYPRODUCT MARKETING
The EPA-sponsored FGD byproduct'marketing system began as a limited
production-marketing model for sulfuric acid in the early 1970's (8).
Several expansions of the methodology led in 1978 to the basis of the
present system (9), a comprehensive analysis of the potential of FGD
-------
EFFECT OF POWER PLANT SIZE ON WASTE DISPOSAL COSTS.
EFFECT OF COAL SULFUR CONTENT ON WASTE DISPOSAL
COSTS.
0.
ra
o
60
40
20
60
I
20
400
800
1200
1600
OJ
!f 3
I
1—Untreated ponding
2—IUCS
3—Sludge-flyash' blending
4—Gypsum
5—DI-HVO landflU
400 800
Pn.-p.r plant si:'^, K.i
^\
1200
I
1600
3
o-
s
41
0)
CO
I
1—Untreated ponding
2--IUCS
3—Sludge-flyash blending
4—Gypsum
5—Dravo landfill
J_
2 3
Sulfur in coal, /;
Ji'!l>., s on disfosal co3t;
-------
EFFECT OF COAL ASH CONTENT ON WASTE DISPOSAL COSTS.
60
40
20
_J
20
12
16
1—Untreated ponding
2—IUCS
3—SJudge-flyash blending
4—Gypsum
5—Dravo landfill
12
I
16
Ash in coal,
20
(co;U
EFFECT OF DISTANCE TO DISPOSAL SITE ON WASTE
DISPOSAL COSTS.
80
60
40
20
63==
4
10
1—Untreated ponding
2—IUCS
3—Sludge-fly ash blending
4—Gypsum
5—Dravo landfill
6—Mine disposal
2 4 6 8
Distance to disposal site, miles
10
-------
byproduct sulfur and sulfuric acid production and marketing by U.S.
electric utilities. Basically the system compares low-sulfur fuel and
regeneration and waste-producing FGD costs for existing and planned U.S.
utility power plants, determines FGD byproduct revenue from sales to
U»S. sulfuric acid plants, and determines the mix of strategies that
results in the least-cost option and the highest total revenue from FGD
byproduct sales. FERC and published utility data, transportation data,
and U.S. sulfuric acid plant data are used. TVA process economics,
scaled to projected power plant operating conditions, determine FGD
costs.
An updated projection of FGD sulfuric acid marketing potential for
1983 was published in 1979 (10), as was a users manual for the com-
puterized system (11). The 1983 projection also contained a manually
prepared forecast of FGD sulfur marketing potential. Several trends
became apparent in the 1983 projection: rapidly evolving FGD technology;
disproportionate fuel cost changes, particularly for petroleum products;
changes in historical patterns of utility coal use and sulfur and sul-
furic acid production; and evolving environmental legislation promised
to influence earlier patterns of FGD byproduct production.
•Developments in FGD, such as the recognition that chloride control
was necessary in some cases for regeneration processes to prevent loss
of absorbent effectiveness, special purge systems, and severe corrosion
problems, altered FGD costs. New technologies, such as spray dryer FGD
and coal reduction, promised further changes. The type of fuel used in
the FGD process was also becoming an important economic factor. The
growing importance of secondary sulfur and sulfuric acid production was
seen to be a potentially important consideration. Legislation such as
RCRA and the 1979 NSPS revisions, restricting waste disposal options and
the use of low-sulfur fuel, would be important in FGD economics in the
1980*s. It was also apparent that the usefulness of these projects
would be increased by extending them further into the future, on a scale
similar to the time period required for power plant planning and
construction.
Beginning in late 1979, a projection for 1985 was started. Although
a 1990 projection would have been more desirable, availability of data,
particularly on power plant construction, coal use, and fuel costs,
precluded a projection beyond 1985 at that time. Numerous system changes
were made, including updated FGD technologies (limestone throwaway,
magnesia to acid, and ACP for sulfur), a general updating of power plant,
transportation, and acid plant data, inclusion of a spray dryer FGD
sulfur-producing process, and inclusion of Canadian sour gas sulfur as a
market factor in the upper United States. The results, which were
published this year (12), showed a number of changes from previous
projections.
The combined sulfur and sulfuric acid market for 1985 was projected
to be 165,000 tons of sulfur from 11 power plants and 554,000 tons of
sulfuric acid from 6 power plants. The total benefits for the electric
65
-------
utility and sulfuric acid industry were about $20,000,000. The results,
shown in Table 4, differ considerably from the 1983 projection, which
showed 1,200,000 tons of sulfuric acid but no sulfur.
Several factors are important in both the total FGD byproduct
production projected and the sulfur-sulfuric acid mix. Most of the
production of both comes from new plants projected for a 1985 startup,
which were assigned to regulation under the 1979 revised NSPS for
modeling purposes. In addition, inclusion of fixation and landfill
disposal in the limestone scrubbing process, used for the waste-producing
FGD option enhances the FGD byproduct option, although limestone scrubbing
remains the predominate FGD option.
Sulfuric acid production was reduced by several factors, among
which increased costs for the magnesia process used in the FGD model
were most important. Inclusion of provisions for chloride control and
the cost of fuel oil in the process were particularly important. The
increase in potential FGD sulfur production stems largely from the use
of a spray dryer recovery FGD process based on the Rockwell International
aqueous carbonate process. Reduced costs in the form of simultaneous
fly ash and particulate sulfur salts collection and the use of coal as
the reducing agent, were important factors. In maximizing the combined
sulfur-sulfuric acid market, all of which is assumed to be sold to
sulfuric acid plants, alternate markets for Sulfur were also more prevalent
than those for sulfuric acid.
The 1985 projection indicates several factors that will have important
influences on FGD byproduct production by the late 1980's. Environmental
legislation affecting waste disposal practices and the restricting use
of low-sulfur coal as a compliance strategy could enhance the economic
attractiveness of regeneration FGD processes. The economics of byproduct
FGD processes that use coal as the fuel in the regeneration-manufacturing
process will be more favorable than those using oil br natural gas.
Similarly, processes that combine flue gas cleaning functions, such 'as
fly ash and sulfur salt collection, will have important economic advantages.
Fuel Oil Price Escalation
An interesting aspect of FGD economics in the past few years, as
the cost basis has been projected into the 1980"s, is the disproportionate
effect of energy costs. This is particularly apparent in the byproduct
marketing studies, which are projected further into the future than most
FGD economic studies. In the 1985 projection a 15% annual inflation
rate for No. 6 fuel oil was used, based on petroleum cost projections
available in early 1980. As an illustration of the effect of this rate
on costs, equivalent cost increases for fuel oil, natural gas, and coal
are shown below.
66
-------
TABLE 4. 1985 PROJECTION OF THE PRODUCTION AND DISTRIBUTION
OF FGD SULFUR AND SULFURIC ACID
Power plant location
Sulfur
Staten Island County, NY
Martin County, FL
Washington County, FL
Sherburne County, MN
Westmoreland County, PA
Montgomery County, MD
Shelby County, AL
Williamson County, IL
Rusk County, TX
Henderson County, TX
Armstrong County, PA
Sulfuric Acid
Person County, NC
Jasper County, IL
Pike County, IN
Northhampton County, PA
Delaware County, PA
Titus County, TX
Tons
7,000
28,000
20,000
8,000
24,000
10,000
12,000
11,000
9,000
7,000
29,000
165,000a
103,000
122,000
51,000
182,000
53,000
43,000
554,000b
Consumer location
Newark, NJ
Pierce, FL
Do than, AL
White Springs, FL
Dubuque , IA
North Bend, OH
Copley, OH
Baltimore, MD
Tuscaloosa, AL
East St. Louis, IL
Fort Worth, TX
Fort Worth, TX
Cleveland, OH
Richmond, VA
Wilmington, NC
Norfolk, VA
Tuscola, IL
Indianapolis, IN
Daepwater, NJ
Edison, NJ
Gibbstown, NJ
Gibbstown, NJ
Shreveport, LA
Tons
7,000
28,000
7,000
13,000
8,000
8,000
16,000
10,000
12,000
11,000
9,000
7,000
29,000
165,000a
36,000
26,000
41,000
122,000
51,000
95,000
74,000
13,000
53,OGC
43,000
554,000b
The potential revenue/savings to both industries combined is
projected to be as much as $10,000,000 for an approximate
average of $60/short ton of sulfur.
The potential revenue/savings to both industries combined is
projected to be as much as $10,500,000 for an approximate
average of $19/short ton of sulfuric acid.
67
-------
Annual price
escalation, %
Equivalent price increase, 1979-1985
No. 6 fuel oil, Natural gas, Coal,
$/gal $/kft3 $/ton
5
15
25
0.20
0.79
1.69
1.37
5.29
11.33
30.13
116.32
249.36
To equal the price increase projected for fuel oil, for example,
the price of coal would have to increase over 100 $/ton. Processes such
as the magnesia process that use fuel oil are thus placed at a dis-
advantage compared with processes such as the AGP using coal.
The effect of fuel oil price escalation on the cost of FGD sulfuric
acid is shown in Figure 5. The effect is twofold, first in FGD costs
and second in the avoidable production costs to acid producers. This is
a cost calculated by the byproduct marketing system to determine the
price of FGD acid at each acid plant. It represents the break-even
point between buying FGD acid to meet marketing requirements and producing
acid. In shutting down an acid plant, however, steam production is lost
and normally must be replaced by a boiler. Because of size, this
logically would be an oil-fired boiler. High fuel oil price escalation
rates thus decrease avoidable production costs, resulting in the need of
a higher acid price margin to make the purchase of FGD acid economical.
90
80
o:
<
5 70
o
< 60
o
cc
p 50
to
H
O
Q
UJ
£E
3°
20
10 -
I Increased FGD Cost
Decreased Acid Plant Avoidable Cost
%
I
10. 15, 20
FUEL OIL PRICE ESCALATION (%) TO 1985
Figure 5. Reduction in potential FGD sulfuric acid margin with
No. 6 fuel oil annual price escalation.
68
-------
FGD AND SOLID WASTE PROCESS EVALUATIONS IN PROGRESS
With the completion of the 1977-1980 series of SOX control and FGD
solid waste process designs and evaluations, plans were made for extension
of the series to other important FGD and waste disposal processes not
yet evaluated. During the planning cycle, dry scrubbing processes were
just beginning to capture strong interest. Therefore, the first new
study for the 1980's was a preliminary economic evaluation of this
technology. The first report on a lime spray dryer system for a western
low-sulfur coal application was published during early 1980 (13). A
second more detailed report summarizing current dry FGD process technology
and the economics for both low- and high-sulfur coal will be published
soon (14). T. A. Burnett will present results from these reports in a
paper to be presented later in the symposium.
A second project is now underway to prepare a report summarizing
the designs and economics of wet limestone-lime processes which have
been studied at the EPA-TVA Shawnee Test Facility. Thirteen different
process variations included in this report are listed below.
1. Turbulent Contact Absorber® (TCA) - Onsite ponding
2. TCA - Forced oxidation - Landfill
3. TCA - Forced oxidation Adipic acid - Landfill
4. TCA - Forced oxidation - MgO - Landfill
5. Spray Tower (ST) - Onsite ponding
6. ST - Forced oxidation - Landfill
7. ST - Forced oxidation - Adipic acid - Landfill
8. ST - Forced oxidation - MgO - Landfill
9. Venturi-Spray Tower (V-ST) - Onsite ponding
10. V-ST - Forced oxidation - Landfill
11. V-ST - Forced oxidation - Adipic acid - Landfill
12. V-ST - Forced oxidation - MgO - Landfill
13. Venturi - Forced oxidation - Adipic acid - Landfill
The final report should be available during 1981.
69
-------
A third project, which is about half completed, is a study of three
leading gypsum-producing FGD systems. The Dowa process, which was
developed in Japan on oil-fired boilers, is being marketed in the United
States by UOP and has been tested on a 10-MW prototype at Shawnee, is
one of the processes. The Saarberg-Holter process, a German-developed
system marketed by Davy-McKee in the United States, is the second process,
The third system is a limestone spray tower using adipic acid addition,
forced oxidation, and gypsum stacking for waste disposal. The report
for this project is expected to be ready for distribution in mid-1981.
There are other gypsum-producing processes being developed for commer-
cial use; it is hoped that these can be evaluated in a future study.
The last defined project now underway in the expanded series is an
evaluation of ash disposal systems and practices for coal-fired power
plants. The draft report for this project has been prepared and publica-
tion is expected shortly.
The ash disposal methods evaluated in this study are represented by
five base-case processes based on major utility ash disposal practices.
Four base cases represent disposal of noncementitious eastern coal ash.
They consist of (1) direct sluicing of combined fly ash and bottom ash
to separate ponds with once-through (nonrecycled) water, (2) the same
system with recycled transportation water, (3) direct sluicing of fly
ash and bottom ash to temporary ponds, followed by excavation and truck-
ing of both to a common landfill, and (4) collection of bottom ash in
dewatering bins from which it is trucked to a separate landfill and
collection of fly ash in dry storage silos from which it is trucked to a
separate landfill.
The fifth base case represents a situation in which the power plant
is burning a western-type coal which contains appreciable calcium,
making the ash subject to spontaneous cementitious reactions that affect
handling properties. The handling and disposal system is designed to
forestall these reactions by keeping the ash dry until shortly before
placement at the disposal site.
NEW PREMISES
The FGD and waste disposal studies that are now in progress are
based on new design and economic premises. During the 1977-1980 series
of studies it was recognized that changing economic conditions, fuel use
patterns, developments in.economic evaluation techniques, and, particu-
larly, developments in FGD technology and environmental legislation
justified revision of the TVA design and economic premises. Consequently.
TVA began studies that led to the adoption of new economic premises in
1979. During this period numerous discussions were held with EPA, EPRI,
and with other TVA organizations concerned with the use of these premises.
70
-------
Design Premises
Essentially the same power plant conditions are retained. For the
base case these are a new, midwestern, 500-MW, pulverized-coal-fired,
dry-bottom boiler. The heat rate is increased from 9,000 to 9,500 Btu/kWh
and the excess air is increased from 33% to 39%, however. The sulfur
content of the coal remains at 3.5% but the heating value is increased
from 10,500 to 11,700 Btu/lb. The operating schedule is also changed to
5,500 hr/yr for 30 years. A constant annual operating time is used to
facilitate levelizing of lifetim^ costs.
Major changes were made in the FGD design premises to reflect
current regulations and to improve process reliabilities. Required S0£
removal efficiency is now based on the 1979 NSPS. For the base-case
coal these require an 89% removal efficiency instead of the 79% needed
to meet the 1971 NSPS used in the old premises. In keeping with current
design trends a spare absorber train and provisions for emergency bypass
of 50% of the total flue gas are included. The old premises contained
no spare absorber or bypass provisions. In addition, ID booster fans,
instead of FD booster fans, are used in the new designs. For nonrecovery
processes both pond and landfill waste disposal methods are revised to
reflect more recent environmental concerns. These are primarily based
on RCRA Subtitle D (nonhazardous waste) guidelines and include provisions
for such factors as seepage and runoff control, security, monitoring,
and reclamation.
FGD process design features are usually based on technology pre-
vailing at the time of the study. The limestone scrubbing process is,
however, somewhat of a premise adjunct since it is used so frequently as
a basis of comparison in FGD studies. This process serves as an example
of the changes in FGD technology that have occurred over the past few
years. The current limestone process differs from the old process used
in the 1977-1980 studies in several respects. A spray tower instead of
a mobile-bed absorber, forced oxidation to gypsum, and landfill waste
disposal are now included in the basic system. The use. of a spray tower
results in a lower gas velocity of 10 ft/SBc instead of the 12.5 ft/sec
used in the old process with a mobile-bed absorber.
The new limestone scrubbing process represents several industry
trends in limestone scrubbing that have become evident in recent years,
The use of a spray tower ^instead of more complicated mobile-bed and
venturi - spray scrubbers has become common. The simpler spray tower is
expected to provide greater reliability and require less maintenance
although these: have not been, quantified in practice. The problem of
waste disposal has also been addressed, both by increasing use of
stabilization, fixation, and landfill disposal techniques and by other
methods of producing a more tractable waste, such as oxidation to gypsum.
The use of a spray tower, air oxidation, and landfill disposal in
the new process recognizes these trends. The process is based in part
on continuing test work on spray towers, forced oxidation, and waste
71
-------
dewatering at the EPA-spbhsored test facility at the Shawnee Steam
Plant. Like the previous limestone scrubbing process, however, it is
generic and incorporates general industry information as well as data
from Shawnee.
Economic Premises
Numerous changes were also made in the economic premises. Specific
provisions for sales tax, freight, and overtime for construction delays
are included. The method of calculating indirect capital investment is
simplified and modified to more accurately reflect complexity of engineering
and construction costs of processes evaluated. Contingencies and allowances
for modification after startup are also defined as process-specific
variables reflecting degree of development and established technology.
Provision for recognition of anticipated royalties is also made. Land
prices and interest during construction are increased.
First-year revenue requirements are now calculated using levelized
capital charges (30-year life, capital recovery factor, 6% per year
inflation and 10% per year cost of money, discounted to the first year)
instead of the average capital charges used in the old premises. In
addition, levelized lifetime revenue requirements are also calculated to
represent inflated and discounted costs over the life of the system.
The base years for capital investment and first-year revenue require-
ments are also advanced to 1982 and 1984 respectively. A project con-
struction period from 1981 to 1983 is now assumed, with plant startup in
early 1984.
COST COMPARISON OF OLD AND NEW PREMISES
The key old and new design and economic premises for evaluation of
the limestone scrubbing process are shown in Table 5. A stepwise cost
transition from the old premises and technology to those for the new
limestone scrubbing evaluation is shown in Table 6 and illustrated in
Figure 6. Overall, the cumulative changes result in nearly doubled
capital investment and first-year revenue requirements. The investment
increases resulting from the new economic premises are related to higher
indirect capital investment costs, particularly in interest during
construction, contractor expense, and working capital. The increase in
first-year revenue requirements stems largely from capital charges based
on the capital investment. New power plant coal and air rates, the
operating profile, and the 1979. NSPS all produce similar increases in
capital investment. In these cases the main factors are increased flue
gas volume, increased lifetime.waste disposal requirements, and the more
stringent scrubbing conditions. The effect on annual revenue require-
ments is similar except, of course, that the reduction in yearly operating
hours results in a reduction in costs. Addition of reliability factors
(a spare scrubber train, emergency bypass, and a spare ball mill) also
cause appreciable increases in both capital investment and first-year
72
-------
TABLE 5. COMPARISON OF OLD AND NEW PREMISE CONDITIONS
USING THE LIMESTONE SCRUBBING PROCESS
Old premises New premises
Design Premises
Coal, Btu/lb
Excess air, %
Heat rate, Btu/kWh
Operating profile
First year, hr/yr
Lifetime, hr (30 years)
FGD
SOX removal, %
Emergency bypass, %
Spare units
Booster fan
Limestone process
Absorber
L/G, gal/kaft3
Gas velocity, ft/sec
AP, in. H20
Forced oxidation
Waste disposal
10,500
33
9,000
7,000
127,000
1971 NSPS
0
0
FD
Mobile bed
50
12.5
8
No
Pond
11,700
39
9,500
5,500
165,000
1979 NSPS
50
1
ID
Spray tower
90
10.0
1.4
Yes
Landfill
Economic Premises
Cost index year
Capital investment
Annual revenue requirements
Indirect capital costs
Land, $/acre
Interest during construction,
Limestone process contingency,
Pond contingency, %
Pond allowance for startup, %
Capital charges
Depreciation
1979
1980
3,500
I 12
% 20
20
8
Average annual
Straight line
1982
1984
Revised
5,000
15.6
10
10
0
Levelized
Sinking fund
73
-------
TABLE 6. COST COMPARISON IN TRANSITION FROM OLD TO NEW PREMISES
AND TECHNOLOGY FOR THE LIMESTONE SCRUBBING PROCESS
Capital investment
Condition
Old premises and technology
Above with new economic
premises and pond
Above with new power plant
design premises
Above with new operating
profile
Above with 1979 NSPS
Above with reliability
factors (spares and bypass)
Above with spray tower
Above with landfill
Above with 1982, 1984 costs
k$
48,700
55,100
57,100
59,800
63,600
77,100
83,300
76,000
96,800
$/kW
98
110
114
120
127
154
167
152
194
% change
_
13
4
5
6
21
8
-9
28
% total change
13
17
23
29
58
71
56
99
First-year revenue requirements
k$
14,100
16,200
17,000
16,500
17,200
20,100
21,500
21,700
27,300
Mills/kWh
4.0
4.6
4.9
6.0
6.3
7.3
7.8
7.9
9.9
% change
_
15
5
-3
4
17
7
1
26
% total change
_
15
21
17
22
43
52
'54
94
-------
#
0°
200-
150-
H
Z
w _
> *
H
100^
5!
50-
10-
o-
s
.c
w
oo
H
fe
SUM OF PREMISE ft TECHNOLOGICAL CHANGES
^ ^
.". C
Yfft
//1
SUM OF PREMISE ft TECHNOLOGICAL CHANGES-
Figure 6. Stepwise conversion of limestone scrubbing costs
from old :to new premises and technology.
75
-------
revenue requirements, The use of a spray tower instead of a mobile-bed
absorber increases costs primarily because of the lower flue gas velocity
and higher slurry recirculation rate, which requires larger ducting and
pumping requirements.
Substitution of landfill for ponding substantially reduces capital
investment by eliminating pond construction costs. The resulting reduction
in capital charges essentially counteracts the increased waste disposal
costs in first-year revenue requirements.
The largest cost increase is a result of advancing the cost index
year from 1978 to 1982 for capital investment and from 1980 to 1984 for
first-year revenue requirements.
Overall, economics in the form of inflation and higher interest
have the largest effect in comparison of the limestone process using the
old and new premises and technology. Technical changes related to
improvements in reliability, such as bypass and redundancy provisions,
also have a large effect. The higher SOX removal efficiency has less
effect than the economic and technical changes.
ADVANCED LIMESTONE SCRUBBING TECHNOLOGY
As stated earlier, TVA is now conducting an. EPA-sponsored economic
evaluation of advanced limestone scrubbing technology. The study encom-
passes recent developments in limestone scrubbing such as chemical
additives, increasing use of spray towers, forced oxidation, and landfill
techniques. The complete results of this project will be published in
1981.
Of particular interest at this time is the advanced limestone
system using a spray tower, forced oxidation, adipic acid addition and
landfill of the gypsum waste. The interest comes from favorable results
at the Shawnee Test Facility. Earlier bench- and pilot-scale studies
were made by TVA and EPA on adipic acid addition and EPA is sponsoring
an adipic acid demonstration unit at the Southwest Plant of Springfield
(Missouri) City Utilities. The advantage of adipic acid (or other
similar additives) lies in its buffering action, which controls the
slurry pH at more favorable reaction conditions. This increases the
reactivity of the slurry, improving S02 removal efficiency and increasing
limestone utilization.
As a special feature, an economic comparison of the advanced process
with the new conventional and old conventional limestone processes is in
order. The design conditions for the three processes are shown in
Table 7-
Tables 8 and 9 show the capital investments and annual revenue
requirements for the three processes based on the base-case conditions
and the new premises that were discussed previously. The cos'cs thus
76
-------
TABLE 7. PROCESS DESIGN CONDITIONS AND PREMISES - LIMESTONE PROCESSES
Type of absorber
Forced oxidation
Adipic acid use
Waste disposal
Scrubber gas velocity,
ft/sec
L/G, gal/kaft3
Limestone stoichiometry
Air stoichiometry
Percent sulfite oxidation
ID fan/FD fan
Spare scrubber
Filter cake solids, %
Pond settled solids, %
Spare ball mill
Reheat
Bypass available
Advanced pro ces s
New
conventional
Old
conventional
Spray tower
Yes
Yes (1000 ppm)
Thickener-filter-landfill
10
80
1.2
2.5
95
ID
Yes
80
Yes
In-line steam
50% emergency
Spray tower
Yes
No
Thickener-filter-landfill
10
90
1.3
2.5
95
ID
Yes
80
Yes
In-line steam
50% emergency
Mobile bed
No
No
Pond
12.5
58
1.3
0
30
ID
Yes
40
Yes
In-line steam
50% emergency
-------
TABLE 8. CONVENTIONAL AND ADVANCED LIMESTONE SCRUBBING PROCESSES
CAPITAL INVESTMENT
Direct Investment
Material handling
Feed preparation
Gas handling
S02 absorption
Reheat
Solids disposal
Total
Services, utilities, and miscellaneous
Total
Landfill or pond construction
Capital
Old
conventional3
3,498
3,485
9,600
19,830
2,851
2.063
41,327
2.480
43,807
.13,960
investment,
New
conventional"
3,497
3,484
11,129
22,988
3,304
2.868
47,270
2.836
50,106
2,076
500
k$
Advanced0
3,503
3,490
10,821
22,351
3,213
2,850
46,228
2,774
49,002
1,983
495
Total
57,767
52,682
51,480
Indirect Investment
Engineering design and supervision
Architect and engineering contractor
Construction expense
Contractor fees
Contingency
Total fixed investment
80,458
75,517
73,790
Other Capital Investment
Allowance for startup and modifications
Interest during construction
Land
Working capital
Total capital investment
$/kW
103,030
206
Basis
Upper Midwest plant location represents project beginning mid-1980, ending
mid-1983. Average cost baais, mid-1982. Spare pumps, one spare scrubbing
train, and one spare ball mill are included. Disposal pond and landfill
located 1 mile from plant. Investment includes FGD feed plenum but
excludes stack plenum and stack.
a. Old conventional process is a mobile bed absorber with onsite ponding
of sulfite sludge.
b. New conventional process is a spray tower, forced oxidation and gypsum
landfill.
c. Advanced system is same as b. but with adipic acid addition for
enhanced S02 removal.
78
-------
TABLE 9. CONVENTIONAL AND ADVANCED LIMESTONE SCRUBBING PROCESSES
ANNUAL REVENUE REQUIREMENTS
Annual cost, k$
Direct Costs - First-Year
Raw materials
Limestone
Adipic acid
Total raw materials cost
Conversion costs
Operating labor and supervision
FGD
Solids disposal
Utilities
Process water
Electricity
Steam
Fuel
Maintenance
Labor and material
Analyses
Total conversion costs
Total direct costs
Indirect Costs - First-Year
Overheads
Plant and administrative (60% of
conversion costs less utilities)
Total first-year operating and
maintenance costs
Levelized capital charges (14.7% of
total capital investment)
Total first-year annual revenue
requirements
Levelized first-year operating and
maintenance costs (1.886 x first-
year 0 and M)
Levelized capital charges (14.7% of
total capital Investment)
Levelized annual revenue
requirements
First-year annual revenue requirements
Levelized annual revenue requirements
Old
conventional
1,128
_
1,128
460
-
35
1,732
1,273
-
3,923
104
7,527
8,655
2.692
11,347
•15.145
26,492
21,401
15,145
36,545
9.63
13.29
New
conventional
1,128
-
1,128
658
529
26
2,018
1,365
199
4,025
104
8,924
10,052
3,057
13,109
14,234
27,343
24,724
14,234
38,958
Mills /kWh
9.94
14.17
Advanced
1,041
216
1,257
658
517
26
1,874
1,367
189
3,937
104
8,672
9,929
2,998
12,927
13,907
26,834
24,381
13,907
38,288
9.76
13.92
Basis
Upper Midwest plant location, 1984 revenue requirements.
New plant with 30-year life.
Power unit on-stream time, 5,500 hr/yr.
Coal burned, 1,116,500 tons/yr.
Boiler heat rate, 9,500 Btu/kWh.
Total capital investment:
Old conventional - $103,030,000
New conventional - $ 96,832,000
Advanced $ 94,608,000
79
-------
incorporate a spare scrubber, emergency bypass, and a 1981-1983, 1984
time period, among other differences from the FGD studies discussed
previously. All of the costs except those for landfill were developed
by the TVA Shawnee Computer Economics Program (15).
Both the new conventional process and the advanced process have
slightly higher direct capital investment costs than the old conventional
process in most areas. The old conventional process has disposal site
(pond) construction costs over ten times higher than the disposal site
(landfill) construction costs than the others, however. The result is a
slightly lower capital investment for the new conventional and advanced
processes. The use of adipic acid in the advanced process produces a
minor increase in material handling costs and much larger decreases in
absorber and disposal costs. The increased reactivity of the limestone
slurry allows both less stringent scrubbing conditions and improved
limestone utilization, resulting in lower limestone consumption and less
unreacted limestone in the waste.
In annual revenue requirements, the old conventional process has
lower conversion costs, primarily because of lower labor and supervision
and electricity costs, resulting in lower overall expense. The increase
in labor and supervision cost for the new conventional and advanced
processes is largely for disposal operations because trucking and earth-
moving operations are required. In comparison of the new conventional
process and the advanced process, adipic acid addition causes a slight
overall reduction in costs, primarily because of lower limestone and
electricity consumption.
80
-------
REFERENCES
1. S. V. Tomlinson, F. M. Kennedy, F. A. Sudhoff, and R. L. Torstrick.
Definitive SOX Control Process Evaluations - Limestone, Double-
Alkali, and Citrate FGD Processes. TVA ECDP B-4, Tennessee Valley
Authority, Office of Power, Emission Control Development Projects,
Muscle Shoals, Alabama. EPA-600/7-79-177, U.S. Environmental Protec-
tion Agency, Office of Research and Development, Washington, D.C.,
1979.
2. K. D. Anderson, J. W. Barrier, W. E. O'Brien, and S. V. Tomlinson.
Definitive SOX Control Process Evaluations: Limestone, Lime, and
Magnesia FGD Processes. TVA ECDP B-7, Tennessee Valley Authority,
Office of Power, Emission Control Development Projects, Muscle
Shoals, Alabama. EPA-600/7-80-001, U.S. Environmental Protection
Agency, Office of Research and Development, Washington, D.C., 1980.
3. J. R. Byrd, K. D. Anderson, S. V. Tomlinson, and R. L. Torstrick.
Definitive SOX Control Process Evaluations: Aqueous Carbonate,
Wellman-Lord, Allied Chemical, and Resox® FGD Technologies. Tennessee
Valley Authority, Office of Power, Division of Energy Demonstrations
and Technology, Muscle Shoals, Alabama. U.S. Environmental Protec-
tion Agency, Office of Research and Development, Washington, D.C.
(In press)
4. G. G. McGlamery, R. L. Torstrick, J. W. Broadfoot, J. P. Simpson,
L. J. Henson, S. V. Tomlinson, and J. F. Young. Detailed Cost
Estimates for Advanced Effluent Desulfurization Processes. TVA
Bulletin Y-90, Tennessee Valley Authority, Office of Agricultural
and Chemical Development, Muscle Shoals, Alabama. EPA-600/2-75-006,
U.S. Environmental Protection Agency, Office of Research and Develop-
ment, Washington, D.C., 1975.
5. J. W. Barrier, H. L. Faucett, and L. J. Henson. Economics of
Disposal of Lime-Limestone Scrubbing Wastes: Untreated and
Chemically Treated Wastes. TVA Bulletin Y-123, Tennessee Valley
Authority, National Fertilizer Development Center, Muscle Shoals,
Alabama. EPA-600/7-78-023a, U.S. Environmental Protection Agency,
Office of Research and Development, Washington, D.C., 1978.
6. J. W. Barrier, H. L. Faucett, and L. J. Henson. Economics of
Disposal of Lime/Limestone Scrubbing Wastes: Sludge/Flyash Blending
and Gypsum Systems. TVA Bulletin Y-140, Tennessee Valley Authority,
National Fertilizer Development Center, Muscle Shoals, Alabama.
EPA-600/7-79-069, U.S. Environmental Protection Agency, Office of
Research and Development, Washington, D.C. , 1979.
81
-------
7. J. D. Veitch, A. E. Steele, and T. W. Tarkington. Economics of
Disposal of Lime/Limestone Scrubbing Wastes: Surface Mine Disposal
and Dravo Landfill Processes. TVA EDT-105, Tennessee Valley Authority,
Office of Power, Division of Energy Demonstrations and Technology,
Muscle Shoals, Alabama. EPA-600/7-80-022, U.S. Environmental Protec-
tion Agency, Office of Research and Development, Washington, D.C. ,
1980.
8. J. I. Bucy, J. L. Nevins, P- A. Corrigan, and A. G. Melicks. The
Potential Abatement Production and Marketing of Byproduct Elemental
Sulfur and Sulfuric Acid in the United States. TVA S-469, Tennessee
Valley Authority, Office of Agricultural and Chemical Development,
Muscle Shoals, Alabama, 1976.
9. J. I. Bucy, R. L. Torstrick, W. L. Anders, J. L. Nevins, and P. A.
Corrigan. Potential Abatement Production and Marketing of Byproduct
Sulfuric Acid in the U.S. TVA Bulletin Y-122, Tennessee Valley
Authority, Office of Agricultural and Chemical Development, Muscle
Shoals, Alabama. EPA-600/7-78-070, U.S. Environmental Protection
Agency, Washington, D.C., 1978.
10. W. E. O'Brien and W. L. Anders. Potential Production and Marketing
of FGD Byproduct Sulfur and Sulfuric Acid in the U.S. (1983 Projection).
ECDP B-l, Tennessee Valley Authority, Office of Power, Emission Control
Development Projects, Muscle Shoals, Alabama EPA-600/7-79-106, U.S.
Environmental Protection Agency, Washington, D.C., 1979.
11. W. L. Anders. Computerized FGD Byproduct Production and Marketing
System: Users Manual. TVA ECDP B-2, Tennessee Valley Authority,
Office of Power, Emission Control Development Projects, Muscle Shoals,
Alabama. EPA-600/7-79-114, U.S. Environmental Protection Agency,
Washington, D.C., 1979.
12. W. E. O'Brien, W. L. Anders, and J. D. Veitch. Projection of 1985
Market Potential for FGD Byproduct Sulfur and Sulfuric Acid in the
U.S. TVA EDT-115, Tennessee Valley Authority, Office of Power,
Division of Energy Demonstrations and Technology, Muscle Shoals,
Alabama. EPA-600/7-80-131, U.S. Environmental Protection Agency,
Washington, D.C., 1980.
13. T. A. Burnett and W. E. O'Brien. Preliminary Economic Analysis of
a Lime Spray Dryer FGD System. TVA EDT-112, Tennessee Valley
Authority, Office of Power, Division of Energy Demonstrations and
Technology, Muscle Shoals, Alabama. EPA-600/7-80-050, U.S. Environ-
mental Protection Agency, Washington, D.C., 1980.
14. T. A. Burnett and K. D. Anderson. Technical Review and Economic
Evaluation of Spray Dryer FGD Systems. Tennessee Valley Authority,
Office of Power, Division of Energy Demonstrations and Technology,
Muscle Shoals, Alabama. U.S. Environmental Protection Agency,
Washington, D.C. (In press)
82
-------
15. C. D. Stephenson and R. L. Torstrick. Shawnee Lime/Limestone
Scrubbing Computerized Design/Cost-Estimate Model Users Manual.
ECDP B-3, Tennessee Valley Authority, Office of Power, Emission
Control Development Projects, Muscle Shoals, Alabama. EPA-
600/7-79-210, U.S. Environmental Protection Agency, Washington,
D.C., 1979.
83
-------
S02 AND NOx ABATEMENT FOR COAL-FIRED BOILERS IN JAPAN
Jumpei Ando
Faculty of Science and Engineering, Chuo University
Kasuga, Bunkyo-ku, Tokyo 112
The total capacity of coal-fired utility boilers in Japan,
which was only 4,300 MW (3.7% of total utility power) in 1979, is
expected to increase to 10,000 MW (5.6%) in 1885 and to 22,000 MW
(10.0%) in 1990. Most of the boilers will apply FGD by the wet
limestone-gypsum process because of its reliability and relatively
low cost. To save energy and water, F.GD systems with a low
pressure drop and small water consumption are preferred. Tests on
FGD by a dry carbon process are under way.
NOx concentrations in flue gases from existing coal-fired boilers
have been lowered to 200 - 350 ppm by combustion modification including
staged combustion and the use of low-NOx burners. For further abatement,
selective catalytic reduction (SCR) has started to be applied to several
coal-fired boilers. The first full-scale combination system of SCR
and FGD was put into operation in April 1980. The plant cost for
SCR is about one-third that for FGD. A new combustion technology has
also been developed in attempts to lower NOx below 100 ppm.
Preceding page blank
85
-------
1. COAL USAGE AND POLLUTION CONTROL IN JAPAN
Most utility power companies in Japan switched fuel from coal to
oil between 1960 and 1974 except Electric Power Development Co. (EPDC)
which was established by the Japanese government jointly with major
power companies to use domestic coal. Due to the recent rise in oil
and gas prices, power companies have started to construct new coal-fired
boilers (Table 1), most of which will use imported coal because the
supply of domestic coal is limited to 20 million tons yearly. Although
Japan has imported over 60 million tons of coal yearly, all of the
imported coal has been for coke production for the steel industry. The
import of fuel coal has been started and is expected to reach 45 million
tons in 1990.
Major problems with coal usage are (1) emissions of SO,,, NOx and
particulates on combustion, (2) handling and storage problems, and (3)
ash disposal. Those problems are serious in Japan where a large
population is concentrated in a small land space. The new boilers
are to be located in regions relatively far from large cities and
industrial districts, where the environmental regulations by the Central
Government are not quite stringent. However, in order to construct
a large plant, it is necessary to make an agreement with local governments,
by which extensive countermeasures for pollution control are necessitated.
All of the new coal-fired boilers will need FGD. NOx concentrations
in flue gas from major coal-fired boilers has been reduced to 200 -
350 ppm while the emission standard by the Central Government is
400 ppm for new boilers and 480 ppm for existing ones. Further reduction
will be needed for new boilers. Some power companies have started to
apply selective catalytic reduction (SCR) which usually removes about
80% of NOx (Table 1).
A new combustion technology to lower NOx concentration below
100 ppm with coal and below 50 ppm with oil has been developed.
(Section 6.2).
Particulates can be removed sufficiently by a combination of
electrostatic precipitator and wet FGD. A bag house has been tested
but has not been considered promising for a large boiler.
In attempts to solve the coal handling problem, coal-oil mixture
(COM) has been studied extensively and may be used for some of the new
boilers. The major drawback with COM is that more than half of the
energy is derived from oil. To save oil, coarse-grain COM has been
tested, which uses up to 6 mm grains of coal which is transported with
oil as a slurry and separated from oil for burning.
The largest problem with coal usage may be ash disposal, because
landspace for discarding is limited. New uses of the ash, as feedstock
86
-------
Table 1 Coal-fired utility boilers in Japan
(Larger than 175 MW)
Year of Completion
Power
company
EPDC
it
it
ii
ii
ii
Chugoku
ii
Hokkaido
ti
Kyushu
ii
Joban Kyodo
Tohoku
Tokyo
Power
station
Isogo
Takasago
Takehara
Matsushima
Matsuura
Mito
Shimonoseki
Misumi
Tomato-Atsuma
Sunagawa
Matsuura
Reihoku
Nakoso
Noshiro
Mito
Boiler
No.
1
2
1
2
1
3
1
2
1
2
1
1
1
1
4
1
2
1
2
8
9
1
2
1
2
Capacity
MW
^265
265
250
250
250
700
500
500
1,000
1,000
1,000
175
700
350
125
700
700
700
700
600
600
600
600
1,000
1,000
a
Boiler FGD SCR
1967 1976
1969 1976
1968 1975
1969 1976
1967 1977 1981
1982 1982 1982
1981 1981
1981 1981
1984b
1986b
1988b
1967 1979 1980
1985b
1980 1980 1980°
1982 1982
1984b
1988b
1987b
1989b
1983d 1983
1983d 1983
1985b
1985b
1988b
1988b
a Selective catalytic reduction of NOx
b Planned.
c Treating one-fourth of the gas.
d Mostly oil will be used with less coal for a while without FGD.
87
-------
for cement production replacing clay, as filler for asphalt, as raw
material for aggregate, etc., have been developed.
Studies have been carried out also on fluidized bed combustion
(FBC), gasification, and liquefaction of coal, but not as extensively
as in the USA. The major problem with FBC in Japan is the difficulty
in disposing of the ash containing lime and calcium sulfate. Tests
have been conducted in search for an S02 absorbent that can be separated
from ash, regenerated and recycled, but so far do not seem promising.
Gasification and liquefaction may not be suitable to Japan which has
to depend on imported coal, since a considerable portion of energy
of coal is consumed by gasification or liquefaction. Although
liquefaction may be important in future, the plant may have to be
constructed abroad and the product imported.
2. STATUS OF FGD FOR COAL-FIRED UTILITY BOILERS
Before 1979, FGD plants for coal-fired utility boilers were limited
to the 5 plants of EPDC. Among the EPDC plants, two at Takasago Station
had an appreciable scaling problem until 1977 mainly at the mist
eliminator which had been washed with a circulating liquor saturated
with gypsum.1) By using fresh water together with the liquor for the
wash, the scaling problem was solved.!»2) Since 1978, all of EPDC's
FGD plants have been operated with virtually 100% operability and
reliability (Table 2).
Table 2 Operation hours of EPDC's boilers and FGD plants
(April 1978 through March 1979)
Operation hours
Boiler Boiler (A)* FGD (B) B/A (%)
Isogo
Takasago
Takehara
No.
No.
No.
No.
No.
1
2
1
2
1
7,705
8,206
7,829
8,167
7,583
7,705
8,206
7,823
8,147
7,580
100.0
100.0
99.92
99.75
99.95
* When an FGD plant is shut down due to its own trouble,
the boiler is operated by using low-sulfur oil.
Therefore, B/A (%) shows operability as well as
reliability.
88
-------
Operation parameters of the plants are shown in Table 3. Although
the plants are highly reliable and removes over 90% of 862 and over
70% of fly ash, they have the following drawbacks: (1) A large gas
pressure drop due to the use of a venturi or perforated plate scrubber
to attain a high dust removal efficiency, which results in a large
power consumption. (2) Requirement of a large amount of water for gas
cooling and also for purging wastewater from the system in order to
maintain chlorine in the scrubber liquor below a certain level for
corrosion prevention. (Usually more than half of the water charged
to the FGD system is volatilized in the prescrubber).
In order to lower the pressure drop, new FGD plants, including
Chugoku Electric's Shimonoseki plant constructed by MHI and two
EPDC plants at Matsushima under construction by Babcock Hitachi and
IHI, use a spray tower for gas cooling and particulate removal. A gas-
gas heater (heat exchanger) is used for the new plants as well as the
Tomato-Atsuma plant of Hokkaido Electric in order to cool the FGD inlet
gas to save water and to heat the FGD outlet for energy conservation.
Dry processes for FGD have received attention as a possible way
for further improvement and also because of the convenience for use
in conjunction with selective catalytic reduction of NOx. An activated
carbon process has been tested at EPDC's Takehara station. (Section 6.1)
The Electric Power Industry Federation also is to make pilot plant
tests on activated carbon processes for coal-fired boilers at 3 power
stations.
3. NOx ABATEMENT AND COMBINATION OF SCR AND FGD
3.1 NOx Regulation and Selective Catalytic Reduction (SCR)
NOx concentration in flue gases from coal-fired boilers has been
restricted by the emission standards by the central government to a
level below 480 ppm for existing boilers and below 400 ppm for new
boilers. The concentration can be achieved by combustion modification
without appreciable difficulty. Most local governments, however,
enforce much more stringent regulations. For example, Yokohama City,
in an effort to lower the ambient N02 concentration from the current
0.06 - 0.07 ppm in daily average to 0.04 ppm, has asked EPDC's Isogo
Station to lower to 169 ppm the NOx concentration in flue gases from
the existing two 265 MW coal-fired boilers. EPDC has lowered the
NOx concentration to 200 ppm by combustion modification including
staged combustion and low-NOx burner and has been making further efforts
to meet the requirement. Isogo Station has a limited landspace in
which they managed to retrofit FGD plants and has no more space to
install a flue gas treatment (FGT) plant for NOx removal. Therefore,
89
-------
Table 3 Operation parameters of FGD plants for coal-fired utility boilers
Power company
Station
Boiler No.
Boiler capacity (MW)
FGD constructor
FGD start-up
Gas treated (1,000 Nm3/hr)
Inlet S02 (ppm)
Inlet dust (mg/Nm3)
Prescrubber (cooler)
Type
L/G (liters/Nm3)
Scrubber (S02 absorber)
, Type
L/G
Outlet S02 (ppm)
Outlet dust (mg/Nm3)
S02 removal efficiency (%)
Dust removal efficiency (%)
Pressure drop (mm I^O)
Wastewater (t/hr)
Energy requirement (%)n
Reliability (%)*
EPDC
Isogo
1
265
IHla
May '76
821
450
1,500
EPDC
Takasago
1
250
Mitsuib
Feb. '75
792
1,500
100
EPDC
Takehara
1
250
BHC
Feb. '77
793
1,730
400
Venturi
7
Venturi
6
Venturi
2.5
Venturi
7
25
50
94.4
96.6
360f
10
2.9
100.0
Venturi
6
100
30
93.3
70.0
325f
7.5
3.2
99.9
ppe
7
100
50
94.2
87.5
615f
12
3.3
100.0
EPDC
Matsushima
1
500
IHIa
Jan. '81
1,826
1,000
300
2
500
BHC
Jan. '81
1,826
1,000
300
Spray
Spray
2.8
Spray
13.4
50
30
95.0
90.0
Spray
15
50
30
95.0
90.0
133§
Chugoku
Shimonoseki
1
175
MHId
July '79
586
1,310
830
Spray
3
Packed
14
55
50
95.8
94.0
120f
15
2.1
100.0
Hokkaido
Tomato
1
350
BHC
Oct. '80
1,268
232
45
Venturi
PPe
23
90.0
a Ishikawajima-Harima Heavy Industries
c Babcock Hitachi K.K.
f By two scrubbers and mist eliminators
h Percent of power generated
b Mitsui Miike Machinery Co.
d Mitsubishi Heavy Industries e Perforated plate
g By two scrubbers
i EGD operation hours percent of desired FGD operation hours
-------
they need to reduce NOx further by improved combustion. Even more
stringent regulations may be applied for new larger boilers, necessitating
FGT.
Among many ways of FGT developed in Japan, selective catalytic
reduction (SCR) that uses ammonia and catalyst at 300 - 400°C is by
far the most advanced method, which has been used in constructing
about 100 commercial plants mainly for flue gas from oil-fired boilers.
The advantages of SCR over other FGT processes are simplicity and
reliability which enable unattended operation, lack of the by-product
disposal problem, and relatively low cost. SCR is conveniently applied
to flue gas leaving a boiler economizer at 300 - 400°C. The major
reaction is shown below:
4NO + 4NH3 + 02 = 4N2 + 6H20
At the early stage of development, SCR encountered the following
technical problems, most of which have been solved by recent improvements:
(1) Catalyst poisoning by SOx in flue gas. (2) Catalyst pluggage
by dust. (3) Catalytic oxidation of a portion of S02 to S03. (4) Leak
ammonia from SCR reactor, which reacts with 863 and H20 to form
ammonium bisulfate deposit in an air preheater.
Many of the catalysts developed recently are based on Ti02 with
small amounts of V20^ and other components, are resistant to SOx,
and oxidize about 1% or less S02. In order to prevent dust plugging
of the catalyst, parallel flow type reactors with honeycomb, plate,
and tube catalysts have been used for dusty gases such as coal-fired
boiler flue gas.
More than 90% of NOx can be removed by using over 1 mol NH3 to
1 mol NOx as shown in Figure 1. However, 80% removal has been generally
applied to utility boilers as the optimum control level, because
compared with 90% removal, it requires about 40% less catalyst resulting
in the reduction of cost as well as pressure drop and also because
it can reduce leak ammonia to a low level (5 ppm or below) to minimize
the deposit of ammonium bisulfate in the air preheater. Over 90%
removal with a low leak NH3 is difficult for a large boiler because
.the gas velocity as well as NOx concentration is not uniform in different
parts of the duct.
Low-temperature catalysts active at 150 - 250°C have also been
developed but have not been used commercially yet because ammonium
bisulfate forms on the catalyst and lowers its activity. Ammonium
bisulfate can be removed by heating the catalyst to over 350°C. The
low-temperature catalyst may not be suitable for boilers for which
economizer outlet gas around 350°C can be treated but may be useful
for other sources for which only cold gas around 200°C is available.
91
-------
g
0)
t-i
100
90
80
70
60
20
_L
10
e
a
a.
cd
0)
1-1
0.6
0.7
0.8 0.9
mole ratio
1.0
1.1
Figure 1 Performance of honeycomb catalyst for coal-fired boiler
flue gas ( Inlet NOx 300 ppm, at 370 °C. SV means space
velocity: flue gas volume per hour divided by catalyst
volume. For high-dust system)
High-dust system
B
320-400
/
SCR
320-400
s
APR
150-160
V
/
ESP
150-160
^
FGD
Low—dust system
B
320-400
Hot
ESP
320-400
SCR
320-400
?
APH
150-160
/
FGD
Figure 2 Systems for coal-fired boiler flue gas treatent (Figures show
gas temperature, °C.
B: Boiler APH: Air preheater)
92
-------
3.2 Combination of SCR and FGD
At an early stage of development, the SCR reactor was placed
downstream of FGD in order to reduce SOx poisoning and dust plugging.
This system, however, requires a large amount of energy for heating
the gas after FGD and has not been used since SOx-resistant parallel
flow type catalysts have been developed. Figure 2 shows two combination
SCR/FGD systems currently used for coal-fired boilers. In both
systems, the economizer outlet gas at 330 - 400°C is treated by SCR,
cooled to 150°C by an air preheater, and then subjected to FGD. The
high dust system treats the gas with full dust load (15 - 25 grams/Nm3)
by SCR, and therefore the catalyst should be hard in order to avoid
erosion by dust and thus is less porous and may not be highly
active. On the other hand, the low dust system uses a hot electrostatic
precipitator (ESP) upstream of SCR, which is suitable for dust removal in
flue gas from low-sulfur coal. The hot ESP usually reduces the dust
to 100 - 200 mg/Nm3 and protects the catalyst from erosion. However,
the dust leaving the hot ESP is finer and richer in alkaline components
and tends to deposit on the catalyst surface. The problem of ammonium
bisulfate deposit in the air preheater is also appreciable with the
low dust system while it is insignificant with the high dust system
(Section 5.3). Therefore, leak ammonia should be kept at a lower level
with the low dust system than with the high dust system.
As shown in Table 4, the Shimonoseki plant, Chugoku Electric
uses the high dust system while the Tomato-Atsuma plant of Hokkaido
Electric and the plants at Takehara, EPDC use the low-dust system.
Two plants at Nakoso, Joban Joint Electric will use the high dust system.
Table 4 SCR plants for coal-fired utility boilers
Company
Chugoku
Hokkaido
EPDC
EPDC
Joban
Station
Shimonoseki
Tomato-Atsuma
Takehara
Takehara
Nakoso
Capacity
(MW)
175
350 x 1/4
250 x 1/2
250 x 1/2
700
600
600
Vendor
MHI
BH
BH
KHIa
ndb
MHI
IHI
bUK
Dust
High
Low
Low
Low
Low
High
High
type
Catalyst
Honeycomb
Plate
Plate
Tube
,b
nd
Honeycomb
Honeycomb
Comp-
letion
1980
1980
1981
1981
1982
1983
1983
a Kawasaki Heavy Industries
b Not decided
93
-------
The flue gas leaving the SCR reactor contains a small amount of
ammonia, which is caught by a prescrubber of the FGD system. Although
ammonia has no adverse effect either on the operation of wet lime/
limestone process FGD or on the quality of by-product gypsum, it is
contained in a small amount in wastewater from the FGD system. If
needed, the ammonia in the wastewater can be removed by a conventional
biochemical treatment (activated sludge process) or by ammonia stripping.
The latter has been used at the Owase plant, Chubu Electric while the
former is to be used at the Takehara plant, EPDC.
3.3 SCR Cost
Examples of SCR plant cost for utility boilers are shown in
Table 5. The cost for the new gas-fired boiler at Chita was 1,860
yen/kW, while that for the new oil-fired boiler at Kudamatsu was
2,860 yen/kW. Those for existing oil-fired boilers at Kudamatsu and
Chita were considerably higher than that for the new oil-fired boiler,
because of complicated duct work for retrofitting (Kudamatsu and Chita)
and the requirement of additional fans (Kudamatsu). The SCR plant for
coal at Shimonoseki is more costly than for oil.
The difference in cost with the fuel type is due mainly to the
amount of catalyst needed. Generally speaking, an active pellet
catalyst can be used for clean gas, while for flue gas from oil
containing 20 - 100 mg/Nm3 of dust, a honeycomb catalyst with a
channel size of 6 - 7 mm and wall thickness of 1 - 1.5 mm consisting of
SOx resistant material has been used in a volume 3-4 times that of
the pellet catalyst. For coal, the catalyst volume may be nearly
double that for oil because of a larger channel size of honeycomb for
dust plugging prevention and a harder structure for erosion prevention
resulting in lower activity.
Estimated SCR costs for new 700 MW utility boilers using coal and
low-sulfur oil are shown in Table 6. Honeycomb catalyst is used for
both oil and coal. The assumed channel size and wall thickness in
millimeters are 6.6 and 1.4 for oil, 7.4 and 1.6 for coal with the
low-dust system, and 8.2 and 1.8 for coal with the high-dust system.
Leak ammonia is maintained below 10 ppm for oil (low sulfur) and coal
with the high dust system while it is kept below 5 ppm for coal with
the low-dust system which is liable to air preheater plugging. Based
on those assumptions, an equal space velocity was assumed for high and
low dust systems of coal. The space velocity is about one-half that
for oil.
The investment cost including civil engineering and test operation
for 80% NOx removal is nearly 4,000 yen/kW for oil and nearly 7,000
yen/kW for coal, while the cost for 90% removal is higher by about 30%
for oil and 40% for coal. The annualized SCR costs in yen/kWhr for
80% removal, assuming 7 years' depreciation, 70% boiler utilization,
94
-------
Table 5 Cost of SCR plants for utility boilers (in battery limits)
NOx
Space . Plant cost Year
in
Power
company
Chubu
Chubu
Chugoku
Chugoku
Chugoku
Power
station
Chita
Chita
Kudamatsu
Kudamatsu
Shimonoseki
Boiler
(MW)
700
700
375
700
175
Fuel
Gas
Oilb
Oilb
oiib
Coal
New or
retrofit
New
Retrofit
Retrofit
New
Retrofit
removal
(%)
Over 80
Over 80
Over 80
Over 80
Over 50°
Const-
ructor
BH
MHI
IHI
IHI
MHI
Catalyst
type
Pellet
Honeycomb
Honeycomb
Honeycomb
Honeycomb
velocity
(hr-1)
20,000
6,000
5,500
5,500
3,000
i
I09yen
1.3
2.2
2.2
2.0
1.7d
yen
kW
1,860
3,570
5,870
2,860
9,710d
com-
plete<
1977
1980
1979
1979
1980
a Flue gas volume per hour divided by catalyst volume
b Low-sulfur oil
c Catalyst for 50% removal has been used to meet the current regulation, while the SCR system has been
designed for 80% removal.
d Including boiler modification for economizer bypass.
-------
Table 6 Estimated SCR cost for new 700 MW utility boilers
Annual power generation 4,292,400 MWhr. 70% utilization.
LeakNH3: 5-10 ppm for oil and coal with high-dust
system. Less than 5 ppm for coal with low-dust system
Coal
Fuel Oil (low S) (high and low dust)
Flue gas, Nm3/hr. (NOx ppm)
NOx removal efficiency (%)
Space velocity (hr )
Investment cost
Q
Catalyst (billions of yen)
Other ( " )b
Total ( " )b
Total (1,000 yen/kW)
Annual cost (billions of yen)
Capital cost
Catalyst
Other6
Total
Annualized cost (yen/kWhr)
(1,000 yen/Nm3 of NOx removed)
2,000,000
80
5,100
1.22
1.50
2.72
3.89
0.50
0.61
0.27
1.38
0.32
1.15
(120)
90
3,400
1.82
1.75
3.63
5.10
0.62
0.91
0.31
1.84
0.43
1.39
2,300,000
80
2,700
2.81
2.00
4.81
6.87
0.78
2.81
0.48
4.07
0.95
1.20
(300)
90
1,700
4.46
2.30
6.76
9.66
1.02
4.46
0.55
6.03
1.40
1.58
3 3
a 3.1 million yen/m for oil, 3.3 million yen/m for coal.
b Including civil engineering and test operation.
c Interest (10%) on initial charge of catalyst and interest and depreciation
(25%) on investment cost excluding catalyst.
d Catalyst life: 2 years for oil and 1 year for coal.
e Ammonia, power, etc.
96
-------
and a catalyst life of 2 years for oil and 1 year for coal, are 0.32
for oil and 0.95 for coal, while the costs per unit amount of NOx
removed is just about equal for oil and coal. Compared with 80%
removal, 90% removal costs about 40% more iri yen/kWhr. Actually
90% NOx removal may be difficult for a large boiler without increasing
leak NH3, because gas velocity as well as NOx concentration may not
be uniform in different parts of the SCR reactor inlet.
For coal, about 70% of the annualized SCR costs is accounted
for by catalyst. If the catalyst is useful fpr 2 years, the costs
will be lowered by about 35%. The catalyst life is usually guaranteed
for 1 year for both oil and coal. Operation experiences have shown
that the catalyst for oil may be useful for over 3 years. It may be
possible to extend catalyst life for coal to 2 years,
4. SHIMONOSEKI PLANT, CHUGOKU ELECTRIC
4.1 Outline
Shimonoseki Station of the Chugoku Electric Power Co. has two
boilers — a 175 MW coal-fired boiler (No. 1) and a 400 MW oil-fired
boiler (No. 2). Regulations for the station are shown in the following
table.
Table 7 Regulations for Shimonoseki Station
Air pollution control
k Value 2.7 (Ground level concentration 0.0047 ppm)
SOx (total) Below 412 Nm3/hr
Particulates Below 130 kg/hr
No. 1 Boiler Below 200 mg/Nm3
No. 2 Boiler Below 40 mg/Nm3
NOx Below 330 Nm3/hr
No. 1 Boiler Below 350 ppm
No. 2 Boiler Below 170 ppm
Floating particulates Below 0.2 mg/Nm3
97
-------
Water pollution control
pH 5.8 - 8.6
Suspended solids | Below 12 kg/day
' Below 15 mg/liter
Normal-hexane-soluble
material
Chemical oxygen demand
J Below 0.8 kg/day
» Below 1 mg/liter
( Below 12 kg/day
\ Below 15 mg/liter
The No. 1 boiler was completed in 1967 and was burning coal and
oil in the ratio of 25 to 75 before a full scale FGD plant was completed
in July 1979 using the MHI wet limestone-gypsum process. After the
FGD plant was put into operation, coal and oil was used in the ratio
of 50 to 50. It was difficult to use larger amounts of coal because
of the NOx regulation (below 350 ppm) . Although the regulation may be
met by combustion modification even with the burning of coal only, it
was likely that further NOx reduction might be required in future.
Chugoku Electric, therefore, decided to install a full-scale SCR unit,
which was completed in March 1980 to allow combustion of coal only.
The SCR unit is the first full-scale plant for a coal-fired boiler
in the world and has the nature of a demonstration plant.
Figure 3 shows the combined system of SCR and FGD for the No. 1
boiler. The flue gas is first subjected to SCR at 330 - 400°C, passed
through two trains of air preheaters and dust collectors (multicyclone
and ESP), and then undergoes FGD after it is passed through a heat
exchanger.
The No. 2 boiler is a relatively new one and has used a high-sulfur
oil with FGD by the MHI wet limestone-gypsum process.
4.2 SCR System
The design basis of the SCR system is shown below:
Boiler capacity 175 MW
Fuel Coal
Gas flow rate 550,000 Nm3/hr
Gas temperature 370°C
Inlet NOx 500 ppm
Outlet NOx 250 ppm (100 ppm in future)
98
-------
Boiler
ID
Stack
Heat exchanger
(Gas-gas heater)
SCR « Selective catalytic reduction of NOx
MC : Multicyclone
Figure 3 Flue gas treatment system for No.l coal-fired boiler (175 MW)
(Shimonoseki Power Station, Chugoku Electric)
-------
NOx removal efficiency 50% (80% in future)
Inlet SOx 1,600 ppm
Reactor One reactor, with downflow of gas
Catalyst Honeycomb. Square type with 10 mm
pitch (about 8.2 mm opening)
Space velocity 3,000 hr"1
The No. 1 boiler is for base load and the gas temperature at
economizer outlet is normally around 360°C, suitable for SCR. The
load is occasionally lowered to 25% of full load, resulting in the drop
of the gas temperature to 300°C. Since ammonium bisulfate may deposit
on the SCR catalyst during the low-load operation, a bypass system was
installed as shown in Figure 3 to control the gas flow by dampers
to mix a portion of hot gas with the economizer outlet gas to maintain
the gas temperature.
An SCR reactor was installed beside the boiler so that the treated
gas is sent to the existing air preheaters. The reactor contains
5 horizontal layers of honeycomb catalyst, through which flue gas is
passed downwards. The flue gas contains about 410 ppm NOx, 360 ppm
SO^ and nearly 20 grams/Nm3 of fly ash. A layer of "dummy" spacer
with the same shape as the honeycomb was placed on top of the first
honeycomb layer, in order to maintain a uniform parallel gas flow and
to prevent catalyst erosion by fly ash.
Planning and design of the SCR system was started in July 1979.
Construction was begun in October 1979. Boiler modification and
reactor connection were performed during the shutdown of the boiler
for annual maintenance between February 1 and March 31, 1980. Since
start-up of operation in April 1980, the boiler, the SCR system and
the FGD system have been operated without trouble.
Current regulations require about 50% NOx removal. Therefore,
a NH3/NOx mole ratio of about 0.56 has been used to reduce NOx
concentrations from 410 to 185 ppm (55% removal) and to maintain leak
NH3 at reactor outlet below 3 ppm. In future, 80% of NOx may be
removed by increasing the amount of catalyst and by using about 0.82 mol
NH3 to 1 mol NOx, keeping leak NH3 below 5 ppm.
A catalyst life of 1 year is guaranteed by MHI, which will take
all of the spent catalyst when fresh catalyst is placed. Replacement
of catalyst will require 15 days with 15 workers working 7 hours a day.
The air preheater has had a soot blow system on the cold side which
has been operated 4 times a day, two hours each time. When the SCR
system was installed, an additional soot blow system was installed on
the hot side of the preheater, which has also been operated 4 times
100
-------
a day, 2 hours each time. The plugging problem of the preheater by
ammonium bisulfate has thus been prevented. The soot blow system will
be used less frequently.
The total investment cost was about 2 billion yen including the
boiler modification of which 1.7 billion was paid to the constructor.
4.3 FGD System
A flow sheet of the FGD system is shown in Figure 4. Flue gas
leaving the air preheater at 160°C is cooled to about 95°c by a
Ljungstrom type heat exchanger and introduced into a semiventuri type
spray scrubber newly developed by MHI for particulate removal, and
then into a grid packed tower with a holding tank at the bottom and
a mist eliminator at the top. Limestone slurry is fed to the tank.
The treated gas at 55°C is heated to 120°C by the heat exchanger
eliminating gas heating by oil firing. About 90% of both S02 and
particulates are removed (Tables 3 and 5). Slurry handling systems —
oxidation of calcium sulfite, gypsum centrifuge, etc., are similar to
those of the standard MHI process.2)
After its startup in July 1979, the FGD plant was operated
continuously without trouble until February 1980, when the boiler was
shut down for annual maintenance. During the operation period, coal
and oil were used in the ratio of 25 to 75 at the beginning and then
in the ratio of 50 to 50. Fresh water, at the rate of 30 tons/hr?
was fed mainly to the syray tower and used partly for mist eliminator
wash. Of the 30 tons/hr, 13 tons were volatilized, 2 tons went into
gypsum as water of crystallization and moisture, and 15 tons were
sent to a wastewater treatment system.
Inspection during the shutdown period detected a little deposit
of particulates in the heat exchanger and a slight erosion of rubber
lining but neither scaling nor corrosion. The soot blow system was
reinforced during the shutdown period in order to eliminate the deposit
formation in the heat exchanger.
Since its restart in April, using coal only this time, the FGD
system has been operated trouble-free again. Because a fan is placed
upstream of the heat exchanger, a small amount of inlet gas at 160°C
leaks in the heat exchanger to mix with the FGD outlet gas, thus
lowering the removal efficiency of SO? and particulates to some extent
(Table 8). Placing the fan between the heat exchanger and the
prescrubber (co.oler) results in the leak of the FGD outlet gas to the
inlet and an increase in removal efficiency, but it may cause corrosion
of the fan due to condensation of sulfuric acid at low temperatures
around 90°C. MHI has been testing a new type of air preheater without
gas leakage.
101
-------
Fan
o
Stack
l"i
VvV
Dust
collector
160°C
Fan
r-
120°C
Heat
exchanger
(Gas-gas
heater)
Water
(22t/hr)
Water
I(St/hr)
Mist
eliminator
Wastewater(10t/hr)
Limestone
^-^ to_ scrubber
l
Oxidizer
Air
rQ
Cooler
Scrubber
Wastewater(5t/hr)
T
Thick-
ener
f
^r-
£
/
*-i" >
Centr
Tank (j
n
Figure 4 Flowsheet of FGT) system for No.l boiler at- Shimonoseki Power Station
-------
Table 8 862 and particulate removal efficiency(Shimonoseki plant)
Concentration and Coal and oil Coal only
Pollutants removal efficiency (50 ; 50) Low S Medium S
S02 FGD inlet (ppm) 1,230 355 1,310
FGD outlet (ppm) 78 20 55
Removal efficiency(%) 93.7 94.4 95.8
HEaoutlet (ppm) 136 38 115
Removal efficiency(%) 89.0 89.2 91.2
Particulates FGD inlet (mg/Nm3) 200 1280 830
FGD outlet(mg/Nm3) 12 80 50
Removal efficiency(%) 94.0 93.8 94.0
HEa outlet (mg/Nm3) 21 130 85
Removal efficiency(%) 89.5 89.8 89.7
a Heat exchanger
Ammonia contained in a small amount in flue gas has had no adverse
effects on FGD and on the quality of fly ash which has been used for cement
and land fill. Also, ammonia has been injected into the flue gas from the
No.2 oil-fired boiler between the air preheater and ESP in order to prevent
corrosion of ESP and to increase soot removal efficiency. Thus ammonia is
contained in the flue gas introduced into the No.2 FGD system, which has
also been operated without trouble.
Chugoku Electric recently decided to install similar SCR and FGD
system for 5 relatively small existing coal-fired boilers.
5. OTHER COMBINED SYSTEMS
5.1 Takehara Plant, EPDC
EPDC has been constructing a full-scale demonstration plant of
SCR combined with FGD at its Takehara Station for the No. 1 boiler
(250 MW). Since various types of coals including low-sulfur coal will
be used, a hot electrostatic precipitator is installed. As shown in
Figure 5, all of the flue gas from the boiler is passed through two
parallel trains of a hot ESP, SCR reactor, air preheater and ID fan.
One of the reactors is constructed by Babcock Hitachi Ltd. using a
plate catalyst developed by Hitachi Ltd., while the other is constructed
by Kawasaki Heavy Industries (KHI) using a tubular catalyst. Over 80%
103
-------
Addition for demonstration test
J
B Boiler, APH Air preheater, ESP Electrostatic precipitator
IDF Induced fan, HESP Hot electrostatic precipitator
Figure 5 Demonstration plant at Takehara, EPDC (250 MW)
Conventional type
Deposit
Hot
Intermediate
Cold
Soot
blow
Modified type
Soot
blow
Hot
Combined Intermediate
and cold
Soot
blow
Figure 6 Arrangement of air preheater elements
104
-------
of NOx will be removed maintaining leak NHj below 10 ppm.
Since the air preheater treats an SOx-rich, dust-lean gas, ammonium
bisulfate may deposit in intermediate and low temperature zones
(Figure 6). Pilot plant tests have shown that the deposit formed
between the two zones is difficult to remove by soot blowing. For
the demonstration plant, a modified design of the air preheater
elements as shown in Figure 6 will be used to reduce the plugging
problem.
The treated gas is sent to an existing FGD plant constructed by Bab-
cock Hitachi using the limestone-gypsum process (Table 3). The leak
NH3 will be caught by the FGD system and contained in the wastewater.
EPDC has installed a wastewater treatment system using a conventional
activated sludge process to remove ammonia, because Takahara Station
faces the Seto Inland Sea which is sometimes plagued by the red tide
problem.
The total additional system for the demonstration as shown in
Figure 5 cost 8 billion yen including control systems and a storage
and injection system of ammonia. The new ID fans are estimated to
consume about 1,500 kW more than does the existing ID fans, which is
equivalent to 0.6% of the power generated by the boiler.
EPDC will construct a full scale combined system for the new
No. 3 boiler (700 MW), for which the low-dust system may also be applied.
5.2 Tomato-Atsuma Plant, Hokkaido Electric
Hokkaido Electric Power Co. has constructed a new 350 MW coal-fired
boiler in a newly opened industrial region near Tomakomai, which has
started test operation in summer 1980 and is scheduled to be put in
commercial operation in October 1980 using a low-sulfur coal (S = 0.3%).
By an agreement with local governments, SOx emissions should be kept
below 180 Nm3/hr (about 140 ppm), NOx below 200 Nm3/hr (about 160 ppm),
and particulates below 200 kg/hr (about 160 mg/Nm3).
For SOx abatement, half of the gas from the boiler is treated by
a wet limestone-gypsum process FGD plant constructed by Babcock Hitachi.
NOx is reduced below 200 ppm by combustion modification including staged
combustion, flue gas recirculation, and dual-register low-NOx burners.
In addition, one-fourth of the gas is treated by SCR for 80% NOx removal
to meet the agreement.
Since a low-sulfur coal is used, a hot electrostatic precipitator
has been installed which reduces the dust content down to 45 mg/Nm3.
One-fourth of the gas passing through the hot ESP is treated by an SCR
reactor containing a plate catalyst developed by Hitachi Ltd. An
economizer bypass system has been installed to maintain the gas
temperature above 300°C.
105
-------
Hokkaido Electric plans to install a 600 MW coal-fired boiler.
If the plan is authorized, Hokkaido Electric plans to reevaluate
the design including the necessity of the bypass and the use of cold
vs. hot ESP.
5.3 Nakoso Plant, Joban Joint Electric Co.
Tokyo Electric Power Co., jointly with Tohoku Electric Power Co.,
Joban Joint Electric Co., and MHI, has carried out pilot plant tests
at Nakoso Station of Joban on combined systems of SCR (high-dust and
low-dust) and wet limestone-gypsum process FGD using 4,000 Nm^/hr
of flue gas from a coal-fired boiler. In 1979, the high dust system
was operated for 5,000 hours while the low-dust system was operated
for 4,000 hours. Further tests are in progress in 1980.
Honeycomb catalysts are used for both systems with downflow of
the gas. With the high-dust system, erosion of the catalyst by dust has
been prevented by placing on top of the honeycomb a dummy spacer which
has the same cross section as the honeycomb. The air preheater has
been kept clean by applying soot blowing once a day; ammonium bisulfate
has not deposited appreciably because of the cleaning effect of fly
ash. With the low-dust system, the dust leaving the hot ESP is in a
small amount but consists of fine particles which are rather sticky
and tend to deposit particularly at the inlet of the honeycomb.
Moreover, the air preheater requires soot blowing 3 times a day to
prevent the deposit of ammonium bisulfate.
The FGD system has been operated without trouble. A semiventuri
type spray scrubber developed by MHI is used for the prescrubbing. Tests
indicated that the dust contained in the gas in concentrations of 100,
200, and 300 mg/Nnr* was reduced to about 20, 30, and 40 ppm, respectively,
by the prescrubber and to about 15, 20, and 30 ppm, respectively by
the S02 absorber.
Joban has started to construct 2 new boilers with a capacity of
600 MW each, which will use low-sulfur oil with a small amount of coal
to start with. Both boilers will have high-dust system SCR units with
a honeycomb catalyst. The units for one of the boilers will be
constructed by MHI and the units for the other boiler by IHI. FGD
plants may be constructed when larger amounts of coal are used.
106
-------
6. OTHER MAJOR ACTIVITIES
3\
6.1 Pilot Plant Tests by Activated Carbon Process
EPDC, jointly with Sumitomo Heavy Industries, has been operating
a pilot at Takehara with a capacity of treating 10,000 Nm3/hr of flue
gas from the No. 1 coal-fired boiler to remove over 90% of S02 and over
30% of NOx by activated carbon and ammonia. A flowsheet of the pilot
plant is shown in Figure 7. The flue gas containing 1,300 ppm of S02
and 320 ppm of NOx at about 150°C is injected with 225 ppm NH3 and is
introduced in a reactor with activated carbon in a moving bed. Over
90% of S02 is adsorbed by the carbon to form sulfuric acid and ammonium
sulfate (reactions 1 and 2) while over 30% of NOx is converted to N2
(reaction 3).
S02 + H20 + 1/2 02 •*• H2S04 (1)
H2S04 + 2NH3 •*• (NH4)2S04 (2)
4NO + 4NH3 + 02 + 4N2 + 6H20 (3)
The char loaded with the sulfur compounds is heated in a separate
moving bed to over 350°C by inert gas produced by incomplete combustion
of LPG gas. Concentrated S02 gas is released by the heating (reactions
4 and 5), then is introduced into a coal-bed reactor and converted
to S by the Resox process developed by Foster Wheeler Co. (reaction 6).
The sulfur vapor is condensed to recover elemental sulfur. The gas leaving
the condenser is incinerated and sent to the existing wet limestone-gypsum
process FGD plant.
H2S04 + 1/2 C -*• S02 + 1/2 C02 + H20 (4)
(NH4)2S04 + 02 •*• S02 + N2 + 4H20 (5)
S02 + C ->• S + C02 (6)
About 1.6% of the carbon is consumed in one cycle which takes
3 days. The sulfur condenser had a plugging problem, which has been
solved by applying a technology used for the Glaus furnace. The remaining
major problem is the low recovery of sulfur at 60 - 70%. Efforts have
been made to improve the recovery.
The low NOx removal efficiency is due to the low temperature. Over
200°C with over 2 mole NH3 to 1 mol NOx may be needed to attain over
80% removal. For commercial application, it may be preferable to use
SCR for the boiler economizer outlet at 300 - 400°C and then apply
the carbon process for S02 removal only without using ammonia. EPDC
107
-------
No.l Boiler
Electrostatic
Air Preheater Precipitator
Pilot plant
Wet Desox
Stack
Coal
i
Regenerator J Condenser
Tall gas
blower
Ammonia injection
unit
Inert gas
generator
Sulfur
Note:
Gas
____Activated
carbon
Incinerator
Figure 7 Pilot plant for FGH by activated carbon process for elemental sulfur recovery 3)
(Takehara f.3«ni-, F.PDC)
-------
is to install a prototype plant of the carbon process at its Matsushima
Station by 1982 to treat one-fourth of the gas from a new 500 MW
coal-fired boiler, while three-fourths of the gas will be treated by
the wet limestone-gypsum process.
6.2 New Combustion Technology
About one-tenth of fuel used for the boiler is injected above the
combustion zone in the boiler to form a reducing atmosphere where NOx
formed by the combustion is reduced to N2- Air is added above the
reducing zone for complete combustion. The technology was originated
by MHI and has been further developed by Tokyo Electric Power Co.
jointly with MHI, Hitachi, and IHI for NOx abatement for boilers.
Tests with pilot plants with a capacity ranging from 5,000 to 8,000 kW
using various fuels have indicated that about 50% of NOx is removed.
By using the process in combination with conventional combustion
modification, NOx concentration has been reduced to 10 - 20 ppm with
gas, 40 - 60 ppm with oil, and 60 - 100 ppm with coal. The boiler is
a little larger than a conventional boiler. Tests on a larger scale
are planned.
REFERENCES
1. Y. Nakabayashi, Plan, Design and Operating Experience of FGD For
Coal Fired Boilers Owned by EPDC, Paper No. 41, EPA FGD Symposium,
March 1979
2. J. Ando, S02 Abatement for Stationary Sources in Japan, EPA-600/
7-78-210, November 1978
3. EPDC and Sumitomo Heavy Industries,.Simultaneous SOx-NOx Removal
System for Coal-fired Boiler, October 1979
189
-------
Session 2: IMPACT OF RECENT LEGISLATION/REGULATIONS
Walter C. Barber, Chairman
Office of Air Quality Planning and Standards
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina
Panel: Impact of Recent Legislation/Regulations
Brief overviews of recent Legislation/Regulation,
under the CAA, CWA, and RCRA, followed by
questions from the audience.
Members: John W. Lum
Office of Water Planning and Standards
U.S. Environmental Protection Agency
Washington, D.C.
Penelope Hansen
Off ice of Solid Waste
U.S. Environmental Protection Agency
Washington, D.C.
No papers or discussions are included for this session.
Preceding page blank
-------
Sessions:- FGD RESEARCH AND DEVELOPMENT PLANS
Julian W. Jones, Chairman
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina
Preceding page blank
113
-------
RECENT TRENDS IN UTILITY
FLUE GAS DESULFURIZATION
by
M. P. Smith, M. T. Melia, and B. A. Laseke, Jr.
PEDCo Environmental, Inc.
and
Norman Kaplan
U.S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Emissions/Effluent Technology Branch
Preceding page blank
115
-------
ABSTRACT
PEDCo Environmental, Inc., under contract to the Industrial
Environmental Research Laboratory-RTF and the Division of Sta-
tionary Source Enforcement of the U.S. Environmental Protection
Agency, has been monitoring the status of utility flue gas
desulfurization (FGD) technology since 1974. Information for
this program is obtained by visits to plants having operational
FGD systems and through periodic contacts with representatives
of utility companies, FGD system and equipment suppliers, system
designers, research organizations, and regulatory agencies.
This paper summarizes the status of utility FGD technology
as of the end of August 1980 and indicates recent trends in both
the design and performance of the FGD systems. The discussion
of current status includes the number and capacity of operation-
al and planned FGD systems, as well as identification of the
systems according to process type, emission control strategy,
S02 inlet concentration levels, and removal efficiencies.
Process design developments and trends are summarized for the
major components and subsystems associated with commercial FGD
systems. In discussing FGD system performance, composite graphs
are included presenting annual system availability data (through
June 1980) for low-, medium-, and high-sulfur coal FGD instal-
lations. A statistical analysis of the data for the years 1978
and 1980 indicates overall trends in FGD system dependability.
Finally, capital and annual cost data (both reported and ad-
justed) are included for the operational FGD systems and cost
model comparisons are made.
The current data indicate that 203 FGD systems are either
operational, under construction, or planned (as of August 1980),
representing a total controlled capacity of about 97,000 MW. Of
the 203, 73 systems are operational, representing 27,155 MW of
controlled capacity. The dependability analysis indicates that
the overall median availability for these operational systems
has increased 1.5%, 16.5%, and 50.6% for low-, medium-, and
high-sulfur coal FGD installations, respectively, between the
years 1978 and 1980.
116
-------
NOTES
1. Company Names and Products.
The mention of company names or products is not to be
considered an endorsement or recommendation for use by the
U.S. Environmental Protection Agency.
2. Consistency of Information.
The information presented was obtained from a variety of
sources (sometimes by telephone conversation) including
system vendors, users, EPA trip reports and other technical
reports. As such, consistency of information on a partic-
ular system and between the several systems discussed may
be lacking. The information presented is basically that
which was voluntarily submitted by developers and users
with some interpretation by the author. The order of
presentation of information or the amount of information
presented for any one system should not be construed to
favor or disfavor that particular system.
3. Units of Measure.
EPA policy is to express all measurements in Agency docu-
ments in metric units. When implementing this practice
will result in undue cost or difficulty in clarity, IERL-
RTP provides conversion factors for the non-metric units.
Generally, this paper uses British units of measure.
The following equivalents can be used for conversion to the
Metric system:
British Metric
5/9 (°F-32) °C
1 ft 0.3048 m
1 ft2 0.0929 m2
1 ft3 0.0283 m3
1 grain 0.0648 gram
1 Ib (avoir.) 0.4536 kg
1 ton (long) 1.0160 m tons
1 ton (short) 0.9072 m tons
1 gal. 3.7853 liters
1 lb/106 Btu 429.6 ng/J
1 Btu/kWh 1055.056 J/kWh
117
-------
SECTION 1
INTRODUCTION
For more than 6 years PEDCo Environmental, Inc., under
contract to the U.S. Environmental Protection Agency (EPA), has
monitored the development and growth of flue gas desulfurization
(FGD) technology for fossil fuel-fired utility boilers in the
United States. The program provides an objective and current
perspective of FGD technology as applied to fossil fuel-fired
utility boilers and facilitates, through information dissemi-
nation, improvements in the design and performance of current
and future systems.
The program addresses performance of operational FGD sys-
tems, process and design characteristics of both operational and
planned systems, projected application and nature of future
processes and systems, and costs associated with both current
and planned systems. The program also includes the monitoring
of particulate matter scrubbers operating on coal-fired utility
boilers in the United States and FGD systems operating on coal-
fired utility boilers in Japan.
Program emphasis is on the performance of the operational
systems. Accurate portrayal of system performance requires data
concerning system/module dependability, operating problems anci
solutions, operating and maintenance costs, and outlet emissions
and removal efficiency. Data on outlet emissions of sulfur
dioxide (SO2), particulate matter, and nitrogen oxides (NO ) and
on removal efficiency of S02 and particulate matter are^ con-
sidered information needs in order to assess actual system
performance with respect to control requirements in the recently
promulgated revised New Source Performance Standards (NSPS) for
electric utility steam generating units.
Utilities, system and equipment suppliers, system design-
ers, research organizations, regulatory agencies, and others all
volunteer the information for this program. This voluntary
approach facilitates timely dissemination of pertinent informa-
tion in this key technological area. All information that is
gathered is stored in a computerized data base known as the Flue
Gas Desulfurization Information System (FGDIS). This system is
discussed in more detail in Appendix A.
118
-------
Information on operational systems is verified solely by
the utilities and reported essentially as received. Any modifi-
cations or adjustments to the reported data are made solely for
purposes of a consistent format that will allow reliable compar-
isons and evaluations to be made.
119
-------
SECTION 2
TECHNOLOGY OVERVIEW
CURRENT STATUS
Table 2-1 lists the number of domestic utility FGD systems
according to status and equivalent electrical capacities as of
the end of August 1980.
TABLE 2-1. NUMBER AND TOTAL CAPACITY OF FGD SYSTEMS,
AUGUST 1980
Status
Operational
Under construction
Planned:
Contract awarded
Letter of intent
Requesting/evaluating bids
Considering only FGD
TOTAL
No. of
units
73
39
29
7
15
40
203
Total
controlled
capacity, MW
27,155
17,855
13,769
5,590
8,424
24,200
96,993
Equivalent
scrubbed .
capacity, MW
24,765
16,854
12,919
5,590
8,424
23,980
92,532
Total controlled capacity (TCC) represents the gross capacities (MW) of
coal-fired units brought into compliance by FGD systems, regardless of
the percent of the flue gas treated.
Equivalent scrubbed capacity (ESC) represents the effective capacities of
the FGD systems (in equivalent MW), based on the percent of the flue gas
treated.
GROWTH TRENDS
Power-Generating and FGD Capacity
As indicated in Table 2-1, 73 coal-fired power-generating
units currently equipped with operational FGD systems represent
120
-------
a total controlled capacity of 27,155 MW.. This compares with a
December 1979 total coal-fired power-generating capacity of
approximately 235,000 MW. Current projections indicate that the
latter will rise to approximately 370,000 MW by the end of 1990.
Based on the known utility commitments to FGD, the percentage of
coal-fired capacity controlled by FGD will increase from its
current level of 11.5% to 26.5% by the end of 1990.
Table 2-2 presents the projected distribution of power-gen-
erating sources (by energy source) in the electric utility
industry. Table 2-3 presents the percentage of current and
projected coal-fired and total power-generating capacities
controlled by FGD.
Based on the requirements of the revised NSPS, actual
FGD-controlled capacity should exceed the levels indicated in
the preceding discussion. Currently, about 50 additional units,
representing a total capacity of approximately 25,000 MW, have
been identified as requiring S02 controls in the decade just
begun; however, identification of these units and information
regarding their status is not ready for public release as a
result of the premature stage of their planning, developments in
ongoing litigation, and the determination of applicable emission
control standards.
Figure 2-1 shows current and projected FGD-controlled
capacity and total power-generating capacity of coal-fired units
through 1990. This figure represents the committed FGD-
controlled capacity (those systems identified in Table 2-1), the
uncommitted FGD-controlled capacity (those units that cannot be
identified at the present time), and current and projected
coal-fired power generating capacities (those values cited in
Table 2-2 and the preceding discussion).
Figure 2-2 shows estimated FGD-controlled capacities at the
indicated month and year. An estimated total of 37,834 KW of
FGD-controlled capacity was identified in November 1974. By
August 1980, this figure had risen to 96,993 MW (see Table 2-1).
This represents an overall growth rate of 156% for the 6-year
period. In addition, the figures reflect a better than 55%
increase in the last 2 years.
Other notable changes that occurred during the 1974 to 1980
growth period include:
0 A 384% increase in the number of operational systems.
0 A 753% increase in operating capacity (ESC).
0 An increase in the average capacity of the FGD-
equipped unit from 170 MW to 340 MW.
121
-------
TABLE 2-2. DISTRIBUTION OF POWER-GENERATING SOURCES
BY ENERGY SOURCE '
December 1979
December 1990
Percent of total
Coal
39
44
Nuclear
9
14
Oil
25
20
Hydro
13
11
Gas
13
10
Other
1
1
Total, GW
603
833
jj Adapted from U.S. Department of Energy (1979) and Rittenhouse (1978).1>2
Figures reflect annual losses of 0.4% of the year-end capacity attributed
to retirement of older units.
TABLE 2-3. FGD-CONTROLLED POWER-GENERATION CAPACITY
(percent of total)
Period
August 1980a
December 1990
Coal -fired
capacity
11.5b
26.5
Total capacity
4.5b
11.6
Represents FGD-committed capacity as of August 1980.
Based on FGD capacity as of August 1980 and total power-generating
capacity as of December 1979.
122
-------
450
400 -
350
300
250
200
150
100
50
I I I I I I
COAL-FIRED CAPACITY
UNCOMMITTED FGD CAPACITY
COMMITTED FGD CAPACITY
1975 76 77 78 79 80 81
82 83 84
YEAR
85 86 87
89 90
Figure 2-1. Projections of coal-fired generating
capacity and FGD capacity from 1975 to 1990.
123
-------
o
o
ro
O
(X
-------
Process Type
FGD systems may be categorized in several ways, some gen-
eral and others more specific. Some general categorizations
used in the survey are:
0 wet vs. dry process
0 throwaway product vs. salable product process
A more specific categorization is by process (e.g., lime, lime-
stone, magnesium oxide, Wellman-Lord).
Tables 2-4, 2-5, and 2-6 summarize the current status of
FGD capacities associated with each of the foregoing process
categories. These tables show that the vast majority of oper-
ating experience to date has been obtained with wet calcium-
based, throwaway-product FGD systems. Of the 68,044 MW of FGD
capacity committed to a specific process (see Table 2-6), 62,541
MW (approximately 92%) are wet calcium-based, throwaway-product
systems.
Table 2-4 shows that all currently operating processes are
wet systems. With the recent advent of spray dryer collection
processes, 10 systems, representing an ESC of 3,523 MW, are
currently committed for future operation with a dry system.
Therefore, dry systems represent almost 12% of the FGD capacity
in the under construction and contract awarded status cate-
gories .
Table 2-5 indicates that approximately 6% of the current
operating FGD-controiled capacity produces a salable product
(elemental sulfur or sulfuric acid). This level of application
of salable product processes is expected to remain relatively
unchanged in the near future, as reflected by the 7% and 9%
levels currently committed in the under construction and planned
status categories. In the planned category, if the 641 MW
scheduled to produce gypsum for sale are not considered (gypsum
may have to be thrown away if a market is not available), the 9%
is reduced to 7%.
Table 2-6 reflects several trends in the industry with
respect to chemical process selection. Direct lime and lime-
stone systems currently account for approximately 89% of the
chemical processes selected, and a comparison of the two shows a
distinct industry preference for the latter, which will get
stronger in the near future as more systems are placed in
service. This trend is evident in that 53% of the lime/
limestone capacity in operation, 59% of the lime/limestone
capacity under construction, and 66% of the planned lime/
limestone capacity are limestone systems.*
* Includes alkaline fly ash limeyiimestone processes.
125
-------
TABLE 2-4. COMMITTED FGD CAPACITY - WET VS. DRY PROCESSES
Wet
Dry
TOTAL
FGD capacity (ESC), MW
Operational
24,767
0
24,767
Under
construction
15,194
1,660
16,854
Contract
awarded
11,056
1,863
12,919
Total
51,017
3,523
54,540
TABLE 2-5. DISTRIBUTION OF FGD SYSTEMS BY END-PRODUCT
FGD capacity (ESC), MW
Salable product
Throwaway product
TOTAL
Operational
1,600
23,167
24,767
Under construction
1,208
15,646
16,854
Planned
2,991a
29,678
32,669b
Total
5,799
68,491
74,290b
This total contains 641 MW of capacity which will produce gypsum for sale
rather than sulfur or sulfuric acid.
This total is less than that reflected in Table 2-1 because a number of
planned FGD systems have not yet been committed to a process.
126
-------
TABLE 2-6. DISTRIBUTION OF FGD SYSTEMS BY PROCESS
Process
Limestone3
Limeb
Lime/spray drying
Lime/limestone
Sodium carbonate
Magnesium oxide
We 11 man Lord
Dual alkali
Aqueous carbonate/
spray drying0
Citrated
Total
FGD capacity (ESC), MW
Operational
11,172
9,869
0
20
925
0
1,540
1,181
0
60
24,767
Under
construction
8,816
4,940
1,120
0
330
574
534
0
540
0
16,854
Planned
16,164
6,035
1,907
475
250
750
0
842
0
Q
26,423e
Total
36,152
20,844
3,027
495
1,505
1,324
2,074
2,023
540
60
68,044
Includes alkaline fly ash/1
, configurations.
Includes alkaline fly ash/1
tions.
Includes nonregenerable dry
d configurations.
This system is operating at
and is listed as a utility
, a 25-MW interchange to the
Because the processes of al
in this status category are
imestone and limestone slurry process design
ime and lime slurry process design configura-
collection and regenerable process design
the St. Joseph Zinc Co., G. F- Wheaton Plant
FGD system because the plant is connected by
Duquesne Light Company.
1 planned systems are not known, the totals
less than those in Table 2-1.-
127
-------
Emission Control Strategy
Emission control strategy refers to the measures used to
control particulate matter and S02 emissions from power plants
firing fossil fuels. At FGD-equipped, coal-fired utility
boilers, three basic combinations of primary particulate
matter/S02 control equipment are used: electrostatic precipi-
tator (ESP)/FGD, fabric filter (FF)/FGD, and two-stage scrub-
bing. Table 2-7 summarizes emission control strategies for the
current and planned FGD-equipped units.
TABLE 2-7. SUMMARY OF EMISSION CONTROL SELECTION'
ESP/FGD
FF/FGD
Two- stage
scrubbing
Total
Operational
No.
46
27
73
MW
16,564
8,203
24,767
Under construction
No.
32
3
4
39
MW
13,890
990
1,974
16,854
Contract awarded
No.
22
7
0
29
MW
10,823
2,096
0
12,919
Total
No. .
100
10
31
141
MW
41,277
3,086
10,177
54,540
Capacities represent ESC.
As indicated in Table 2-7, several industry preferences emerge
with respect to selection of a control strategy. The most
obvious is the strong preference to use an ESP for primary
particulate matter control upstream of the FGD system. Second,
a small but increasing preference for FF's is influenced by the
advent of the spray dryer/dry collection FGD technology. The
suppliers of most of : the dry processes offered commercially
recommend a FF as the preferred collection device. All the
FF/FGD combinations presented in this table are spray dryer/dry
collection systems. Third, a preference for the use of two-
stage scrubbing system for SO2 and particulate matter control is
diminishing. The units under construction that will use two-
stage scrubbing are either retrofit applications where the
existing particulate matter control devices (E'SP's) need lip-
grading or new applications where the alkalinity of the
collected fly ash will be used as a source of reagent.
128
-------
APPLICATION CONSIDERATIONS
New/Retrofit Units
Figure 2-3 shows a comparison of FGD-controlled capacities
with new and retrofit FGD systems. As indicated in this figure,
many of the original FGD systems were retrofits (e.g., retrofits
accounted for 62% of the operating capacity in service in 1975).
As- of August 1980, new systems accounted for 75% of the oper-
ating capacity. This trend toward application of FGD systems on
new sources is a result of the NSPS promulgated, pursuant to the
Clean Air Act Amendments. By 1990, FGD systems installed on new
boilers are expected to comprise 86% of the total.
Design SO9 Removal, Coal Sulfur Content, and Inlet S09 Level
Tables 2-8 and 2-9 summarize the FGD systems in service,
under construction, and planned according to design values for
S02 . removal, coal sulfur content, and inlet S02 level.
Table 2-8 presents a breakdown of the FGD systems that are
operational, under construction, and planned (contract awarded)
according to level of S02 removal efficiency versus coal sulfur
content. Some general statistics from the table are evident.
First, more than 70% of the FGD capacity is designed for S02
removal efficiencies of 80% or greater (almost evenly dis-
tributed between efficiencies of 80 to 89% and the 90% or
greater). Second, more than 85% of the FGD capacity installed
or planned is for boilers burning low- and high-sulfur coals,
with the capacities almost equally distributed between the two.
Table 2-9 presents a breakdown of FGD capacity by status
category according to design inlet SO2 levels. Establishing 4
lb/106 Btu as the break- off level between low- and high-inlet
S02 leads to the conclusion that FGD systems are used to a
greater extent on low-level S02 inlets than on high-level
inlets. Since 56% of present operational capacity is applied to
low-inlet SO2 levels, as are 62% of the systems under construc-
tion, and 64% of the planned systems, it appears that more of
the future coal fired utility units are expected to use low- or
medium-sulfur coal with FGD than high-sulfur coal and FGD. This
may be because there is more coal-fired utility growth where
low- or medium-sulfur coal exists.
Note that the preferences and trends cited in Tables 2-8
and 2-9 virtually exclude any impact that may be brought about
by the revised NSPS of June 1979. This discussion is therefore
limited to technological preferences and trends that developed
largely in response the Federal, state, and local regulatory
standards under the original NSPS of December 1971.
129
-------
CO
o
o
<
a.
co
CD
a:
UJ
a.
o
1975 76 77 78 79 80 81 82 83 84 85 86 87 88 89 SO
YEAR
Figure 2-3. Committed FGD operating capacity for new and retrofit
installations through 1990.
130
-------
TABLE 2-8. DESIGN S02 REMOVAL EFFICIENCIES OF FGD SYSTEMS
WITH RESPECT TO COAL SULFUR CONTENT
Design
removal
efficiency
< 70
Total
70-79
Total
80-89
Total
> 90
Total
TOTAL
Coal sulfur
content3
Low
Medium
Hi ah
Low
Medi urn
High
Low
Medium
Hiqh
Low
Medium
Hiah
Low
Medium
High
Operational
No.
7
7
0
14
6
1
4
11
13
2
12
27
6
3
12
21
32
13
28
MWb
3,066
1,306
0
4,372
2,359
800
1.180
4,339
3,938
918
4,181
9,037
2,044
749
4,225
7,018
11,407
3,773
9,586
Under
construction
No.
2
1
0
3
3
1
1
5
2
3
8
13
6
3
9
18
13
8
18
MWb
767
280
0
1,047
1,262
382
500
2,144
1,017
1,080
3,557
5,654
3,200
544
4,265
8,009
6,246
2,286
8,322
Contract awarded
No.
0
0
0
0
7
0
0
7
6
2
4
12
2
2
6
10
15
4
10
MWb
0
0
0
0
3,273
0
0
3,273
3,303
1,000
1,955
6,258
800
530
2,058
3,388
7,376
1,530
4,013
Total
No.
9
8
0
17
16
2
5
23
21
7
24
52
14
8
27
49
60
25
56
MWb
3,832
1,586
0
5,419
6,894
1,182
1,680
9,756
8,253
2,8io
9,693
20,949
6,044
1,823
10,548
18,415
25,029
7,58r
21,921
Low-sulfur content is less than 1%; medium-sulfur content is 1 to 2.5%
.sulfur; high-sulfur content is greater than 2,5%.
Capacities represent ESC.
TABLE 2-9. FGD SYSTEM S02 INLET LEVELS
FGD system S02
inlet
(Tb/106 Btu)
< 1.9
2.0 - 3.9
4.0 - 5.9
> 6.0
TOTAL
Operational
No.
26
18
8
21
73
MWa
8,636
5,235
4,260
6,635
24,766
Under
construction
No.
10
10
4
15
39
MWa
5,039
2,933
1,204
7,678
16,854
Contract awarded
No.
12
5
10
2
29
MWa
5,856
2,520
3,743
800
12,919
Tota^
No.
48
33
22
38
141
MWa
19,53''
10,688
9,207
15,113
54,539
Capacities represent ESC<
131
-------
SECTION 3
PROCESS DESIGN DEVELOPMENTS
This section addresses preferences and trends in the
process design development of commercial FGD systems.
CHEMICAL ADDITIVES
Chemical additives are used to improve the chemistry of
lime- and limestone-based FGD systems. For example, magnesium-
promoted processes have been used to reduce scaling, to increase
sulfur dioxide removal, and to improve reagent utilization.
Table 3-1 lists the number and generating capacity of units
that now have or will have FGD systems with magnesium-promoted
processes.
TABLE 3-1. NUMBER AND CAPACITY OF UNITS USING MAGNESIUM-PROMOTED
FGD PROCESSES
Process
Lime
Limestone
Lime/alkaline fly ash
Total
Operational
No.
7
0
0
7
MWa
4,433
0
0
4,433
Under construction
No.
2
1
2
5
MWa
860
670
1,400
2,930
Contract awarded
No.
0
1
0
1
MWa
0
650
0
650
Equivalent scrubbed capacity.
The introduction of magnesium into lime- -and limestone-
based FGD processes has been of great interest over the last 10
years, but most full-scale magnesium-promoted systems actually
began operations in the mid to late 1970's. Table 3-1 shows
that the trend in the use of magnesium promotion is declining.
132
-------
SYSTEM ENERGY CONSUMPTION
Table 3-2 shows the range and average of energy require-
ments of lime and limestone processes as a percentage of gross
generating capacity for new and retrofit systems. As shown in
the table, there is no significant difference between new and
retrofit systems.
TABLE 3-2. ENERGY CONSUMPTION FOR OPERATIONAL WET LIME AND
LIMESTONE SCRUBBING SYSTEMS3
Process
Lime
Limestone
Newb
Range
1.6 - 6.0
1.1 - 5.5
Average
3.8
3.2
Retrofit13
Range
1.5 - 3.5
3.4 - 5.6
Average
2.6
4.6
Overall
Average
3.1
3.4
Excluding flue gas reheat.
Electrical energy consumption of the FGD installation as a percentage of
gross.
FANS
Table 3-3 shows the trends in fan preference used on FGD
systems. Although most of these fans are centrifugal, utilities
are considering more innovative designs. Because early FGD
systems were considered separate from the rest of the generating
plant, separate booster fans provided draft for the scrubbing
systems. Newer power plants have fans sized to provide draft
for the entire boiler/scrubber installation as a unit. Where
ESP's or baghouses provide particulate matter removal prior ro
the scrubbing system, forced-draft fans (with respect to the
scrubber) are used extensively. These fans operate on dry flue
gas. Most induced-draft (ID) fans operate on dry flue gas as
well because they are often installed downstream from reheaters.
Carbon steel is now and will continue to be the primary con-
struction material for fans.
ABSORBERS
Table 3-4 is a breakdown of the number and capacity of
units equipped with FGD systems according to generic absorber
type and status. Combination absorbers include spray/packed and
tray/packed absorbers as well as concentric venturi/spray tower
absorbers. Impingement towers are fixed-baffle or fixed-vane
133
-------
TABLE 3-3. NUMBER AND CAPACITY OF UNITS BY FAN SPECIFICATION
AND INITIAL STARTUP YEAR
Fan specification
Design
Centrifugal
Axial
NRe
Function
Unit
Booster
NRe
Application
IDC
FDe
NRe
Service
Wet
Dry
NR
Materials
Alloy
Carbon steel
Rubber- lined
carbon steel
NR
a
Year of actual or projected FGD system initial startup
1971-1974
No.
8
0
3
3
5
3
7
2
2
1
8
2
1
8
0
2
MWa
2,198
0
145
191
1,199
945
2,073
250
20
408
1,915
20
408
1,915
0
20
1975-1978
No.
31
1
1
21
11
1
10
23
0
3
30
0
3
28
2
0
MWa
12,529
185
200
9,623
2,849
442
4,041
8,873
0
2,344
10,570
0
2,344
9,850
720
0
1979-1982
No.
34
4
37
13
22
40
11
40
24
6
49
20
3
47
0
25
MWa
12,880
1,313
14,315
5,417
7,482
15,609
3,833
16,411
8,264
1,820
19,433
7,255
1,141
18,443
0
8,924
Equivalent scrubbed capacity.
With respect to the FGD system
Induced draft.
Forced draft.
Not reported.
134
-------
absorbers, such as the disc contactor design. Fixed- or
static-bed, mobile-bed, and rod-deck absorbers are considered
packed towers. Systems in which flue gas is contacted with a
slurry or solution such that the flue gas is adiabatically
humidified and the slurry or solution is evaporated to apparent
dryness are defined as spray dryers. Both horizontal and
vertical spray absorber modules, which use radial, central,
cocurrent, countercurrent, or crosscurrent spray arrangements,
are considered spray towers. Impingement, sieve, and valve tray
absorbers are considered tray towers. Fixed- and variable-
throat venturi scrubbers as well as other absorber designs that
operate on a venturi principle are grouped under venturi
absorbers.
TABLE 3-4. NUMBER, CAPACITY, AND STATUS OF UNITS EQUIPPED WITH FGD
SYSTEMS BY ABSORBER TYPE
Absorber type
Combination absorbers
Impingement tower
Packed tower
Spray dryer
Spray tower
Tray tower
Venturi absorber
Operational
No.
10
1
19
0
20
15
8
MWb
3269
265
6265
0
7181
4396
3391
Under construction
No.
6
0
8
5
16
3
1
MWb
2871
0
3211
1660
7075
1802
235
Contract awarded
No.
3
2
2
6
15
1
0
MWb
1391
842
750
1863
8008
65
0
a These totals include S02 absorbers. Parti cul ate matter scrubbers are
b excluded.
Equivalent scrubbed capacity.
Table 3-4 indicates that spray towers have retained their
popularity and that spray dryers will become more prominent in
the 1980fs. Except for Venturis, which are on the decline, and
these two prominent designs, the other absorbers show no marked
change in commercial acceptability.
MIST ELIMINATORS
Utilities and system designers apparently prefer mist
eliminators of the chevron design, particularly when they are
preceded by a bulk separator. The primary material of construc-
tion is plastic, although some mist eliminators are made of
135
-------
alloys. None of those in the contract awarded status and only
one unit now under construction will be constructed of materials
other than plastic.
Most mist eliminators are horizontal, that is they are
installed perpendicular to the vertically rising gas stream of
conventional vertical absorbers. Vertical mist eliminators are
used in horizontal absorber modules and some vertical absorbers
that have a 90-degree turn of the duct (and thus a horizontal
duct section before entry into the stack). The advantage of a
vertical mist eliminator is that the liquid collected is removed
perpendicular to the gas flow rather than opposite to it, thus
improving the liquid removal efficiency. These patterns are
somewhat evident in Table 3-5.
TABLE 3-5. NUMBER AND CAPACITY OF FGD-EQUIPPED
UNITS BY MIST ELIMINATOR TYPE, CONFIGURATION, AND INITIAL STARTUP YEAR
Type
Chevron
Mesh-pad
Radial -vane
Configuration
Horizontal
Vertical
Year of actual or projected FGD system initial startup
1971 - 1974
No.
10
1
2
11
0
MWa
2,202
no
250
2,323
0
1975 - 1978
No.
34
1
1
28
6
MWa
12,106
360
125
10,355
1,418
1979 - 1982
No.
38
0
1
17
5
MWa
14,929
0
475
6,663
1,793
Equivalent scrubbed capacity.
REHEATERS
Four reheat strategies are currently in use or planned for
domestic utility FGD systems: flue gas bypass, direct-
combustion, hot-air-injection, and in-line reheat. In direct-
combustion systems, fuel oil or gas is burned and hot combustion
products are mixed with the wet scrubbed gas before it enters
the stack. Hot-air-injection systems heat ambient air on the
shell side of a steam tube heat exchanger and inject it into the
flue gas stream. In-line reheaters heat the flue gas as it
passes through the duct and contacts the reheater tubes. Both
of the latter two systems use steam tubes with circulating steam
136
-------
or pressuri2ed hot water for heat transfer. In some instances a
unit will combine reheat systems. For example, where the per-
cent of gas scrubbed can be made to vary with coal sulfur con-
tent/ the flue gas is reheated by bypassing the particle-cleaned
gad around the scrubbing system to the scrubber exit ductwork
until the amount of allowable bypassed gas becomes inadequate
for the required degree of reheat (when the percent sulfur is
high), at which point a backup hot-air-injection reheater is
activated.
Another variation of the basic reheater is the waste-heat
recovery reheater. A waste-heat recovery reheater on a system
currently under construction is an in-line reheater that
includes two heat transfer areas. In the first transfer area,
upstream of the scrubber, heat is absorbed from the flue gas;
water circulating through heat exchanger tubes transfers the
heat to a second transfer area downstream from the scrubber.
Table 3-6 is a breakdown of the reheat processes reported
by number and capacity of units where these systems are in-
stalled or planned.
TABLE 3-6. NUMBER, CAPACITY, AND STATUS OF UNITS USING
FLUE GAS REHEAT STRATEGIES
Reheat type
Bypass
Bypass/hot air injection
Di rect-combusti on
Hot-ai r- i n jecti on
In-line
Waste- heat recovery
Operational
No.
19
1
10
21
14
0
MWa
7,149
447
2,589
6,738
5,441
0
Under construction
No.
10
1
1
6
3
2 J
MWa
4,661
447
240
2,570
1,375
1,408
Contract awarded
No.
2
0
0
3
3
0
MWa
1,320
0
0
1,475
286
0
Equivalent scrubbed capacity.
Five units (1687 MW) that are operational,, one unit (110
MW) that is under construction, and five units (1416 MW) for
which contracts have been awarded do not include reheaters.
STACK FLUES
Table 3-7 is a breakdown of units according to materials of
construction of the stacks, status, and whether or not they have
reheat. The flues of most stacks are and continue to be made of
137
-------
TABLE 3-7. NUMBER, CAPACITY, AND STATUS OF UNITS EQUIPPED WITH FGD SYSTEMS
ACCORDING TO FLUE/LINER TYPE AND REHEAT APPLICATION
Flue/liner
Alloy
ARBMb
Carbon steel
C.S. /inorganic lining
C.S.c/organic lining
Fiberglass
HCBCd
Operational
With reheat
No.
0
19
5
2
7
2
10
MW°
0
6103
2976
1834
2369
455
2370
Without reheat
No.
1
6
0
1
2
0
0
MW
917
2015
0
98
514
0
0
Under construction
With reheat
No.
0
11
0
0
0
3
1
MWa
0
5472
0
0
0
1220
242
Without rehe.at
No.
0
4
0
0
0
0
0
MWa
0
1455
0
0
0
0
0
Contract awarded
With reheat
No.
0
4
0
0
0
2
0
MWa
0
1426
0
0
0
1000
0
Without reheat
No.
0
3
0
0
0
0
0
MW
0
1687
0
0
0
0
0
. Equivalent scrubbed capacity.
Acid resistant brick and mortar.
j Carbon steel.
Hydraulic-cement-bonded concrete.
-------
acid-resistant brick and mortar (ARBM). Information regarding
materials of construction in the units under construction or on
which contracts have been awarded is lacking partially because
utilities often do not finalize stack design until late in the
construction stage.
SLUDGE DISPOSAL
Table 3-8 is a breakdown of units equipped with FGD accord-
ing to sludge treatment, transportation, disposal method, site,
and operational status. As in the case of stacks, information
on units under construction and on which contracts are awarded
is incomplete because final disposal strategies are often not
finalized until plant construction is nearly complete. Also,
when a separate contract is arranged for sludge disposal, it is
often-not awarded until after initial plant construction.
Most disposal sites are and will continue to be on the
plant site. One trend is to increase sludge solids content by
fly ash addition and/or using vacuum filters so the material can
be landfilled. Another trend is to provide some sort of sludge
treatment before final disposal; primary methods are fly ash
stabilization, forced oxidation, and proprietary fixation. As
more systems produce sludge with higher solids content, waste
transport by truck and/or conveyor belt will become more
prominent.
139
-------
TABLE 3-8. NUMBER, CAPACITY, AND STATUS OF UNITS EQUIPPED WITH FGD
SYSTEMS ACCORDING TO SLUDGE DISPOSAL SPECIFICATIONS AND STATUS
Disposal specification
Sludge treatment type
Bottom ash addition
Fly ash/lime stabili-
zation
Fly ash addition0
Fly ash mixing
Forced oxidation
Proprietary fixation
Sludge transportation
Conveyor
Pipeline
Rail
Truck
Sludge disposal method
Landf i 1 1
Lined pond
Mine fill
Unlined pond
Disposal site
Onsite
Off site
Operational
No.
1
5
9
3
4
11
4
29
3
12
21
30
2
14
55
14
MWa
490
956
3,494
1,785
2,025
5,615
1,070
10,666
1,785
3,526
9,011
9,408
632
3,971
15,915
4,899
Under construction
No.
0
0
0
3
6
6
2
1
1
6
16
9
3
0
17
4
MWa
0
0
0
1,219
3,430
2,686
1,140
280
500
2,733
7,858
3,943
1,421
0
8,002
1,397
Contract awarded
No.
0
1
0
2
1
2
0
0
0
5
10
1
0
0
2
2
MWa
0
65
0
1,000
166
1,370
0
0
0
2,146
3,824
50
0
0
1,067
120
b Equivalent scrubbed capacity.
c FGD wastes and bottom ash are disposed of together.
d FGD wastes and fly ash are disposed of together.
FGD wastes and fly ash are mixed before final disposal.
140
-------
SECTION 4
PERFORMANCE TRENDS
OPERATING EXPERIENCE
In the past 5 years, FGD has become the most commercially
(developed means of control of S02 emissions from coal-fired
boilers, and operating experience has increased significantly.
At the end of 1975, 20 units were either on line (or had been),
and approximately 198,000 hours of on-line experience had been
accumulated. By August 1980, 85 FGD systems had been operated
on utility boilers, and more than 460,000 hours of operation had
been logged. This represents a 425% increase in the number of
FGD systems operated and a 230% increase in total hours logged.
The operational hours above reflect the number of hours
reported by the utilities. Because hours of operation often are
not available for such periods as initial system startup or per-
formance testing, the actual number of operational hours is
greater than reported, as is the corresponding percentage
increase.
DEPENDABILITY
For characterization of system performance, four dependa-
bility parameters have been developed: availability, operabil-
ity, reliability, and utilization. Table 4-1 defines these
parameters.
The FGD survey program includes monitoring the performance
of the operating FGD systems and logging monthly operating-
parameters (e.g., boiler and FGD system operating hours, forced
outage times, scheduled outage times). If the data permit,
monthly dependability parameters are calculated for both the
entire FGD system and its respective modules (where applicable).
When modular operating parameters are known, total FGD system
dependability parameters are derived by averaging all the
modular figures, except in those cases where the FGD system
design includes spare capacity. In these instances, a spare
capacity factor is included in the calculation of the total
system parameter, which ensures that the overall FGD system
dependability is not penalized as a result of equipment
redundancy.
141
-------
TABLE 4-1. PARAMETERS OF DEPENDABILITY
Availability index
Operability index
Reliability index
Utilization index
Hours the FGD system is available for operation
(whether operated or not) divided by the hours in
the period.
Hours the FGD system was operated divided by the
boiler operating hours in the period.
Hours the FGD system was operated divided by the
hours it was called upon to operate.
Hours the FGD system operated divided by the
total hours in the period.
142
-------
Figures 4-1 and 4-2 reflect the availability history of
four FGD installations on boilers firing low- or medium-sulfur
(<2.5%) coal, and Figures 4-3 and 4-4 show the availability of
four FGD installations on boilers firing high-sulfur (>2.5%)
coal. These units represent systems for which sufficient data
jare available for analysis. In each case, the data points
represent 12-month rolling averages of the monthly total system
availabilities. The rolling averages are calculated by
averaging the availability data for the first 12 months of
operation, dropping the first data point, and adding the 13th
for a second average, and so on.
Figures 4-5, 4-6, and 4-7 are composites of the availabil-
ities of individual FGD systems. They show average annual
availabilities (through June 1980) for operating units firing
low-sulfur (<1%) coal, medium-sulfur (1-3%) coal, and high-
sulfur (>3%) coal, respectively.* Some newly operational sys-
tems were not included (even though data were available) because
they had been in operation for less than 1 year and yearly
availability averages were not available.
Figure 4-8 provides statistical analyses of the data con-
tained in the three composite graphs for 1978 and 1980. In each
case, the availability points for these two years were plotted,
and the median of each array was determined. Note that the
median FGD system availability for those systems applied to
units firing high-sulfur (>3%) coal .has shown a better than 50%
increase in the 2-year period, and is approaching that of the
low- to medium-sulfur coal units. This indicates a rising trend
in the overall dependability of FGD systems for high-sulfur coal
application. The median availability for units firing medium-
sulfur (1-3%) coal increased 16.5% and, for units firing low-
sulfur coal, 1.5%. The lower percent change for these two
categories is attributable to their higher median availability
in 1978 and attests to the stable and reliable operating
histories experienced by FGD systems on these low- and medium-
sulfur coal units.
S02 REMOVAL EFFICIENCY
Table 4-2 presents S02 removal efficiency performance test
results and total system design removal efficiency values for
some of the operational FGD systems. Table 4-3 presents contin-
uous monitoring data for some of these systems. All but two of
the systems represented in these tables are commercial lime/
*
These categories were used to provide a more even graphical
distribution; however, they differ slightly from those used in
previous sections.
143
-------
3
i
40
I I I I I I I1
I I I I I
I I I I I I I 1 I I 1 I I I I I 1 I I I I I I I 1 1 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 1 I I
JFHAHJJASOKOJfHAMJJASONDJfMAHJJAiONDJFMAHJJASONDJFHANJJASOND
1476 1977 1976 1979 1980
ItAK
Col strip 1
100
I I I I I I I I I I I I I I I I I I I I I I I I 1 I I 1 I l I 1 I I l I l i ! l I l I l l
JfKAHJJASOHOJfM»HJJASOIIOJIHAKJJA50NOJFHAMJJASOI(D
•»» 1978 1979 1960
Colstrip 2
Figure 4-1. Availability histories for FGD installations at the
Colstrip Station of Montana Power (£2.5% S coal).
144
-------
100
M>
E
§
I
20
l i i i l i i i i i i I l l l i l i i i i i t I t i i i i i i t l l t I i 1 t l | I I 1 i I
JfHANJJAiONOJFHAMJJAiUNOJINAHJJAiONOJFMAHJJASONl)
1977 197* 1979 19UO
YUR
Sherburne 1
100
S M
40
I I I l I I 1 i l l l I I I I I I 1 I l I I I I I I I I I I I I 1 I
JfMAMJJASOkOJFMAMJJAiONUJFMANJJASOND
1*7* 1979 I960
TtAt
Sherburne 2
Figure 4-2. Availability histories for FGD
installations at the Sherburne Station
of Northern States Power (<2.5% S coal).
145
-------
100
BO
.. i i !..i i .... i I .. . . . i i i i i i i i i i i i i i i i i I Bruce Mansfield 1
JFHAHJJASOIIOJfHAHJJASONDjrMAMJJASONOJFMANJJASOIIO
U77 1978 1979 I960
I 1 I I I I I I I I I I I I I I
I I I I I
l l l . l l l l l t l I l i l i | 'l \ . i i | I i i l l i l l 1 l I
JfHAHJJASONOaFKAHJJASONDJFMAKJJASOND
1978 1979 1980
Bruce Mansfield 2
100
p
i
i i i
JFMAHJJASONOJfHAMJJASOHOJFMAHJJASONO
'»'• 1979 mo
Widows Creek 8
Figure 4-3. Availability histories for FGD installations
at the Bruce Mansfield Station of Pennsylvania Power
and Widows Creek Station of Tennessee Valley Authority (>2.5% S coal)
146
-------
too
80
60
20
I I I
i i
i I i i i i i
•JFHAKJJASON
1974
DJfNAMJJASOHLJF
1975
MAMJJASONOJf
1976
MAHJJASONDjr
1977
MANJJASONDJ
1978
r H A
i .1 J • s n
1S79
pjr HAUJJASOIIP
Figure 4-4.
Availability history for the FGD Installation at the LaCygne Station
of Kansas City Power and Light (>2.S% S coal).
-------
-LAWRENCE 5
100
REID GARDNER 1
/
80
S 60
en
«=t
40
20
SAN JUAN 1
SAN JUAN 2
APACHE 2
COLSTRIP 1
1974
COAL CHj|
1975
JIM I
'INITIAL SYSTEM AVAILABILITY AVERAGE.
1 I 1 '
1976
1977
YEAR
1978
1979
1980
Figure 4-5. Annual average availability histories for
low sulfur (<1%) coal FGD installations.
148
-------
100 -
80
5 60
CO
40
20
®=INITIAL SYSTEM AVAILABILITY AVERAGE.
I I I
R.D. MORROW, SR. 2
R.D. MORROW. SR. 1
1974
1975
1976
1977
YEAR
1978
1979
1980
Figure 4-6. Annual average availability histories for
medium sulfur (1-3%) coal FGD installations.
149
-------
d 60
CO
<:
40 -
20 -
LA CYGNE 1
GREEN RIVER 1-3
A.B. BROWN 1
INITIAL SYSTEM AVAILABILITY AVERAGE.
1 _J I
1974 1975 1976
1977
YEAR
1978 1979
Figure 4-7. Annual average availability histories for
high sulfur (>3%) coal FGD installations.
150
-------
100
80
•*
i-
1
M 40
cn
*-•
20
1 r- 1
-
•
— «•
-
-
1978 1979 1980
100
80
M
I 60
|
40
20
i l i
* !
t •
•
* •
-
-
1978 1979 1980
100
80
•*
| 60
1
40
20
i i l
I
_
: i
-
i i i
1978 1979 1980
YEAR YEAR YEAR
MEDIAN = 93.45 MEDIAN = 94.85
* CHANGE IN
MEDIAN =1.5%
MEDIAN * 77.65 MEDIAN ' 90.50
X CHANGE IN
MEDIAN = 16.51
MEDIAN = 53.50 MEDIAN = 79.05
S CHANGE IN
MEDIAN = 50.61
Low sulfur coal installations.
Medium sulfur coal installations. High sulfur coal installations.
Figure 4-8. Statistical analyses of the annual availability data for the years 1978 and 1980.
-------
TABLE 4-2. S02 REMOVAL EFFICIENCIES; PERFORMANCE TEST DATA
tn
Utility name/
unit name
Arizona Public Service
Choi la 1
Duquesne Light
ohillips 1-6
Kansas City Power & Light
LaCygne 1
Kansas Power & Light
Lawrence 4
Kentucky Utilities
Green River 1-3
Louisville Gas & Electric
Can Run 4
Can Run 5
Cane Run 6
Montana Power
Colstrip 1
Colstrip 2
it i nuprl^
Unit rating,
HW (gross)
119
408
874
125
64
188
200
299
360
360
Process
type
Limestone
Lime
Limestone
Limestone
Lime
lime
Limestone
Dual alkali
Lime/alkalins
f lyash
Lime/a 1 kal in*
flyash
Fuel sulfur
content, %
0.5
1.5
5.4
0.6
4.0
3.8
3.8
4.8
0.8
0.8
Design removal
efficiency, %
92a
83b
80
73
80
85
85
95
60
60
Date
10/73
10/73
1975
3/75
5/75
8/77
10/77
10/78
3/77
8/77
7/79
7/80
2/76
1/77
5/77
6/77
10/76
11/76
l?/76
3/77
r>/77
Performance
test results, %
92
58.5
86-93
77
80
77
96-98
83
95
86-89
88
94
75
81
88
81
68
83
83
86
83
Remarks
Test results are based
on testing of Module
A only
Test results are based
on the average of
tests from October 2,
to October 21 , 1973
Tests results are from
two-stage scrubbing
train
Test results were
taken from a 4-hour
full load test
Results are based on
an 8-hour maximum
continuous load test
Summary of a 4-hour
full load test
Summary of overall
results from accept-
ance tests
Results are the
average of six test
runs
Results of a 7- to
10-day test period
Performance test re-
sults
The result is an aver-
age of three emission
tests
The result is from
compliance test per-
formed over an 11-
day period
Tests were EPA Method
6 procedures
Tests were EPA Method
6 procedures
-------
TABLE 4-2 (continued)
Utility name/
unit name
Northern Indiana Public Service
O.H. Mitchell 11
South Mississippi Elect. Power
R.D. Morrow. SR. 1
Springfield City Utilities
Southwest 1
Texas Utilities
Martin Lake I
Unit rating,
MW (gross)
115
200
194
793
Process
type
Wellman Lord
Limestone
Limestone
Limp stone
Fuel sulfur
content, %
3.5
1.3
3.5
0.9
Design removal
efficiency, X
90
85C
80
95d
Date
9/77
3/80
4/80
9/77
6/77
8/78
Performance
test results, X
91
92
90
92
98-99
98-99
Remarks
Tests commenced on
Aug. 29, 1977, and
were completed on
Sept. 14. 1977. test
period included 12
days at 92 MW flue
gas equivalent and
3-1/2 days at 1 10 MW
flue gas equivalent
Results of five EPA
Method 6 tests across
the absorber
Results of seven EPA
Method 6 tests across
the absorber
Average result of co»-
pliance test runs
Preliminary acceptance
test results at 750
MW
Acceptance test
results
b Module A removal efficiency; overall unit design removal efficiency is 59%.
Design removal efficiency of the two-stage scrubbing trains.
. Absorber design removal efficiency; overall removal efficiency is 53%.
Absorber design removal efficiency; overall removal efficiency is 71%.
-------
TABLE 4-3. S02 REMOVAL EFFICIENCIES: CONTINUOUS MONITORING DATA
Uti lity name/
unit name
Colorado Ute
Craig 2
Kansas City Power &
Light
LaCygne 1
Kansas Power & Light
Lawrence 4
Louisville G&E
Cane Run 4
Cane Run 5
Cane Run 6
Mill Creek 3
Montana Power
Colstrip 1
Northern Indiana Public
Service
D.H. Mitchell 11
Northern States Power
Sherburne 2
Pennsylvania Power
Bruce Mansfield 1
Philadelphia Electric
Eddystone 1A
South Carolina Public
Service
Winyah 2
South Mississippi
R.O. Morrow, SR, 1
R.O. Morrow, SR. 2
Unit rating,
MW (gross)
455
874
125
188
200
288
442
360
115
740
917
120
280
200
200
Process
type
Limestone
Limestone
Limestone
Lime
Lime
Dual alkali
Lime
Lime/alkaline
flyash
Wellman Lord
Limestone/
alkaline
flyash
Lime
Magnesium
oxide
Limestone
Limestone
Limestone
Fuel
sulfur
content,
%
0.5
5.4
0.6
3.8
3.8
4.8
3.8
0.8
3.5
0.8
3.0
2.6
1.7
1.3
1.3
Design removal
efficiency, %
85
80
73
85
85
95
85
60
90
50
92
90
69
85a
a
85
Date
5/80
6/80
7/80
8/80
9/77
10/77
2/79
7/77
8/77
10/77
11/77
7/80
7/80
6/80
6/80
4/76
7/76
9/76
12/76
8/77
10/77
Actual removal
efficiency, %
65
66
66
66
81
97
94
81
84
84
84
87
85
95
85
86
90
89
81
90
90
11/77 ; 91
4/77
10/77
9/77
11/77
6/79
7/79
8/79
4/80
5/80
6/80
7/80
8/80
9/79
5/80
6/80
7/80
8/80
58
81
97
85
80
84
80
80
80
90
90
80
95
85
90
85
80
(continued)
154
-------
TABLE 4-3 (continued)
Utility name/
unit name
Tennessee Valley Author'
Widows Creek 8
Unit rating,
MW (gross)
ty
516
Process
type
Limestone
Fuel
sulfur
content,
%
3.7
Design- removal
efficiency, %
89
Date
11/77
12/77
1/78
2/78
3/78
4/78
5/78
6/78
7/78
8/78
9/78
5/79
6/79
7/79
8/79
9/79
10/79
11/79
12/79
1/80
2/80
3/80
4/80
5/80
6/80
7/80
Actual removal
efficiency, %
91
94
89
85
92
90
89
92
88
89
91
80
84
86
88
83
87
88
86
84
84
83
86
83
82
87
a Absorber design removal efficiency; overall removal efficiency is 53%.
155
-------
limestone installations. The two exceptions are demonstration
systems utilizing dual alkali and magnesium oxide processes.
The available data, although not extensive, indicate that actual
removal efficiencies of these systems generally meet or exceed
design values at both low-sulfur and high-sulfur coal instal-
lations. This would seem to indicate that meeting or exceeding
design S02 removal efficiency has not been a significant problem
for FGD systems on units firing high-sulfur coal. For example,
the FGD installation at the La Cygne power station (the FGD-
equipped unit currently firing the highest-sulfur coal) success-
fully passed performance testing early in 1975. Results from 10
days of continuous monitoring in late September 1977 indicated
that the system was continuing to exceed its design removal
efficiency of 80%.
PERFORMANCE CONSIDERATIONS
Because of the widely varying conditions at stations where
FGD systems are applied (e.g., differences in plant size, coal
sulfur content, and required removal efficiencies), it is dif-
ficult to pinpoint specific variables affecting overall FGD
system performance. Certain general considerations can be
identified, however, and are discussed below.
S02 Inlet Levels and Removal Requirements
In general, FGD systems operating on units with low to
medium S02 inlet levels have demonstrated a higher level of
overall dependability than those operating on units with higher
inlet levels. This is illustrated in the statistical analyses
of the overall FGD system availability (Figure 4-8) for low-
sulfur coal units. Obviously, the lower SO2 removal requirement
contributes to this difference.
Unit Load Profile and Coal Characteristics
Higher dependabilities have resulted from a reduction in
the number of chemical and mechanical problems on FGD systems
applied to new, base-loaded boilers designed to fire coal from
one or several specific sources. The flue gas generated by such
units generally has more relatively constant and stable charac-
teristics, and overall system dependability apparently improves
because the system does not have to respond to dramatic varia-
tions in flue gas flow rates and composition. In FGD systems
retrofitted to cycling, and peak-load units, these systems often
must respond to conditions that reach or exceed their process
control capabilities, and problems result from the variations
that occur in reagent feed rate and loss of chemical control.
156
-------
System Redundancy and Bypass Capability
FGD systems are now considered an integral part of the
power generating plant, and more stringent regulations prevent
many utilities from bypassing the FGD system. Thus, the current
design trend is toward incorporation of spare absorber modules
and ancillary equipment. Systems so designed have greater
dependability because the failure of a single component does not
necessarily force the entire system off line. Spare capacity
also promotes a more flexible operating and maintenance strategy
by allowing some routine maintenance to be performed without
removing the system from service. The result is an overall
reduction in FGD system downtime.
Utility Experience
As utilities continue to gain more experience with FGD
system operation, overall system dependabilities are expected
rise. In the early stages of FGD operation, utility staffs had
little experience with the chemical processes involved in FGD
operation, and the chemical and mechanical problems that are
inevitable with complex processes such as these were difficult
to rectify. The steadily increasing commercial operating hours
will allow system operators and maintenance personnel to gain
the experience necessary for more efficient and expeditious
analysis of system problems and implementation of solutions. In
addition, utilities are employing more chemical engineers and
other personnel familiar with gas/liquid systems to deal with
these problems.
Operating and Maintenance Philosophy
A general trend in plant philosophy regarding operation and
maintenance (O&M) is the dedication of specific crews to handle
this responsibility, rather than considering it a secondary
function of the power plant O&M personnel. This change will
permit faster and more precise changing of system parameters to
meet varying load conditions, and overall system reliability
should improve as problems are attended to expeditiously.
System Design Generation
Building on experience gained in the operation of first-
generation systems, system suppliers and designers are now
providing better process design configurations and materials of
construction. Indicative of this trend are the broader guaran-
tees system suppliers are now offering with respect to S02
removal efficiency, mist carryover, waste stream quality/
quantity, power consumption, reagent consumption, and availabil-
ity. Many of the newer systems should exhibit fewer of the
traditional operating problems, especially during the critical
startup and debugging phases of operation.
157
-------
SECTION 5
CAPITAL AND ANNUAL COSTS
INTRODUCTION
Another important function of the utility survey program is
the acquisition and analysis of cost data. In this program,
emphasis is on costs associated with operational systems because
of the availability of meaningful and complete data. These data
are adjusted only to ensure their completeness and accuracy and
to facilitate comparison. The approach and methodology used in
analyzing these costs and the results of these analyses are
briefly described in the following subsections.
APPROACH
Capital and annual cost data on operational FGD systems
have been obtained continuously since March 1978. Costs for
each system are obtained directly from the utilities and from
published sources, and then itemized by individual FGD cost
element. The itemized costs are then adjusted to a common basis
to enhance comparability. This adjustment includes factors for
estimating costs not given by the utilities and escalating all
costs to common dollars (mid-1980). All adjusted cost data and
computations are reviewed and verified with the appropriate
utility.
It is important to note that the costs analyzed here are
real costs, not cost model projections. When a particular
itemized cost is not reported by the utility, an estimated cost
based on known system design and operating factors is included.
The use of estimates is not arbitrary; they are used only when
cost items are unavailable or are judged to be unrepresentative.
ADJUSTMENT PROCEDURE
Capital Costs
All costs associated with control of particulate
matter emissions are deducted.
158
-------
Capital costs for modifications necessitated by
installation of an FGD system are added if they were
not included in the reported costs.
Sludge'disposal costs are adjusted to reflect a 20-
year life span for retrofit systems and a 30-year life
span for new systems.
Any unreported direct and indirect costs incurred are
estimated and included.
All capital costs are escalated to mid-1980 dollars.
All $/kW values reflect the gross generating capacity
of the unit.
Annual Costs
All costs are adjusted to a common 65% capacity
factor.
Direct costs that were not reported are estimated and
added.
Overhead and fixed costs that were not reported are
estimated and added.
All annual costs are escalated to mid-1980 dollars.
All mills/kWh values are based on a 65% capacity
factor and the net generating capacity of the unit.
RESULTS
Table 5-1 summarizes both reported and adjusted costs for
all 45 operational FGD systems on which cost data were obtained.
This table also summarizes the results by application (new/
retrofit) and by sulfur content of the coal (high sulfur/low
sulfur). Table 5-2 lists the results by process type. A
plant-by-plant listing of the reported and adjusted costs for
the operational FGD systems addressed in this study is provided
in Appendix B.
COST MODEL COMPARISON
During the past few years, various organizations have
conducted major cost studies of the capital and annual costs
associated with different FGD processes. Reasons for these cost
studies range from comparing the economics associated with
159
-------
TABLE 5-1. CATEGORICAL RESULTS OF THE REPORTED AND ADJUSTED
CAPITAL AND ANNUAL COSTS FOR OPERATIONAL FGD SYSTEMS
Results
All
New
Retrofitted
High sulfur
Low sulfur
Reported
Capital
Range,
($/kW)
23.7-174.8
23.7-174.8
29.3-157.4
29.3-157.4
23.7-174 8
Average,
($/kW)
78.9
78.4
79.6
75.1
82.3
Annual
Range,
(mills/kWh)
0.29-13.02
0.29- 5.81
0.46-13.02
0.92-13.02
0.29-11.32
Average,
(mills/kWh)
2.97
2.19
4.54
3.71
2.09
Adjusted
Capital
Range,
($/kW)
35.1-258.9
35.1-242.1
57.5-258.9
57.5-233.6
35.1-258.9
Average,
($/kW)
116.2
107.4
131.4
106.3
122.6
Annual
Range,
(mills/kWh)
1.80-18.64
1.80-13.44
4.36-18.64
3.70-18.37
1.80-18.64
Average,
(mills/kWh)
7.64
6.49
9.38
7.48
7.40
(Ti
O
-------
TABLE 5-2. ADJUSTED CAPITAL AND ANNUAL COSTS FOR OPERATIONAL
FGD SYSTEMS BY PROCESS TYPE
Process
Limestone
Lime
Dual alkali
Lime/
alkaline
flyash
Sodium
carbonate
Wellman-
Lord
Limestone/
alkaline
flyash
Reported
Capital
Range
(J/kW)
23.7-168.0
29. 3-122. 8
47.2-174.8
92.5-101.4
42.9-113.6
132. 8-1 57- 4
Average
($/kH)
68.8
71.0
97.8
98.4
72.4
142.4
49.3
Annual
Range
(mills/kWh)
0.29- 7.80
0.92-11.32
1.25- 2.97
0.23- 0.46
Average
(mills/kUh)
2.47
3.69
1.30
2.40
0.38
13.02
0.75
Adjusted
Capital
Range
($/kH)
35.1-148.7
57.5-192.7
80.6-242.1
131.0-133.8
79.9-138.5
233.6-258.9
Average
($/kW)
99.6
104.5
134.6
132.9
101.7
249.1
94.5
Amu
Range
(mills/kWh)
1.80- 8.56
3.70-10.82
5.10-13.44
5.99- 7.79
5.29- 6.78
1.7.86-1.8.37
al
Average
(mills/kWh)
6.02
6.91
8.11
7.19
6.02
. V8.10
4.63
-------
commercial and emerging FGD processes to determining the cost
impact of increasingly stringent S02 standards. Table 5-3
presents the results of several representative cost studies
recently completed and the assumptions on which they are based.
In this table, the capital and annual cost estimates and
their underlying assumptions are summarized for a number of
"base cases." In this context, "base case" refers to a con-
ventional wet limestone slurry FGD process such as that typical-
ly installed on a new 500-MW (net) boiler firing high sulfur
eastern coal. This table shows that capital and annual costs
vary widely, with the capital values ranging from $94.5 to
$194.4/kW and the annual values ranging from 4.03 to 16.91
mills/kWh. These wide variations in estimated costs for es-
sentially the same case result from differences in the intent of
the studies and in the assumptions on which each is based. With
respect to the latter, variations can be noted for virtually
every key assumption.
By use of the reported and adjusted capital and annual
costs for the operational FGD systems presented in Appendix B,
it was possible to compare the estimated costs in these cost
studies with actual costs. For this comparison, only limestone
systems have been analyzed, as this was the "base case" of all
the aforementioned cost studies.
Table 5-4 presents the adjusted capital and annual costs of
the limestone systems currently in service on coal-fired utility
boilers. Generally, these costs represent the technology of
first-generation limestone systems that have been operational
for several years. Many have bypass capabilities. Most of
these systems scrub less than 100% of the flue gas and therefore
do not require a separate reheat system. A significant number
of units have total removal efficiencies of less than 70%. Few
systems have spare components and few have oversized components
to provide spare capacity. Sludge is typically disposed of in
ponds without fixation or treatment.
A comparison shows that capital and annual costs of actual
systems approach the costs developed by the Tennessee Valley
Authority (TVA) and Beychok cost studies. The actual average
capital cost for limestone FGD systems is $99.6/kW; average
annual cost is 6.03 mills/kWh. The TVA cost study arrived at a
capital cost of $97.5/kW and an annual cost of 4.03 mills/kWh;
the Beychok cost study, a capital cost of $94.5/kW and an annual
cost of 6.61 mills/kWh. The criteria used in developing the
costs in these two studies are also based on early FGD tech-
nology.
Assumptions used in the other cost studies reflect future,
more advanced FGD system designs. They also reflect inclusion
162
-------
TABLE 5-3. BASE CASE CAPITAL AND ANNUAL COST ESTIMATES
01
Category
General criteria:
Sponsoring organization
Year prepared
Plant size
Plant location
Plant capacity, HW (net)
Plant capacity factor, i
Plant application
Plant heat rate, Btu/kWh (net)
Fuel (source) ;
Fuel characteristics, HV/S1/AJ
Emission standard ,
SO? emissions, lb/10 Btu
S02 removal efficiency, X
Process design criteria:
Process
Number of modules
Number of spares
Gas bypass capability
Reheat AT, °F
Water loop
Solids dewatering, X solids
Sludge treatment
Sludge disposal
Economic criteria:
Capital cost basis
Annual cost basis
Battery limits
Price level
Cost estimates:
Total capital cost, $/kW
Total first year annual costs,
mills/kWh
FGD economic studies
3
Bechtel
EPRI
1979
2-unit
North Central
1000
70
New
9986
Coal (Illinois)
10.100/4.0/16.0
Revised NSPSa
1.2
87
Limestone
8
2
Complete bypass
56
Closed
50
Fly ash/lime
Truck/landfill
Total
Total first
year revenues
Gas inlet to
sludge disposal
July 1978
157.3
8.9
PEDCo 4
Environmental
EPA
1977
1-unit
Midwest
500
65
New
9000
Coal (Eastern)
12,000/3.5/14.0
Revised NSPSb
0.6
90
Limestone
5
1
Complete bypass
50
Closed
50
Fly ash/lime
Pumping/ pond
Total
Total first
year revenues
Gas inlet to
sludge disposal
July 1980
160.2
10.5
5
Steams-Roger
EPRI
1979
1-unit
North Central
500
70
New
9724
Coal (Illinois)
10,100/4.0/16.0
Revised NSPSC
0.8
90
Limestone
4
1
Complete bypass
50
Closed
45
Fly ash/lime
Truck/landfill
Total
Total first
year revenues
Gas inlet to
sludge disposal
July 1978
179.7
7.86
Beychok/ (
Stone & Webster
EPRI
1977 to 1978
1-unit
Midwest
500
70
New
9000
Coal (Eastern)
12,000/3. 5/NA
Revised NSPSd
0.5
90
Limestone
HA
MA
NA
Yes
Closed
Yes
Fly ash/line
NA
Total
Total first
year revenues
Gas inlet to
sludge disposal
First quarter 1977
94.5
6.61
7
SRI/Radlan
EPR!
1979 to 1980
1-unit
NA
499
70
New
NA
Coal (Eastern)
10.100/4.0/16.0
Revised NSPS6
0.5
93
Limestone
5
1
Complete bypass
50
Closed
60
Fly ash/lime
Truck/ landfill
Total
Total first
year revenues
Gas inlet to
sludge disposal
January 1979
194.4
16.91
8
TVA
EPA
1979
1-unit
Midwest
500
80
New
9000
Coal (Eastern)
10.500/3.5/16
NSPSf
1.2
80
Lines tone
4
0
No bypass
50
Closed
None
None
Pumping/ pond
Total
Total first
year revenues
Gas inlet to
sludge disposal
Mid-1979 (capital)
Mid-1980 (annual)
97.5
4.03
NA = Not available.
'Proposed standard of September 1978.
Evaluated sti"d?rds in anticipation of revision to NSPS.
Promulgated st ajlgited NSf 3
sled standard as stringent e?
efcvalui'-cj standards more string--/
Pre--'1- -•
<"i ol 1971 .
-------
TABLE 5-4. ADJUSTED CAPITAL AND ANNUAL COSTS OF
OPERATIONAL LIMESTONE FGD SYSTEMS
Utility name
unit name
Alabama Electric Coop
Tombigbee 2 and 3
Arizona Public Service
Choi! a 1
Choi la 2
Central Illinois Light
Duck Creek 1
Indianapolis Power & Light
Petersburg 3
Kansas City Power & Light
LaCygne 1
South Carolina Public Service
Winyah 2
South Mississippi Electric Power
R.D. Morrow, SR. 1 and 2
Southern Illinois Power Coop
Marion 4
Springfield City Utilities
Southwest 1
Tennessee Valley Authority
Widows Creek 8
Average
$/kW, capital
35.1
74.6
148.7
121.3
148.4
81.4
43.1
108.7
110.8
133.5
145.1
99.6
mills/kWh, annual
2.91
4.36
7.64
7.96
8.59
6.89
1.80
6.01
7.12
7.66
8.56
6.03
The variability of these figures occurs in part because FGD systems in-
stalled on some boilers do not accommodate 100% of the boiler flue gas.
The costs for such systems are proportionately lower than those for full
capacity FGD systems. This is magnified by the conventional use of gross
kW for the $/kW figure and net kW for the mills/kWh figure, regardless of
the % of the flue gas scrubbed. These figures represent -the capital and
annual costs required to bring the individual units into compliance.
164
-------
of a separate reheat system, the effects of more stringent S02
emission standards, more elaborate sludge disposal strategies,
and one spare scrubber module for extra capacity.
165
-------
SECTION 6
CONCLUSIONS
The discussion in the preceding sections of this paper
indicates that the significant rate of growth observed in the
development and application of FGD technology for coal-fired
utility boilers has been matched by the considerable improve-
ments observed in the performance of the operational systems.
With respect to the latter, the most significant improvement in
the performance of the operational systems involves the in-
creased level of dependability observed for the high sulfur coal
units. During the past 2 years, the dependability of these
systems has improved to a level which approaches that observed
for the low sulfur coal units. It is anticipated that this
trend will continue and will be reflected in less startup and
commercial operating problems for systems now being placed in
service or planned for service.
Promoted by the requirements set forth in the Clean Air Act
Amendments and the pursuant NSPS, application of FGD to all new
coal-fired utility boilers constructed in the near future is
anticipated.
166
-------
REFERENCES
1. U.S. Department of Energy. Energy Information Administra-
tion. Office of Coal and Electric Power Statistics.
Electric Power Statistics Division. Inventory of Power
Plants in the United States, December 1979. Publ. No.
DOE/EIA-0095(79).
2. Rittenhouse, R. C. New Generating Capacity: When, Where,
and By Whom. Power Engineering 82(4):57, April 1978.
3. Bechtel National, Inc. Economic and Design Factors for
Flue Gas Desulfurization Technology. EPRI CS-1428, April
1980.
4. PEDCo Environmental, Inc. Particulate and Sulfur Dioxide
Emission Control Costs for Large Coal-Fired Boilers.
EPA-450/3-78-007 (NTIS PB 281271), February 1978.
5. Augustine, F., S. D. Severson, and J. L. Winter. Economics
of Four FGD Systems. Prepared for Electric Power Research
Institute by Stearns-Roger Engineering Corporation under
EPRI Research Project No. 1180-3, Draft Report, November
1979.
6. Beychok, M. R. Comparative Economics of Advanced Regener-
able Flue Gas Desulfurization Processes. EPRI CS-1381,
March 1980.
7. Oliver, E. D. and K. Semrau. Investigation of High SO?
Removal Design and Economics, Volume 2: Economics. EPRI
CS-1439, June 1980.
8. Tomlinson, S. V-, F. M. Kennedy, F. A. Sudhoff, and R. L.
Torstrick. Definitive SO Control Process Evaluations:
Limestone, Double-Alkali, and Citrate FGD Processes.
EPA-600/7-79-177 (NTIS PB 80-105828), August 1979.
167
-------
APPENDIX A
FLUE GAS DESULFURIZATION INFORMATION SYSTEM
BACKGROUND
The most significant product of EPA's utility FGD survey
program is a quarterly summary report generated from a computer-
ized data base known as the Flue Gas Desulfurization Information
System (FGDIS). This data base represents the latest develop-
ment in this program. Previously, manually updated and semi-
automated data files were used to store and retrieve informa-
tion. The increased emphasis on FGD for SO2 control (resulting
from its commercial development) necessitated a more efficient
data storage/retrieval system for processing and transmitting
these data. In the fall of 1978, FGDIS was developed to meet
this need.
DESCRIPTION
Design and performance data for both the operational and
planned domestic utility FGD systems are stored in the FGDIS.
Also stored are data on operational domestic scrubbers for re-
moval of particulate matter and data on operational FGD systems
applied to coal-fired utility boilers in Japan.
The design data contained in FGDIS encompass the entire
emission control system and the power-generating unit to which
it is applied. Descriptions include location, standards limit-
ing emissions of S02 and particulate matter, power-generating
capacity, boiler and stack information, average fuel analyses,
and other more general data. Input of design data specific to
FGD systems ranges from general information such as process
type, system supplier, and initial system startup date to more
specific component design information and operating parameters
such as absorber type, gas and liquid flow rates, and pressure
drop. Also included in the data are descriptions of the methods
of solids concentrating and waste disposal, flue gas reheat, and
mist elimination, and information on capital costs and annual
revenue requirements of FGD systems.
For operational FGD systems, the FGDIS stores comprehensive
performance data, including periodic dependability parameters
158
-------
and the service times (operating, forced-outage, and scheduled
outage) from which they are calculated. Where available, actual
system S02 and particulate matter removal efficiencies are
included (and qualified). Problems encountered with system
operation and solutions implemented to correct them are des-
cribed. The performance of the FGD-equipped boiler is described
in terms of service time, production (kWh), and capacity factor.
Figure A-l presents a complete FGDIS structure diagram
illustrating all of the information areas and some of the key
data entries contained in the system, as well as the hierarchy
associated with them in the data base. General unit data are at
level 0, whereas most of the specific component data are at
Level 3.
CAPABILITIES
In addition to being used to generate a quarterly report,
FGDIS is also available for direct on-line access. This im-
portant function not only provides interested parties with an
opportunity to examine data that are too specific for convenient
inclusion in the quarterly report, but it also provides immedi-
ate access to information that has been loaded into the system
but not yet published (i.e., information that has become avail-
able during the period between quarterly reports). Information
is gathered, reduced, verified, and loaded into the FGDIS on a
continual basis to ensure that the files remain current and
complete.
Access to the FGDIS data files and -manipulation of these
data are accomplished via MRI System 2000 . This comprehensive
data base management system offers extensive data retrieval
capabilities. The set of user-oriented commands provided are
flexible enough to satisfy virtually any information need. The
PRINT command will produce the compilation of a simple sequen-
tial list, or a set of report writer commands will produce a
tabulation of the requested data in a predetermined report
format. Utilization of system functions (average, standard
deviation, summation, maximum value, minimum value) will elicit
statistical analyses of the numerical data in the files. In
addition, the data requested through the available commands can
be selectively limited by a set of criteria included in the
commands. This feature facilitates examination of design or
performance parameters for a specific unit or a specific process
type, and so on. The retrieval possibilities are limited only
by the needs and imagination of the user.
The FGDIS files are stored at EPA's National Computer
Center (NCC) in Research Triangle Park, North Carolina, and are
accessible via a nationwide communications network consisting of
169
-------
(-STACK
SUPP.
TYPE
OW RATE
HflUfl
LINER
TYPE
»RADE
ONTENT
( - 1
PARIICUIATE
£OJL
MECHANICAL
M.C. TYPE
SUPPLIER
REMOVAL EFF.
DESIGN AP
3>L
1
FAM1C fUTER
F.F. TYPE
SUPPLIER
flEMOVAL EFF.
GAS/CLOTH RAT
„, PARTICIPATE
*** SCRISMR
csp scausKR rm
wnnit sumiCR
DESIGN If R0OYAL EFf.
F.W GfNERAL
PROCESS TYPE
START DATE
STATUS
1
otsia
INFORMATION
SPARE SPARE
CAPACITY COtffONENT
1HPEK INDEX
ABSORBER ABSORBER
FAN FAN
PUMP PUMP
SYS
DEPENDA
DATE
unir c.f
UNIT AVI
iCOCTlCS
PILOT PUNT DATA
REPORTED/
PARTICIPANTS ADJUSTED
PROCESS CAPITAL COST
TEST PERIOD ANNUAL COST
SYSTtM
PERfQt
MODULE
OPER.'
REL. -
UTIL.
QUEhCHEH/
PfiESfiTURATOR
QUENCHER TYPE
SUPPLIER
DESIGN iP
L/G RATIO
AJSORBEBS
ABSORBER TYPE
SUPPLIER
US FLOV
L/G RATIO
REMOVAL EFF.
H.i. TYPE
SUPPLIER
HORIZ./VERT.
STAGES
PASSES/STAGE
8EME4UB
REHEATER TYPE
LOCATION
HEATING HED
FLUE GAS IT
FANS
FAN DES1GK
SUPPLIER
AFTLICAT11*
GAS FLOU RATE
WXWACTURFJ1
H30UUTION
DUCTHOR'.
LOCATION
CO«flGUILiTIC«
D1WNS10SS
REAGENT
DEUCE TYPE
FUNCTION
FEED CAPACITY
TANKS
TANK TYPE
LOCATION
CONFIGURATION
CAPACITY
PU»S
F-IW TYPE
MANUFACTURER
CAPACITY
SERVICE
COII
SOL IDS
ICENTRATl
DEUAIER1NG
DEVICE TYPE
CAPACITY
INLET SOLIDS
XJTLET SOLIDS
END PRODUCT
RECOVERABLE
PRODUCT
TYPE
CJJANTITY
DT5POSTTIOH
INSTHUHENTMfOH
-CHEH1CA1.
•PHYSICAL
WATER BALANCE
WATER LOSSES
WATER AOD,
WATER SOURCE
POINTS OF AOD,
CHEN. FUNCTION
CHEH. NAPE
CONSOHPTION
cn»t«
moncn
joiuuaa
FUCBIEII
SLUDGE
QUANTITY
COMPOSITION
I
TREATHEHT
METHOD
DEVICE TYPE
INLET CHAR.
OUTLET CHAfl,
DISPOSAL
INTEftlK/FIKAL
im
LOCATION
SITE CAPACITY
Figure A-l. FGDIS structure diagram.
-------
local telephone numbers in 21 cities and WATS services.
Arrangements are currently being made so that persons interested
in gaining access to the FGDIS can obtain account numbers,
training, and additional information from the National Technical
Information Service (NTIS), Springfield, Virginia, in addition
to providing continual on-line access capability, NTIS also can
process selective information requests for limited information
needs that do not warrant acquisition of. ,a permanent computer
account number (single requests for specific tabulated informa-
tion).
171
-------
APPENDIX B
REPORTED AND ADJUSTED CAPITAL AND ANNUAL COSTS FOR
OPERATIONAL FGD SYSTEMS
Alabama Electric
Tcmbigbee 2
Tonbigbee 3
Arizona Public Service
Choi la 1
ChollJ 2
Central Illinois light
Duck Creek 1
Central Illinois Public Service
Newton 1
Columbus i. Southern Ohio Electric
Conesville 5
Conesville 6
Ouquesne Light
El rang 1-4
Phillips 1-6
Indianapolis Power » Light
Petersburg 3
Kansas City Power t Light
Hawthorn 3
Hawthorn 4
LaCygne 1
tofttuckv Utilities
Sreen River 1-3
Louisville Gas t Electric
Cane Run 4
Cane Run 5
Cane Run 6
Hill Creek 3
Paddy's Run 6
Mlnnkota Power Cooperative
Hilton R. Young 2
Monongahela Power
Pleasants 1
Montana Power
Col strip 1
Col strip 2
Nevada Power
Reid Gardner 1
Reid Gardner 2
Reid Gardner 3
Northern Indiana Public Service
Dean H. Mitchell 11
Northern States Power
Sherburne 1
Sherburne 2
Pacific Power I Light
Jim Bridger 4
Pennsylvania Power
Bruce Mansfield 1
Bruce Mansfield 2
Philadelphia Electric
Eddys tone 1
Public Service Company of New
Mexico
San Juan 1
San Juan 2
South Carolina Public Service
Authority
Winyah 2
South Mississippi Electric
R.D. Morrow 1
R.O. Morrow 2
Southern Illinois Power Coop
Marion 4
Southern Indiana Gas * Electric
A.B. Brown 1
Springfield City Utilities
Southwest 1
Tennessee Valley Authority
Widows Creek B
Utah Power I Light
Hunter 1
Huntington 1
KB
Capital cost
6,992.100
6.992,100
6,550,000
44,352.000
30.583,000
107,831,000
22,836,000
22,836.000
59,541,000
50,356,000
55,724,000
3,220,000
3.220,000
46,900.000
4,500.000
12.647,000
12.481,000
20.596.900
18,846,880
3,700,000
44,119,500
65,693,400
36,500,000
36,500.000
5,363.378
5,363,378
14,200.565
18.192,040
34,982,000
34,982,000
49,643,000
110,639,000
110.639,000
30.856,000
47,944,410
47,985,000
6,646,000
10.896,000
10,696,000
15,200,930
12,495,000
16,744,500
47.900,000
24.400,000
27.090,000
Total/kW
27.4
27.4
52.0
168.0
73.5
174.6
55.6
55.6
116.8
122.8
99.5
29.3
29.3
53.7
70.3
66.6
62.4
71.5
42.6
52.9
92.5
106.3
101.4
101.4
42.9
42.9
113.6
157.4
49.3
49.3
90.3
120.6
120.6
285.7
132.8
137.1
23.7
53.7
53.7
87.9
47.2
86.3
87.1
56.7
63.0
oorted
Annual
217,464
217,464
NA
1.003,568
10,851,000
NA
9,132,726
9,132.726
21,027.000
18,301,000
NA
346,441
M6.441
7,413,047
364,005
960.301
763.443
NA
321 ,463
N«
1.779,375
9,015,879
6,128,000
6,128,000
251,514
251,514
131,824
2,414,589
2,716,758
2,716,758
NA
9,979,850
9,979,850
3,808,000
NA
NA
527,000
NA
NA
859,453
1,850,565
778,749
14,576,400
OA
2,946,400
Mills/kWh
0.33
0.33
NA
0.75
5.54
NA
5.81
5.81
7.18
11.32
NA
1.15
1.15
4.99
5.20
1.29
0.92
NA
1.25
NA
1.25
2.73
2.97
2.97
0.46
0.46
0.23
13.02
0.75
0.75
NA
3.28
3.28
6.37
NA
NA
0.29
NA
NA
1.03
1.30
1.20
7.80
NA
1.27
Capital
8,949,850
8,949,850
9,400,764
39,748,800
50,452,200
149,388,600
76.423,700
76,423,700
87,852,700
78,993,100
78,967,000
6,329,500
6,329,500
71.124,100
7.682,400
20.045,000
17,146,000
23,205,000
26,751,200
7,288,000
62,872,500
70,058,000
48,183,500
48,183,500
9,992,150
9,992,150
17,307,000
26,999,900
67,996,450
67,996,450
59,732,500
121,270,800
121,270.800 .
20,206,400
92,034,400
90,608,200
12,060,300
22,056,750
22,056,750
19.177,750
21,477,900
25,904.900
79,785,300
29,625,000
Adjusted
i/kW
35.1
35.1
74.6
148.7
121.3
242.1
93.0
93.0
172.3
192.7
148.4
57.5
57.5
81.4
120.0
105.5
85.7
80.6
60.5
104.1
131.0
113.4
133.8
133.8
79.9
79.9
138.5
233.6
94.4
94.4
108.6
132.3
132.3
187.1
254.9
258.9
43.1
108.7
108.7
110.8
81.1
133.5
145.1
68.9
Annual
3,893,050
3,893,050
3,130,900
10.221,000
17,143,200
44,003,900
26,288,970
28,288,970
30,006,600
35,558,600
25,189.600
2,436,200
2.436,200
32,189,700
2,817.900
5,334,000 '
4.975,500
8,867,600
8,855,500
3,746,200
13,914,300
26,148,300
14,719,250
14,719,250
3.314,600
3,314,600
4,247,300
9,832,000
18,990,800
18,990,800
19,440,100
44,890,750
44,890,750
6,296,400
31,930,100
31,483,100
2,648,100
6,162,250
6,162,250
6,525,600
7,252,100
^Hills/Mi
2.91
2.91
4.36
7.6S
7.96
13.44
6.62
6.62
10.82
18.64
8.59
4.39
4.39
6.89
8.25
5.15
4.56
5.79
3.70
10.36
5.99
7.92
7.79
7.79
5.29
5.29
6.78
18.37
4.63
4.63
6.71
9.5c
9.56
•,0.5j
. 17.86
18.07
1.80
6.01
6.01
7.12
5.13
7,413,800 7.66
25,140.300
9,492,200
8. 56
4.17
5.28
NA » Not available.
'172
-------
The Department of Energy's
Flue Gas Desulfurization
Research and Development Program
Edward C. Trexler, P. E.
U. S. Department of Energy
Office of Coal Utilization
The Department of Energy's flue gas desulfurization (FGD) research and
development activities are conducted as part of the Advanced Environmental
Control Technology Program (AECT) which is managed within the organization
of the Assistant Secretary for Fossil Energy. This new AECT program was
initiated in FY 1979 with a goal to identify, research, develop, refine and
demonstrate cleanup equipment that will clean flue gas for compliance with
existing and anticipated environmental pollution regulations, and equipment
that will remove the undesirable components from coal derived gas streams
to assure reasonable life for utilization equipment such as gas turbines
and fuel cells. The flue gas cleanup portion of the AECT program budget
amounted to $2.7 million in FY 1979 and $20.1 million in FY 1980.
The FGD project is divided into two parallel efforts identified by the sched-
uled completion dates as very near-term (end 1983) and near-term (end 1986).
The very near-term effort aims at improving the SC^ removal efficiency and
reducing the waste disposal problems of conventional lime/limestone scrubbers,
This is being done in coordination with EPRI and EPA scrubber improvement
programs, through private sector scrubber instrumentation and analysis, by
tests at TVA and other utility prototype and full-scale scrubber facilities,
and by transfer of process improvement information. The near-term effort is
aimed at supporting newer technology S02 removal processes that include non-
regenerable (throwaway) and regenerable systems that produce potentially
marketable by-products such as sulfur and sulfuric acid. These technologies
are, or will, be under experimental test at Fossil Energy Technology Centers,
under prototype testing by DOE and EPRI at TVA and other sites, and under
initial commercial use evaluation by DOE at power stations and industrial
plants. As these technologies mature, private industry will be encouraged
to cost-share development with the Government. Information on progress will
be disseminated via reports, symposia, plant visitations, demonstrations
and workshops.
173
-------
The Department of Energy's
Flue Gas Desulfurization
Research and Development Program
NATIONAL PRIORITIES IN ENERGY AND ENVIRONMENT
The Nation's entrance into the 1980's is characterized by the need to
solve difficult and interrelated problems. High on this priority list
are the needs to significantly reduce oil imports, to protect and enhance
the environment, and to improve the economic posture of the Nation through
increased national productivity. That these needs are important to the
Nation is evidenced by the abundance of contemporary legislative activity
which promotes both the diminished use of oil and gas through coal utili-
zation and the enhancement of the environment. Explicit in these legis-
lative acts is the need for achieving these goals within the bounds of
economic constraint. Meeting these goals will require the coordinated
effort of both the private and public sector. This paper seeks to pro-
mote such coordinated effort by presenting the Department of Energy's Flue
Gas Desulfurization Research and Development Plans. It is our desire that
this summary serve as a focal point for new and improved communication, and
that the end result will be success through a better coordinated effort.
The approach in this paper is to identify the energy challenge in terms of
flue gas desulfurization system needs, to introduce you to our new cleanup
technology efforts and how the FGD program is oriented to other DOE pro-
grams, to'note our special relationships with EPA, TVA and EPRI, and to
discuss in some detail particular programs which we are pursuing. In
addition, we would also have you join with us in examining the challenges
and opportunities of the future.
THE ENERGY CHALLENGE
The oil importation reduction challenge perhaps can be best appreciated by
observing our recent energy flow from supply through consumption, and by .
comparing consumption with domestic supply. Domestic and imported supply^
in 1977 was:
Supply Quads/Yr.
Domestic Coal 15.9
Domestic Natural Gas 22.7
Domestic Oil 16.68
Imported Oil 18.91
Consumption]/ in key sectors in 1977 was:
Consumption (Quads/Yr.)
Sector CoalN. GasOTl
Electric Energy Generation 10.64 3.26 3.45
Residential/Commercial .22 7.21 5.99
Industrial 3.14 8.65 7.60
Transportation 0.0 .54 20.0
174
-------
A comparison on a percentage basis between domestic reserves and con-
sumption is given by Figure 1.
Figure 1
U.S. RESERVES VS. U.S. CONSUMPTION
2% OIL 49%
2% NUCLEAR 3%
2% GAS 26%
94% COAL 18%
OTHER4%-
MEASURED U.S. RECOVERABLE CONSUMPTION PATTERN
ENERGY RESERVES
TOTAL = 10,600 QUADS TOTAL (1977) = 76.56 QUADS
Clearly, it can be seen that the Nation needs a substantial shift in
consumption from oil and gas which are not abundant, to coal which is.
Because of the nature of the respective markets, it would appear easier
to accomplish this shift initially from the more centralized consumers
such as the utilities and the major industrial plants. The adminis-
tration has set as a goal that the oil and gas consumption of this
sector be reduced fifty percent (50%) of present consumption by 1990
and legislation has been enacted accordingly.
The interrelationship of our energy challenge with environmental goals
was previously noted. In the near-term, we must burn more coal arid we
must burn it cleanly and economically, and this means we need additional
FGD options. Key environmental regulations affecting coal utilization
are outlined in Figure 2.
Figure 2
ENVIRONMENTAL REGULATIONS
AFFECTING COAL UTILIZATION
Clean Air Act - 1977
o National Ambient Air Quality .'.Standards
o New Source Performance Standards
175
-------
o Prevention of Significant Deterioration Regulations
o Nonattianment Policy
o State Implementation Plans
Resource Conservation and Recovery Act - 1976
Toxic Substances and Control Act - 1976
Clean Water Act - 1977
Safe Drinking Water Act - 1974
Further, it is to be noted that the acid rain phenomena has been receiv-
ing considerable attention recently. This interest could result in
new legislation and the need for retrofitting a new breed of low cost FGD
systems into many existing coal burning installations if such sources are
proven to be major contributors to the problem.
In summary, in terms of R&D objectives, we need the early supply of an
assortment of systems which enable utilities and major industrial users
to operate reliably and economically on coal or coal derived fuels while
meeting all present and anticipated environmental regulations. Further,
it is important that some of these systems be particulary oriented toward
retrofit applications.
ORIENTING THE DOE FGD ACTIVITY AND PROGRAM
The Department of Energy, Fossil Energy Assistant Secretary, pursues these
R&D goals with a broad based program of which the Flue Gas Desulfurization
Program is a part. The DOE program is basically a private sector assis-
tance program. The Department seeks to identify technologies with high
potential public benefit and seeks to promote their accelerated develop-
ment and demonstration by assuming some of the financial burden and risk.
The orientation of the FGD program to certain other FE programs can be
seen by Figure 3.
Figure 3
CLEANUP TECHNOLOGY CONTROL OPTIONS
-------
The Flue Gas Desulfurization Program Is operated from the Office of Coal
Utilization's Division of Cleanup Technology Development. Other programs
operated from the Division are crosshatched in the Figure. Cleanup tech-
nology development is pursued through DOE Field Technology Centers as
shown below by Figure 4.
Figure 4
CLEANUP TECHNOLOGY DEVELOPMENT
DEPARTMENT OF ENERGFY
ASSISTANT SECRETARY FOR FOSSIL ENERGY
OFFICE OF COAL UTILIZATION
DIVISION OF CLEANUP TECHNOLOGY DEVELOPMENT
I I I 2
Morgantown Energy Pittsburgh Energy Grand Forks Energy Laramie Energy
Technology Center Technology Center Technology Center Technology Center
(METO* (PETC) (GFETC) (LETC)
*Lead Center
The Cleanup Technology Division has sought, since its creation in 1979, to
build on the excellent FGD technology foundation layed down by the private
sector and by EPA, TVA and EPRI. I am personally grateful for the many
reports from them which have afforded us the opportunity to understand
and assess the technological choices. Much of our initial effort has
been in providing support to programs initiated by these organizations,
and we intend to continue this approach along with our modest in-house
efforts, and to significantly expand our joint efforts with the private
sector. The importance seen for this program within DOE is evidenced by
its growth from a modest $2.7M in FY 1979 to a requested $21.OM in FY
1981. Our FY 1982 request maintains the momentum of this rapidly growing
effort. We believe we will contribute by bringing the energy perspective
into FGD development.
DOE FGD R&D PROGRAM
Although the DOE FGD Program includes some effort aimed at improving the
reliability, operability and performance of conventional lime/limestone
scrubbers, and includes some attention to new FGD approaches, the majority
of our effort is going into what might be called the emerging or advanced
FGD systems. This affords us the opportunity to select and pursue those
particular efforts which would appear to offer the mose benefits for the
markets which need to be served in order that coal utilization can be
maximized in the shortest time.
,177
-------
Figure 5 describes those technologies which we have tentatively chosen
to evaluate and how and when these evaluations might lead to large scale
utility and industrial demonstration.
Figure 5
EMERGING FGD SYSTEMS
UTILITY DEMO (lOOmW)
TECHNOLOGY
EVALUATIONS
DOWA
DRY SCRUBBING
AQUEOUS
CARBONATE
DUAL ALKALI
MAGNESIUM
OXIDE
CHIYODA 121
INDUSTRIAL DEMO
PILOT READY
TECHNOLOGIES (10mW)
SUB-PILOT
TECHNOLOGIES (ImWI
FY79
FY80
'_|i°HN)ic1iS
}
jii
1 N
El
1
SUPPORT
FY81
]Lfi)| PO
ft EVALL
FY82
•
N|| D
ATION
I
| WESTERN UTILITY
1 |HI-S PILOT PLANT
EW YORK ESEERCO DEMO
(c
d
»RI PILOT)
I STUC
SHAWNI
n
FY83
ESIGN. CC
3
I]
J
1
IPS ft SIGECO EVAL|
ORNL I
[ EPRI C
3IES |©|
E ©
ITERI4J |P
TVA-JOH
)EMO $
-vC
PON 1| I
FACILITY
RDA||
MSONVILI
DESIGN. C
© TE
TESTING
FY 84
NSTRUC1
E{
ONSTRUC
ST$ Cg)
{&
FY85
•
.TEST
CO
T, TEST
\
J
FY86
MMERCI/
D
FY87
LIZATION
Figure 6 identifies our key near-term needs and the candidates which we
are evaluating to meet these needs.
Figure 6
KEY NEAR-TERM NEEDS/CANDIDATES
Need
Reliable, low cost, retrofitable FGD
systems for eastern coal applications
which produce an easily disposable
dry or gypsum waste product.
Reliable, low cost FGD systems for
western coal applications with low
water consumption and manageable
waste products
Candidates
o Forced Oxidation Systems
o Systems such as:
- Chiyoda 121
- DOWA
o Spray Dryers
o Dry Injection
o Spray Dryers
178
-------
o Reliable, low cost regenerate systems o PETC technology data base for:
which utilize coal for reduction or - Improved steam stripping
regeneration - Direct coal reduction
- Improved copper oxide/
gasifier system
Detailed discussions of these needs and.the primary candidates which are
being evaluated are included in the paragraphs which follow.
EASTERN NEEDS
Foremost on our list of key needs 1s the need to provide by the early 1980's
reliable, low cost, retrofitable FGD systems for eastern coal applications,
which produce easily disposal waste products such as gypsum or dry insolu-
able solids rather than sludge. The candidates for this need would appear
to include the newer forced oxidation systems, systems such as CHIYODA 121
and DOWA and spray dryer systems. We are particularly encouraged by the
recently reported improved stoichiometrics for spray dryer processes which
show them to have economic advantages even with higher sulfur coal. Our
evaluation programs are as follows:
0 EVALUATION OF FORCED OXIDATION SYSTEMS
DOE will study data from recent full-scale commercial forced
oxidation systems and compare them with the projected quali-
ties of CHIYODA 121 and DOWA.
We expect to complete this study in January and the results
might lead us to initiate an evaluation effort.
o EVALUATION OF GYPSUM WASTE SYSTEMS
- DOE has tentative plans to join with EPRI in evaluating a full
size CHIYODA 121 module.
- DOE may support additional DOWA efforts at the TVA Shawnee
Test Facility.
We also believe that much can be learned by carefully studying the
results of the recently completed CHIYODA 121 pilot scale (23 MWe)
tests.
0 EASTERN COAL SPRAY DRYER EVALUATIONS
- Eastern Coal Spray Dryer development/evaluation at pilot
scale (RFP - early FY 1981 award)
- Spray Dryer evaluation at PETC 500 #/hr coal-fired boiler.
- Spray Dryer performance characterization at ANL on 170,000
#/hr steam boiler (Preliminary)
179
-------
- Joint EPA/DOE/EPRI Spray Dryer characterization at 100 MWe
utility unit (Preliminary)
As noted previously, the optimistic projection for the application of
spray dryers to eastern coals is recent and, accordingly, our program to
increase emphasis in this area is not completely in place.
Our primary approach is to pursue this evaluation through the private
sector and, accordingly, we have been preparing an RFP for such an evalu-
ation. This RFP, which is scheduled for release in October, offers to
fund pilot scale testing of eastern coal optimized spray dryer on a slate
of eastern coals, and offers further to fund the conceptual design and
economic evaluation of commercial scale units.
Parallel with this effort, we propose to obtain parametric performance
data on a subpilot unit at our Pittsburgh Energy Technology Center (PETC),
and to take advantage of the installation of a spray dryer being installed
on a 170,000 Lb/Hr. steam boiler firing eastern coal at the Argonne National
Laboratory (ANL).
Further, it is to be noted that DOE, in conjunction with EPA and EPRI, have
been discussing Spray Dryer characterization testing on a 100 MWe utility
unit and such a unit could be used to verify, at a large scale, the per-
formance projections derived from pilot scale evaluations.
WESTERN NEEDS
For western markets, we see the need for reliable, low cost systems, with
low water consumption and manageable waste products. Key facets of our
western applications program are as follows:
o EVALUATION OF DRY SCRUBBERS FOR WESTERN APPLICATIONS (GFET.C)
- Field testing of full-scale utility Spray Dryers with lime
and sodium reactants.
- Continued testing and evaluation of dry injection of alkaline
ash, nahcolite and trona and the regeneration of reactants.
REGENERATION WITH COAL
We see the need for reliable, low cost regenerable systems which can
utilize coal for reduction or regeneration, and we are approaching this
need at this time with in-house laboratory tests and studies. This
program is as follows:
o PETC TECHNOLOGY BASE FOR IMPROVED REGENERABLE FGD SYSTEMS
- Model the reaction dynamics for direct reduction of S02
with coal. Verify at bench scale.
- Measure $62 partial pressures for prospective organic
absorbants to optimize absorption/steam stripping systems.
180
-------
- Evaluate at PDU scale a fluid bed reactor copper oxide
system employing coal gas as a reducing agent.
It is our desire to produce a data base from which the private sector
might create or optimize improved regnerable systems which use coal
instead of oil or gas for reduction or regeneration.
POTENTIAL NEW CHALLENGES/OPPORTUNITIES
As noted previously, the attention being given to existing pollution
problems such as acid rain could precipitate the development of a new
breed of low cost, retrofitable, less than NSPS capture systems to which
our present plans are not addressed. We are looking carefully at the
work being sponsored by EPA in this area, and we will be joining them in
the Limestone Injection Multistage Burner (LIMB) effort. In addition,
we have been evaluating burners, such as the staged slagging combustors,
under other FE programs, which might employ limestone injection and
which might lead to workable systems for such applications.
A related challenge might come from the proposed Powerplant Fuels Conser-
vations Act of 1980 (S. 2470). While the major thrust of this proposed
legislation is to mandate the conversion from oil to coal of approximately
18,000 MWe of powerplants primarily along the eastern coast, it also con-
tains a very important "offset" provision. The offset provision seeks
to offset the approximate 110,000 tons/hr of S02 additional emission
caused by the conversion to coal, by funding the addition of advanced S09
removal systems to approximately 3,000 MWe of existing coal-fired units."
To DOE, this is both a challenge in terms of being able to make wise
choices as to appropriate systems by late 1982, and an opportunity for
increased development and demonstration at a large scale. Our tentative
plan for implementing the offset provision is shown below in Figure 7.
Figure 7
GENERAL PLAN
Program Definition
• Track Legislative Action
• Characterize SO2 Dlstr.
• Characterize Regions
• Characterize Candidates
• Develop Plan
. Budget
• Input To FY 82 +
Procurement
• Prepare PON
• Establish SEB
• Solicit/Eval/Salect
• Negotiate Grants
Design/Construction
Monitor
Test Program
• Establish Test Prog.
• Monitor Tests
• Analyze Results
• Disseminate Results
Resource
Supt.
Supt.
Supt.
Supt
OCU/Supt.
OCU
OCU
OCU
OCU/Proc
HDQ Proc
METC/OCU
OCU/METC
METC/Supt.
METC/Supt
OCU/RA/
EPRI
SCHEDULE
FY80 FY81 FY82 FY83 FY 84 FY 85 FY86
181
-------
SUMMARY
In summary, DOE looks forward to joining with the flue gas cleanup
community in pursuing jointly both our energy and environmental goals,
and to contribute to the overall success through our perspective of
the nations energy needs.
We are pleased with the opportunity to share with you our plans and our
thinking, and we look forward to the opportunity to get to know all of
the participants better, and to work together toward these important
national goals.
182
-------
EPRI RESEARCH RESULTS IN FGD: 1979 - 1980
S. M. Dalton, C. E. Dene, R. G. Rhudy, and D. A. Stewart
Electric Power Research Institute
3412 Hill view Avenue
Palo Alto, California 94303
ABSTRACT
EPRI has a research effort of approximately $10M/year in
flue gas desulfurization covering engineering evaluations,
field testing, bench testing, pilot plants, prototypes and
demonstrations. This paper reports selected results from
projects on FGD water integration, gypsum crystallization,
limestone dissolution, wet stack operation, sulfur produc-
tion via RESOX, absorption/steam stripping, cyclic reheat,
and integrated emission control. A brief review of current
demonstration plans and program emphasis is also included.
183
-------
EPRI RESEARCH RESULTS IN FGD
1979 - 1980
EPRI'S WORK IN FGD
The Electric Power Research Institute (EPRI), as the research arm of the U.S.
electric utility industry, has established a research and development program
in flue gas desulfurization. In this area, the Institute will fund approxi-
mately ten million dollars of R&D work each year over the next five years.
Projects include engineering evaluations, laboratory testing, pilot plant
work, prototype development, demonstration installations and field testing.
CONTENTS OF THIS PAPER
In this paper are presented recent data from selected EPRI projects in the
areas of FGD field testing, economic evaluations, limestone dissolution, wet
stacks operation, FGD water integration, cyclic reheat, crystallization,
sulfur production via RESOX, absorption/stream stripping, and integrated
emission control. Also included in the paper is a discussion of EPRI's R&D
program emphasis in the next few years. Each project that has significant
recent results is discussed under a separate heading for that project.
OBJECTIVES OF THE PROGRAM
EPRI's FGD research efforts are designed to meet one or more of the following
objectives:
Reduce costs: Reduce capital, operating, maintenance .and
disposal costs.
Improve reliability: Identify reliable systems or components;
develop improved materials; identify
mechanisms and modes of failure, and
repair requirements.
Improve resource utilization: Improve energy efficiency; reduce depen-
dence on oil, electricity and gas; reduce
water consumption and discharges; improve
by-product utilization.
184
-------
The specific projects discussed in this paper represent recent EPRI funded
work not reported separately at this symposium. Papers are being presented in
other sessions covering successful testing of a 23 MU Chiyoda Thoroughbred 121
system with gypsum stacking at Gulf Power's plant Scholz, joint EPRI/TVA/UOP
testing of the 10 MW Dowa prototype, and EPRI solid waste disposal efforts.
These topics will not be covered further in this paper*
185
-------
Projects discussed in this paper:
Subject
Absorption
Steam
Stripping
Character-
ization of
FGD
Chemistry
Corrosion
Crystalli-
zation
Cyclic
Reheat
Economics
EPRI Project
Number
RP1402-2
RP1258-1
-2,3,4
RP1410-3
IPS79-747
RP982-21
RP982-14
RP982-19
RP1031-2.3
RP1652-1
TPS78-760
TPS78-767
RP1180-9
Integrated RP1646-1
Emission 1870-2
S0? Reduction RP784-2
RP1257-1
Water TPS80-730
Integration
Wet Stack
Designs
RP1653-1
Project Description
Lab and pilot development
of Flakt Boliden citrate
and novel steam stripping
processes
Test two FGD units compre-
hensively to establish
operating capabilities
and material and
energy balances.
Mg dissolution from
1imestone to improve
scrubber performance.
Lab testing of corrosion
and erosion in FGD.
Bench Scale sulfate
crystallization
Economic and field
evaluation of the cyclic
reheat concept (using inlet
heat to reheat)
High S02 removal
Design and Economics
Vol 1 Design
Vol. 2 Economics
Economic and Design Fac-
tors for FGD Technology
Build and test pilot
2-1/2 MW integrated
facility
RESOX pilot and prototype
development.
Material balance to show
effect of different water
sources on various FGD
systems
Entrainment and engineer-
ing for wet stacks
Contractors
U of Texas at Austin
(Dr. Rochelle), TVA,
Steams-Roger, Radian
Black & Veatch,
MR I, PEDCO, TRW
Radian Corp.
Battelle Columbus
SumX Corp.
U. of Arizona
(Dr. Randolph)
Bechtel National Corp.
(Companion studies)
Radian Corp.
SRI International
Bechtel National
Stearns-Roger,
et al
Foster Wheeler Energy
Corp., et al
Radian Corp.
Dynatech R/D Co.
EPRI
Contact
D. A. Stewar
R. G. Rhudy
D. A. Stewar
R. G. Rhudy
C. E. Dene.
D. A. Stewar
R. G. Rnucy
R. G. Rhudy
C. R. McGowir
D. V. Giovanni
T. M. Morask)
T. M. Moras*)
D. A. Stewart
R. Kosage
C. E. Dene
186
-------
RESULTS SUMMARY HIGHLIGHTS
Absorption/Steam Stripping
Characterization
Chemistry
Corrosion
Laboratory work has confirmed the Flakt-Boliden
citrate process data, identified several poten-
tial stream stripping process improvements, and
set the stage for 1 MW pilot plant testing at
TVA's Colbert plant facility.
Two FGD units have been tested, the Col strip
Unit 2 of Montana Power Co. and the Conesville
Unit 5 of Columbus and Southern Ohio Electric
Co. Some details of the test results are given
in the attached writeup.
Certain magnesium-containing limestones may be
more reactive than high-calcium stones depend-
ing on the minerology. Three promising stones
have been identified for further screening.
Surveyed installations and manufacturers and
identified downstream ductwork, stacks, dampers
and expansion joints as special problem areas.
Evaluated chemical additives as corrosion
inhibitors and identified N-lauroylsarcosine
for further evaluation.
Crystallization
Developed calcium sulfate crystal growth pre-
dictive equations, evaluated certain crystal
habit modifiers, and found a crystal!izer con-
figurations which may help in controlling
gypsum crystal size.
187
-------
Cyclic Reheat
Economics
S02 Reduction
Water Integration
The use of inlet heat to reheat exhaust gases
appears to break even economically with steam
reheat for moderate steam costs when using
high-cost alloy for the cyclic reheat system.
The low sulfur coal full-scale cyclic reheat
system was tested and found to be working
well. High sulfur coal cyclic reheat economics
depend on construction material, degrees of
reheat, and inlet flue gas temperature.
Several special purpose evaluations were per-
formed. Regenerable processes generally are
more expensive for the specific cases
studied. High S02 removal design studies
(TPS 78-760/1767) identified potential for
effect of Mg in reducing high S02 removal costs
in conventional FGD. Generalized case studies
(RP1180-9) identified spray drying as a cost
saving technique for western FGD and CT-121 as
having low lifetime costs. Under RP784-1, the
possible benefits of absorption/steam stripping
combined with RESOX were identified (though
these were not verified in later work).
Pilot work at 1 MW scale has verified RESOX
suitability for different types of coals and
for different S02 feed stream concentrations.
German 42 MW prototype efforts have not shown
high sulfur yields or sustained operating
times.
Over forty material balance cases have been
evaluated. Several cases show increased
scrubber scaling potential with certain sources
of water.
188
-------
Wet Stack Design Based on a literature search and theoretical
calculations a significant portion of the water
present in the stack appears due to carryover
from mist eliminators. Design criteria from
existing wet stacks and the problems
encountered are identified in the attached
write-up.
189
-------
ABSORPTION/STREAM STRIPPING PROCESS
RP1258 and RP982-20
OBJECTIVE
Regenerate processes for S02 removal from stack gases are being investigated
in order to hasten development of an economically feasible FGD process alter-
native to the throwaway systems. Initial cost comparisons of several pro-
cesses indicated that absorption/steam stripping may be an economically com-
petitive FGD process.
PROJECT DESCRIPTION
A project was initiated in 1978 to study the Flakt-Boliden absorption/steam
stripping process as it was the most technically advanced. Laboratory con-
firmation of basic process and pilot plant construction were conducted conc-
urrently followed by pilot plant tests to obtain firm data for design and cost
studies.
RESULTS
In the absorption/steam stripping process, S02 is absorbed in a buffered
solution and then stripped from the solution with steam. The stability of the
dissolved S02 in the buffered solution is an important factor in S02 recov-
ery. Loss of the dissolved S02 may result by disproportionate or by reac-
tion with another component, such as the buffer or oxygen. The results of a
study of the stability of S02 in the two most important absorbents, sodium
citrate and diethylenetriamine (DETA), are shown in Table 1, along with the
stripping steam requirement.
190
-------
TABLE 1 Loss of SOg from Buffer Solutions
Loss Rate
x 103, M/hr
1.2
3.3
8.9
23»0
2.9
6.2
12.0
Buffer
Citrate
DETA
Buffer
Cone., M
0.5
0.5
0.5
0.75
2.0
2.0
2.0
Initial
dissolved
S0?
Concf M
0.2
0.2
0.2
0.2
0.08
0.2
0.2
Temp, °C
140
150
158
163
139
145
155
Estimated
Stripping
Steam Rate,
Kg/Kg S02
40
20
DETA solutions do not appear to retain SOp as easily as citrate solutions.
The savings In steam costs are the primary reason for continued investigation
of DETA. Comparisons of stability of the absorbents, citrate and DETA, are
being made.
FUTURE WORK
The pilot plant study of the Flakt-Boliden process (citrate absorbent) is
currently underway at the Colbert Station of the Tennessee Valley Authority.
Following analysis of the data from this test program, further pilot tests
will be conducted with either citrate or DETA. The extent of the test program
with DETA depends on the results of the laboratory work on DETA stability.
191
-------
CHARACTERIZATIONS OF FULL-SCALE SCRUBBERS
RP1410-3
OBJECTIVE
The full-scale scrubber characterizations project was initiated in order to
provide the needed data base to enable the utilities to optimize their exist-
ing systems effectively, to aid them in selecting new systems, and to provide
informed utility responses to possible new emission standards.
PROJECT DESCRIPTION
The project is directed at performing extensive and detailed characterizations
of the capabilities of selected, representative, currently operating, full-
scale lime and limestone wet scrubbing systems. The program characterizes the
performance of the selected scrubber system with respect to the following:
o Meeting emission standards and performance guarantees, with emphasis
on sulfur dioxide removal.
o Quality and quantity of selected unregulated discharges for such
species as organic compounds, volatile metals, fine particulates, and
trace elements.
o Actual costs compared to estimated costs, including both capital and
operating costs.
o Reliability, availability, and operability.
The initial scrubber systems selected for characterization are Columbus and
Southern Ohio Electric Company's Conesville Unit 5 and Montana Power Company's
Colstrip Unit 2, burning high sulfur eastern coal and low sulfur western coal,
respectively.
RESULTS AND CONCLUSIONS
Work has been completed at Conesville and a draft final report is in review.
Field testing has been completed at Colstrip and the data are being analyzed
prior to preparation of a final report.
192
-------
Conesville Unit 5 is a 411 megawatt power plant which began operation in
January, 1977. The flue gas cleaning system for this unit consists of a cold-
side electrostatic precipitator followed by a turbulent contact absorber
scrubbing system capable of greater than 90 percent sulfur dioxide removal
from a high sulfur coal. Magnesium containing lime is used as the scrubber
additive and the scrubber sludge is stabilized by a commercial "fixation"
process and stored onsite in a landfill operation.
With respect to regulated emissions, greater than 95 percent S02 removal was
measured across one module of the two-module system. Although the net S02
removal is decreased for Unit 5 because of a system bypass and these measure-
ments were short term (8 hours), the presence of a high level of dissolved
alkalinity provided by the magnesium in the lime would allow a reduction of
one third in pumping power (3 pumps to 2 pumps) with only a 1 to 2 percent
change in the S0£ removal.
The particulate removal across the module measured was always positive. The
particulate removals may not be representative because the inlet values were
higher than expected (suggesting either high inlet ESP loadings or non-optimum
ESP operation) and the outlet values may be affected by SOj condensation
across the scrubber. However, no evidence was found to indicate a significant
scrubber related particulate emission increase. Removal of NOX was insignif-
icant.
The condensation of 503 across the scrubber created problems in the particu-
late size distribution measurements. The only particulate penetration
measured, in the 0.1 to 0.2 ym range, was attributed to sulfuric acid conden-
sation based on the size, appearance, and elemental composition of the mater-
ial captured. The trace element data is still being reviewed and it is too
early to present the results. The measurements of organic emissions indicated
few were present and what was measured was well below its toxic level.
Average availability of the Conesville Unit 5B scrubber module from January
1979 through August 1979 was 39.2 percent. If major outages which resulted
from labor problems, failure of major equipment components, and design changes
193
-------
are subtracted for this period, maximum expected availability would increase
to about 68 percent.
The remaining 32 percent of module unavailability is due to a variety of
maintenance requirements, such as cleaning plugged lines, cleaning scrubber
modules, and repairing equipment which had malfunctioned. A vigorous record
keeping plan has been initiated by the operating utility which will allow
identification of individual maintenance problems in the future. Maintenance
levels on the unit have been substantially increased and the current availab-
ility of the unit is close to the boiler availability.
194
-------
CHEMISTRY OF MAGNESIAN LIMESTONE DISSOLUTION
RP982-21 and TPS80-730
OBJECTIVE
The presence of dissolved magnesium In lime or limestone FGD systems generally
Improves the S02 absorption due to an associated increase in disolved
alkali. However, the magnesium present naturally in limestone is usually in
the form of dolomite, which is too slowly soluble to significantly increase
the magnesium ion concentration. Recently, a few limestones containing
greater than 1% MgCOj (magnesian limestone) have been tested which appear to
have a portion of the magnesium in soluble form.
To determine if magnesian limestones containing soluble magnesium compounds
are common, a survey of the literature was conducted to locate limestone
formations containing greater than 3% MgCOg but less than pure dolomite (46%
MgCOg). These formations have been sampled for chemical and mineralogical
analyses and solubility and rate of dissolution determinations.
PROCEDURE
Samples of 12 different magnesian limestones have been taken directly from *ne
quarries. These quarries are mostly in the east and midwest. Samples of sane
western U.S. limestones are also available for study. Characterization of the
stones includes chemical analyses for major constituents, X-Ray diffraction to
determine mineral content, and optical and electron microscopy to determine
grain size. Selected stones were tested for equilibrium solubility in water
by mixing a ground sample with water, agitating at a constant temperature, and
analyzing with time to a constant composition. The rate of dissolution of
ground limestone 1s determined by adding limestone to simulated FGD liquors
and analyzing with time. The effects of limestone particle size, temperature,
rate of agitation, pH, and Initial solution composition on solution rate are
being studied.
195
-------
RESULTS AND CONCLUSIONS
Ten magnesian limestone samples from different formations have been subjected
to X-Ray diffraction analysis. Of these stones, three contain portions of the
magnesium in a form other than dolomite. A comparison of the rate of dissolu-
tion if MgC03 from these three samples will be compared to rates of a dolomite
stone and calcite stone. Preliminary results from a study of the effects of
variables such as particle size, temperature, rate of agitation, pH and solu-
tion concentrations on the rate indicate that these variables affect the rate
of solution by different degrees for the different stones. For example,
increasing temperature from 50°C to 60°C increases the rate of solution of
CaC03 from Fredonia limestone but has little effect on rate from Maysville
Limestone (a magnesian limestone).
FUTURE WORK
The experimental procedures described here will be used on additional
limestone samples to determine if variables studied have any major effects on
rate of solution of either magnesium or calcium compounds in the limestone.
If the effects on solubility are not the same for each limestone, further
characterization of the limestone properties will be made in an attempt to
correlate limestone variables with differences in solubility behavior.
196
-------
CORROSION INHIBITORS
RP982-17
OBJECTIVE
Corrosion complicated by erosion has been a continuing problem in wet scrub-
bing systems for flue gas desulfurization (FGD). The attempts to solve these
problems have been made using coatings, linings, and various metal alloys.
Despite these attempts, maintenance and replacement coats have remained very
high.
Techniques such as the use of corrosion inhibitors have not been seriously
investigated for corrosion prevention in FGD systems. SumX Corporation has
undertaken a study for EPRI designed to determine the feasibility of using the
absorption type corrosion inhibitors in lime or limestone scrubbing solutions.
The major objective of this study is to determine if absorption inhibitors can
be used to lessen corrosion in FGD equipment.
PROJECT DESCRIPTION:
The work consists of laboratory experiments using electrochemical techniques
to detect changes in the corrosion potential of the metal in scrubber liquors.
Data from literature as well as recommendations from inhibitor suppliers were
used to select inhibitors for preliminary screening. The effect of these
inhibitors on the corrosion of mild steel, 304L stainless steel and 316L
stainless steel under one set of solution conditions was measured.
RESULTS AND CONCLUSIONS:
To date 10 compounds have been tested with mild steel. N-Lauroylsarcosine has
shown the best inhibitor properties. Sulfite concentrations appear to have a
major influence on corrosion. The formation of a reaction film can be crit-
ical to the corrosion rate.
197
-------
Tests conducted with stainless steels are incomplete; however, uniform cor-
rosion rates are very low. The major influence on corrosion observed thus far
has been temperature.
Future work on this project will involve compounds related to N-Lauroylsarco-
sine, completing tests with 304L and 316L stainless steels. In addition,
tests to determine sensitivity to inhibitor concentration and other solution
characteristics will be conducted. Coupon tests with slurry solutions will be
performed for extended periods with the most promising inhibitor compounds.
198
-------
CORROSION AND MATERIALS OF CONSTRUCTION
RP982-14
OBJECTIVE
The objective of the materials of construction in wet scrubbers project was to
comprehensively document and analyze the utility experience with materials of
construction in full-scale lime and limestone wet FGD systems on boilers
burning eastern or western coals. The result will be a summary of materials
experience.
PROJECT DESCRIPTION
Information on field performance of construction materials was collected
primarily by site visits, but also by telephone and letter contacts with FGD
system operators and equipment vendors, and by literature searches. Informa-
tion was collected for the following FGD system components: prescrubbers,
absorbers, spray nozzles, mist eliminators, reheaters, fans, ducts, expansion
joints, dampers, stacks, storage silos, ball mills, slakers, pumps, piping
valves, tanks, thickeners, agitators, rakes, vacuum filters, centrifuges, and
pond linings.
Materials documentation and analysis include successes, failures, reasons for
success or failure, failure mechanisms, and relative costs of various mater-
ials. Detailed trip reports on each site visit are included in an appendix.
The results are designed to be a first step in aiding utilities and FGD equip-
ment suppl iers in selecting materials that will perform satisfactorily at
minimum expense.
RESULTS AND CONCLUSIONS
Stack linings and outlet ducts (beyond outlet dampers) are the scrubber com-
ponents that have a significant history of materials problems and are critical
components in that failures may require complete boiler shutdown and loss of
generating capacity for lengthy periods due to the lack of standby components
or bypass capability.
199
-------
The performance of a stack lining depends on whether the scrubbed gas is
delivered to the stack wet or reheated, and whether or not the stack is also
used for hot bypassed gas. These factors appear to have a strong effect on
the performance of lining materials, in spite of differences in fuel sulfur,
application techniques (e.g., surface preparation or priming), operating
procedures (e.g., thermal shock), design (e.g..annulus pressurization), and
other factors which can affect performance.
Inlet and bypass ducts are generally not a major problem area for utilities
with scrubbers. However, the outlet duct has been a major problem area,
particularly for units which have duct sections which handle both hot and wet
gas. These sections are for the most part beyond the bypass junction on units
which do not have reheat. Acidic conditions developed during scrubber opera-
tion become more severe on bypass as the temperature is raised and other
corrosive species in the unscrubbed flue gas (chlorine and fluorine) are
introduced.
Research efforts for these two components need to be directed to:
1. Compiling and maintaining general materials performance data
2. Characterizing environmental conditions where failures are occurring
3. Post-testing materials exposed to FGD environments to determine
and/or verify failure mechanisms
4. Laboratory testing of commercial materials to verify proprietary
data, and
5. Developing new or improved materials and designs based on the above
information.
Prescrubbers, absorbers, reheaters, outlet ducts ahead of the outlet dampers,
dampers, pumps, and piping and valves have a moderate history of materials
problems but failures may not require complete boiler shutdown. Spray noz-
zles, mist eliminators, fans, inlet and bypass ducts, expansion joints, stor-
age silos, ball mills, slakers, tanks, thickeners, agitators, rakes, vacuum
filters, centrifuges, and pond linings have a relatively low history of mater-
ials problems and/or are amenable to rapid repair or replacement.
200
-------
CRYSTALLIZATION OF GYPSUM
RP1031-2
OBJECTIVE
Forced oxidation with gypsum crystallization is being proposed as a means of
solving or reducing the problems of handling sulfite sludge. Production of
gypsum offers two alternatives to landfill sludge disposal. One alternative
is to produce a gypsum of sufficient purity and consistency to be used in the
manufacture of wall board. The other is to produce gypsum of sufficient size
to result in easy dewatering allowing "stacking" as another means of disposal.
In order to design FGD systems which will consistently produce a product of
the desired properties, basic crystallization data are necessary. To obtain
these data, a study of the nuclcation rate and growth rate of gypsum has been
completed. In addition, the effects of some operating conditions and
additives on these properties were determined.
EXPERIMENTAL PROCEDURE:
Determination of nucleation and growth rates of gypsum were made in the "mini-
nucleator" developed at the University of Arizona. The crystal!izer in this
apparatus is a one-liter, draft-tube-baffle, jacketed, glass vessel.
Provisions are made to control temperature, liquor flow, and
supersaturation. A particle counter by Particle Data, Inc. connected to a
PDP-8 mini-computer is used to count particles and analyze data.
Supersaturation is normally developed by dissolving a desired compound at one
temperature and crystallizing at a lower temperature for systems where
solubility is temperature dependent. However, CaS04 has a low solubility and
supersaturation was maintained by chemical reaction. Liquors were both
simulated and actual limestone scrubbing liquors. Process variables studied
were pH, agitation rate, and seed crystal concentration. The additives
studies were sodium dodecylbenzenesulfonate, Calgon® CL246, adipic acid, and
201
-------
citric acid. These data were used in a computer simulation program to predict
gypsum crystal size distribution from various crystallizer designs.
RESULTS AND CONCLUSIONS
Qypsum nucleates by secondary nucleation mechanisms of the collision breeding
type when large (>150yu.m) parent crystals are retained in the crystal magma.
High supersaturation and/or an absence of parent seed can result in bursts of
excessive primary nuclei which degenerate particle size.
Although pH does not appear explicitly in the nucleation/growth rate kinetics
expressions, the ratio of nucleation to growth shifts at low pH levels to
produce smaller crystals. Regions of low pH (or sudden decrease in pH) in the
scrubber system would be expected to reduce particle size.
Of the additives studied (sodium dodecylbenzenesulfonate, polyacrylate, adipic
acid, citric acid), only citric acid had a beneficial effect on the size and
shape of the crystals grown.
Computer simulations utilizing the nucleation/growth rate kinetics expressions
developed in this study, together with assumed crystallizer configurations,
indicate that particle size could be nearly doubled using a double-drawoff,
classified removal crystalizer configuration in which mixed underflow and
partially settled overflow streams are removed from the crystallization
tank. Such operation could be achieved simply by installation of an internal
settling baffle.
PLANS FOR FUTURE WORK
Since only one liquor composition was used in these studies, the effects of
other ion concentrations (e.g., chloride and magnesium) in both lime and
limestone system liquors will be studied. These data and the predicted size
improvments are to be verified in a bench-scale crystallizer system.
202
-------
.CYCLIC REHEAT
RP1652-1
OBJECTIVE
A significant power plant operating cost savings is achievable if the cost of
steam or oil used to reheat flue gas downstream of SOg wet scrubbing can be
reduced or eliminated. One method of accomplishing this goal is by use of a
cyclic reheat system which extracts heat from the flue gas entering the
scrubber and uses that heat to reheat the stack gas. This study was performed
to achieve a better understanding of cyclic reheat and to fill in information
gaps regarding its application. The specific objectives are:
o To publicize the status of research work on cyclic reheat.
b To characterize the only existing U.S. full-scale cyclic reheat
installation at Southwestern Public Service's (SPS) Harrington
Station Unit 1 near Amarillo, Texas.
o To provide an economic comparison between cyclic reheat and conven-
tional stack gas reheat schemes.
PROJECT DESCRIPTION:
The approach used to accomplish these objectives and develop the study infor-
mation can be summarized as follows:
o Information on cyclic reheat research activities was obtained by
literature search and by discussions with users, vendors, and
research and engineering institutions regarding equipment types and
systems used or considered for this application.
o Characterization of Harrington Station's cyclic reheat system was
conducted by collecting historic design, cost, operating, and main-
tenance data; by performing gas sampling, component analyses, and
temperature and pressure measurements on selected streams; and by
analyzing these test data for system performance.
o Economic comparisons of cyclic reheat and three conventional stack
gas reheat systems (in-line steam, hot-air injection, and oil-fired
203
-------
reheat) were made on a comparable basis for both low and high sulfur
coal cases. The comparisons are based on two 500 MW units operating
with a reheat level of 50°F and with inlet flue gas (after dust
collection) temperatures of 300°F and 270°F, respectively for the
high and low sulfur cases. Materials of construction of exchanger
tubes are chosen to take into account the sulfuric acid dewpoint and
the temperature level of the reheat medium. Capital and operating
costs are presented on a 30-year levelized basis. Cost sensitivity
analyses were performed to determine the effect of certain design and
energy value parameters.
RESULTS AND CONCLUSIONS
Results of the review of cyclic reheat research activities indicated that
considerable effort has been and is being conducted on different approaches
and equipment types. Most experience in the United S,tates has been with
gas/liquid/gas (Harrington) type systems, while in Japan it has been primarily
with regenerative gas/gas (Ljungstrom) type cyclic reheat. Other approaches
and equipment types in use or being studied include: (1) a heat pipe concept
consisting of a closed tube containing a heat transfer medium which vaporizes
during heat extraction and is condensed in reheating the scrubber outlet gas,
(2) a borosilicate glass tube exchanger for gas/gas'type cyclic reheat, and
(3) a cast iron finned tube exchanger for low level heat recovery in a
gas/liquid/gas system.
Characterization of cyclic reheat at Harrington Station which uses low-sulfur
coal (0.3 to 0.5% sulfur) indicated superior operating experience with no
serious corrosion or plugging problems despite carbon steel construction of
the heat extractor and reheat exchangers. Results of the field test program
indicated performance of the cyclic reheat and FGD systems are reasonably
close to design. Sulfur trioxide (S03) content measured in the flue gas feed
to the heat extractor was found to be considerably less than expected and
indicates probable absorption and neutralization by the alkaline fly ash
either in the flue gas or in the sampling system. Average finned area heat
transfer coefficients for the heat extractor and reheat exchangers were found
experimentally to be 6.3 and 10.3, respectively, as compared with values of
204
-------
9 to 10 from accepted correlations when assuming the same flow conditions and
a clean surface.
Results of screening-type capital estimates for both the Ljunstrom and heat
pipe schemes indicated that these cyclic reheat approaches may have some cost
advantage over the gas/liquid/gas system. However, a gas/liquid/gas system
similar to that operating at SPS1 Harrington Station and to that being
installed at TVA's Paradise Steam Plant was chosen as a base case for com-
parison with conventional reheat methods because of the greater experience and
availability of information.
Simplified EPRI Class 1 (±20%) capital and operating cost estimates were made
for the gas/liquid/gas type cyclic reheat system and for the three conven-
tional reheat systems, each with both high and low sulfur coal. Results are
summarized in Table 2.
TABLE 2 REHEAT SYSTEM CAPITAL AND OPERATING COST SUMMARY
Basis: 2x500 MW coal-fired plant, Midwest
location, 30-year plant life, pricing
level-EOY 1979
HS - High-sulfur coal, 4.0% avg.
LS - Low-sulfur coal, 0.48% avg.
Capacity factor 70%
In-Line Hot-Air Oil-Fire:
Cyclic Reheat Steam Reheat In. Reheat Reneat.
HS
LS HS
LS
HS LS
HS LS
Process Capital, $/kW 17.7 22.8 6.2 6.8 3.1 3.2 2.2
Total Capital, $/kW
First Year O&M Cost,
$/kW
Level 1 zed Capital
Charges, milIs/kWH
Level1 zed O&M Cost,
mills/kWH
Total 30-year Levelized
Cost, mills/kWh
23.2 29.4 8.8 9.6 4.9 5.1 3.7
1.9
0.68
0.60
2.2 4.1 4.4 9.5 10.6 5.0
2.4
4.0
5.5
0.86 0.26 0.28 0.14 0.15 0.11 0.12
0.69 1.30 1.42 3.04 3.40 1.97 2.17
1.28 1.55 1.56 1.70 3.18 3.55 2.08 2.29
205
-------
At the energy values of $2.90 per 1000 pounds ($6.39 per 1000 kg) for
70-250 psia (483-1724 kPa) steam, $4.45 per 1000 pounds ($9.80 per 1000 kg)
for 250-600 psia (1724-4137 kPa) steam, $4.40 per million Btu ($4.17 per GO)
for oil, and 31 mills per kWh for electric power assumed in this study, cyclic
reheat is estimated to have the lowest 30-year levelized total cost for stack
gas reheat. Cyclic reheat has the highest capital requirement; direct
combustion reheat has the lowest.
Study conclusions and recommendations include the following:
o Considerable research activities on cyclic reheat are being conducted
with promising results. For low-sulfur coal application, operating
experience has been good at the Harrington Station of Southwest
Public Service. No serious corrosion or plugging problems are
reported. For medium-sulfur coal application (1-2% sulfur) satis-
factory operating experience is reported in Japan using the regenera-
tive type of heat exchanger (Ljungstrom type). For high-sulfur coal
application there is currently no operating experience; however,
TVA's Paradise Steam Plant FGD system using cyclic reheat with high-
sulfur coal is expected to start operation in 1982.
o The major advantage of cyclic reheat is energy savings. This is
realized at the expense of higher capital cost. A simplified (EPRI
Class I) estimate indicates that when medium-pressure steam costs
$2.30 per 1,000 pounds ($5.10 per 1000 kg) or more, cyclic reheat has
an economic advantage over conventional in-line steam reheat for
high-sulfur coal. The breakeven point for low-sulfur coal application
(based on an inlet flue gas temperature of 270°F (132°C) instead of
300°F (149°C)) is $2.60 per 1,000 pounds ($5.73 per 1000 kg).
o The capital cost of cyclic reheat is quite sensitive to the inlet
flue gas temperature, which influences heat extractor size and mater-
ials of construction. Higher flue gas temperatures mean lower capi-
tal cost, but penalize power plant thermal efficiency. The compara-
tive economics of a cyclic reheat system are also sensitive to the
energy cost. Therefore, each plant should be independently
evaluated.
206
-------
Cyclic reheat eliminates steam (or other fuels) consumption. How-
ever, this energy saving is partially offset by additional gas-side
pressure drop across the exchanger surface. The exchanger size is
influenced by the allowable pVessure drop. High pressure drop
improves heat transfer and reduces exchanger size, but consumes
electrical energy in fan horsepower. Pressure drop and heat
exchanger size must therefore be properly balanced to arrive at an
economic optimum. In a gas/liquid/gas system, the design liquid
temperatures must also be chosen to minimize the exchanger cost
overall (extractors and reheaters).
Cyclic reheat reduces scrubber inlet gas temperature. This has two
effects on the main FGD system: (a) lowering the adiabatic satura-
tion temperature of the gas, and (b) reducing the process water
makeup requirement. Lowering the adiabatic saturation temperature
may improve SOg removal efficiency, depending upon the particular FGD
system. For the advanced concept of citrate absorption/steam strip-
ping, a lower operating temperature means reduced steam consumption
for S0£ stripping (Ibs steam/lbs $02). Reduced process water makeup
may be beneficial in some arid areas; however, it also reduces the
water available for mist eliminator wash. Both these factors are
significant in FGD system design. Less water content in the scrubbed
gas may enhance visibility by reducing the vapor plume.
The rapidly escalating cost of energy has made cyclic reheat an
increasingly attractive alternate to conventional reheat methods for
FGD systems using wet scrubbing. However, before large-scale adop-
tion of this reheat scheme takes place, several areas of uncertainty
such as corrosion, plugging, and cleaning of the heat extractor
should be investigated to minimize design errors and optimize equip-
ment cost.
207
-------
Future studies should include in-depth studies of the Ljunstrom-type
heat exchanger and of the heat pipe for cyclic reheat application.
This would involve close monitoring of operating Ljunstrom-type
systems in Japan.
208
-------
ECONOMICS OF HIGH SOg REMOVAL DESIGNS
TPS 78-760, 78-767
OBJECTIVES
This project is a team effort by Radian Corporation and SRI International. The
objective of Radian's work was to define representative cases and develop
process designs and material balances that could be used to determine costs
for each case. The process designs were developed using a process simulation
computer program developed by the contractor. Cases were selected to span:
o Coal--eastern and western
o S02 removal— 84%, 93% and 99%
0 Alkali—magnesia, limestone and lime
The objective of SRI's work was to use the results of the Radian work to
develop a cost estimate for each case and then analyze the results. The
latest vendor cost information was used to prepare the economic estimates.
PROJECT DESCRIPTION
This project is composed of two separate technical planning studies that were
undertaken to predict the effect of potential increasingly strict S02 emission
limits on the economics of wet scrubbing. In the first study, Radian
Corporation performed process designs and material balances as input to the
second half of the study, an economic evaluation performed by SRI
International.
RESULTS
Process Designs. The major variables that were investigated in these
designs were the liquid-to-gas ratio (L/G) in the scrubber and the volume of
the process slurry holding tank. The former affects the S0£ removal effi-
ciency and the latter affects the scaling potential in the scrubber.
209
-------
Under the study assumptions, higher S02 removals required moderate increases
in L/G and were found to be dependent on the magnesium and chloride levels in
the slurry liquors. This information is useful in gaining an understanding of
the magnitude of the process changes required for high S02 removals.
Cost Estimates. The study results are presented in Table 3. For low-
sulfur coal systems, the design coal chosen meets the 1971 New Source
Performance Standard (NSPS) for S02 without any further S02 removal.
Increasing the design S02 removals to 93% and 99% results in a levelized cost
of 8.5 and 8.9 mills/kWh, respectively. Magnesia scrubbing was about 7-8%
more expensive than limestone scrubbing on a levelized basis for the low-
sulfur western coal cases. For eastern higher-sulfur coal, increasing the
removal requirements to 93% and 99% removal increases the levelized revenue
requirement by 8% and 18%, respectively. Costs are significantly affected by
chloride and magnesium levels in the coal. For high-sulfur coal, magnesia
scrubbing is about 15% cheaper than limestone scrubbing on a levelized revenue
basis.
The significance of the results of this study lies in the comparative numbers
and not in their absolute magnitude. The increased costs are significant for
higher S02 removals but they do not change by an order-of-magnitude as origin-
ally anticipated.
Probably the most significant unanticipated result of the study was the large
effect that the Mg and Cl content of the scrubbing liquor has on system design
and costs for lime and limestone systems. It is clear that this area should
receive more attention in system design.
Finally, although the magnesia system appears less expensive than conventional
lime and limestone systems for high-sulfur coals, it is still not well
developed and its reliability remains uncertain.
Generalized cost estimates such as these are only an aid in planning either a
research program or the selection of a flue gas desulfurization (FGD)
process. It is not appropriate to generalize these comparisons or assume they
represent manufacturers' current selling prices.
210
-------
TABLE 3 ECONOMIC STUDY RESULTS
System
Limestone
Limestone4
Limestone
Limestone (High Cl)
Limestone (Low Mg)
Limestone (High Mg)
Lime
Limestone
Limestone
Magnesia1*
Magnesia
Magnesia
Magnesia
Limestone
Percent S02
Removal
84
93
99
93
93
93
93
93
99
93
99
93
99
93
Type
of Coal1
Eastern
Eastern
Eastern
Eastern
Eastern
Ea stern
Eastern
We stern
Western
Eastern
Eastern
Western
Western
Eastern
Leveli zed
Revenue
Requirement of
FGD. MilTs/kWh2
13.0
14.1
15.4
14.6
13.8
12.9
14.1
8.5
8.9
12.1
13.1
9.1
9.6
14.4
Total Capital
Requirement,
$/kW
165
194
213
204
189
178
178
123
128
193
207
155
163
181
1. Eastern coal, 4.0 sulfur; western coal, 0.48% sulfur; uncontrolled emissions
would be 7.5 and 1.1 Ib/million Btu, respectively. Eastern coal 0.1% Cl in base
case, 0.3% in High Cl case.
2. Assuming an inflation rate of 6.0% per year and a fuel cost increase of 6.2% per
year; 30-year levelized revenue requirement at 1 eve!1 zed capacity factor of
0.7. Methodology standardized by EPRI.
3. Base cases.
4. Variation of base case design.
211
-------
ECONOMICS OF FGD
RP1180-9
OBJECTIVES
The overall objective of this project was to prepare a general and consistent
review of FGD technology economics. Specific objectives were to: (1) review
reasons for variations between published FGD cost estimates, (2) recommend a
consistent methodology for estimating FGD costs, and (3) prepare design and
cost estimates for alternative FGD technologies using this methodology.
PROJECT DESCRIPTION:
An economic evaluation of flue gas desulfurization (FGD) processes was
prepared by Bechtel National, Inc. The report presents a review of published
FGD cost estimates, a discussion of the reasons for variations between
published FGD costs, a recommendation of a methodology for improving the
consistency of FGD cost estimates, and conceptual design and cost estimates
for eight regenerable and nonregenerable FGD technologies, based on the
recommended methodology.
FGD cost and performance estimates are presented for a new 2 x 500 megawatt
unit plant located near Kenosha, Wisconsin and fired by either a 4-percent
sulfur Illinois coal or a 0.48 percent sulfur Wyoming coal. Other major
assumptions include 85 percent sulfur dioxide (S02) removal, four 33-1/3
percent scrubber modules, and redundancy in critical equipment. The evalua-
tion was completed before promulgation of the final revised new-source per-
formance standards for S02 in June 1979. Thus, the 70 to 90 percent S02
removal requirement was not used.
The FGD costs and other data presented in this report have also been used in a
chapter on FGD economics in a report on sulfur oxides control technology being
prepared by the National Research Council's Commission on Sociotechnical
Systems.
212
-------
EPRI intends to update and report FGD cost estimates on a regular basis, as
technologies change.
RESULTS
A review of nine published FGD cost estimates exhibited wide variations in
both estimated and actual FGD costs. These variations often reach factors of
three to five times the costs at the lower end of the cost range. The major
causes of the cost variations are differences in S02 emission standard; scope
of estimate; equipment redundancy; degree of design conservatism; purpose and
level of detail of estimate; and design and economic assumptions, including
coal type, plant location and capacity, and year of estimate.
The standard design and economic assumptions and methodology suggested in the
report are expected to reduce the magnitudes of the differences between esti-
mates. This methodology is already being used in other EPRI-sponsored FGD
evaluations.
Conceptual designs and cost estimates are presented for the limestone slurry,
lime slurry, double alkali, Chiyoda Thoroughbred 121, Wellman-Lord magnesia
slurry, absorption/steam stripping/RESOX, and the lime slurry/spray
drier/fabric filter processes.
For both low and high sulfur coal applications, the alkali-based non-regener-
able processes exhibited the lowest capital and levelized revenue requirements
and the lowest parasitic energy consumptions. The Chiyoda Thoroughbred 121
process appears particularly attractive. It exhibits low total capital and
level ized revenue requirements and also produces a stackable gypsum byproduct
in lower volumes than the sulfite sludge byproducts produced by the other
limestone and lime slurry processes. The spray drier/fabric filter process
using a lime slurry is also attractive, but has not yet been demonstrated for
high sulfur coal applications. These costs are represented in the Figures 1
42.
The economics of absorption/steam stripping and other regenerate processes
are adversely affected by high energy consumption, principally for
213
-------
ro
-
2
LU
200 i—
150
.
«
D.
O
T>
in
O
O
« 100
Q.
to
50
HS
LS
Process capital
General facilities
engineering and
contingency
Owner's cost
I
HS
LS
I
HS
1
LS
HS
LS
HS
vvv
LS
HS
LS
R&D
Successful
HS
LS
HSl
LS
HS
LS
Basis:
Limestone
Slurry
Lime
Slurry
Double
Alkali
Chiyoda
CT-121
Wellman-
Lord
Magnesia
Slurry
New 2 X 500 MW coal fired plant, midwest location,
30 year plant life
Mid 1980 plant startup
High sulfur coal - 4.0% sulfur (avg)
Low sulfur coal - 0.48% sulfur (avg)
Capacity factor 70%, 6132 hrs/yr
85% SO2 removal
Absorption/
Steam
Stripping/
Resox .
Spray Dryer/
Fabric Filter
HS = High sulfur coal
LS = Low sulfur coal
-------
I
0)
e
'3
CT
0)
cc
0)
cc
o
ro
M
01
0>
o>
Basis:
20
15
10
iL 5
HS
LS
Limestone
Slurry
Plant investment
Owner's cost
Total
Captial
Requirements
:LS
Fixed O&M cost
Variable O&M cost
•
•
HS
LS
1
HS
I
R&D
Successful
HS:
"LS:
Lime
Slurry
Double
Alkali
Chiyoda
CT-121
Wellman-
Lord
Magnesia
Slurry
New 2 X 500 MW coal fired plant, midwest location,
30 year plant life, 1979-2008
Mid 1980 plant startup
High sulfur coal - 4.0% sulfur (avg)
Low sulfur coal - 0.48% sulfur (avg)
Capacity factor 70%, 6132 hrs/yr
85% SO2 removal
Absorption/
Steam
Stripping/
Resox
Spray Dryer/
Fabric Filter
HS = High sulfur coal
LS = Low sulfur coal
Figure 2 30 Year levelized FGD revenue requirements, 1979-2008.
-------
regeneration of the scrubbing reagent. Under RP1258, EPRI is evaluating
improvements in the absorption/steam stripping/RESOX process that could reduce
both energy consumption and equipment costs.
Generalized cost estimates such as those presented in this report should be
used only as comparative estimates for research and development planning and
FGD process screening. Since the estimates are based on a specific set of
assumptions, it is not appropriate to generalize these estimates or assume
they represent manufacturers' current, site-specific selling prices.
216
-------
INTEGRATED EMISSION CONTROL PILOT PLANT
RP1646-1
RP1870-2
OBJECTIVE
The objective of EPRI's Integrated Emission Control (IEC) pilot plant research
effort is to provide utilities with engineering guidelines for the specifica-
tion of cost effective, reliable integrated emission control systems for coal
fired plants.
PROJECT DESCRIPTION:
An integrated series of 2-1/2 MW pilot plant modules have been or are being
constructed at EPRI's Arapahoe test facility in Denver, Colorado. The facil-
ity extracts flue gas from Public Service Company of Colorado coal fired
unit. The catalytic NO control module and airheater are currently being
A
tested. Additional modules to be tested include a spray dryer, a wet scrubbing
system, a cooling tower, fabric filter and an electrostatic precipitator. The
following elements or testing are planned:
o Complete characterization of each module and of integration effects.
o Implement a plant water chemistry program including integrating the
water loop.
o Investigate impact of flue gas temperature.
o Determine effect of ammonia on air preheater, scrubber and fabric
filters (baghouses).
o Test baghouse and ESPs as a final collection device.
RESULTS AND CONCLUSIONS:
The catalytic NOX reactor has been operating since March 1980 with NOX reduc-
tion performance close to original design. The test program is not far enough
along to allow for detailed evaluation. Mr. Dan Giovanni, Program Manager of
EPRI's Air Quality Control Program, can answer any general question on per-
formance to date. The performance specifications for spray drying and wet
217
-------
scrubbing modules have been released for bids. A test sequence has been
defined for several equipment configurations. All these activities amount to
a multi-year R&D program that will represent the first attempt to integrate
all of the best available control technologies into a single facility.
218
-------
SULFUR PRODUCTION BY RESOX
RP784-2
RP1257-1
OBJECTIVE
The objective of EPRI's S02 reduction development efforts is to develop a
regenerable FGD process that produces elemental sulfur without using a reduc-
ing gas such as methane (natural gas) or producer gas (CO,H2). The RESOX
process originally developed by Foster Wheeler Energy Corp. takes concentrated
S02 produced by various FGD absorption systems and converts it to elemental
sulfur by reaction with hot crushed coal.
PROJECT DESCRIPTION
Several related projects have contributed to the EPRI sponsored development of
the RESOX technology. Early cost estimates developed for EPRI pointed to
RESOX as a promising regeneration technique. The development effort is two
fold with a U.S. 1 MW laboratory effort in Livingston, New Jersey, and a 42 MW
demonstration effort in Lunen, Federal Republic of Germany. Initial German
results were presented in a paper given in April 1979 at the ACS meeting in
Honolulu.
RESULTS AND CONCLUSIONS:
Initial sulfur production was in July, 1978. Major equipment problems caused
extended outages and little run time from August to March 1979. A total run
time of approximately 900 hours was obtained, with the most productive runs in
May and June of 1979. Low yields of 65-74% elemental sulfur based on a sulfur
material balance caused EPRI to postpone further German efforts until problems
were resolved in the lab. Lab runs were undertaken in October-November, 1979
attempting to reproduce German conditions and to find an improved method to
correct the problem. Both goals were met in the lab program and the problem
was diagnosed as overconversion of S02 to H2S and COS. This was caused by an
imbalance in gas flows, coal flow, and coal reactivity that led to high
temperatures and, thus, low sulfur yields. The success of the improvement is
219
-------
causing EPRI to seek patent protection for the invention. Yields of 70% were
recorded when reproducing Lunen conditions, and yields of 82.1, and 83.8$ were
recorded using the improved method of RESOX operation.
In order to increase the applicability of RESOX, additional lab work has been
done to ascertain that bituminous or subbituminous coals as well as anthracite
can be used as a reductant. Testing using these types of coals was performed
with gases simulating Bergbau Forchung, Wellman-Lord, and magnesia off gas
(Chemico-Basic). This testing was done without the improvement mentioned
earlier, which leads us to believe that yields and sulfur purity can be
increased. Even without the improvement, it still appears noncaking
subbituminous and bituminous coals can be used in the RESOX process and that
relatively dilute rich gas streams, such as magnesia off gas, can be
processed. Coal types tested and results from this earlier testing are
summarized in Table 4 & 5.
TABLE 4 RESOX TEST COALS
Mine/Seam
Black Mesa/Yellow
Seneca/Wadge
Sophia Jacoba
County/State
Navajo, Arizona
Bituminous
Routt, Colorado
Ruhr, Germany
ASTM Ranking
High Vol C
Subbituminous A
Anthracite
220
-------
TABLE 5 RESOX TEST RESULTS FOR VARIOUS
COALS AND FRONT-END PROCESSES
H20/S02
Mol Ratio
S02 In
Feed (Mol %)
Inlet S02
Conversion (%)
El oriental
Sulfur Yield (%)
INJ
Bergau-Forschung Process
Sophia Jacoba coal
Black Mesa coal
Seneca coal
Wellman Lord Process
Sophia Jacoba coal
Black Mesa coal
Seneca coal
Chemlco-Basic Process
Sophia Jacoba coal
Black Mesa coal
Seneca coal
2.2
2.2
2.2
2.5
5.0
6.0
5.0
5.0
4.0
20.7
12.0
15.3
24.4
14.0
11.8
8.3
8.3
9.0
90.0
92.1
86.4
91.3
88.7
84.5
91.6
88.3
82.6
79.5
85.2
71.8
80.0
82.7
75.0
79.7
69.4
68.1
-------
WATER INTEGRATION SIMULATION FOR LIME AND LIMESTONE FGD SYSTEMS
TPS80-730
OBJECTIVE
Efficient utilization of water in power plants has become increasingly
important particularly where water is scarce. For those plants which operate
a wet scrubbing flue gas desulfurization (FGD) system, minimizing water usage
requires careful study of overall water requirements with possible integration
of water treatment and disposal in power plant and FGD systems. It may be
possible, for example, to use some power plant waste streams in an FGD
systems.
PROJECT DESCRIPTION
To determine the effects of various water streams on the operation of the FGD
system, a computer model which calculates stream compositions for lime or
limestone wet scrubbing has been used. This model will accept two different
water compositions per material balance calculation. Approximately 40 differ-
ent cases using raw water and waste streams such as cooling tower blowdown,
and water treatment wastes in various combinations in the FGD systems have
been done.
Four different raw water sources ranging in total dissolved solids (TDS) from
60 to 3400 ppm were chosen for study. Other variables are coal supply (one
eastern and one western), FGD system (lime and limestone), prescrubber (with
and without), and S02 removal efficiency.
These data were used in the Inorganic Process Simulator program of Radian
Corp. to obtain reference material balances assuming that raw water was the
only source of water for the FGD systems. Various plant streams (cooling
tower blowdown, lime softening waste streams, etc.) were also calculated using
these raw water sources in a computer program simulating cooling tower
operations.
222
-------
To determine if an FGD system could utilize any of the cooling tower waste
streams as makeup water, combinations of raw water, cooling tower blowdown,
and treatment wastes were used in the Inorganic Process Simulator was water
sources. Material balances, scaling potential, operating conditions, scaling
potential, operating conditions for the desired S02 removal, and stream com-
positions are determined by this computation program. Feasibility of using
the cooling tower waste streams was judged by comparing the simulator waste
stream data to those of the reference raw water data.
RESULTS AND CONCLUSIONS
Preliminary results of simulations using an eastern coal are shown in Table 6,
TABLE 6 Simulation of Desulfurization of Eastern Coal Flue Gas
S02 in Flue Gas 3000 ppm
S02 Removal Efficiency 90%
FGD Absorbent Limestone
Makeup
Water Source
Lake Sakajawea
Santee River
Mississippi River
Cooling Tower Blowdown
(Miss. River)
TDS
of Water, ppm
3470
66
458
8460
L/G Requi
gal/kft3
, i
96
129
129
86
red
(1/m3)
(12.8)
(17.2)
(17.2)
(11-5)
CaS03
Relative
Saturation
2.5
1.4
1.5
2.8
Simulations using a western low-sulfur coal are given in Table 7.
223
-------
TABLE 7 Simulation of Desulfurlzation of Western Coal Flue Gas
S02 in Flue Gas
S02 Removal Efficiency
FGD Absorbent
Makeup
Water Source
Lake Sakajawea
Mississippi River
Cooling Tower (Miss.
River)
Cooling Tower
(miss. River)
Cooling Tower
(Lake Sakajawea)
TDS
of Water, ppm
400 ppm
70%
Limestone
L/G Required
Absorber Effluent
CaS03 Relative
Saturation
3470
458
1530
8480
10,700
gal/kft3
15
62
53
25
9
0/m3)
(2.0)
(8.3)
(7.1)
(3.3)
(1.2)
0.7
0.1
0.5
1.8
1.5
The preliminary results show that the quality of the water used can affect
major variables such as L/G. Effects of water quality on lime slaking, lime
and limestone availability and utilization, scaling, crystallization and mist
elimination can also be indicated by these studies and wil be included in
evaluations of the data as the work continues.
FUTURE PLANS
The simulator studies are to be continued using various combinations of water
and waste water streams. Other coal, lime, and limestone compositions are to
be combined in the system calculations. Build-up of impurities (such as
chloride, sodium, and magnesium) will be calculated. Laboratory tests will
then be completed to determine the effects, if any, on phase relationships,
crystallization of calcium sulfite and sulfate, reagent utilization, and
corrosion potential.
224
-------
ENTRAINMENT IN WET STACKS
RP1653-1
OBJECTIVES
The history of wet stacks in the utility industry indicates two major problems
relative to their operation: increased materials corrosion and mist genera-
tion. This project' has been directed toward the problem of mist generation.
PROJECT DESCRIPTION
The work involves the collection and evaluation of historical data and
laboratory pilot research on aerosol emission (reentrainment) from stack
walls. The latter work involves experimental measurement of the critical
velocity where water droplets are removed from the condensate film for the six
different combinations of stack liner materials and construction roughness
shown below.
STACK LINER MATERIALS TESTED FOR ENTRAINMENT FROM CONDENSATE FILM
1. Acid resistant brick (Custodis)
o
Radical tolerance of construction (3.3 x 10"Jm or 0.13 in)
2. Acid resistant brick (Custodis)
Radical tolerance of construction 0.0
3. SI units CXL - 2000 coating
Col bran division of Pullman Power Products
4. Plastic coating No. 4005
Vinyl Ester, Wisconsin Protective Coating Corp.
5. Inconel alloy welded
6. FRP (Fiberglass reinforced plastic) Alcore division of Custodis
225
-------
Included 1n the study is an evaluation of choke design and operation on a wet
stack. (The choke in a stack is the narrowing of the stack diameter at the
top or exit to increase velocity and aid in plume dispersion.) This work
includes experimental evaluation of two choke systems designed with water-film
collectors. Separation and reentrainment prevention techniques for wet
systems are also being evaluated using a mathematical model.
Based on the operating experience, mathematical modeling and experimental
work, a set of guidelines for acceptable wet stack system designs will be
formulated. The guidelines documents will include criteria for the selection
of duct size and stack diameter, and a discussion for the trade-off between
liner construction and critical reentrainment velocity, and the need for
reentrainment prevention techniques or entrainment separation devices. The
information is intended for use by A&E firms and utility owners to select or
review wet stack system designs.
RESULTS AND CONCLUSIONS:
Tables fi and Pwinddcate-'swneoof t^eeinfornra-tiow^thatthJ^Lbeen-iodat&iTffitlJih-i the
survey of wet stack operating experience. Based on literature and laboratory
measurements a properly operating mist eliminator carryover rate is
0.23-2.3 g/m3 (0.1-1 gr/ft3). Under upset conditions this can reach as high
o O
as 9.2 g/nr (4 gr/ft°). Theoretical estimates of stack condensation range
from 0.11-0.55 g/m3 (0.05-0.24 gr/ft3). If the measurements and estimates are
accurate, the mist eliminator carryover is a significant portion of the
condensed moisture in the stack and thus is a very Important variable to be
considered in the design of wet stack systems. The validity of these
laboratory measurements needs to be confirmed by comprehensive field
measurements.
226
-------
TABLE 8 WET STACK DESCRIPTION
N>
ro
Plant Number
1
2
3
4
5
6
7
Entrainment Condition
Entralnment 1s a big
probl an
Moderate entrainment
Noticeable Entralnment
during:
- absorber overload
- dirty demisters
- plugged drains
No known entrainment
No known entrainment
No known entrainment
Slight noticeable entrain-
ment during humid weather
No.
Units
2
1
2
2
2
2
2
Participate
Removal
VentuM
ESP
Venturi
ESP
ESP
ESP
ESP
Absorber
Mobile Bed
Mobile Bed
Venturi
Venturi
Venturi
Mobile Bed
Mobile Bed
I.D.
Fans
Dry
Dry
Wet
Wet
Wet
Dry
Dry
Secondary
Demlsters
None
None
Tried retro-
fitting 3 con-
figurations with
no success
Yes
Chevron 4-pass
Yes
Chevron 4-pass
None
None
-------
TABLE 9 STACK BREECHING DUCT DATA
Stack Data
Breeching Duct Data - Entrance to
Stack
Plant
1
2
ro
ro
00 3
4
5
6
7
Height
50m
(165 ft)
183m
(600 ft)
290m
(950 ft)
119m
(390 ft)
104m
(340 ft)
244m
(800 ft)
137m
(450 ft)
Diameter
(base-top)
4.9m
(16 ft.)
5.9m
(19-1/2 ft)
5.8m
(19 ft)
8.8-7. 9m
(29-26ft)
8. 8-7. 9m
(29-26 ft)
13.4-7.9m
(44-26 ft)
3.4m
(11 ft)
Gas Velocity
(max)
7.6m/s
(25 ft/s)
27.4 m/s
(90 ft/s)
27.4 m/s
(90 ft/s*)
7.3m/s
(24 ft/s)
llm/s
(36 ft/s)
24.4 m/s
(80 ft/s)
30 m/s
(95 ft/s)
Gas Temperature
(average)
43.3°C
(110°F)
48.9°C
(120°F)
48.9°C
(120°F)
48.9°C
(120°F)
48.9°C
(120°F)
51.7°C
(125°F)
54.4°C
(130°F)
Liner Material
Carbon Steel
w/ Precrete
Mild steel w/
Ceil coat
Carbon Steel
w/ Heil Rigi-
flake
Acid proof
Brick & Mortar
Acid proof
Brick & Mortar
Acid Proof
Brick & Mortar
Acid Proof
Brick & Mortar
Liner
Insulation
None
2-3 Inches
fiberglass
None
None
None
None
None
No.
Flues
1
1
4
1
1
1
1
Height
Width
Scrubber & demlster
share stack flow
enters at bottom of
stack.
7.6m
(25 ft)
12.2m
(40 ft)
12.2m
(40 ft)
9.1m
(30 ft)
5.8m
(19 ft)
3.7m
(12 ft)
3.7m
(12 ft)
3.7m
(12 ft)
4.6m
(15 ft)
2.3m
(7-1/2
ft)
Distance from
dust to stack
exit
From demlster
23m
(75 ft)
247m
(810 ft)
88m
(290ft)
72m
(2325 ft)
213m
(700 ft)
101m
(330 ft)
* Secondary source gives 60 ft/sec velocity.
-------
FUTURE PROGRAM EMPHASIS
Future R&D emphasis will be on the following:
o Demonstrations
« Chiyoda 121
~ RESOX
— Aqueous carbonate process
o Pilot Plant
— Absorption/steam stripping improvements
~ Spray dryer testing
— Integrated emission control
o Field Testing
— Continuous emission monitor testing
— Materials testing
— Spray dryer chacterization
o Evaluations
~ Reliability improvements
— Cyclic reheat feasibility
o Laboratory Testing
-- Corrosion inhibitors
-- Limestone dissolution
-- Crystallization
~ Additives
229
-------
Tech Transfer
-- Revised Lime FGD Systems Data Book
— Issue limestone data book
-- Continuous emission monitor guidelines
— Workshops and seminars
CONCLUSION
EPRI research and development has attempted to address problems in FGD which
have led to the high cost, low reliability and inefficient resource use in
current systems. EPRI's efforts are aimed at near term solutions to problems
in system chemistry, corrosion, cost, energy use, by-product character, and
system design. The results reported in this paper are documented more fully
in individual reports that are either in print or in the process of publica-
tion. EPRI welcomes and encourages comments, criticisms ,or inquiry regarding
its FGD programs and asks that such calls be directed to Stu Dalton, Program
Manager, Desulfurization Processes Program. (415) 855-2467.
230
-------
Session 4:'UTILITY APPLICATIONS
H. William Elder, Chairman
Tennessee Valley Authority
Muscle Shoals, Alabama
231
-------
TEST RESULTS ON ADIPIC ACID-ENHANCED LIMESTONE SCRUBBING
AT THE EPA SHAWNEE TEST FACILITY
-THIRD REPORT-
D.A. Burbank, S.C. Wang, and R. R. McKinsey
Bechtel National, Inc.
50 Beale Street
San Francisco, California 94105
and
J.E. Williams
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
ABSTRACT
Adipic acid has been demonstrated as a powerful scrubbing additive for
enhancing SO? removal in lime and limestone wet scrubbing tests both
at the EPA/IERL pilot plant at Research Triangle Park, North Carolina,
and at the EPA-sponsored Shawnee Test Facility near Paducah, Kentucky.
Improved limestone utilization and operating reliability have also been
demonstrated.
Earlier test results using adipic acid, from July 1978 through October
1979, were reported at the Fifth Symposium on Flue Gas Desulfurization
in Las Vegas, Nevada, March 5-8, 1979, and at EPA's Fifth Industry Briefing
in Raleigh, North Carolina, December 5, 1979. This is the third report
on the recent adipic acid test results at the Shawnee Test Facility from
October 1979 through May 1980.
The recent tests with adipic acid were conducted only on the venturi/spray
tower system. All tests were made with limestone slurry. These included:
(1) partial factorial tests to characterize the effects of pH, adipic acid
concentration, and other operating parameters on S0? removal; (2) single-
loop (one-tank) tests without forced oxidation at low pH and high (4000
ppm) adipic acid concentration; (3) tests with a venturi only to determine
the limits of SO? removal; (4) single-loop forced oxidation tests, with
both one tank and two tanks; and (5) bleed stream oxidation tests at low
pH and high (4000 ppm) adipic acid concentration.
Major efforts during the recent test period were directed toward investiga-
tion of the effect of pH on the degradation of adipic acid. It was found
that the adipic acid degradation is minimized when the scrubber is operated
at low (below 5.0-5.1) inlet pH. Forced oxidation and poor limestone utili-
zation tend to increase the deqradation.
Preceding page blank
233
-------
ACKNOWLEDGMENT
The authors wish to express their appreciation to the following Bechtel
personnel who contributed to the preparation of this paper:
C. L. DaMassa H. K. Pate!
T. M. Martin D. T. Rabb
D. P. McGrath C. H. Rowland
M. S. Mihalik V. C. Van der Mast
The authors wish to thank Gary Rochelle of the University of Texas at
Austin, consultant to the project, for his constant input of ideas and
assistance in the adipic acid development program.
Further acknowledgment and appreciation are extended to personnel of TVA's
Energy Design and Operations Branch, both at the Shawnee Test Facility
and in Muscle Shoals, Alabama, who are responsible for operation, main-
tenance, and engineering modification of the facility.
The contribution and support of the Department of Energy are also acknowledged.
NOTE
Although it is the policy of the EPA to use the metric system for quantitative;
descriptions, the British system is used in this report. Readers who are more
accustomed to metric units are referred to the conversion table in the Appendix;
234
-------
TEST RESULTS ON ADIPIC ACID-ENHANCED
LIMESTONE SCRUBBING AT THE EPA
SHAWNEE TEST FACILITY
- THIRD REPORT -
Section 1
INTRODUCTION
Since October 1977 one of the primary objectives of the Environmental Protection
Agency (EPA) alkali wet scrubbing test program has been to enhance S02 removal
and improve the reliability and economics of the lime and limestone wet scrubbing
systems by use of adipic acid as a chemical additive.
Testing of adipic acid-enhanced limestone scrubbing began in October 1977 at
the EPA 0.1 MW pilot plant at Industrial Environmental Research Laboratory,
Research Triangle Park (IERL-RTP), North Carolina (Reference 1). As a logical ;
progression, larger scale testing was conducted beginning in July 1978 at EPA's 10
MW prototype Shawnee Test Facility located at the Tennessee Valley Authority (TVA)
Shawnee Steam Plant near Paducah, Kentucky. Test results from the Shawnee Test
Facility from July 1978 through October 1979 were presented in two previous reports
(References 2 and 3). As part of EPA's continuing program of technology transfer.,
to further demonstrate the effectiveness of adipic acid, and to encourage its use,
the EPA contracted with Radian Corporation in the spring of 1980 to carry out a
full-scale demonstration program of adipic acid-enhanced limestone scrubbing. The
program is being conducted at the Springfield City Utilities' Southwest Station
near Springfield, Missouri. Testing in the full-scale units began in the late
summer of 1980.
This report is the third presenting the test results with adipic acid from the
Shawnee Test Facility. The report covers the period from October 1979 through Mcy
1980. During this period, adipic acid testing was conducted only on the venturi/
spray tower system (Train 100). All tests were conducted with limestone slurry
and with flue gas having high fly ash loading (3 to 6 grains/scf dry).
During the report period, Train 200 (TCA) was operated by EPRI/UOP/TVA on a
DOWA basic aluminum sulfate process, and Train 300 was operated by EPRI/TVA
on a cocurrent, high-velocity scrubber configuration.
THEORY AND ADVANTAGES OF ADIPIC ACID-ENHANCED SCRUBBING
Adipic acid is a dicarboxylic organic acid [HOOC(CH2)4COOH] in powder form,
which is commercially available and used primarily as a raw material in
the nylon manufacturing industry. Initial tests with adipic acid at the
235
-------
IERL-RTP pilot plant were undertaken as a result of theoretical analyses by
Rochelle (Reference 4). Adipic acid effectively buffers the pH in limestone/
lime SOo absorbers and improves the S02 removal efficiency. The buffering
action limits the drop in pH at the gas/liquid interface during S02 absorp-
tion, and the resultant higher concentration of S02 at the interface accele-
rates the liquid-phase mass transfer. The capacity of the bulk liquor for
reaction with S02 is also increased by the presence of calcium adipate in
solution. Thus, the S02 absorption becomes less dependent on the dissolution
rate of limestone or CaSOo in the absorber to provide the necessary alkalinity.
In the case of limestone scrubbing, it follows that a given S02 removal effi-
ciency can be achieved at a lower limestone stoichiometry.
Further analysis by Rochelle (Reference 5) suggested that the use of additives
would be most attractive economically when used in scrubbers employing forced
oxidation. If no decomposition or volatilization of the additive occurs, the
makeup requirements of the additive would be minimized by the more tightly
closed liquor loop achievable due to the better dewatering properties of the
oxidized sludge.
Several advantages of adipic acid over other additives, such as MgO, have been
cited previously (References 1, 2, and 3). Further, the optimum concentration
of adipic acid for effective improvement in S02 removal is only 700 to 1500 ppm
at a scrubber inlet pH above about 5.2. The preliminary economic evaluations
(Reference 2) have shown that adipic acid can reduce both the capital investment
and the operating cost of limestone systems while simultaneously increasing the
performance, even under those conditions in which the actual addition rate is
3 to 5 times the theoretical requirement due to the degradation of the acid.
This report shows that the degradation of adipic acid can be minimized when the
scrubber inlet pH is lowered to below about 5.0. Although higher adipic acid
concentration is needed at the lower pH to achieve the same degree of S02 removal
efficiency, overall adipic acid consumption is reduced compared to the higher pH
operation. For this reason, and with the further improvement in limestone utili-
zation at low pH, the low pH operation should be more economically attractive.
Section 11 presents an update of the economic evaluations given in Reference 2.
TEST FACILITY AND PROGRAM
Readers who are unfamiliar with the Shawnee Test Facility and the earlier adipic
acid test programs are referred to References 2 and 3. A summary of the earlier
work is given in Section 2. This report covers the adipic acid test results from
October 1979 through May 1980 on the venturi/spray tower system. The following
adipic acid tests were conducted during this period:
* Partial factorial tests to characterize the venturi/spray tower
performance using a single tank without forced oxidation
• Investigation of the effect of pH on adipic acid degradation
with and without forced oxidation
236
-------
t SOg removal capability of the venturl scrubber alone
• Forced oxidation within the scrubber loop using a single tank
• Forced oxidation within the scrubber loop using two tanks in series
t Forced oxidation of the bleed stream
All tests were conducted using limestone slurry and flue gas containing 3 to 6
grains/dry scf of fly ash. Sections 3 to 8 discuss and summarize these tests.
Section 9 describes scrubber system behavior during limestone blinding and
the conditions leading to it. Recommended solutions for eliminating or
avoiding limestone blinding are also given. Section 10 gives updated data
on the dewatering properties of adipic acid-enhanced limestone slurry.
237
-------
Section 2
SUMMARY OF PREVIOUS WORK
Based on the earlier test results through October 1979 (References 1, 2 and 3),
both at the IERL-RTP pilot plant and at the Shawnee Test Facility, the charac-
teristics of adipic acid as a lime/limestone scrubber additive can be summarized
as follows:
BENEFICIAL ASPECTS
® Adipic acid significantly enhances SOo removal. At a scrubber inlet
pH above about 5.2, at which most of the adipic acid is in ionized
form, the optimum concentration range is only 700 to 1500 ppm.
• At the minimum effective pH of 5.2, the corresponding limestone
utilization is normally about 80 percent or higher; thus the quantity
of waste solids generated is reduced. High limestone utilization
also contributes to reliable scrubber operation.
« With proper pH control and sufficient adipic acid concentration
(sufficient buffer capacity), steady outlet S02 concentrations can
be maintained even with wide fluctuations of inlet S02 concentrations.
• Adipic acid-enhanced limestone scrubbing has lower projected capital
and operating costs than unenhanced limestone or limestone/MgO scrubbing
(Reference 2). This is primarily due to the reduced limestone consump-
tion, the associated grinding cost, and the reduced quantity of waste
sludge generated with adipic acid-enhanced scrubbing.
« Since limestone dissolution is not a rate-controlling step in S02
absorption for an adipic acid-enhanced limestone system, adipic
acid should promote use of less expensive and less energy-intensive
limestone than lime.
« The effectiveness of adipic acid is not affected by forced oxidation
and it can be used with both lime and limestone in systems with or
without forced oxidation.
• The effectiveness of adipic acid is not adversely affected by chlorides
as is the 1imestone/MgO process. Thus it is especially attractive for
very tightly closed liquor-loop operation.
« When used with lime, both good S02 removal and sulfite oxidation can
be achieved in a single-loop scrubbing system using within-scrubber-
loop forced oxidation.
NEGATIVE ASPECTS
» Adipic acid decomposition, and the indications of its being adsorbed
on solids or occluded in solids (Reference 6), require adding up to
238
-------
5 times that amount theoretically required (Reference 2). However,
the consumption over the ranges anticipated has negligible economic
impact.
• Some decomposition products, such as valeric acid, have an unpleasant
odor. However, this has not been a problem in testing to date.
OTHER CONSIDERATIONS
Tpxicity. No further work in this area has been conducted by the EPA since
the last report (Reference 3). Preliminary results from Level 1 chemical
and bioassay analyses showed no measurable difference in toxicity or mutage-
nicity of samples with and without adipic acid addition. These samples were
taken in February 1979 from a limestone/adipic acid forced-oxidation run and
a base case limestone run without forced oxidation. It should be noted that
adipic acid is a food additive.
Limestone Blinding and Calcium Sulfite Scaling. Adipic acid buffers the pH
drop across the scrubber, and therefore increases the potential of calcium
sulfite scaling at the bottom part of the scrubber. At a constant liquid-to-
gas ratio, addition of adipic acid increases the S02 make-per-pass and
similarly increases the sulfite scaling tendency at the bottom of the scrubber.
In the case of limestone scrubbing, blinding of limestone by calcium sulfite
could occur, resulting in low pH and poor limestone utilization. This would
be particularly true with forced oxidation in the scrubber loop (or in a
system with a high level of natural oxidation); such conditions deplete calcium
sulfite solid seeds. Operating and design considerations for avoiding limestone
blinding are presented in Section 9.
239
-------
Section 3
FACTORIAL TESTS ON THE VENTURI/SPRAY TOWER
SYSTEM WITH LIMESTONE/ADIPIC ACID SLURRY
Fifty 1 linestone/adipic acid partial factorial tests, Runs VAA201 through
VAA250, were conducted on the venturi/spray tower system. All tests were
made without forced oxidation and with a common effluent hold tank as shown
in Figure 3-1.
The tests examined the effect of spray tower liquid-to-gas ratio, scrubber
inlet liquor pH, and adipic acid concentration on S02 removal. Table 3-1
summarizes the test results. The operating conditions held constant during
these tests were:
Fly ash loading: High (3-6 grains/dry scf)
Flue gas rate: 35,000 acfm @ 300°F (except Run VAA 207 @ 20,000 acfm)
Hold tank level: 8 ft 6 in. (9.1 - 38 minutes residence time)
Slurry solids concentration: 15 percent
Venturi pressure drop: 9 inches H20 for runs with 600 gpm,
plug wide open for runs with 125 gpm
Spray header configuration (top header is No. 4):
For 400 gpm Header 4
For 800 gpm Headers 3 and 4
For 1200 gpm Headers 2,3, and 4
For 1600 gpm All four headers
Solids dewatering equipment: Clarifier and centrifuge
OVERALL S02 REMOVAL BY VENTURI AND SPRAY TOWER
Equation 3-1 for predicting S02 removal has been fitted to the 10 venturi/spray
tower runs (Runs VAA201 through VAA206 and VAA234 through VAA237) for which the
slurry flow rate to the venturi was held at 600 gpm and the venturi pressure
drop was 9 inches H20.
Fraction S02
Removal = 1 - exp [-0.0019 (L/G)0'55 exp(0.8pH + 8xlO'4 A}] (3-1)
where:
L/G = spray tower liquid-to-gas ratio, gal/mcf (saturated)
pH = scrubber inlet liquor pH
A = adipic acid concentration in scrubber liquor, pprc
The ranges of operating variables covered by the 10 correlated runs are:
L/G = 15-57 gal/mcf
pH = 5.2-5.8 (limestone stoichiometry
controlled at 1.2)
A = 600-1400 ppm
Gas flow rate = 35,000 acfm at 300°F
Inlet S02 concentration = 1500-2200 ppm
240
-------
MAKEUP WATER
CLARIFIED LIQUOR FROM
SOLIDS DEWATERING SYSTEM
BLEED TO SOLIDS
DEWATERING SYSTEM
G-104
Figure 3-1. Flow Diagram for the Venturl/Spray Tower System
With One Tank and Without Forced Oxidation
-------
Table 3-1
RESULTS OF LIMESTONE/ADIPIC ACID FACTORIAL TESTING
ON THE VENTURI/SPRAY TOWER USING ONE TANK WITHOUT FORCED OXIDATION
Run
Mo.
VAA201
VAA202
YAA203
VAA204
VAA205
VAA206
VAA207
VAA208
VAA209
VAA210
VAA211
VAA21-2
VAA213
VAA214
VAA215
VAA216
VAA217
VAA218
VAA219
VAA220
VAA221
VAA222
VAA223
VAA224
VAA225
VAA226
VAA227
VAA228
VAA229
VAA230
VAA231
VAA232
VAA233
VAA234
VAA235
VAA236
VAA237
VAA23S
VAA239
VAA240
VAA241
VAA242
VAA243
VAA244
VAA245
VAA246
VAA247
VAA248
VAA249
VAA250
Liquor Rate
(qpm)
Venturl ST
600 1200
600 800
600 400
600 1600
600 800
600 1200
125 1200
125 800
125 1200
125 400
125 1600
125 1200
125 800
125 1200
125 1200
125 1200
125 1600
125 1600
125 1200
125 1600
125 800
125 400
125 1200
125 1600
125 800
125 400
125 1200
125 1200
125 1200
125 800
125 800
125 1200
125 800
600 800
600 400
600 1200
600 800
125 800
125 1200
125 400
125 1600
125 800
125 1200
125 400
125 800
125 800
125 800
125 1200
125 1200
125 1200
ST
L/G
(gal/mcf)
43
29
15
57
29
43
75(1>
29
43
15
57
43
29
43
43
43
57
57
43
57
29
15
43
57
29
15
43
43
43
29
29
43
29
29
15
43
29
29
43
15
57
29
43
15
29
28
29
43
43
43
Pressm
(Inch
Venturl
8.9
8.7
8.8
8.9
8.9
8.7
0.9
2.3
3.1
2.4
3.7
2.fi
2.5
3.1
2.7
2.7
2.7
2.3
2.6
2.7
1.9
3.3
3.0
3.4
3.1
4.0
3.1
3.6
3.7
4.6
4.1
4.2
3.9
9.0
9.2
8.9
9.0
3.1
3.2
3.7
3.7
3.1
2.9
3.R
3.0
3.0
3.8
2.9
3.0
3.8
•e Drop
H,0)
Total
13.5
14.4
13.6
14.1
15.0
14.6
3.1
7.7
7.9
7.0
8.8
7.9
7.7
8.6
9.0
8.1
7.7
7.6
8.0
7.4
8.6
7.9
7.8
8.2
7.9
9.0
8.0
8.9
8.9
9.7
9.6
8.9
8.9
14.0
15.2
12.2
11.3
8.1
8.1
8.5
9.3
7.6
7.7
8.6
7.5
7.9
9.0
7.7
7.7
9.2
Inlet
Liquor
PH
5.70
5.45
5.85
5.75
5.80
5.65
5.75
5.85
5.70
5.80
5.85
5.60
5.65
S.35
5.05
4.65
5.35
5.00
5.00
4. 300°F for Run VAA207.
• C« 300°F.
-------
Venturl Hqu1d-to-gas ratio • 21 gal/mcf
Venturl pressure drop • 9 Inches H20
Equation 3-1 explains 90 percent of the variation in the data for S02 removal
with a standard error of estimate of 2.7 percent S02 removal (see Figure 3-2)
S02 REMOVAL BY SPRAY TOWER ONLY
Equation 3-2 for prediction of S02 removal has been fitted to the 40 spray
tower runs (minimum effect of venturi - 125 gpm for flue gas humidification):
Fraction S02
Removal = 1 - exp [-2.2xlO~4 (L/G)0'75 exp (pH + 6.2xlO~4 A)] (3-2)
where L/G, pH, and A have the same definitions as for Equation 3-1.
The ranges of variables covered by the 40 correlated runs are:
L/G = 15-75 gal/mcf
pH = 4.6-5.9
A = 600-2400 ppm
Gas flow rate = 35,000 acfm at 300°F (one test at 20,000 acfm)
Inlet S02 concentration = 1600-2900 ppm
Venturi slurry flow rate = 125 gpm
Venturi pressure drop = 2-4 inches H20 (wide open plug)
Equation 3-2 explains 93 percent of the variation in the data for S02 removal
with a standard error of estimate of 4.3 percent S02 removal (see Figure 3-3).
It is important to note that the S02 removal predicted by Equation 3-2
includes the effect of the venturi operating at the minimum conditions defined
above. The magnitude of this effect is discussed later.
Figures 3-4 through 3-6 illustrate the effects of spray tower liquid-to-gas
ratio and inlet liquor pH on S02 removal at adipic acid concentrations of
600, 1300, and 2000 ppm, respectively. The lines on the figures represent
the predictions of Equation 3-2 with actual data points also shown.
Note that the S02 removals for a pH of 4.6 and 2000 ppm adipic acid in
Figure 3-6 are similar both to those in Figure 3-5 for a pH of 5.0 and 1300 ppm
adipic acid, and to those in Figure 3-4 for a pH of 5.4 and 600 ppm adipic
acid. These values are more clearly seen in the following:
Percent S02 Removal at
Scrubber Adipic Spray Tower L/G of
Inlet Acid,
pH ppm
5.4 600
5.0 1300
4.6 2000
30 gal/mcf
60
61
62
243
50 gal/mcf
74
75
76
70 gal/mcf
82
83
84
-------
100
70
100
MEASURED PERCENT SO2 REMOVAL
Figure 3-2. Measured vs. Predicted (Eq. 3-1) S02 Removal
by the Venturi/Spray Tower
90
100
MEASURED PERCENT SO2 REMOVAL
Figure 3-3. Measured vs. Predicted (Eq. 3-2) S0? Removal
by the.Spray Tower
244
-------
20
30 40 SO 60
LIQUID-TO-GAS RATIO, gel/mrf (uturatad)
70
80
Figure 3-4. Effect of Spray Tower Liquid-to-Gas Ratio and Inlet pH
on Spray Tower S02 Removal at 600 ppm Adi pic Acid
100
ApH-5.7
QpH-5.4
VPH-5.0
D pH - 4.6
LIQUID-TO-GAS RATIO, gal/mcf (saturated)
Figure 3-5. Effect of Spray Tower Liquid-to-Gas Ratio and Inlet pH
on Spray Tower SO? Removal at 1300 ppm Adi pic Acid
245
-------
i
UJ
cc
S
i-
UJ
O
K
ut
a.
ApH = 5.7
O pH = 5.4
V pH = 5.0
D pH = 4.6
30 40 50 60
LIQUID-TO-GAS RATIO, gal/mcf (saturated)
Figure 3-6. Effect of Spray Tower Liquid-to-Gas Ratio and Inlet
pH on Spray Tower S02 Removal at 2000 ppm Adipic Acid
246
-------
Thus, within the ranges tested, each 0.4 unit drop in scrubber inlet pH
requires a 700 ppm increase in adipic acid concentration to achieve similar
percent S02 removal.
S02 REMOVAL BY VENTURI ALONE AT MINIMUM SLURRY FLOW RATE AND PRESSURE DROP
For a 2 to 4 hour period at the end of each of seven factorial tests
with the spray tower alone (Runs VAA210, VAA213, VAA222, VAA231, VAA238,
VAA239, and VAA240), the spray tower slurry flow was shut off in order to
determine the S0£ removal achieved by the venturi alone at a minimum slurry
flow rate of 125 gpm, minimum pressure drop of 2 to 4 inches H?0 (wide open
plug), and 35,000 acfm gas flow rate (venturi L/G = 4.5 gal/mcf). These
tests indicated that, at these conditions, the venturi scrubber obtains about
20 percent S02 removal at 600 ppm adipic acid concentration and an inlet pH
of 5.7, 22 percent SO? removal at 1300 ppm adipic acid and a pH of 5.3, and 42
percent S02 removal at 2000 ppm adipic acid and a pH of 5.4. Equation 3-2
does not Include any corrections for SOo removal in the venturi. This should
be taken Into consideration when using Equation 3-2 with Figures 3-3
through 3-6.
247
-------
Section 4
EFFECT OF pH ON ADIPIC ACID CONSUMPTION
During both the earlier factorial tests with adipic acid addition (Reference
3) and the latest factorial tests (Section 3), it was noticed that the
rate of adipic acid addition required to maintain a desired concentration in
the scrubber liquor was substantially reduced when the scrubber inlet pH was
controlled at 5.0 or lower. At higher pH operation, it is necessary to add
adipic acid at up to about 5 times the theoretical addition rate (as defined
below), either because of degradation or decomposition of the adipic acid.
Apparently, the degradation or decomposition process is inhibited under low
pH conditions.
Although the exact mechanism of adipic acid degradation is still not under-
stood, it was decided to investigate the effect of pH on the adipic acid
consumption rate in more detail.
Early in the adipic acid-enhanced lime/limestone testing, it was noted that
the S02 removal enhancement by the adipic acid is maximized when the scrubber
inlet pH is maintained at about 5.2 or higher under the prevailing scrubber
conditions (chloride concentrations). This is because most of the adipic acid
is ionized and its buffering capacity more fully utilized at these higher
inlet pH levels (Reference 7).
Operations at lower pH therefore require higher adipic acid concentrations
to maintain the same degree of S02 removal efficiency (Section 3), because
the ionization and buffer capacity of adipic acid are reduced. However,
experience at Shawnee shows that the total adipic acid consumption at a
scrubber inlet pH below 5.0 and concentration as high as 4000 ppm is actually
lower than at a pH of about 5.4 and 1500 ppm when significant degradation was
noted. Potential advantages of low pH operations are obvious:
• Lower operating cost due to lower adipic acid consumption.
• Easier forced oxidation, in-loop or bleed stream, and less air (and
compressor energy).
0 Essentially complete limestone utilization and improved scrubber
operating reliability.
• Reduced sensitivity to limestone type and grind; fine grinding of
limestone is probably not required.
« Lower sulfite scaling potential.
0 Better prospects (sensitivity) for automatic pH control.
• Greater flexibility for S02 emission control; high sensitivity of
S02 removal to pH allows raising p'H to increase the buffer capacity
and S02 removal when needed.
248
-------
• Improved acceptance of the concept by plant owners because of the
reduced quantity of adlplc add degradation products.
• Applicability to low-sulfur subbituminous and lignite coals containing
alkaline ashes which are extractable only at low pH.
• Probable lower cost due to all of the above factors.
Seven runs were conducted on the venturi/spray tower system to investigate
the effect of pH on the adipic acid consumption rate. These tests were made
with a single effluent hold tank and without forced oxidation. The flow
configuration for these tests is the same as that shown in Figure 3-1.
DISCUSSION OF TEST RESULTS
Table 4-1 summarizes the major test conditions and the run-average results
for the seven tests made in this series. The scrubber inlet pH and the adipic
acid concentration were varied in the tests. All other conditions were held
constant.
Theoretical Adipic Acid Consumption Rate. The theoretical adipic acid
consumption rate is defined as the rate of adipic acid leaving the scrubber
system in the liquor which is entrained in the discharged sludge (filter
cake, centrifuge cake, or clarifier underflow) in a closed-liquor-loop
operation. The theoretical consumption rate is calculated from the material
balances for solids discharged from the scrubber system, solids (or liquor)
concentration in the discharged sludge, and adipic acid concentration in the
liquor.
Since some adipic acid decomposes to lower-carbon carboxylic acids and the
analytical method employed at the Shawnee laboratory determines the total
carboxyl group, "adipic acid concentration" as reported throughout this
report means "total carboxylic acid expressed as adipic acid." Note that
most of the degradation products are also effective as enhancing agents for
SO2 removal.
Effect of pH on Adipic Acid Consumption Rate. As can be seen in Table 4-1,
the ratios of actual-to-theoretical adipic acid consumption were all 1.0 at
a scrubber inlet pH of 4.60 and 4.85 for Runs 926-1A, 926-1G, and 926-1B,
when the adipic acid concentrations were controlled at 4090, 2270, and 1435
ppm; respectively. This indicates that there was essentially no degradation
of adlpic acid, within the accuracy of the material balance calculations.
Further increase in the scrubber inlet pH to 5.05, 5.25, and 5.50 during Runs
926-1C, 926-1H, and 926-1D resulted in actual-to-theoretical adipic acid
consumption ratios of 1.17, 1.24, and 1.59, respectively.
Despite the higher adipic acid concentration required at the lower pH operation,
the total adipic acid consumption can be actually less, as can be seen in
Table 4-1, in terms of actual adipic acid consumption per ton of SO^ absorbed.
249
-------
Table 4-1
RESULTS OF VENTURI/SPRAY TOWER LIMESTQNE/ADIPIC ACID
TESTS USING A SINGLE TANK WITHOUT FORCED OXIDATION
Major Test Conditions
Fly ash loading
Gas rate, acfm'e 300°F
Venturi liquor rate, gpm
Spray tower liquor rate, gpm
Percent solids recirculated (controlled)
EHT residence time, m1n
EHT tank level, ft
Scrubber inlet pH (controlled)
Adipic acid concentration, ppm
Venturi pressure drop, Inches HoO
Run-Average Results
Start-of-run date
Onstream hours
Percent SO^ removal
Inlet S02 concentration, ppm
Adipic acid concentration, ppm
Actual adipic acid consumption, Ibs/tons S0£ abs.
Adipic acid consumption ratio ( actual /theor)
Percent solids recirculated
Scrubber Inlet pH
Scrubber outlet pH
SO^ make-per-pass, mmole/1
Limestone utilization, %
Scrubber Inlet sulfite concentration, ppm
Scrubber outlet sulfite concentration, ppm
Sulfite oxidation, %
Scrubber inlet gypsum saturation, %
Centrifuge cake solids, wt%
Mist eliminator restriction, %
926-1A
High
35,000
600
1600
15
9.1
8.5
4.6
(1)
9
12/18/79
297
90
2650
4090
—
1.0
15.5
4.60
4.30
8.05
97
1540
1550
49
130
69
926-1G
High
35,000
600
1600
15
9.1
8.5
4.8
(1)
9
1/3/80
244
91
2250
2270
4.3
1.0
15.1
4.85
4.50
6.90
96
875
1440
51
116
70
926-1 B
Hiqh
35,000
600
1600
15
9.1
fi.5
4.8
1300
9
1/16/flO
184
84
2115
1435
3.0
1.0
14.9
4.85
4.55
6.00
95
965
1545
49
127
69
0
926-1 C
Hiqh
35,000
600
1600
15
9.1 ,
8.5
5.0
1300
9
1/24/PO
116
91
2150
1290
5.7
1.17
15.1
5.05
4.65
6.60
95
325
695
47
129
66
926-1H
High
35,000
600
1600
15
9.1
8.5
5.25
1300
9
1/29/80
169
93
2150
1285
8.0
1.24
14.9
5.25
4.85
6.75
92
180
305
30
118
61
926-10
Hiqh
35,000
600
1600
15
9.1
8.5
5.5
1300
9
2/8/80
116
96.5
2410
1330
9.6
1.59
15.5
5.50
5.00
7.85
80
135
185
17
112
60
926-1 E
Hiqh
35,000
600
1600
15
9.1
8.5
5.0
700
9
2/15/80
119
77
2450
735
6.0
1.0
15.2
5.05
4.60
6.35
95
325
710
32
113
60
0
on
O
Notes: (1) Adipic acid concentration controlled at a level to provide 92% S02 removal.
-------
Effect of pH and Adipic Acid Concentration on SO? Removal. As mentioned in
Section 3, the results of factorial tests show that higher adipic acid con-
centration is required at low pH than at high pH to achieve similar S02 removal.
This trend is also evident from the results of Runs 926-1A, 926-1G, 926-1C and
926-1H:
926-1A 926-1G 926-1C 926-1H
Scrubber inlet pH 4.60 4.85 5.05 5.25
Adipic acid cone., ppm 4090 2270 1290 1285
Percent S02 removal 90 91 91 93
Inlet SO? cone., ppm 2650 2250 2150 2150
Percent limestone utilization 97 96 95 92
Thus, the optimum scrubber inlet pH appears to be 5.0 to 5.1 (Run 926-1C) where
adipic acid concentration required is only about 1300 ppm to achieve 91 percent
S02 removal. More importantly, the adipic acid degradation is insignificant
at this pH level (1.17 actual-to-theoretical consumption ratio for Run 926-1C).
Note that S02 removal is more sensitive to pH and inlet. S02 concentrations at
the scrubber inlet pH levels of 4.6 to 4.85 tested because the buffer capacity
of adipic acid is reduced at the lower pH levels.
Limestone Utilization. One of the benefits of the low pH operation is that
very high limestone utilization can be realized. Limestone utilizations
were 95 percent or higher at the scrubber inlet pH of 5.05 or lower and
9.1 minutes residence time in the effluent hold tank.
Sulfite Oxidation and Centrifuge Cake Solids. Another important benefit of
low pH operation is the ease of forced oxidation of sulfite. A natural
oxidation level of about 50 percent was achieved at the scrubber inlet pH of
5.05 or lower, as compared to 15 to 20 percent oxidation at a normal inlet
pH of about 5.5. The resulting centrifuge cake solids concentrations were
almost 10 percentage points higher for the lower pH operation.
SUMMARY
The following is a summary of the test results:
• Apparent degradation of adipic acid is inhibited at low pH, with or
without forced oxidation (see Sections 6 and 7). Without forced
oxidition, the critical pH appears to be about 5.0 at the scrubber
inlet, below which degradation is minimized (actual-to-theoretical
consumption ratio equals 1.0).
• Because of reduced ionization and buffer capacity of acipic acid at low
pH, the required adipic acid concentration is 2 to 3 times higher at a
scrubber inlet pH of 4»6 to 4.85 than at 5.05 to 5.25 to achieve a similar
degree of S02 removal (about 91 percent).
• Operation at low pH and high adipic acid concentration results in lower
total adipic acid consumption than at high pH and low concentration.
251
-------
• The optimum scrubber inlet pH for the venturi/spray tower with a single-
tank configuration appears to be 5.0 to 5.1 with respect to total adipfc
acid consumption, limestone utilization, and the sensitivity of SOo
removal to pH and inlet S02 concentration.
e Other benefits obtained when the scrubber inlet pH was held at 5.05 or
lower include: high limestone utilization (95 percent or higher),
high natural sulfite oxidation (about 50 percent), and the resultant
high centrifuge cake solids (near 70 percent).
252
-------
Section 5
VENTURI SCRUBBER S0? REMOVAL
WITH LIMESTONE/ADIPIC ACID SLURRY
A series of 12 runs (Runs 927-1A through 927-1L) were made using only the
venturi scrubber to determine its maximum S02 removal capability with adipic
acid-enhanced limestone scrubbing.
While.it is recognized that S02 removal with the venturi alone would not meet
the S02 emission standard for high-sulfur coal, even with very high concentra-
tions of adipic acid, scrubbing with the venturi alone could be attractive
economically for low-sulfur coal applications where only 70 percent S02
removal is required.
A single tank was used without forced oxidation for all tests. The flow
configuration for these tests is the same as that shown in Figure 3-1,
except the slurry flow to the spray tower (Pumps G-101 and G-204) was turned
off.
The slurry flow to the venturi was held constant at 600 gpm for all runs.
Variables investigated were adipic acid concentration, gas rate (or venturi
liquid-to-gas ratio at a constant slurry flow rate), venturi pressure drop,
and inlet pH. Operating conditions common for all runs were:
Fly ash loading: High (3-6 grains/dry scf)
Effluent hold tank level: 8.5 ft
Effluent hold tank residence time: 33.4 minutes
Slurry solids concentration: 15 percent
Solids dewatering equipment: Clarifier and centrifuge
DISCUSSION OF TEST RESULTS
Table 5-1 summarizes the major test conditions and the run-average test
results.
Effect of Adipic Acid Concentration. Runs 927-1A, 927-1D, and 927-1E were
an operated at a gas rate of 35,000 acfm (@ 300°F), a liquid-to-gas ratio
of 21 gal/mcf, a venturi Inlet pH of 5.1, and at a pressure drop of about
8.3 Inches H20. Average S02 removal increased from 34 to 41 and 65 percent
when the adipic acid concentration was raised from 815 to 1335 and 3985 ppm,
respectively. Hourly S02 removal data for these three runs are plotted in
Figure 5rl. It appears that the S02 removal levels off at about 65 percent,
suggesting that the overall rate of S02 absorption may have been limited by
the gas-phase mass transfer above 3500 ppm adipic add.
Effect of Llquid-to-Gas Ratio. During Runs 927-1B, 927-1C, and 927-1G, the
Tiquld-to-gas ratio was increased to 37 gal/mcf. Average S02 removal increased
only marginally to 39, 47, and 68 percent, respectively. For these runs,
venturi pressure drop was 6 inches H20.
253
-------
Table 5-1
RESULTS OF VENTURI LIMESTONE/ADIPIC ACID TESTS USING A SINGLE TANK
WITHOUT FORCED OXIDATION
X
Major Test Conditions
Fly ash loading
Gas rate, acftn 0 300T
Veoturl liquor rate, gpm
Percent solids redrculat^d (controlled)
EHT residence time, rain
EHI tank level, ft
Venturi inlet Ifouor pH (controlled!
Adipic acid concentration, ppra
Venturi pressure drop, Inches HUG
Run-Average Results
Start-of-run date
Onstreaoi hours
Percent S02 removal
Inlet S02 concentration, ppm
Adipic acid concentration, ppm
Scrubber percent solids reclrculated
Scrubber inlet pH
Sulfite concentration in Inlet liquor, ppm
SO, Hdke-per-pass, nwle/1
Linestone utilization, %
Sulfite oxidation, I
Inlet liquor gjpsura saturation, I
Centrifuge cake solids, wtt
Hist eliminator restriction, %
Ventur! pressure drop. Inches H?0
927-1A
High
35,000
600
15
33.4
8.5
5.1
700
6
2/JO/30
43
3«
19QO
B15
14.6
5.05
365
8.35
92
36
115
66
8.2
927-18
Hlqh
20,000
600
15
33. 4
8.5
5.1
700
F
2/22/80
24
39
2470
705
14.5
5.15
365
6.80
92
34
115
67
5.9
927-1J
Hiqh
27.500
600
15
33.4
8.5
5.1
700
z/22/80
24
33
2390
795
14.6
5.10
110
7.65
91
29
135
68
5.9
927-1C
Hi oh
20,000
600
15
33.4
8.5
5.1
1300
6
2/24/80
48
47
2595
1360
14.0
5.15
330
8.65
91
32
120
66
6.0
927-10
High
35,000
600
15
33.4
8.5
5.1
1300
6
2/26/80
11
41
2445
1335
16.3
5.15
255
12.4
83
32
120
68
8.4
927-1E
High
35.000
600
15
33.4
8.5
5.1
4000
6
2/27/80
69
65
2360
3985
16.4
5.10
285
19.0
85
28
125
61
8.3
927-1F
Hi in
27,500
600
15
33.4
8.5
5.1
4000
6
3/1/80
26
67("
2255
3990
15.3
5.10
460
14.7
85
23
130
63
5.9
927-1G
Hi ah
20.000
600
15
33.4
8.5
5.1
4000
6
3/1/80
13
68
2790
4030
14.3
5.05
235
13.4
85
22
145
63
6.0
927-1H
Hi oh
27,500
600
15
33.4
8.5
5.1
4000
9
3/2/80
13
69
3030
4005
15.0
5.10
485
20.3
R8
20
125
63 '
8.7
927-11
High
27,500
600
15
33.4
8.5
5.1
4000
12
3/2/80
21
65
2945
4015
15.3
5.10
555
18.6
91
32
63
11.1
927-1K
Hinh
27,500
600
15
33.4
8.5
4.R
4000
9
3/4/f?0
32
59
2245
4050
14.8
4.85
010
12.9
93
26
140
65
R.8
927-1L
Hinh
20,000
600
15
33.4
8.5
«.8
4000
6
3/5/80
16
62
2100
4340
13.6
4.80
1010
9.20
90
30
125
68
6.2
Note: (1) SOp removal dropped to 60? when inlet SC? concentration increased to ?ft70 ppn under replicate conditions.
-------
ro
tn
en
100
90 -
80 -
~ 70 -
UJ
£
O
ca
£S
t/1
O
a:
40 -
30
20 -
gP
°
0800
°°
00 O
o o
AVERAGE OPERATING CONDITIONS
INLET S02
INLET LIQUOR pH
PERCENT SOLIDS
PRESSURE DROP
FLUE GAS RATE
LIQUID-TO-GAS RATIO
RESIDENCE TIME
1960 - 2540 ppm
5.0 - 5.2
15.4
8.3 inches
35,000 acfm
21 gal/mcf
33 minutes
HoO
10
1000
Figure 5-1
2000 3000
AD I PIC ACID CONCENTRATION, ppm
4000
Percent S02 Removal vs. Adipic Acid Concentration
During Limestone Runs 927-i/;, ID and IE
5000
-------
Although S02 removal was below 70 percent with high inlet S02 concentra-
tion, the venturi-only mode of operation with limestone/adipic acid slurry
may be viable for low-sulfur coal applications where inlet S02 concentrations
are less than 1000 ppm.
Effect of Venturi Pressure Drop. During Runs 927-1F, 927-1H, and 927-11, the
venturi pressure drop was varied at 5.9, 8.7, and 11.1 inches H>>0, respectively,
For these runs, adipic acid concentration was maintained at 4000 ppm, liquid-
to-gas ratio was controlled at 27 gal/mcf, and the inlet pH was controlled at
5.1. SOo removal was 60 percent at 5.9 inches HoO pressure drop, and appeared
to level off at 65 to 69 percent at 8.7 and 11.1 inches H20.
Effect of Venturi Inlet pH. Run 927-1K was made under the same conditions as
Run 927-1H, except for the scrubber inlet pH. S02 removal increased signi-
ficantly from 59 percent at 2245 ppm inlet S02 concentration and at 4.85
inlet pH to 69 percent at 3030 ppm inlet S02 concentration and at 5.10 inlet
pH. Similar sensitivity of S0? removal to pH can be observed by comparing
Runs 927-1G and 927-1L.
SUMMARY
Based on the test results, the following conclusions can be made:
• At a liquid-to-gas ratio of 21 gal/mcf, a venturi inlet pH of 5.1,
and a venturi pressure drop of 8.3 inches H20, S02 removal appears
to level off at 65 percent above 3500 ppm adipic acid. (S02 removals
greater than 65 percent may be possible at pH higher than 5.1.)
• Increasing the liquid-to-gas ratio to 37 gal/mcf (with a somewhat
reduced pressure drop of 6 inches H90) improves SOo removal mar-
ginally. L 2
• With low-sulfur coals producing less than 1000 ppm inlet S0? concen-
tration, 70 percent S02 removal should be acnievable at 5.1 inlet
pH, 4000 ppm adipic acid, 6 to 8 inches H?0 pressure drop, and 21-
37 gal/mcf liquid-to-gas ratio.
e.
S02 removal is sensitive to inlet pH (4.8 to 5.1) and adipic acid
concentration (700 to 3500 ppm), but is insensitive to liquid-to-
gas ratio (21 to 37 gal/mcf) and venturi pressure drop (6 to 11
inches H20).
256
-------
Section 6
LIMESTONE/ADIPIC ACID TESTING ON THE VENTURI/SPRAY TOWER
WITH ONE TANK AND FORCED OXIDATION
Following the venturi-only testing, the venturi/spray tower system was modified
to allow testing in a single-tank forced-oxidation mode. Seven runs were made,
including four runs with only the venturi.
Although sulfite oxidation of 99 percent or higher was achieved for the runs
with forced oxidation, limestone blinding was encountered as evidenced by
poor limestone utilization. The long (50 ft) crossover line which routed the
venturi and spray tower effluent slurries to the oxidation tank apparently
behaved as an effective plug-flow reactor in which calcium sulfite precipitated
preferentially on the alkaline limestone particles in the effluent slurry
deficient in calcium sulfite solid crystal seeds.
SYSTEM DESCRIPTION
Figure 6-1 is a schematic flow diagram of the venturi/spray tower system using
a single tank (D-208) in which compressed air is injected through a 3-inch
diameter open-ended pipe ell. The venturi and spray tower effluent slurries
are routed to the oxidation tank via a 16-inch diameter crossover line about
50 ft long. This crossover line is operated full (490 gallons) of slurry
because nearly its entire length is below the oxidation tank liquid level.
The line acts as a plug-flow reactor as previously mentioned. It is emphasized
that this setup is necessitated by the limited availability of space and
is unique to the Shawnee Test Facility*
A severe cavitation problem in the slurry recirculation pumps during initial
startup was solved by installing a baffle near the pump suction nozzles and by
moving the air injection point higher, to between the two agitator turbines.
Both turbines propel the slurry downward. Figure 6-2 shows the arrangement of
the modified oxidation tank.
DISCUSSION OF TEST RESULTS
Tables 6-1 and 6-2 summarize the results of the single-tank forced oxidation
tests with both the venturi and the spray tower in operation, and with the venturi
alone, respectively. The initial test plan called for variations of the scrubber
inlet pH and adipic acid concentration, to observe the effects on adipic acid
consumption under forced oxidation conditions (to compare with the results
presented in Section 4 without forced oxidation). However, the original test
objectives were modified in favor of a more thorough study of the limestone
blinding phenomenon when it was encountered.
257
-------
FLUE GAS
IVJ
£
LIMESTONE SLURRY
ADIPIC ACID
MIST
ELIMINATOR
VENTURI
FLUE GAS
MAKEUP WATER
SPRAY TOWER
CROSSOVER
(NOT
USED)
COMPRESSED
AIR
LINE (~50 FT.)
EFFLUENT HOLD TANK
D-101
\i \<
D
CLARIFIED LIQUOR
FROM SOLIDS DEWATERING SYSTEM
BLEED TO SOLIDS
DEWATERING SYSTEM
OXIDATION TANK
D-208
G-1Q5
G-1W
Figure 6-1. Flow Diagram for the Venturi/Spray Tower System
With One Tank and Forced Oxidation
-------
NEW BAFFLE
BAFFLE SUPPORT
OVERHEAD VIEW OF TANK BOTTOM
TANK-MALI
PIPE
SUPPORT
TANK WALL
NEW
BAFFLE
SIDE VIEW
SIDE VIEW
Figure 6-2. Arrangement of Modified Venturi/Spray Tower
Oxidation Tank (D-208)
259
-------
Table 6-1
RESULTS OF VENTURI/SPRAY TOWER LIMESTONE/ADIPIC ACID TESTS
WITH ONE TANK AND FORCED OXIDATION
Ma.1or Test Conditions
Fly ash loading
Gas rate, acfm 9 300°F
Venturl Hquor rate, gpm
Spray tower Hquor rate, gpm
Percent solids redrculated (controlled)
Oxidation tank residence time, m1n
Oxidation tank level, ft
Scrubber Inlet pH (controlled)
Ad1p1c add concentration, ppm
A1r rate to oxldlzer, scfm
Venturl pressure drop, Inches H20
Run-Average Results
Start-of-run date
Onstrem hours
Percent SO, removal
Inlet S02 Concentration, ppm
Ad1p1c add concentration, ppm
Ad1p1c add consumption ratio (actual/theor.)
Actual adlplc add consumption, Ibs/ton
SOj absorbed
Percent solids reclrculated
Scrubber Inlet pH
Sulflte concentration 1n Inlet liquor, ppm
SO, make-per-pass, mmole/1
Limestone utilization, %
Sulflte oxidation, %
Gypsum saturation 1n Inlet liquor, %
Centrifuge cake sol Ids, wt*
Air stolchlometry, atom 0/mole S02 abs.
M1st eliminator restriction, %
914-1A
High
35,000
600
1600
15
2.9
17
4.6
4000
200
9
3/13/80
106
91.6
19SO
4040
3.41
64.1
15.9
4.60
1250
6.1
46
98.7
145
79
1.9
914-1B
High
35,000
600.
1600
15
2.9
17
5.1
4000
2CO/300
9
3/19/80
11
(1)
9H-1C
H1ah
35,000
600
1600
15
2 A
.y
17
4.6
4000
0
9
4/7/80
47
92.6
1955
4225
1.59
43.3
14.9
4.66
862
6.1
93
32
120
65
0
3
(1) No steady state was established due to severe limestone blinding.
Table 6-2
RESULTS OF VENTURI LIMESTONE/ADIPIC ACID TESTS
WITH ONE TANK AND FORCED OXIDATION
Major Test Conditions
Fly ash loading
Gas rate, acfm 9 300"F
Venturi liquor rate, gpm
Percent solids redrculated (controlled)
Oxidation tank residence time, m1n
Oxidation tank level, ft
Venturl inlet Hquor pH (controlled)
Venturl Inlet Hquor limestone
stolchlometry (controlled)
Adipic add concentration, ppm
A1r rate to oxldizer, scfm
Venturl pressure drop, Inches H20
Run-Average Results
Start-of-run date
Onstream hours
Percent S02 removal
Inlet S02 concentration, ppm
Adipic add concentration, ppm
Adlpic add consumption ratio (actual/theor.)
Actual adipic acid consumption, Ibs/ton
S02 absorbed
Percent solids redrculated
Scrubber Inlet pH
Sulflte concentration in Inlet Hquor, ppm
S02 make-per-pass, mmole/1
Limestone utilization, %
Sulflte oxidation, %
Gypsum saturation in inlet Hquor, %
Centrifuge cake solids, wtt
Air stoichlometry, atom 0/mole S02 abs.
Mist eliminator restriction, *
927-1M
High
30,000
600
15
10.6
17
5.0
—
4000
300
9
3/21/80
115
71.5
2260
4170
2.19
32.0
15.0
5.05
• 75
17.1.
50
99.4
105
78
3.8
"
927-1 N
High
20,000
600
15
10.6
17
5.0
..
4000
300
91)
3/26/80
13
77.4
2030
3960
3.0
50.8
16.1
5.15
44
11.1
35
99.2
105
79
5.9
~
927-10
High
20,000
600
15
10.6
17
--
1.2
4000
300
9n
3/30/80
75
67.4
2070
4130
1.93
28.6
15.0
4.55
28
9.9
85
99.2
110
78
6.6
""
927-1 P
High
30,000
600
15
10.6
17
5.0
__
4000
0
9
4/2/80
108
69.6
2225
3960
2.26
62.5
14.9
5.07
349
16.4
54
23
125
63
0
(1) Actual pressure drop was about 7 Inches H20 because:of a problem with the adjustable plug
mechanism and low gas flow rate.
260
-------
Initial Tests. In Run 914-1A, a total slurry flow rate of 2200 gpm resulted
fn Z.9 minutes residence time in the oxidation tank (Table 6-1), 98.7 percent
sulfite oxidation in the solids, high inlet liquor sulfite concentration
(1250 ppm), and poor limestone utilization of 46 percent despite a low scrubber
inlet pH of 4.6.
To reduce the high inlet liquor sulfite concentration, Run 914-1B was first
run at higher pH (5.1 vs. 4.6) and then at Increased oxidation intensity (air
rate 300 scfm vs. 200 scfm). However, no indication of increased limestone
utilization was noted and the run was terminated.
Venturi-Only Test. The low oxidation tank residence time of 2.9 minutes was
increased to 10.6 minutes during Runs 927-1M, 927-1N, and 927-10 (Table 6-2)
by operation of the venturi only (600 gpm). This necessarily raised the S02
make-per-pass to 17.1 m-moles/liter (Run 927-1M) which was reduced to 11.1
m-moles/liter in Run 927-1N. The limestone utilization was still low and a run
at a controlled limestone stoichiometry of 1.2 (Run 927-10) confirmed that lime-
stone blinding was occuring in the crossover line because the scrubber inlet pH
of 4.55 was lower than expected. This line is in effect a 50 second residence time
plug-flow reactor to which is fed slurry depleted in calcium sulfite seed crystals
and in which a favorable environment is provided for the liquor sulfite to
precipitate on limestone particles before reaching the oxidation tank.
Base Case Tests Without Forced Oxidation. Run 927-1P was made under the same
conditions as Run 927-1M except that the air to the oxidizer was shut off to
provide a base case run without forced oxidation. Limestone utilization remained
poor (54 percent) due to the combined effect of continued high S02 make-per-pass
(16.4 m-moles/liter) and long residence time (near 50 seconds) in the crossover
line.
Run 914-1C was made under the same conditions as Run 914-1A except without
forced oxidation. With an S02 make-per-pass of only 6.1 m-moles/liter and 13
seconds residence time in the crossover line, combined with sufficient calcium
sulfite solid crystal seeds (32 percent oxidation), limestone utilization
improved to 93 percent.
Effect of pH and Limestone Utilization on Adi pic Acid Consumption. Section 4
mentioned that essentially no degradation of acipic acid occurs at a scrubber
inlet pH below 5.0 when oxidation is not forced. In these tests adipic acid
degradation appeared to increase with forced oxidation. In addition, it was
observed that poor limestone utilization increases the degradation. These
observations are more clearly seen in the following table:
927-1M 927-1N 927-10
Venturi inlet pH 5.05 5.15 4.55
Percent limestone utilization 50 35 85
Adipic acid consumption ratio 2.19 3.0 1.93
(Actual/Theoretical)
Percent unaccounted loss,of 54.3 66.7 48.2
adipic acid
Actual adipic acid consumption, 32.0 50.8 28.6
Ibs/ton S02 absorbed
Unaccounted adipic acid loss, 17.4 33.9 13.8
Ibs/ton S02 absorbed
261
-------
SUMMARY
The following is a summary of the test results and findings:
ft Good sulfite oxidation of 99 percent or higher was achieved in
the solids despite poor limestone utilization.
» Limestone blinding occurred in the 50 ft long crossover line which
transfers the venturi and spray tower effluent slurries to the oxidation
tank and which behaved as an effective plug-flow reactor for calcium
sulfite precipitation. This peculiarity in flow configuration is
unique to the Shawnee Test Facility.
« Limestone blinding caused by the long crossover line and high SO^ make-
per-pass could not be prevented by increasing the oxidation intensity
in the oxidation tank to reduce the sulfite concentration in the scrubber
inlet liquor, even at SO? make-per-pass values as low as 6.1 m-moles/
liter, and was compounded by depletion of calcium sulfite seed crystals
in the scrubber effluent.
• Actual-to-theoretical adipic acid consumption ratio and total actual
adipic acid requirement (Ibs per ton S02 absorbed) increase with forced
oxidation, increasing pH, and decreasing limestone utilization.
262
-------
Section 7
LIMESTONE/ADIPIC ACID TESTING ON THE VENTURI/SPRAY
TOWER WITH TWO TANKS AND FORCED OXIDATION
Operation with two tanks in series, with forced oxidation in the first tank
and limestone added to the second tank, has several advantages over the single-
tank operation with forced oxidation:
• Low pH (scrubber-effluent pH) in the first tank (oxidation tank)
promotes sulfite oxidation.
• The possibility of limestone blinding by calcium sulfite is
decreased because fresh limestone is added after the oxidation
tank.
• Limestone utilization is increased with two tanks in series which
approximate a plug-flow reactor for limestone dissolution.
• Extra residence time for calcium sulfate crystallization is provided
by the second tank.
• The second tank provides air-free suction for the slurry recircu-
lation and bleed pumps, thus avoiding pump cavitation.
Earlier test results from the TCA system with limestone/adipic acid and forced
oxidation have shown two-tank operation to be superior to the single-tank
mode (Reference 3). Eight runs (Runs 916-1A through 916-1H) were made to
confirm this conclusion on the venturi/spray tower system using two tanks in
series. A schematic flow diagram is shown in Figure 7-1. Air is injected
into the first tank (D-208) while limestone and adipic acid are added to the
second tank (D-101). The detailed arrangement of the oxidation tank (8 ft
diameter) is shown in Figure 6-2.
DISCUSSION OF TEST RESULTS
Table 7-1 summarizes the results of the eight runs made with two tanks in series,
including one run (Run 916-1H) without forced oxidation. In general, good
S02 removal was achieved with excellent oxidation of the solids for all the
forced oxidation tests. However, as in the tests with forced oxidation
using a single tank (Section 6), calcium sulfite blinding of limestone in
the crossover line continued to reduce the limestone utilization below
the level normally expected with two-tank operation. This remained true
despite the efforts to increase limestone utilization by either increasing the
oxidation intensity or lowering the S02 make-per-pass.
263
-------
FLUE GAS
FLUE GAS
cn
LIMESTONE SLURRY
ADIPICACID
REHEAT
COMPRESSED
AIR
G-104 EFFLUENT HOLD TANK
~D-101
MAKEUP WATER
CLARIFIED LIQUOR
FROM SOLIDS
DEWATERING SYSTEM
BLEED TO
SOLIDS
DEWATERING
SYSTEM
Figure 7-1. Flo» Diagram for the Venturi/Spray Tower System
Wi Ui "iwo Tanks and FoT^d- Oxidation
-------
Table 7-1
RESULTS OF VENTURI/SPRAY TOWER LIMESTONE/ADIPIC ACID TESTS
WITH TWO TANKS AND FORCED OXIDATION
Major Test Conditions
Fly ash loading
Flue gas rate, acfm 1? 300°F
Venturi liquor rate, gpm
Spray tower liquor rate, gpm
Percent solids reclrculated (controlled)
EHT res. time (mini/tank level (ft)
Ox1
—
7.5-7.7
916-10
High
30,000
600
1200
15
11.1/8.5
3.8/18
5.1
1500
1.5
6
5/6/80
163
92.4
2220
1490
—
--
15.0
5.12
5.03
72
99.8
105
81
1E9l>
84(3)
—
7.3
916-1E
High
20,000
600
1600
15
9.1/8.5
3.1/18
5.1
1500
1.5
7
5/13/80
145
98.0
1880
1540
4.89
8.91
14.Q
5.13
5.00
86
P9.5
105
60
90m
88(3)
—
3.6
916-1F
High
30,000
600
1200
15
11.1/8.5
3.8/18
5.4
1500
1.5
6
5/19/80
71
89.2
2260
1510
2.30
7.75
15.9
5.14
5.06
46
99.4
145
583
#«
7.1
916-16
High
30,000
600
1200
15
11.1/8.5
3.8/18
5.4
1500
2.5
6
5/22/80
61
93.7
2150
1550
3.32
8.76
15.6
5.33
5.24
61
99.6
120
33
?46
85(3)
7.1
916-1H
High
30,000
600
1200
15
11.1/8.5
3.8/1R
5.1
1500
0
6
5/24/80
P4
85.5
2500
1440
2.03
11.4
15.6
5.12
4.77
96
50.4
120
317
°59»>
1
7.6
(1) Venturi operated with plug wide open for all rufis except for Run 916-1A where pressure drop was
controlled at 9 inches H20.
(2) System operated with clarifier only.
(3) Drun filter used in place of centrifuge.
-------
Forced Oxidation Testing. During the testing covered by Runs 916-1A through
916-1G, several measures were taken to eliminate or minimize the effect of the
crossover line. Operating parameters explored included:
• Liquid-to-gas ratios in the spray tower of 17.8 to 100 gal/mcf
« S02 make-per-pass of 3.6 to 12.0 m-moles/liter
• Adi pic acid concentrations of 1490-4015 ppm
e Scrubber inlet pH of 4.77 to 5.33
• Air stoichiometry to the oxidizer of 1.5 to 2.5 atoms 0/mole S02
absorbed
However, the overriding tendency of the crossover line to act as a plug-
flow reactor, as described in Section 6, could not be eliminated.
Base Case Test Without Forced Oxidation. Run 916-1H was made under the same
conditions as Run 916-ID except that the air to the oxidizer was turned off.
Significantly, the limestone utilization improved to 96 percent because suffi-
cient calcium sulfite solid seeds were available (50.4 percent oxidation) and
blinding of limestone by calcium sulfite in the crossover line was eliminated.
SUMMARY
The following is a summary of the test results:
• Good S02 removal and excellent sulfite oxidation (99.4 to 99.8 percent)
were achieved with the two-tank forced oxidation system.
• Limestone utilization for the two-tank operation was higher than for
single-tank operation (Section 6) but below that expected with two-tank
operation without limestone blinding.
• As in the single-tank operation with forced oxidation (Section 6),
limestone blinding caused by the crossover line and high S02 make-
per-pass cannot be eliminated by increasing the oxidation intensity
to reduce sulfite concentration in the scrubber inlet liquor.
• Reducing the S02 make-per-pass (Run 916-1E), and hence the scrubber
effluent sulfite concentration, improved limestone utilization but
not to the expected 1evel.
266
-------
Section 8
BLEED STREAM OXIDATION OF LIMESTONE/ADIPIC ACID
SLURRY FROM THE VENTURI/SPRAY TOWER SYSTEM
In April 1979, prior to this reporting period, five bleed stream oxidation
tests were made on the venturi/spray tower system using limestone slurry with
1500 ppm of adipic acid (Reference 3). At that time, good sulfite oxidation
of 99 percent was achieved when the slurry pH in the oxidation tank was kept
below about 6.0 by recycling 60 gpm of oxidation tank slurry back to the ef-
fluent hold tank. Satisfactory oxidation (95 percent) was also obtained
without the recycle, but at the high oxidation tank residence time of about
7.5 hours for the bleed stream.
Recent tests with adipic acid additive have demonstrated several advantages of
operating at low pH and high adipic acid concentration (see Section 4). There-
fore, three tests (Runs 915-1A, 915-1B, and 915-1C) were conducted in April
1980 to see if operating at reduced pH was conducive to bleed stream oxidation.
The flow diagram of the bleed stream oxidation tests on the venturi/spray tower
system is shown in Figure 8-1. The same oxidation tank used in one-tank and
two-tank in-loop forced oxidation (Sections 6 and 7) was used in these three
tests. The detailed arrangement of the oxidation tank is shown in Figure 6-2.
DISCUSSION OF TEST RESULTS
The'results of bleed stream oxidation tests at low pH are given in Table 8-1.
All tests achieved better than 95 percent S02 removal at 4.8 to 5.1 scrubber
inlet pH and about 4000 ppm adipic acid. Average limestone utilizations were
88 to 91 percent.
Good sulfite oxidation of 98 percent was achieved only in Run 915-1C when the
scrubber inlet pH was controlled at 4.8 with an air stoichiometry of 1.80 atoms
oxygen/mole SO? absorbed. Oxidation was only about 70 percent at 5.0 scrubber
inlet pH and 1.55 air stoichiometry (Run 915-1A), or at 5.1 scrubber inlet. pH
and 2.10 air stoichiometry (Run 915-1B).
The oxidation tank pH was 5.4, 5.7, and 4.8 for Runs 915-1A, 915-1B, and 915-.C,
respectively, as compared with 5.5 to 5.6 for runs made earlier in April 1979 when
good oxidation was achieved at 1.50 to 1.85 air stoichiometry. The lower oxida-
tion efficiency for the recent tests may be attributed to the poor oxidizer
arrangement shown in Figure 6-2.
As has been observed previously, adipic acid degradation increased with pH
during these runs. For Runs 915-1C, 915-1A, and 915-1B, under similar lime-
stone utilizations, the actual-to-theoretical adipic acid consumption ratios
were 1.26, 3.33, and 5.20, respectively, when the scrubber inlet pH increased
from 4.8 to 5.0 and to 5.1, and the oxidation tank pH increased concurrently
from 4.8 to 5.4 and to 5.7. Actual adipic acid consumption increased from 15,4
to 40.1 and to 44.5 Ibs/ton S02 absorbed, respectively.
267
-------
REHEAT
G.1M EFFLUENT HOLD TANK
D-101
MAKEUP WATER
CLARIFIED LIQUOR
FROM. SOLIDS
DEWATERING SYSTEM
BLEED TO
SOLIDS
DEWATERING
SYSTEM
G-105
Figure 8-1. Flow Diagram for Bleed Stream Oxidation in the
Ventur-1/Spray Tower System
-------
Table 8-1
RESULTS OF VENTURI/SPRAY TOWER LIMESTONE/ADIPIC ACID TESTS
WITH BLEED STREAM OXIDATION
fro
cr>
(£>
Major Test Conditions
Fly ash loading
Flue gas rate, acfm P 300°F
Venturi liquor rate, gpm
Spray tower liquor rate, gpm
Percent solids recirculated (controlled)
EHT Res. time (min)/tank level (ft)
Oxid. Tk. Res. time (mini/tank level (ft)
Scrubber inlet pH (controlled)
Adipic acid concentration, ppm
Air rate to oxidizer, scfm
Venturi pressure drop, inches 1^0
Run-Average Results
Start-of-run date
Onstream hours
Percent S02 removal
Inlet SOg concentration, ppm
Adipic acid concentration, ppm
Adipic acid consumption ratio, (actual/theor.)
Actual adipic acid consumption, Ibs/ton Sf^ absorbed
Percent solids recirculated
Scrubber inlet pH
Oxidation tank pH
Limestone utilization, %
Sulfite oxidation in oxidation tank, %
Sulfite oxidation in scrubber inlet, %
Gypsum sat'n. in oxidation tank, %
Gypsum sat'n. in scrubber inlet, %
Oxidation tank liquor SO? concentration, ppm
Air stoich., Ib atoms 0/1 h mole S02 absorbed
Centrifuge cake solids, wti
Mist eliminator restriction, %
915-1A
High
35,000
600
1600
15
9.1/8.5
-m
5.1
4000
200
9
4/10/80
98
97.6
2340
38^0
3.33
40.1
15.2
4.99
5.40
91
69
26
105
120
115
1.55
70
915-1B
Hiah
35,000
600
1600
15
9.1/8.5
-/17
5.1
4000
300
9
4/14/80
24
98.0
2550
4045
5.20
44.5
15.5
5.09
5.70
90
73
25
105
115
95
2.10
79
91 5- 1C
High
35,000
600
1600
15
9.1/8.5
-/17
4.8
4000
200
9
4/15/80
127
96.0
2030
4140
1.26
15.4
15.6
4.82
4.80
88
98
54
100
105
140
1.80
79
3
-------
Previous Shawnee data indicated that the dewaten'ng properties of slurries
from bleed stream oxidation are better than those of unoxidized slurries
but inferior to those from in-loop forced oxidation. For Run 915-1C, with
98 percent sulfite oxidation and 4140 ppm adipic acid, the initial settling
rate of solids averaged only 0.3 cm/min, somewhat better than the 0.2 cm/min
settling rate for unoxidized slurry (see Section 10). For the bleed stream
oxidation runs made in April 1979, the average settling rate was much higher
at 0.8 cm/min for slurries with good oxidation (95 percent or higher) and
with lower 1500 ppm adipic acid concentration. These values for bleed stream
oxidation are in the lower range of 0.3 to 1.6 cm/min reported in Table 10-1
for all the oxidized limestone slurry with adipic acid.
SUMMARY
At a scrubber inlet pH of 4.8 and about 4000 ppm adipic acid concentration,
98 percent oxidation of sulfite was achieved in the bleed stream oxidation tank
(4.8 pH) with an air stoichiometry of 1.8 atoms oxygen/mole SO? absorbed.
The S02 removal was 96 percent at 2030 ppm inlet S02 concentration and the
limestone utilization was 88 percent. The actual-to-theoretical adipic acid
consumption ratio was 1.26 and the actual adipic acid consumption was 15.4
Ibs/ton S02 absorbed (8.7 Ibs/ton limestone fed).
270
-------
Section 9
LIMESTONE BLINDING BY CALCIUM SULFITE
Blinding of limestone as evidenced by low limestone utilization has been
encountered during limestone tests with and without adipic acid enhancement.
The limestone blinding is most common under in-loop forced oxidation condi-
tions, where the recirculated slurry is deficient in solid calcium sulfite
crystal seeds and the calcium sulfite in the liquor preferentially preci-
pitates on, and blinds, the alkaline limestone particles. This section describes
system behavior during limestone blinding, the conditions leading to it, and
recommended solutions for eliminating or avoiding limestone blinding.
SYSTEM BEHAVIOR DURING LIMESTONE BLINDING
Limestone blinding in a scrubber system is normally characterized by the
following phenomena:
• Severe drop in slurry pH
• Very insensitive pH response to limestone addition at low pH
• Poor limestone utilization
• High sulfite concentration in the liquor
The first indication of limestone blinding is a precipitous drop in the pH
of the recirculating slurry for no apparent reason. In order to control system
pH, the operator normally begins to increase the limestone feed rate, leading
to poor limestone utilization. Limestone utilization as low as 20 to 25
percent has been observed at Shawnee. While the pH response to the limestone
feed rate is normally more sensitive at a low pH range of 4.5 to 5.5 (less
limestone buffer) than at a high pH range of 5.5 to 6.5 (more limestone
buffer), the response is typically sluggish even at low pH when limestone
blinding occurs.
CONDITIONS LEADING TO LIMESTONE BLINDING
The necessary conditions for blinding to occur are:
• Slurry solids deficient in calcium sulfite crystal seeds
• High sulfite concentration and/or supersaturation in the slurry
liquid
271
-------
The slurry solids deficient in calcium sulfite crystal seeds (i.e., high
gypsum content) can be a result of forced oxidation or high natural oxidation,
Experience at Shawnee indicates that limestone blinding does not occur at
sulfite oxidation levels in solids below approximately 85 percent under most
of the operating conditions.
When the slurry solids contain an insufficient amount of calcium sulfite
crystals, the saturated sulfite in the liquor tends to precipitate pre-
ferentially on alkaline solid particles such as limestone, because the
solubility of calcium sulfite is a strong function of pH and decreases with
increasing pH. Thus, even if the bulk liquor is not supersaturated with
sulfite, as may be the case with low bulk liquor pH, supersaturation and
precipitation could occur in the high pH region in the vicinity of the lime-
stone particles, causing blinding.
High sulfite concentration or supersaturation can be caused by:
« Insufficient oxidation intensity (affecting both scrubber inlet and
outlet)
• High S02 make-per-pass (affecting scrubber outlet)
The use of additives, such as adipic acid, enhances the S02 removal and
increases the S02 make-per-pass, thus increasing the potential for limestone
blinding.
RECOMMENDED SOLUTIONS
Operating Considerations. Limestone blinding by calcium sulfite is the result
of calcium sulfite-deficient slurry solids (high gypsum content) and high cal-
cium sulfite supersaturation (or high sulfite concentration) in the liquor. The
latter can be caused by insufficient oxidation intensity, high S02 make-per-pass
or both.
Therefore, any measures that can reduce these effects will reduce the chance of
limestone blinding. Better oxidation can be obtained by:
o Increasing the air stoichiometry
» Increasing the oxidation tank level to provide a longer air bubble
residence time
« Increasing the oxidation tank agitation
This would reduce the sulfite saturation and concentration at the scrubber
inlet.
The level of sulfite in the scrubber effluent liquor can be reduced by reducing
the inlet liquor sulfite as above or by reducing the S02 make-per-pass.
Lower S02 make-per-pass can be obtained by lowering the flue gas throughput,
increasing the slurry flow rate, or both.
272
-------
Design Considerations. If an in-loop forced oxdiation system with a single
tank is desired, then provision should be made:
• To provide an adequate oxidation intensity to minimize sulfite
saturation at the scrubber inlet
• To reduce S02 make-per-pass (outlet sulfite concentration)
A better solution appears to be the use of two tanks in series, which provide
several advantages over the single-tank mode listed in Section 7.
Limestone blinding in the long scrubber effluent line (Sections 6 and 7), which
acts as a plug-flow reactor, is unique at Shawnee. In full-scale design, the
scrubber effluent piping should be as short as possible to minimize the poten-
tial for limestone blinding.
273
-------
Section 10
DEWATERING CHARACTERISTICS OF
ADIPIC ACID-ENHANCED LIME/LIMESTONE SLURRIES
Cylinder settling tests and vacuum funnel filtration tests are routinely
conducted in the Shawnee Laboratory to monitor the settling and dewatering
characteristics of slurry solids.
In the previous reports (References 2 and 3), a comparison of the results
of these monitoring tests from July 1978 through October 1979 was presented
for lime and limestone slurry with and without adipic acid addition. It
was found that adipic acid has an insignificant effect on the quality of
solids (settling rate and filterability), except that the settling rate of
oxidized limestone slurry may be retarded.
Table 10-1 has been updated to include additional data, obtained from October
1979 through May 1980, for limestone slurry with adipic acid addition both
with and without forced oxidation.
The updated data show a higher average initial settling rate of 0.9 cm/min
(0.3 to 1.6 cm/min range) for oxidized limestone slurry with adipic acid,
compared to the 0.6 cm/min (0.3 to 0.9 cm/min range) previously reported for
the same type of slurry (Reference 3). The average initial settling rate for
oxidized limestone slurry without adipic acid remains the same at 1.1 cm/min.
The settling rate of unoxidized limestone slurry again shows essentially
no effect from adipic acid. The average settling rate is 0.2 cm/min with
or without adipic acid.
274
-------
Table 10-1
COMPARISON OF SHAWNEE WASTE SLURRY DEWATERING
CHARACTERISTICS WITH AND WITHOUT ADIPIC ACID ADDITION
ro
-j
en
Alkali
Limestone' '
Limestone. .
Limestone131
Limestone
Lime
Lime
Lime
Lime
Fly Ash .
Loading11'
High
High
High
High
High
Hioh
High
Hi(jh
Forced
Oxidation
Yes
Yes
No
No
Yes
Yes
No
No
Adi Die
Acid(z)
Yes
No
Yes
No
Yes
No
Yes
No
Initial
Solids
Cone., wt %
IF
IS
15
15
8
8
8
ft
Initial
Rate,
Avg.
0.9
1.1
0.2
0.2
1.7
_
1.5
1.5
Settlino
cm/mi n.
Ranne
0.3-1.6
n.fi-1.4
0.1-0.4
0.1-0.5
1.5-1.9
_
1.2-1.9
O.P-2.2
Ultimate
Solids
Avg.
70
73
50
51
55
_
51
49
Settled
, wt %
Ranpe
57-P3
62-flfi
37-69
41-67
50-60
_
44-63
43-57
Funnel Test Cake
Solids,
Avg.
70
74
56
57
61
-
62
56
wt %
Range
59-77
65-8P
48-73
48-66
54-69
-
48-73
50-62
(1) Slurries with hioh fly ash loadinn contain about 40 percent fly ash in solids.
(2) Adipic acid concentration range is 300 to 4500 ppro.
(3) Data have been updated to include test results from October 1979 throunh May
-------
Section 11
ECONOMICS OF ADIPIC ACID-ENHANCED
LIMESTONE SCRUBBING
The economics of adipic acid-enhanced limestone scrubbing has been projected
for forced-oxidation systems designed to achieve an average of 90 percent
S02 removal from high sulfur flue gas. The results indicate that, for the
cases studied, both capital and operating costs are approximately 4 to 6
percent lower for adipic acid-enhanced limestone systems than for a lime-
stone system without additive. The major savings are in the reduced lime-
stone requirement and the associated grinding equipment. Additional 1 to 2
percent savings in operating cost result from the reduced quantity of waste
solids that need to be disposed of in the adipic acid-enhanced limestone system.
The operating conditions for four study cases, including a limestone case
with MgO additive, were prepared by Bechtel and are presented in Table 11-1.
The capital investment and revenue requirement were calculated by the
Economics Evaluation Section of TVA's Energy Design and Operations Branch
using a TVA/Bechtel Design-Economics Computer Program (Reference 8). The
results are listed in Table 11-2. The evaluations are based on a 500-MW
scrubbing facility incorporating forced oxidation and operating on flue gas
from a boiler burning 4 wt % sulfur coal. The capital investment and revenue
requirement in Table 11-2 include the dewatering equipment (thickener and filter)
but exclude the waste sludge (filter cake) disposal area.
The cases evaluated are:
Case 1 - A limestone base case without additive operated at rela-
tively high limestone stoichiometry and liquid-to-gas
ratio to achieve 90 percent S02 removal. It should be
noted that long-term reliability with this mode of opera-
tion has not been demonstrated at Shawnee.
Case 2 - A limestone case with MgO addition. Oxidation of the
scrubber bleed stream was chosen because in-loop
oxidation is incompatible with magnesium-enhanced
scrubbing. As in Case 1, long-term reliability has not
been demonstrated at Shawnee for this mode of operation.
Case 3 - A limestone case with adipic acid addition operated at
high pH. Although only 800 ppm of adipic acid is required
to obtain 90 percent S02 removal, degradation of adipic
acid at high pH requires about five times the theoretical
adipic acid addition rate.
Case 4 - A limestone case with adipic acid addition operated at low
pH. For this case, 2000 ppm adipic acid is required.
However, the low pH operation requires only 1.4 times the
theoretical adipic acid addition rate and 1.05 limestone
stoichiometry.
276
-------
Table 11-1
CONDITIONS FOR ECONOMIC ANALYSIS OF ADIPIC ACID-
ENHANCED LIMESTONE SCRUBBING WITH FORCED OXIDATION
Capacity:
Coal:
Scrubber:
SO* Removal Efficiency:
Superficial Gas Velocity:
Number of Trains:
Solids Dewatering:
Onstream Factor:
Effluent Hold Tank Residence Time;
Oxidation Tank Residence Time:
Oxidation Tank Level:
Air Sparger Pressure Drop:
Oxidation Tank Agitator Hp:
Solid Sulfite Oxidation:
Air Stoichiometry:
Number of Tanks:
500 MW
4 wt% sulfur
TCA with 3 beds, 4 grids, and 5 inches
of static height of spheres per bed
90%
12.5 ft/sec
5, including 1 spare train
To 80 wt% solids by thickener and rotary
drum vacuum filter
5500 hours/year
5 minutes
5 minutes
18 ft
5 psi
0.002 Brake Hp/gal
99%
1.7 Ib-atpms 0/1b mole S02 absorbed
2 (effluent hold tank and oxidation tank)
Case No.
Alkali-
Additive
Additive Concentration, ppm
Additive Rate, Ib/hr
L/G, gal/mcf
Limestone Stoichiometry,
moles Ca/mole SOo absorbed
TCA Inlet pH '
Mode of Oxidation(d)
1
Limestone
-
_
-
58
1.52
5.8
in loop
2
Limestone
MgO
5500 (a)
104
50
1.20
5.4
bleed
stream
3
Limestone
Adi pic
Acid
800 fhl
83.3(b)
50
1.20
5.6
in loop
4
Limestone
Adi pi c '•
Acid
2000,
53.6^ i
50
1.05
4.8
in loop
Notes:
(a)
(b)
(c)
(d)
Effective Mg++ concentration.
Five times theoretical consumption.
1.4 times theoretical consumption.
In-loop oxidation with two tanks uses an oxidation tank
followed in-series by an effluent hold tank where alkali
is added. Bleed stream oxidation uses one effluent hold
tank in the scrubber loop and one bleed stream tank where
air is injected.
277
-------
As shown in Table 11-2, both the total capital investment and the first
year revenue requirement are the lowest for adipic acid-enhanced limestone
scrubbing at low pH (Case 4). The total capital investment is reduced by
4.8 percent and the first year revenue requirement reduced by 5.8 percent
for the limestone/adipic acid/low pH case (Case 4) compared with the con-
ventional limestone case (Case 1). The revenue requirement includes 14.7
percent annual capital charge.
Total capital investment and operating cost for adipic acid-enhanced lime-
stone at high pH (Case 3) are higher than those for limestone/adipic acid
at low pH (Case 4), but are still lower than those for the conventional
limestone (Case 1) or the 1 imestone/MgO case (Case 2). Total capital
investment is lower by 3.9 percent and the first year revenue requirement
is lower by 4.0 percent for Case 3, compared with Case 1.
Note that the capital investment and revenue requirements shown in Table 11-2
are significantly different from those presented by TVA in an earlier session.
This is due to differences in process equipment and operating parameters
itemized in Table 11-3. The most significant factors are coal sulfur content,
scrubber type, superficial gas velocity, L/G, hold tank residence time, and
landfill investment (not included in Bechtel comparison). The operating
parameters in Table 11-1 were selected to represent conditions which have been
tested at Shawnee. If the basic design parameters for the comparison in
Table 11-1 were adjusted to be the same as those used in the earlier TVA
comparison, the same relative results (i.e., limestone scrubbing with
adipic acid addition is slightly more economical than standard limestone
scrubbing) would be obtained.
Table 11-4 illustrates the additional savings that result from adipic
acid addition. Because of the lower pH operation, and thus lower
limestone consumption, the amount of waste solids produced is lower for
limestone/adipic acid cases (Cases 3 and 4) than for a limestone case
(Case 1). Assuming a landfill disposal cost of $10/dry ton, including
14.7 percent annual capital charge, the first year revenue requirements
for the sludge disposal area are 0.97, 0.83, and 0.77 mills/kWh for
Cases 1, 3, and 4, respectively. Thus, the total first year revenue
requirement is 9.34 mills/kWh for Case 4 compared with 10.06 mills/kWh
for Case 1. This is a reduction of 7.2 percent, compared with 5.8 percent
when the sludge disposal cost is not included.
It should be noted that the differences in total capital investments and
operating costs amoung these cases are small. Furthermore, the cost
figures are not meant to be accurate or representative of actual scenarios.
The principal conclusion from these evaluations is that adipic acid addition
does not increase costs but decreases them slightly on the same comparison
basis.
278
-------
Table 11-2
RESULTS OF ECONOMIC ANALYSIS OF ADIPIC ACID-
ENHANCED LIMESTONE SCRUBBING WITH FORCED OXIDATION
Case No. Additive
1
2 MgO
3 Adipic Acid
4 Adipic Acid
Total Capital
Additive Investment^'
Cone., ppm $MM(1982)
87.40
5500 85.26
800 83.97
2000 83.22
$7kW
174.8
170.5
167.9
166.4
First Year
Revenue , w ,
Requirement(bMc)
$MM(1984) r
25.01
24.15
24.01
23.56
1il 1 s/kWh
9.09
8.78
8.73
8.57
Notes: (a) Effective Mg+ concentration.
(b) Does not include waste sludge disposal area.
(c) Includes 14.7% annual capital charge.
Raw Material Costs (1984): Limestone - $8.5/ton
MgO - $460/ton
Adipic Acid - $1200/ton
Table 11-3
LIMESTONE PROCESS COMPARISON
Item
Type scrubber
Superficial gas velocity
fc sulfur in coal (as fired)
Effluent hold tank residence time
Oxidation tank residence time
Air stoichiometry
Landfill investment
Adipic acid
L/C
Limestone stoichiometry
Adipic acid consumption ratio ( actual /theor.)
Bechtel
(Case 3)
TCA
12.5 ft/ sec
4.0
5 min
5 min
1.7
Not included
800 ppm
50
1.2
5
TVA
Spray tower
10 ft/ sec
3.36
12 min
5 min
2.5
Included
1000 ppm
80
1.2
3
279
-------
Table 11-4
REVENUE REQUIREMENT IN
SLUDGE DISPOSAL AREA
Filter Cake,
Case No. dry tons/hr
1 48.7
2 41.6
3 41.6
4 38.3
First Year Revenue
Total Excluding. .
Sludge Disposal ID)
9.09
8.78
8.73
8.57
Rquirement,
Sludge
Disposal10'
0.97
0.83
0.83
0.77
Mills/kWh(a)
Total
10.06
9.61
9.56
9.34
Notes: (a) Includes 14.7% annual capital charge.
Costs are based on 1984 dollars.
(b) From Table 11-2.
(c) Assumes $10/dry ton, including 14.7% annual capital charge.
280
-------
Section 12
SUMMARY OF CURRENT WORK
Important test results of adipic acid-enhanced .limestone scrubbing (with
high fly ash loading) on the venturi/spray tower system from October 1979
through May 1980 are summarized below. The summary of previous work
through October 1979 has been presented in Section 2.
• Factorial tests with spray tower only (venturi plug wide open
with 125 gpm slurry flow) showed that, within the operating
ranges of 15 to 75 gal/mcf spray tower liquid-to-gas ratio, 4.6
to 5.9 scrubber inlet pH, and 600 to 2400 ppm adipic acid con-
centration, each 0.4 unit drop in the scrubber inlet pH requires
a 700 ppm increase in adipic acid concentration to achieve a
similar percent S02 removal.
• Apparent degradation of adipic acid is quenched at low pH.* Without
forced oxidation, essentially no degradation occurs at a scrubber
inlet pH below about 5.0. Both forced oxidation and high limestone
stoichiometry (due to limestone blinding at low pH conditions) con-
tribute to higher adipic acid degradation.
• Operating a scrubber at low pH and high adipic acid concentration can
actually result in lower total adipic acid consumption than operation
at a high pH and low concentration for the same S02 removal.
• The optimum scrubber inlet pH appears to be 5.0 to 5.1 with respect
to adipic acid consumption, limestone utilization, and the sensitivity
of S02 removal to pH and inlet S02 concentration.
• Operation with the venturi alone (slurry flow to the spray tower
turned off) without forced oxidation indicated that the SO? removal
levels off at a maximum value of about 65 percent at 2000 to 3000
ppm inlet SO? concentration, with 3500 to 4500 ppm adipic acid, 21
gal/mcf liquid-to-gas ratio, 5.1 venturi inlet pH, and 8.3 inches
HoO venturi pressure drop. This mode of operation could, however, be
attractive for low-sulfur coal having less than 1000 ppm inlet S02
concentration where only 70 percent S02 removal is required.
• In an in-loop forced oxidation system, or in a system with high natural
oxidation, blinding of alkaline limestone particles by calcium sulfite
could occur because of the deficiency in calcium sulfite seed crystals.
Operation with two tanks, with forced oxidation in the first tank and
limestone added to the second tank, minimizes the potential for lime-
stone blinding.
* Recent laboratory test results at the University of Texas at Austin
(Reference 9) have shown that the adipic acid degradation decreases in
the presence of Mn ion, and also decreases with pH when Mn is present.
The IERL-RTP pilot plant test results (Reference 10) have identified Mn
and Fe ions as possible inhibitors of adipic acid degradation.
281
-------
« A long scrubber slurry-filled effluent pipeline (a flow configuration
which exists at Shawnee on the venturi/spray tower system due to
system constraints) is detrimental in that it could act as an effect
tive plug-flow reactor for calcium sulfite precipitation and increase
the potential for limestone blinding. Limestone blinding in this manner
cannot be totally eliminated by increasing the oxidation intensity
because calcium sulfite precipitates before being oxidized.
• Additional data showed that adipic acid only slightly reduces the
settling rate of oxidized limestone slurry, to 0.9 cm/min (vs. 0.6
cm/min previously reported) from 1.1 cm/min for oxidized limestone
slurry without adipic acid.
• Economic analyses for a TCA system with 90 percent S02 removal from
4 percent sulfur coal show that both capital and operating costs,
excluding the waste solids disposal area, are approximately 4 to 6
percent lower for limestone scrubbing systems with 800 to 2000 ppm
adipic acid than for a limestone system without additive. Additional
savings for limestone systems with adipic acid can be realized in the
waste solids disposal area because of lower solids production rate.
282
-------
Section 13
FUTURE SHAUNEE TEST PLAN
The test program for the Shawnee Test Facility, as presently conceived for
the remainder of 1980 and 1981, is presented below. The major effort will
still be placed on the adipic acid-enhanced limestone scrubbing.
In late-May and the first-half of June 1980, Train 100 was converted from a
venturi followed by a spray tower to a spray tower-only system. In addition,
the spray tower piping and the internal headers were modified in August 1980
to increase the maximum slurry flow rate from 1600 gpm to 2400 gpm. The
following test activities with the spray tower only are either in progress,
planned, or suggested:
t Factorial tests with limestone slurry with or without forced oxida-
tion, and with or without adipic acid addition, to expand the existing
data base and computer models for predicting $62 removal.
• Long-term (500 hours) demonstration tests with the spray tower only
using limestone/adipic acid slurry with and without forced oxidation.
• Tests to develop design criteria for the spray tower internals.
• Tests with packings having low pressure drop, high efficiency, and
low plugging and scaling potential, such as Glitsch Grid packing.
t Tests with other organic acid additives such as dibasic acid, which
is a byproduct of adipic acid manufacture consisting primarily of
adipic, glutaric, and succinic acids.
• Tests with low S02 during the Boiler No. 10 baghouse acceptance
testing.
• Integrated power plant water management testing, such as water reuse
and additive recovery.
• Testing with other alkalis, such as water treatment sludge, partially
calcined limestone, and hydrated dolomitic lime.
The TCA system (Train 200) was restored from a DOWA basic aluminum sulfate
process operating configuration in late-June 1980. The following activities
are either proceeding, planned, or suggested:
t Simulation of the two full-scale TCA units operating with adipic
acid-enhanced limestone at the Southwest Station of the Springfield
City Utilities at Springfield, Missouri, as part of the EPA full-
scale adipic acid demonstration program.
• Automatic limestone feed control testing.
• Testing with sodium thiosulfate as an oxidation inhibitor.
283
-------
• Tests with Glitsch Grid packing in lieu of spheres.
« Tests with other organic acid additives, such as dibasic acid.
« Development of a magnesium or calcium adipate clear liquor
scrubbing process.
» Development of other forced oxidation methods.
9 Tests with low S02 during the Boiler No. 10 baghouse acceptance
testing.
• Investigation of the effects of limestone type and grind on SO?
removal and limestone utilization.
Some of the tests listed above are interchangeable between Train 100 and
Train 200.
284
-------
Section 14
REFERENCES
1. Borgwardt, R. H., "Significant EPA/IERL-RTP Pilot Plant Results," in
Proceedings: Industry Briefing on EPA Lime/Limestone Wet Scrubbing
Test Programs (August 1978), EPA-600/7-79-092 (NTIS PB 296517),
March 1979 (pp.1-9).
2. Head, H.N., et al., "Recent Results from EPA's Lime/Limestone Scrubbing
Programs - Adipic Acid as a Scrubber Additive," in Proceedings: Symposium
on Flue Gas Desulfurization - Las Vegas, Nevada, March 1979; Volume I,
EPA-600/7-79-167a (NTIS PB 80-133168), July 1979 (pp. 342-385).
3. Burbank, D.A., and S.C. Wang, "Test Results on Adipic Acid-Enhanced
Lime/Limestone Scrubbing at the EPA Shawnee Test Facility - Second
Report," in Proceedings: the Fifth Industry Briefing on IERL-RTP
Lime/Limestone Wet Scrubbing Test Programs (December 1979),
EPA-600/9-80-032 (NTIS PB 80-199813), July 1980 (pp 27-113).
4. Rochelle, G.T., "The Effect of Additives on Mass Transfer in CaC03 or
CaO Slurry Scrubbing of S02 from Waste Gases," Ind. Eng. Chem. Fundam.,
Vol. 16, No. 1, pp. 67-75, 1977.
5. Rochelle, G.T., "Process Alternatives for Stack Gas Desulfurization by
Throwaway Scrubbing," in Proceedings of Second Pacific Chemical Engineering
Congress, Vol.1, p. 264, August 1977.
6. Radian Corporation, Further Study of Adipic Acid Degradation in FGD
Scrubbers, draft final report for EPA Contract 68-02-2608, Task 72,
April 18, 1980.
7. Cavanaugh, C.M., Buffer Additives for Flue Gas Desulfurization Processes,
M.S. Thesis, The University of Texas at Austin, December 1978.
8. Stephenson, C.D., and R.L. Torstrick, "The Shawnee Lime-Limestone Computer
Program," in Proceedings: the Fifth Industry Briefing on IERL-RTP Lime/
Limestone Wet Scrubbing Test Programs (December 1979), EPA-600/9-80-032
(NTIS PB 80-199813), July 1980 (pp 167-222).
9. Rochelle, G., Buffer Additives for Stack Gas Desulfurization by CaO/CaCOo
Slurry, September 1980 Monthly Progress Report, EPA Grant R806743, EPA/IERL,
Research Triangle Park, N.C. (R.H. Borgwardt, Project Officer).
10. Borgwardt, R.H., et al., Limestone Scrubbing of SO? at EPA/RTP Pilot Plant,
Progress Report 46, August 1980.
285
-------
Appendix
CONVERTING UNITS OF MEASURE
Environmental Protection Agency policy is to express all measurements
in Agency documents in metric units. In this report, however, to avoid
undue cost or lack of clarity, English units are used throughout.
Conversion factors from English to metric units are given below:
To Convert From
scfm (60°F)
cfm
Op
ft
ft/hr
ft/sec
ft?
frvtons per day
gal/mcf
9Pm
gpm/ftz
gr/scf
in.
in. HoO
Ib
Ib-moles
Ib-moles/hr
Ib-moles/hr ft2
Ib-moles/min
psi
To
nm3/hr (0°C)
nrVhr
°C
m
m/hr
m/sec
m;
FIT/roe trie tons per day
1/m3
1/min
1/min/itr
g/m3
cm
mm Hg
9
g-moles
g-moles/min
g-moles/min/m2
g-moles/sec
kPa
Multiply By
1.61
1.70
(°F-32)/1.8
0.305
0.305
0.305
0.0929
0.102
0.134
3.79
40.8
2.29
2.54
1.87
454
454
7.56
81.4
7.56
6.895
286
-------
COCURRENT SCRUBBER TESTS
SHAWNEE TEST FACILITY
By
S. B. Jackson
Division of Energy Demonstrations and Technology
Office of Power
Tennessee Valley Authority
Muscle Shoals, Alabama
ABSTRACT
Prototype cocurrent limestone scrubber tests were performed at the
Shawnee Test Facility. The initial cocurrent prototype tests consistently
achieved greater than 90% S02 removal while operating with inlet flue gas
S0£ concentrations ranging from 1500 ppm to 3000 ppm. Although the
prototype scrubber tower was reliable, total system reliability was not
achieved during the initial tests at 27 ft/sec superficial scrubber gas
velocity, primarily because of solids deposits in the mist eliminator and
the inline, indirect steam reheater. At a 20 ft/sec superficial gas velocity
and with low fly ash loading in the inlet flue gas there were no signifi-
cant solids deposits in the mist eliminator or reheater. Mist eliminator
operation was reliable during operation with high fly ash loadings and a
20 ft/sec superficial gas velocity, but the inline reheater continued to
plug with slurry solids. During forced-oxidation tests with a single
scrubber hold tank and multiple hold tanks, operating conditions were
identified which consistently removed greater than 90% of the S0£ and
oxidized greater than 95% of the calcium sulfite in the scrubber slurry
to gypsum.
287
-------
COCURRENT SCRUBBER TESTS
SHAWNEE TEST FACILITY
INTRODUCTION
In 1978 the Hydro-Filter scrubber train at the Shawnee Test Facility
was modified to demonstrate the cocurrent scrubber concept. The design
of the modification and original test program plan were based upon results
from pilot cocurrent scrubber tests conducted at the Tennessee Valley
Authority (TVA) Colbert Pilot Plant. The initial equipment modification
and the 12-month test program (August 1978-July 1979) were funded by the
Electric Power Research Institute (EPRI) and implemented by TVA. A second
period of cocurrent tests (August 1979-July 1980) was funded by TVA,
the Environmental Protection Agency (EPA), and the Department of Energy (DOE),
These tests were conducted to demonstrate reliable operating conditions and
limestone cocurrent scrubber operation with forced oxidation.
This paper summarizes the results of the TVA cocurrent scrubber tests.
The highlights of the Colbert pilot plant tests and the EPRI prototype
cocurrent scrubber tests are presented as background for this discussion.
COCURRENT SCRUBBER
Background
The cocurrent scrubber design as illustrated in Figure 1 has several
potential advantages over other commercial FGD scrubber arrangements.
• The equipment configuration is more compatible with most power
plant duct and fan arrangements. The gas enters the scrubber
at a high elevation and leaves near ground level. The entrainment
separator and reheat systems (likely to require the most maintenance)
can be near ground level. Likewise, the induced draft (ID) fans
can be on the ground and the connecting ductwork to the stack can
be shorter and probably less complex.
• The physical arrangement of the cocurrent scrubber causes the gas
to change direction in the base of the unit before it enters the
mist eliminator. Both the change in direction of the gas and the
vertical position of the entrainment separator promote good liquid
separation and drainage. Also, a separate mist eliminator wash
loop may be used, if needed.
288
-------
. ELEVATION
Figure 1. Cocurrent limestone slurry process plan and elevation.
289
-------
• Scrubbing liquid should coalesce into larger droplets before
disengaging from the gas stream near the base of the scrubber
and further facilitate efficient operation of the mist eliminator.
• Flooding of the unit with the associated high pressure drop and
excessive entrainment of scrubbing slurry (even if grids are added
to improve gas-liquid contact) is less likely. Also, during
normal cocurrent operation the gas-side pressure loss is lower
because some liquid-side energy is recovered.
• Higher gas velocities (small scrubbers) are achieved because of
the reduced tendency to flood and because more efficient mist
elimination is likely. Therefore, smaller or fewer scrubber
modules would be required in a full-scale system.
These potential advantages provided incentive for TVA and EPRI to
conduct pilot scrubber studies of the cocurrent scrubber concept with
flue gas from a coal-fired boiler at the TVA Colbert pilot plant.
Representative results from the Colbert limestone cocurrent scrubber
tests are given in Table 1. These results and preliminary economic
studies justified prototype testing of the cocurrent scrubber at Shawnee.
TABLE 1. LIMESTONE COCURRENT SCRUBBER TEST RESULTS
COLBERT PILOT PLANT
Inlet S02 concentration, ppm 2,461
Outlet S02 concentration, ppm 242
Percent S02 removal 90
Scrubber superficial gas velocity, ft/sec 28
L/G, gal/k£t3 69
Limestone stoichiometry, mol Ca/mol inlet S02 1.26
Height of scrubber, ft 30
Number of grids 5
Depth of each grid, in. 9
Scrubber pressure drop, in. ^0 15.4
A flow diagram of the Shawnee cocurrent scrubber train as installed
for the EPRI cocurrent test program is shown in Figure 2. The scrubber
system was designed for operation over a wide range of conditions, which
are summarized in Table 2. Figure 3 is a schematic of the Shawnee
cocurrent scrubber arrangement.
290
-------
FLUE CAS INLET
MIST
ELIMINATOR
I
TO INDUCED
DRAFT FAN
ro
V
INLINE
REHEATER
COCURRENT
SCRUBBER
PRE3ATURATOR PUMP
RIVER WATER
DISPOSAL FILTER CAKE
PUMP RESLURRY
TANK
1
LIMESTONE
SLURRY
PREPARATION
TANK
RIVER WATER
LIMESTONE SLURRY
FEED PUMP
A TOR
PUMP
C5-K3 C»*«3 6—
MIST ELIMINATOR SCRUBBER
-CIRCULATION CIRCULATION
TANK TANK
bCiHUHBLK LJ L_ '
CIRCULATION PUMPS 1 THICKENER
<
11
„,„, „-->. i A ,
CAKE THICKENER
f /- ,. T UNDERFLOW PUMP
BELT OR DRUM!
FILTER 1 J-~Q k
1 ) — t"J! 'I
FILTRATE PUMP
Figure ?.. Cocurrenf scrubber Shawnnc Steam Plant Test Facility - flow diagram.
RECYCLE LIQUOR RECYCLE LIQUOR
SURGE TANK RETURN PUMP
-------
FLUE GAS INLET,
t
10
1
38
\
1
ft.
1
ft.
SLU
INL
2.5
RRY
ET
ft.—
i
4.2 ft.
1
— • -^z.
/\
"A
I! M 1 1
1
J
X
i
*^
J— 40" DIAMETER
- PRESAT
t
4 ft.
6 ft.
4 ft.
« >
4 .
6 ft.
-4-
4 ft.
4 > 4 .s n
/ t
/(p 4.5ft. _^
t-r-t-n «•
^"3ft- ^
BAFFLE PLATE
TO REHEATER
MIST ELIMINATOR
HOUSING
-.0
Figure 3.
SCRUBBER EFFLUENT
Cocurrent scrubber, schematic.
292
-------
TABLE 2. SHAWNEE PROTOTYPE COCURRENT SCRUBBER
MAJOR DESIGN PARAMETERS
Design parameter Range
Scrubber superficial gas velocity, ft/sec 18-31
L/G (at 32 ft/sec gas velocity), gal/kft3 12-100
Scrubber height, ft 25-45
Number of spray headers 1-4
Number of spray nozzles/header 4-8
Scrubber circulation tank retention time
at maximum recirculation rate 6-17
Extensive testing with sodium carbonate, lime, and limestone absorbents
was performed during the EPRI-funded program. Detailed results of this
program were presented at the EPA Fifth Industry Briefing Conference on
Lime/Limestone Wet Scrubbing. Representative results with each of these
absorbents are shown in Table 3.
Highlights of the lime/limestone tests included in the EPRI test
program follow:
• The gas/liquor contact efficiency of the cocurrent open spray tower
(no grids) was inadequate for S02 removal greater than 85%.
• Installation of grids in the tower provided effective gas and liquor
contact, which increased the S02 removal efficiency to greater than
90%.
• Slurry distribution through a single spray header at the top of the
scrubber provided higher S02 removal than slurry distribution throughout
the tower with multiple spray headers.
• Scrubber operating conditions that strongly affected S02 removal were
gas residence time, recirculated slurry rate, and absorbent stoichiometry,
Gas residence time had the strongest effect. For example, at 27 ft/sec
gas velocity and 1.0 mol Ca/mol inlet S02, the recirculated slurry
rates required to maintain 85% S02 removal with a 25, 35, and 45. foot
scrubber were 2370, 1780, and 1175 gpm respectively.
• S02 removal by lime absorbent was slightly lower than that achieved
in the Colbert pilot lime tests. At similar operating conditions, the
Shawnee scrubber achieved 93% S02 removal while the Colbert scrubber
achieved 96% SO^ removal.
293
-------
TABLE 3. SHAWNEE PROTOTYPE COCURRENT SCRUBBER TEST RESULTS
EPRI TEST PROGRAM.
Absorbents
Major test conditions Sodium, carbonate Lime Limestone
Scrubber physical configuration
Height, ft 25 45 35
Spray header location(s)
(ft from scrubber sump) 25, 15 45 35
Flue gas
Flow rate, aft3/min at 300°F 25,000 25,000 25,000
Scrubber superficial gas
velocity, ft/sec 26.7 26.7 26.7
Slurry recirculation rate, gpm 1,440 1,200 2,400
L/G, gal/kft3 72 60 120
Open tower or grid tower Open tower Grid Grid
Scrubber pressure drop, inches ti^O 2-3 3 3
Absorbent stoichiometry
Mols Na/mol S02 absorbed 2.24
Mols Ca/mol inlet S02 - 1.1 1.3
Inlet S02 concentration, ppm 2)400 2,800 2,400
S02 removal efficiency, % '92 93 90
294
-------
• During a 350-hour limestone scrubbing test, the prototype cocurrent
scrubber consistently averaged 90% S02 removal efficiency for each
successive 24-hour period. Major scrubber operating conditions for
this test were 2500 ppm inlet S02 concentration, 27 ft/sec superficial
gas velocity, L/G equal to 90 gal/kft3, 8.3 inches H20 scrubber pressure
drop and limestone stoichiometry equal to 1.3 mol Ca/mol inlet S02.
• The scrubber tower with grids operated without scaling and plugging
of the tower internals; however, a soot blower was required at the
tower inlet to remove solids deposits at the wet/dry interface.
• The total scrubber train did not operate reliably at 27 ft/sec
scrubber superficial gas velocity because slurry solids deposits
plugged the mist eliminator and reheater.
TVA Cocurrent Scrubber Test Program
After completion of the EPRI program, TVA continued limestone cocurrent
scrubber tests with emphasis upon improvement of the mist eliminator and
reheater reliability and tests with forced oxidation. The TVA tests were
conducted from August 1979 to July 1980. EPA and DOE provided funds for
the test program after June 1. The test program was separated into two
primary test blocks:
• Mist eliminator reliability tests
• Forced-oxidation tests
Mist Eliminator Reliability Tests. These tests were performed to
determine operating conditions that would provide reliable mist eliminator
and reheater operation. Velocity profile determinations upstream and
downstream of the mist eliminators indicated that the gas distribution
at the mist eliminator entrance was very poor (see Figure 4). The plans
for the reliability tests were based primarily upon the hypothesis that
improvement of the gas distribution at the outlet of the scrubber and the
entrance to the mist eliminators would improve the mist eliminator relia-
bility and efficiency. Improved mist eliminator efficiency would in turn
improve the reheater reliability.
Scrubber operating conditions and scrubber equipment were revised
during this test block as follows:
• The scrubber superficial gas velocity was lowered from 27 to 20
ft/sec.
• The flue gas source was changed from upstream of the boiler ESP to
downstream of the ESP.
• The solids concentration in the recirculated scrubber slurry was
reduced from 15% to 10%.
295
-------
rss
Average = 23
Second-Stage
Mist Eliminator
Four-Pass
First-Stage
Mist Eliminator
Three-Pass
NOTE: The calculated velocity is 18.5 ft/sec In the mist eliminator duct
and 22 ft/sec in the vertical duct.
Average = 28
7VX
I '/ y 3x
\HX68yf29;^22
18
A
s
iO
Figure A. Cocurrent scrubber mist eliminator velocity profile (ft/sec). Air only at a
scrubber superficial gas velocity of 27 ft/sec.
-------
• Presaturator spray nozzles and a soot blower for solids cleaning
were installed at the inlet of the scrubber.
• A 3-pass, open-vane mist eliminator was installed in the outlet
duct of the scrubber sump and turning vanes, were installed in
the 90-degree turn immediately upstream of the mist eliminator.
All of these revisions, except the presaturator installation, were made
to decrease the amount of solid and liquid entrainment leaving the scrubber
(entering the mist eliminator) and to improve the entrainment removal
efficiency of the mist eliminator. Operating conditions and results of
this test series are summarized in Table 4. All tests, except test LS-4100C,
were performed with low fly ash loading in the flue gas. Tests LS-5000C,
5001C, and 5002C were performed with a 20 ft/sec scrubber gas velocity and
tests LS-5010C and 4100C were performed with a 27 ft/sec gas velocity. The
presaturator sprays were installed before test LS-5001C. The open-vane
mist eliminator and turning vanes were installed before test LS-4100C.
Highlights of these tests are summarized as follows:
• Reduction of the scrubber superficial gas velocity to 20 ft/sec
and the fly ash loading to 'U).! gr/sft^ essentially eliminated
solids deposits in the mist eliminator and reheater.
• Maximum localized gas velocities in the mist eliminator were
reduced from 50-60 ft/sec to 35-40 ft/sec when the scrubber gas
velocity was reduced from 27 to 20 ft/sec. (The mist eliminator
vendor claims high entrainment removal efficiency at 35 to 40 ft/sec.)
. • Solids deposited in the mist eliminator and reheater while operating
at 27 ft/sec scrubber gas velocity and low fly ash loadings.
• Solids deposits at the wet/dry interface were controlled by periodically
cleaning the area around the presaturator spray nozzles with a soot
blower.
• The open-vane mist eliminator and the turning vanes that were installed
at the outlet of the scrubber did not impr.ove the mist eliminator
and reheater reliability while operating the scrubber at a 27 ft/sec
gas velocity.
Further testing at 27 ft/sec scrubber gas velocity was postponed to
permit forced-oxidation tests at 20 ft/sec* Future tests at a higher gas
velocity (V30 ft/sec) will be conducted after the scrubber outlet duct is
modified to provide better gas distribution in the mist eliminator.
Limestone Cocurrent Scrubber Tests with Forced Oxidation (Single Tank
Mode). Limestone scrubbing tests with forced oxidation began in October
1979. The first series of tests was performed with air sparging and limestone
297
-------
ID
00
TABLE 4. HIGHLIGHTS OF TVA LIMESTONE MIST ELIMINATOR RELIABILITY TESTS
COCURRENT SCRUBBER
Test Nuinber
On-stream time, hr
Scrubber operating conditions
Physical configuration
Height, ft
Number of grids3 >D
Header location0
Pressure drop, in. ^0
Flue gas
Flow rate (inlet), aft3/min (300°F)
Superficial velocity at 125°F, ft/sec
Inlet S02 concentration, ppm
Slurrye»f
Recirculation rate, gpm
L/G, gal/kft3
Laboratory results
Recirculated slurry
Solids concentration, wt %
pH
Total dissolved solids, ppm
Solids
Stolchlometry, mols Ca/mol inlet 803
Thickener underflow
Solids concentration, wt %
Filter cake
Solids concentration, wt %
S02 removal efficiency, %
LS-5000C
319
38
6
B
3.5-A.O
18, 750
20.0
1,340-2,120
1,400
93
8.3-11.5
6.0-6.4
4,100-9,900
. 1- 12- 1.46:
18.8-27.4
54.9-77.2
88-94
LS-5001C
470
38
6
B
3.3-4.3
18,750
20.0
1,320-2,320
1,495
100
8.2-11.5
5.93-6.43
2,440-11,295
1,16-1.52
16.1-23.0
54.7-78.7
92-94
LS-5002C
348
38
6
B
3.3-4.3
18,750
20.0
1,840-2,440
1,495
100
8.7-11.4
5.83-6.40
7,618-17,708
1.18-1.52
18.1-27.4
56.5-67.3
92-94
LS-5010C
238
38
6
B
6.6-7.9
25,000
27.0
1,360-2,680
1,895
95
8.7-11.4
5.85-6.29
9,423-15,144
1.16-1.47
13.6-21.9
54.7-62.5
93-97
LS-4100C
170
38
6
B
7.0-8.5
25,000d
27.0
2,000-2,680
1,895
95
14.8-16.2
5.84-6.17
5,500-9,545
1.23-1.44
24.9-30.0
54.3-59.6
92-95
a. Grid elevations: 402, 398, 392, 388, 382, and 378 ft.
b. Depth of grids was 3-3/4 inches/elevation.
c. Header elevation: B, 407 ft.
d. Flue gas with full loading of fly ash. All other tests used flue gas with low loading of fly ash.
e. Includes presaturator slurry.
f. Scrubber recirculation tank slurry depth was 6 ft for all tests except test LS-4100C which was 16.
5 ft.
-------
addition in a single scrubber circulation tank as shown in the flow diagram
in Figure 5. The scrubber internal arrangement and the range of major
operating conditions for this first series are summarized in Table 5.
TABLE 5. MAJOR COCURRENT SCRUBBER OPERATING CONDITIONS
LIMESTONE SCRUBBING TESTS WITH FORCED OXIDATION
SINGLE TANK MODE
Flue gas flow rate (inlet) 18,750 a.ft3/min at 300°F
Scrubber gas velocity 20 ft/sec at 125°F
Recirculated slurry rate, gpm 1500
L/G, gal/kft3 100
Recirculated slurry solids
concentration, % 10-15
Scrubber height, ft 38
Limestone stoichiometry,
mols Ca/mol inlet S02 1.3
Grids 6
Depth of each grid, inches 3-3/4
Air stoichiometry,
Ib-atoms 0/lb-mol S02 absorbed 2.0-4.0
The objective of this test block was to define operating conditions
that would simultaneously achieve 90% SOo removal and oxidize greater than
90% of the calcium sulfite in the recirculated slurry to calcium sulfate
dihydrate. Initially screening tests were made to study the effect of
oxidation air rate on the scrubber S02 removal efficiency and the degree
of oxidation. The operating conditions and results of these tests are
presented in Table 6. The S02 removal efficiency and percent oxidation
during several of these tests are briefly summarized below:
Test number LS- 5120C 5121C 5122C 5140C 5130C
Limestone stoichiometry,
mols Ca/mol inlet S0£ 1.3 1.3 1.3 1.1 1.1
Oxidation air stoichiometry,
Ib-atoms 0/lb-mol S02 absorbed 1.6-2.4 2.6-3.2 2.5 2.3-3.1 2.8-3.7
S02 removal efficiency, % 93-97 87-96 88-96 72-85 88-94
Slurry solids oxidation, % 60-81 95-99 66-87 99-100 99-100
299
-------
FLUE GAS INLET
TO INDUCED A-.
DRAFT FAN\J,
co
o
o
PRESATURATOR PUMP
'
LIMESTONE
SLURRY
PREPARATION
TANK
RIVER WATER
WIST. ELIMINATOR
CIRCULATION PUMP
RIVER WATER
+ '
o-
*
0
f
ST .ELIMINATOR
CIRCULATION
^ t
J *
sc
CIR
1
*
'o
-------
TABLE 6. HIGHLIGHTS OF TVA LIMESTONE COCURRENT SCRUBBER TESTS WITH FORCED OXIDATION - SINGLE TANK MODE
Operating Period
Test Number
Ons Cream time, hr
Scrubber operating conditions
Physical configuration
Height, ft
Number of stages'*
Number of grids per stagec
Header location''
Pressure drop, in. H-0
Flue gas ,
Flow rate (inlet), af t /mln at 300°F
Superficial velocity ^ ft/sec at 125°F
Inlet S0_ concentration, ppm
S lurry f
Reclrculation rate, gpm
L/G, gal/kaft3
Scrubber circulation tank conditions
Physical configuration
Slurry depth, ft
Agitator speed, rpm
Oxidation air
Rate, sft3/min
Stoichiometry, Ib-atoms 0/lb-mo] SO^ absorbed
Laboratory results
Recirculated slurry
Solids concentration, vt %
PH
Liquor
Total dissolved solids, ppm
Sulfite concentration, ppm
Oxidation, %S
Solids
Stoichiometry, mols Ca/mol inlet SQ2
Oxidation, %8
Thickener underflow
Solids concentration, wt %
Filter cake
Solids concentration, wt %
S0_ removal efficiency, %
Oct. 24-26
LS-5100C*
47
38
6
3
B
3.2-3.6
18,750e
20.0
1,360-1,680
1,495
100
10
45
100
2.1-3.0
4.8-8.0
5.9-6.1
5,072-10,107
63-145
92-96
1.20-1.38
41-48
12.0-13.7
-
92-94
Dec. 13-23
LS-5110C"
250
38
6
3
B
3.4-4.3
18,750
20.0
2,140-3,300
1,495
100
16.5
68
250
2.8-4.0
13.6-17.0
5.6-6.1
5,724-15,618
16-152
92-99
1.18-1.50
98-100
18.4-27.3
54.9-89.0
90-94
Dec. 27-Jan. 2
LS-5120(y
136
38
6
3
B
3.5-4.0
18.750
20.0
2,120-3,200
1,495
100
16.5
68
150
1.6-2.4
13. 1-1*. 7
5.9-6.4
5,368-7,253
9-163
90-99
1.31-1.47
60-81
17.4-30.0
65.4-80.3
93-97
Jan. 2-10
LS-5121C"
192
38
6
3
B
3.7-4.3
18, 750
20.0
2.260-3,320
1.495
100
16.5
68
200
2.6-3.2
13.4-17.0
5.5-6.2
5,428-8,652
45-253
85-98
1.18-1.35
95-99
22.6-39.0
63.9-85.3
87-96
Jan. 10-15
LS-5130C8
118
38
6
3
B
3.8-4.5
18,750
20.0
1,960-2,800
1,695
113
16.5
68
200
2.8-3.7
13.0-17.1
5.6-5.9
7,242-9,642
7-76
95-100
0.98-1.12
99-100
16.3-33.5
78.8-89.7
88-94
Jan. 15-21
LS-5140C3
140
38
6
3
B
3.6-4.3
18,750
20.0
1.900-2,600
1,695
113
16.5
68
1-50
2.3-3.1
13.5-16.6
5.0-5.5
7,488-11,278
416-1,402
49-94
1.08-1.34
99-100
19.5-27.3
76.0-92.4
72-85
Jan. 23-31
LS-5122C
178
38
6
3
B
3.3-4.0
18,750
20.0
1,820-2,720
1,495
100
16.5
68
_
2.4-2.7
14.0-16.1
5.7-6.3
5,500-8,398
34-136
91-98
1.11-1.43
66-87
18.2-31.7
67.8-79.2
88-96
Turning vanes and open-vane mist eliminator were installed in the 90-degree elbow upstream of the mist eliminator housing.
b. Grid elevations: 402, 398, 392, 388, 382, and 378 ft.
c. Depth of each grid was 1-1/4 inches.
d. Header elevation: B, 407 ft.
e. Flue gas with low-loading of fly ash. All other tests were with flue gas with full-loading of fly ash.
f. Includes 85-95 gpm for presaturator; remaining slurry distributed through six 3-inch, 60-degree spray angle Bete nozzles (ST128TTCN) located at
B-header.
g. Percent of total sulfur present aa sulfate.
(continued)
-------
TABLE 6 (continued)
Operating Period
Test Number
Onstream time, hr
Scrubber operating conditions
Physical configuration
Height, ft
Number of stages3
Number of grids per stage1"
Header- location0
Pressure drop, in- H^O
Flue gasd ,
Flow rate (inlet), aft /min at 300°F
Superficial velocity, ft/sec at 125°F
Inlet S0~ concentration, ppm
Slurry6
Recirculation rate, gpm
L/G, gal/kaft3
Scrubber circulation tank conditions
Physical configuration
Slurry depth, ft
Agitator speed, rpm
Oxidation air
Rate, sft3/min
Stoichiometry , Ib-atoms 0/lb-mol 862 absorbed
Laboratory results
Recirculated slurry
Solids concentration, wt %
pH
Liquor
Total dissolved solids, ppm
Sulfite concentration, ppm
Oxidation, %f
Solids
Stoichiometry, tnols Ca/mol inlet S02
Oxidation, Xf
Thickener underflow
Solids concentration, wt %
Filter cake
Solids concentration, wt %
SO removal efficiency, %
Jan. 31-Feb. 5
LS-5123C
119
38
6
3
B
3.4-3.9
18,750
20.0
2,200-2,560
1,495
100
16.5
68
_
2.8-3.1
14.3-16.7
6.0-6.2
5,714-10,531
•36-100
93-97
1.30-1.38
82-92
20.8-39.4
53.7-78.6
91-95
Feb. 7-28
LS-5124C
449
38
6
3
B
3.5-4.6
18,750
20.0
2,100-3,000
1,495
100
16.5
68
_
3.2-3.6
12.3-17.1
5.5-6.1
4,235-9,935
18-470
80-99
1.18-1.51
92-99
19.2-31.1
73.7-87.7
87-93
Feb. 29-Mar. 11
LS-5150C
228
38
6
6
B
6.9-8.2
18,750
20.0
2,140-3,120
1,495
100
16.5
68
-
2.3-2.7
13.1-16.7
5.0-6,1
4,523-10,401
145-1,031
69-92
0.98-1.32
85-96
15.6-39.4
70.1-87.5
73-90
Mar. 11-13
LS-5151C
64
38
6
6
B
8.8-9.4
18,750
20.0
1,900-2,480
1,885
126
16.5
68
_
2.4-2.7
13.9-16.4
5.0-5.3
9,524-11,144
14-510
75-99
0.98-1.18
99
23.2-33.7
78.4-80.7
89-93
Mar. 14-21
LS-5160C
118
38
6
6
B
8.3-9.1
18,750
20.0
1,880-2,540
1,885
126
16.5
68
-
3.3-3.9
13.9-17.4
5.1-5.7
8,800-12,890
7-118
90-99
0.99-1.21
98-100
16.4-31.0
85.8-89.0
90-95
Mar. 21-26
LS-5161C
115
38
6
6
B
8.4-9.1
18,750
20.0
2,160-2,600
1,885
126
16.5
68
_
2.9-3.2
13.9-16.3
5.4-5,8
8,725-12,852
18-45
96-99
1.03-1.26
99-100
22.3-34.2
78.1-90.6
94-96
a. Grid elevations: 402, 398, 392, 388, 382, and 378 ft.
b. Depth of each grid was 1-1/4 inches,
c. Header elevation: B, 407 ft.
d. Flue gas with full-loading of fly ash.
e. Includes 85-95 gpm for presaturator; remaining slurry distributed through six 3-inch, 60-degree spray angle Bete nozzles (ST128FFCN) located at
B-header.
f. Percent of total sulfur present as sulfate.
-------
These tests demonstrated the degree of difficulty associated with simul-
taneously achieving greater than 90% S02 removal and greater than 90%
oxidation. In several of the tests, particularly LS-5140C, limestone
blinding apparently occurred and the S02 removal efficiency of the scrubber
decreased. This phenomenon has been explained by the hypothesis that
high liquor sulfite concentration (1400 ppm 803 in test LS-5140C) and low
solid-phase sulfite concentration combine to promote precipitation of
calcium sulfite on the surface of the limestone. If this occurs, the
limestone dissolution rate, the overall rate-controlling mechanism of
this process, decreases and, consequently, the S02 removal efficiency
drops.
Laboratory analyses (scanning electron microscope examination) clearly
indicate that the limestone in the slurry from these tests is not physically
blinded. Limestone addition to the scrubber circulation tank during these
tests did not, however, provide the expected increase in slurry pH and
S02 removal efficiency. Further laboratory investigation is needed in an
attempt to fully explain this process problem.
The next series of tests (LS-5155, -5200, -5210, and -5201) were
performed with the air stoichiometry controlled at 3.0 Ib-atoms 0/lb-mol S0£
absorbed. The limestone stoichiometry was varied from 1.1 to 1.3 mol
Ca/mol inlet S02- Also, after test LS-5155, the depth of the.grids in the
•scrubber was increased from 3-3/4 to 7-1/2 inches. Operation with thicker
grids permitted a decrease in L/G to 85 gal/kft3 without reducing the S02
removal below 90%. The major operating conditions and results of this test
series are summarized in Table 7. The performance of the scrubber is briefly
summarized below:
Test number LS- 5155 5200 5210 5201
.Mode Single-tank Single-tank Single-tank Single-tank
Inlet S02 concen-
tration, ppm 2,060-2,600 1,960-2,600 1,800-2,640 1,560-2,480
S0£ removal
efficiency, % 97 90-93 91-94 90-94.
Limestone stoichiometry-
mols Ca/mol inlet S02 1.11-1.16 1.08-1.18 1.20-1.49 1.06-l.lj
Air stoichiometry, Ib-atoms
0/,lb-mol S02 absorbed 2.9-3.2 2.9-3.2 2.8-3.2 2.7-3.2
Liquor to gas ratio,
gal/kft3 126 87 85 85
Scrubber pressure drop,
inches H20 8.4-10.7 6;3-7.0 6.4-7.3 6.2-7.0
Oxidation, %
1 Liquor phase 87-98 88-98 90-100 86-100
Solid phase 99-100 98-100 99-100 99-100
All of these tests consistently achieved greater than 90% S02 removal and
greater than 97% oxidation of the slurry solids. Apparently the higher air
stoichiometry prevented limestone blinding during these tests.
303
-------
TABLE 7. HIGHLIGHTS OF TVA LIMESTONE TESTS WITH FORCED OXIDATION IN A SINGLE TANK MODE
Operating Period
Test Number
Apr11 3°-May 5 *** 5~May 9 May ')-™sy 16 May "'^ 20
LS-5155 LS-52°° LS-521° LS-5201
Onstream time, hr
Scrubber operating conditions
Physical configuration
Height, ft
Number of stages8
Number of grids per stage"
Header location0
Pressure drop, in. ^0
Flue gas**
Flow rate (inlet) . af t3/min at 300°F
Superficial velocity, ft/sec at 125°F
Inlet SC>2 concentration, pptn
Slurry6
Recirculation rate, gpm
L/G, gal/kaft3
Make-per-pass, milli-g-mol S02 absorbed /liter slurry
Scrubber circulation tank conditions
Physical configuration
Slurry depth, ft
Agitator speed, rptn
Oxidation air
Rate, sft3/min
Stoichiometry, Ib-atoms 0/lb-mol S02 absorbed
Laboratory results
Recirculated slurry
Solids concentration, wt %
pH
Liquor
Total dissolved solids, ppm
Sulfite concentration, ppm
'Oxidation, %f
Solids
Stoichiometry, mols Ca/mol inlet S02
Limestone utilization, %
Oxidation, %f
Thickener underflow
Solids concentration, wt %
Sludge cake
Solids concentration, wt %
S02 removal efficiency, %
a. Grid elevations: 402, 398, 392, 388, 382, and 378 ft.
b. Depth of each grid was 1-1/4 inches.
c. Header elevation: B, 407 ft.
d. Flue gas with full loading of fly ash.
e. Includes 65-100 gpm for presaturator; remaining slurry
nozzles (ST128FFCN) located at B-header.
f. Percent of total sulfur present as sulfate.
7 1
3$
6
8.4-10.7
18,750
20.0
2,060-2,600
1,885
126
4.3-5.1
16.5
68
190-210
2.9-3.2
13.8-16.8
5.2-5.8
8,040-9,953
45-167
87-98
1.11-1.16
99-100
25.5-33.7
81.5-89.3
97
distributed
38
&
g
g
6.3-7.0
18,750
20.0
1,960-2,600 1
1,300
87
5.5-7.1
16.5
68
160-200
2.9-3.2
1
12.1-17.3
5.2-5.6
5,013-13,917 8,
23-158
88-98
1.08-1.18
76.9-83.3-
98-100
15.Z-37.9
77.5-89.7
90-93
through six 3-inch,
38
6
5
B
6.4-7.3
18,750
20.0
,800-2,640
1,270
&5
4.9-7.3
16.5
68
150-200
2.8-3.2
11.2-16.6
5.4-5.9
530-15,366
0-136
90-100
1.20-1.49.
62.5-76.9
99-100
18.7-41.1
79.3-87.7
91-94
, 60-degree
38 .
6
6-
B
6.2-7.Q .
18,750" -
20.0'
1,560-2,480
1,265
85
4.9-7.0
14.5
68
140-190
2.7-3.2
13.3-16.8
5.4-5.8
10,479-15,922
0-158
86-100
1.06-1. i:i
71.4-85.1
99-100
19.6-26.3
76.4-82.0
90-94
spray angle Bete.
304
-------
The percent solids in the gypsum filter cake produced during these
tests varied from 76% to 90%. A typical composition of the solids produced.
is summarized below:
Weight % (dry)
59.5
CaS03-l/2H20 0.1
CaC03 13.4
Fly ash 27.0
The settling rate of the solids varied from about 0.1 cm/min at 65% oxidation
to 1.0-2.5 cm/min near complete oxidation. Figure 6 is a plot of the solids
settling rate versus percent oxidation, which was generated from solid
settling tests performed with oxidized slurry from all of the cocurrent
scrubber forced-oxidation tests.
Limestone Cocurrent Scrubber Tests with Forced-Oxidation (Multiple
Tank Mode). Following the single tank mode tests, the scrubber circulation
equipment was modified to permit operation with multiple hold tanks. A
flow diagram of the scrubber train in the multiple tank mode is shown in.
Figure 7. Potential advantages of this mode of operation are:
• A lower pH for the oxidation reaction
• Liquor compositions less likely to promote limestone blinding
• Improved limestone utilization
In this operational mode air is sparged into the first tank, which receives
the scrubber effluent. The lower pH of the effluent should provide improved
oxidation air utilization because calcium sulfite solubility increases as
fhe pH decreases. Addition of limestone in the second tank after the slurry
liquor is oxidized and the liquor sulfite concentration is low should prevent
limestone blinding. The multiple tank arrangement partially simulates plug
flow and should improve the limestone utilization.
Five forced-oxidation tests that were performed in this test series
are summarized in Table 8. Although additional parametric tests should
have been performed, an extended period of operation was required for the
reliability demonstration test, LS-6150, before the test program was discon-
tinued in July.
In tests LS-6100, -6110, and -6120, the oxidation air stoichiometry was
controlled at 3.0, 2.5, and 2.0 Ib-atoms 0/lb-mol S02 absorbed, respectively,
while other process control points remained constant, including the limestone
stoichiometry at 1.1 mpl Ca/mol S02 absorbed. (Test LS-6100B was a repeat of:
305
-------
•H
"a
u
g
oo
5«
3
g
PQ
§
O
oo
0.5 —
70
80
OXIDATION, %
90
100
Figure 6. Limestone scrubbing slurry settling rate versus percent
oxidation.
306
-------
FLUE GAS INLET
TO INDUCED
DRAFT FAN
t*>
O
\
LIMESTONE
SLURRY
PREPARATION
TANK
RIVER WATER
BACK MIX PUMP PRESATURATOR PUMP
THICKENER
UNDERFLOW PUMP
RIVER WATER
TO POND DISPOSAL
DISPOSAL FILTER CAKE
PUMP RESLURRY
TANK
RECYCLE LIQUOR RECYCLE LIQUOR
SURGE TANK RETURN PUMP
Figure 7. Cocurrent scrubber, Sha-vnce Test Facility - flow diagram for forced-oxidation
test <= - lu^l-ciple hold tan" -•
-------
TABLE 8. HIGHLIGHTS OF TVA LIMESTONE COCURRENT SCRUBBER TESTS
WITH FORCED OXIDATION
MULTIPLE TANK MODE
Test number LS-
On-stream time, hr
Limestone stoichiometry,
mols Ca/mol inlet S02
Air stoichiometry, Ib-atoms
0/lb-mol S02 absorbed
Scrubber L/G, gal/akft3
Limestone utilization, %
Scrubber outlet slurry,
pH
Liquor
Sulfite concentration,
ppm
Oxidation, %
Solids
Oxidation, %
Recirculated slurry,
PH
Solids oxidation, %
Liquor oxidation, %
Liquor sulfite concentration.
ppm
S02 removal efficiency, %
6100
143
1.0-1.2
2.8-3.2
85-97
5.2-5.4
160-500
74-92
99.6-100
5.8-6.3
98-100
88-100
t
0-181
89-92
6110
150
1.07-1.10
2.3-2.7
98
83-91
5.3-5.9
68-588
71-96
99.6-100
5.9-6.3
99-100
95-100
0-172
89-93
6120
24
1.04-1.2
2.0
98
77-83
5.3
814
68
99.6
5.6-6.1
99.6
83-97
45-339
89-91
6100B
40
1.07-1.1
2.9-3.1
98
83
5.3-5.4
339-452
79-84
99.5-100
6.0-6.2
99.6-100
98.5-99.4
9-68
89-91
6155
692
1.18-1.42
2,8-3.2
. 98
; 67-77
5.3-5.6
279-598;
72-87
98-100
5.7-6.4
94-100
87-10C
0-139
92-95
308
-------
LS-6100.) Operating conditions for the reliability demonstration were
selected to ensure that the scrubber SC>2 removal efficiency and percent
oxidation were maintained above 90% and 95%, respectively.
Major conclusions and observations from the multiple tank forced-
oxidation .tests and the reliability demonstration include:
• 90% S02 removal efficiency and percent oxidation greater than
98% can be consistently achieved.
• The multiple tank arrangement for these tests does not provide
improved limestone utilization. (There appeared to be a slight
decrease in limestone utilization, compared with single tank
mode with forced oxidation. Additional tests are needed to
determine the cause of this unexpected result.)
• Oxidation air utilization is improved in the multiple tank mode.
Greater than 95% oxidation was achieved in both the liquor and
solid phases of the slurry while operating with 1.1 mol Ca/mol
inlet S02 and 2.5 Ib-atoms 0/lb-mol S02 absorbed. The single tank
mode required 3.0 Ib-atoms 0/lb-mol S02 to achieve these conditions.
• Conditions which promote limestone blinding (high 803 concentration
in the slurry liquor and high percent oxidation of the slurry solids)
did not develop until the oxidation air stoichiometry was reduced
to 2.0. Limestone blinding occurred with the air stoichiometry
controlled at 2.5 during single mode tests.
• The demonstration confirmed the long-term reliability and efficiency
of the scrubber tower. The S02 removal efficiency was 92% to 95%
and the percent oxidation was 98% to 100% during, this 700-hour period.
• Although there were no significant solids deposits in the scrubber
tower or mist eliminator, the reheater plugged with slurry solids.
The deposits in the reheater apparently were caused by higher fly
ash loadings in the flue gas (and the resulting higher recirculated
slurry density) than were present in the earlier successful relia-
bility demonstration. Additional tests are needed to define operating
conditions that do not cause plugging of the inline reheater.
Conclusions. Table 9 is a summary of major design criteria for a
cocurrent scrubber system. These criteria apply primarily to the scrubber
area and are based upon the results of tests at Shawnee.
309
-------
TABLE 9. MAJOR DESIGN CRITERIA FOR LIMESTONE
COCURRENT SCRUBBER WITH FORCED OXIDATION
Parameter
Inlet S02 concentration, ppm 2,400
Percent S02 removal 90
Scrubber superficial gas velocity, ft/sec 20
L/G, gal/kft3 98
Limestone stoichiometry, mol Ca/raol inlet S02 1.3
Number of grids 6
Height of each grid, inches 7.5
Scrubber AP, inches H20 7.0
Total system AP, inches ^0 13.0
Scrubber height, ft 38
Grid spacing, ft 5
Oxidation tank residence time, min 5
Hold tank residence time, min 10
Air stoichiometry, mols 02/mol S02 removed 1.5
Oxidation efficiency, % 99
Percent solids in throwaway gypsum sludge 80
310
-------
DOWA PROCESS TESTS
SHAWNEE TEST FACILITY
By
S. B. Jackson
Division of Energy Demonstrations and Technology
Office of Power
Tennessee Valley Authority
Muscle Shoals, Alabama
C. E. Dene
Coal Combustion Systems Division
Electric Power Research Institute
Palo Alto, California
D. B. Smith
Air Correction Division
UOP, Inc.
Des Plaines, Illinois
ABSTRACT
Dowa dual-alkali process tests at the Shawnee Test Facility were
the first application of the Dowa process with flue gas from a coal-fired
boiler. The operating conditions were based on operating experience at
Dowa facilities at smelter plants, sulfuric acid plants, and oil-fired
steam generator plants in Japan.
The initial tests utilized the existing Turbulent Contact Absorber
(TCA) in the Shawnee train 200. The maximum S02 removal efficiency by the
TCA was 85% to 90%. During the TCA testing, problems with gas flow distri-
bution in the absorber were observed. Subsequently, the mobile sphere
packing in the TCA was replaced with rigid packing to improve gas flow
distribution and gas liquid contact. A factorial absorption test series
was conducted using the rigid packing. As a result, operating conditions
which will consistently achieve greater than 90% S02 removal efficiency
were identified.
Performance of the neutralization and gypsum dewatering process steps
was generally satisfactory during the absorption tests.
Extensive reliability tests were not conducted; however, no significant
reliability problems were identified during the factorial absorption tests.
311
-------
DOWA PROCESS TESTS
SHAWNEE TEST FACILITY
INTRODUCTION
The Dowa process is a dual-alkali flue gas desulfurization (FGD)
process which utilizes basic aluminum sulfate solution for SC^ absorption
and limestone for regeneration of the absorbent. The process was developed
by the Dowa Mining Company of Tokyo, Japan, and will be marketed in the
United States by the Air Correction Division of UOP, Inc. The process
is now in commercial operation in Japan at an oil-fired boiler, smelters,
and sulfuric acid plants. The Shawnee prototype Dowa installation is
the first test of the Dowa process with flue gas from a coal-fired
boiler.
Potential advantages of the Dowa process over the conventional
limestone scrubbing process which were justification for the Shawnee
tests are:
• Utilizes clear solution scrubbing versus slurry scrubbing to eliminate
erosion of equipment and slurry solids buildup on mist eliminator
and absorber internals.
• Requires lower limestone stoichiometry.
• Produces a gypsum byproduct which has better dewatering characteristics
than unoxidized limestone scrubbing sludge. The Dowa gypsum may be
used for wallboard production.
Additional requirements of a Dowa system as compared to a conventional
limestone scrubbing system are as follows:
• Includes more equipment and is more complex than a conventional
single-loop scrubbing system (excluding sludge mixing and fixation
equipment when required as a part of a limestone system).
• Uses an absorbing solution pH of approximately 3 compared to 5 to 6
for a limestone scrubbing system. At the lower pH more acid-
resistant materials of construction are required. In areas where
carbon steel or a low alloy steel is used in a limestone system,
316L or 317L stainless steel or epoxy resin-lined carbon steel is
required.
312
-------
The Shawnee Dowa process test program was a jointly funded project
by the Electric Power Research Institute (EPRI) , the Tennessee Valley
Authority (TVA), and UOP, Inc. The final month of tests was funded by the
Environmental Protection Agency (EPA). The primary purpose of the
program was to demonstrate that the process can effectively treat flue
gas from a coal-fired boiler. Shawnee train 200, a TCA scrubber system,
was modified to the Dowa process configuration and an 8-month test
program was conducted. The original program plan included:
1, A 1-month process equipment shakedown and process demonstration at
operating conditions recommended by Dowa and UOP.
2. Factorial tests of the absorption process step.
3. Factorial tests of the neutralization and dewatering process
steps.
Due to problems with process control and major equipment problems which
caused lengthy delays, the neutralization and dewatering tests were
eliminated from the test program. However, the neutralization and
dewatering sections were operated continuously during all of the tests.
Process Chemistry
The overall chemical reactions in each of the major process steps
are:
• Absorption: A12(S04)3-A1203 + 3S02 ->• A12(S04)3-A12(S03)3 (1)
• Oxidation: A12
-------
Al+3 + x OH' J Al(OH)+(3-x) (7)
+(3~X) + H+ J H0 + AKOH)" (8)
Al(OH)
x -
Oxidation
+ 1/202 + H+ + S04~2 (9)
Neutralization
SO"2 + CaC00(s) + 2H00 •*• CaSO.-2H90 (s) + + CO "2 (10)
4 J £ q- / -5
C03~2 + H20 J HC03~ + OH~ (11)
HC03~ + H20 J H2C03 + OH~ (12)
H2C03 (diss.) t H20 + C02(diss.) ? C02(g) (13)
The last reaction goes to completion at pH 3.
In summary, sulfur dioxide is absorbed in a solution of basic
aluminum sulfate at a pH of approximately 3 [reactions (4) through (8)].
The resultant sulfite in the liquor is oxidized to sulfate by oxygen in the
flue gas and in the air which is sparged into the liquor [reaction (9)] .
The oxidized liquor is regenerated to basic aluminum sulfate by neutrali-
zation with limestone [reactions (10) through (13)]. The gypsum byproduct
from the neutralization step is removed by gravitational settling and
filtration. The filtrate and clarified liquor are returned to the process.
High S02 removal by the process requires the equilibrium of reaction (5)
be shifted to the right to allow more HS03~ in solution. This is accomplished
by more efficient oxidation of the absorber liquor [reaction (9)].
The concentrations of chloride and magnesium in the process liquor
are controlled by a purge stream. The aluminum content of the purged liquor
is recovered by adding excess limestone to precipitate the aluminum as
aluminum hydroxide. The precipitated aluminum is separated from the super-
natant liquor and returned to the process. (Equipment for aluminum recovery
was not installed at Shawnee.)
Control of the process chemistry requires measurement and control of
the total aluminum concentration in the process liquor and the percentage
of this aluminum available for the S02 absorption reactions. The aluminum
concentration is monitored by routine laboratory analysis and controlled by
addition of aluminum sulfate solution to the absorber hold tank. The
percentage of the aluminum available for absorption is controlled by
measurement and control of "% basicity." The concept of % basicity is
defined in the following discussion.
314
-------
Basic aluminum sulfate solution, the absorbent for the Dowa process,
is prepared by reacting solutions of aluminum sulfate with limestone to
remove sulfate as precipitated gypsum. The. limestone is added in less
than stoichiometric amounts to prevent converting the aluminum to aluminum
hydroxide, which would precipitate. Curve A of Figure 1 shows the pH
behavior of an aluminum sulfate solution titrated with either standard
acid or standard base. Curve B of Figure 1 shows the behavior of an
aluminum sulfate solution that is titrated by incremental additions of
powdered calcium carbonate. These results are plotted using the same
abscissas as Curve A. The differences between Curve A and Curve B are
caused by the presence in the latter case of bicarbonate and carbonate
species from the dissolution of the calcium carbonate. These species
affect the pH and buffering capacity of the basic aluminum sulfate
solution.
The flat portion of the pH curve is the region of interest in the
application of basic aluminum sulfate as a scrubbing reagent. For
scrubber applications the range of compositions is limited to (NQH/NAI)
values of about 0.3 to 1.2, where (NQR/NA!) is the ratio of moles of
hydroxide ion per mole of aluminum ion present. The lower limit is
chosen to prevent completely exhausting the scrubbing capacity of the
liquor, and the upper limit is chosen to prevent potential precipitation
of aluminum from the liquor, which would lead to the loss of the aluminum
in the gypsum end product produced in the process.
Within the composition range of interest, the liquor pH only changes
by 0.2 to 0.5 pH units. This small pH change precludes the use of pH as
a process control mechanism. Therefore, in the Dowa process, process
control is based upon liquor composition using basicity, B, which is
defined as follows:
B
As examples of the concept of basicity, consider the following:
Compound Basicity (%)
0
Al(OH)"t'(3~x) IQOx
x 3
A1(OH)3 100
Three independent means of determining liquor basicity can be used
in process control. The liquor basicity is monitored by an in-line
basicity analyzer which determines the liquor basicity automatically on a
continuous basis. In addition, the liquor basicity can be determined by
direct titration in the laboratory or calculated from the results of an
analysis of a lj>—»r sample.
315
-------
co
M
en
I I
10
9
8
7
6
5
4
3
2
I
I I I T
i \ r i
i i
I
A-20 ml O.IOM AI2(S04>3 with O.I N HCI and 0.1 N NaOH as titrants
B-IOO ml O.IOM AI2CS04>3 titrated with 3.45g of CaC03
I
I
I
2.8 2.4 2.0 1.6 1.2 0.8 0.4 0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 4.4
0 25 50 75 100
25 50 75
Basicity <%)
Figure 1. Relationship of pH to basicity and ratios of mols of OH and H to mol of Al
during acid and base titrations of A]_2 (80^)3 solutions.
-------
SHAWNEE DOWA PROCESS EQUIPMENT DESCRIPTION
Figure 2 is a flow diagram of the Shawnee train 200 after installation
of process equipment. The aluminum recovery step was not included in
the process demonstration due to limited funds. The major process
equipment utilized included:
1. The existing TCA scrubber complete with nitrile foam spheres and
a single absorber hold tank (spheres were replaced with rigid
packing during the test program)
2. An air sparger system, including a blower and pipe sparger located
near the bottom of the absorber hold tank
3. Two neutralization tanks installed in series
4. An existing thickener utilized for initial gypsum dewatering
5. An existing horizontal-belt vacuum filter for final gypsum dewatering
6. A reclaimed absorbent hold tank
7. An aluminum sulfate solution preparation and feed system
8. All process pumps and agitators associated with the above equipment
Sulfur dioxide absorption occurs in the TCA absorber. The oxidation
process step occurs in both the absorber and the absorber hold tank. A
bleedstream of absorbent is pumped to the neutralizer tanks, where the
limestone required for neutralization is added. The neutralizer product
overflows from the second neutralizer into a conventional thickener.
The thickener overflow is collected in the reclaimed absorbent tank, and
the thickener underflow is pumped to the filter for final dewatering of
the gypsum byproduct. The filtrate is returned to the reclaimed absorbent
tank. A portion of the thickener underflow is recycled to the first
neutralization tank to provide gypsum seed crystals for the neutralization/
gypsum precipitation step.
The basicity of the absorbent in the absorber loop and the reclaimed
absorbent is continuously monitored with an automatic basicity analyzer
and routinely analyzed in the test facility laboratory. The basicity of
the reclaimed absorbent is controlled by varying the limestone feedrate
to the neutralizer tanks. The basicity of the liquor in the absorber
loop is controlled by varying the rate of the absorbent purge to the
neutralization section.
The aluminum concentration in the absorbent is monitored by laboratory
analysis and controlled by aluminum sulfate solution addition to the
absorber hold tank.
317
-------
TO REHEATER
00
l->
00
LIMESTONE SLURRY
HOLD TANK
ALUMINUM SULFATE
MAKE-UP TANK
ABSORBER LIQUOR
HOLD TANK
PURGE
RECLAIMED ABSORBENT
TANK
Figure 2. Dowa process demonstration flow diagram.
-------
TEST RESULTS
Construction of process equipment changes for the Dowa demonstration
was completed in November 1979. Following the completion of construction,
numerous equipment-related startup problems plus boiler outages prevented
continuous operation of the process demonstration until January 1980.
The major operating conditions selected for the 1-month demonstration
are summarized in Table 1. Problems continued to hinder stable continuous
operation during the 1-month demonstration.
TABLE 1. DOWA PROCESS DEMONSTRATION
SELECTED ABSORBER OPERATING CONDITIONS
Operating condition
Inlet flue gas rate, aft^/min at SOO'F 25,000
Inlet flue gas fly ash loading, gr/sft3 'V-O.IO
TCA sphere bed static height, inches
Bed 1 5
Bed 2 5
Bed 3 3
Basicity, %
Absorber loop 10
Reclaimed absorbent 27
Aluminum concentration, g/& 20
Absorber recycle rate, gal/min 1,250
L/G, gal/kft5 58
Oxidation air stoichiometry,
Ib-atoms 0/lb-mol S02 absorbed 4.0
These problems included:
1. Freezing and plugging the basicity analyzer sample lines caused
by poor location of the sample lines and the failure of heat trace
material.
2. Unstable standard solutions for calibration of the basicity analyzer
caused the process to be controlled either above or below the
desired basicity set points. (This problem was not resolved until
near the end of the demonstration and may be responsible for scattered
test data.)
3. The method for determination of dissolved sulfite in the scrubber
liquor was inaccurate. (This problem was solved by addition of
iodine to samples to stabilize the sulfite concentration prior to
analysis and elimination of the filtration step in the analytical
procedure.)
319
-------
4. Following the initial startup, inspection of the TCA absorber walls
and absorber spray nozzles revealed that calcium sulfate scale
deposits (from previous limestone scrubbing tests in the TCA) were
dissolving in the Dowa liquor, breaking loose from the scrubber
internals, and plugging the absorber spray nozzles. Testing was
delayed while scale was manually removed from the absorber internals.
Despite the resolution of the above stated problems, SCL removal
efficiency still did not match the design expectations. In lieu of the
continuing process demonstration, a series of TCA screening tests to
determine the S02 removal efficiency of this absorber over a wide range
of flue gas rates, absorbent recirculation rates, and oxidation air
flowrates were performed. Attempts to improve control of basicity and
oxidation continued during these tests. The.862 removal efficiency of
these tests continued to be lower than expected from the Dowa process
operating with a TCA scrubber in Japan. The TCA static sphere bed depth
was increased to 8 inches with little effect on SC>2 removal. The maximum
sustained SC>2 removal during this test was 87%. Observation of the
sphere action during absorber operation and. sphere distribution in the
beds after the absorber shutdown indicated that the gas distribution in
the TCA was poor. Consequently, poor gas/liquor contact was suspected
to be the cause of the low removal efficiency.
The nitrile foam spheres were replaced with fixed-bed packing, and
the remainder of the test program was dedicated to absorption studies
with this type of packing. Three series of tests were conducted: two
series with a 9-foot packing height, and a third with a 6-foot packing
height. The ranges of major operating conditions during these tests
included:
Flue gas rate, aft3/min at 300°F 13,000-27,000 (low fly ash loading)
Superficial gas velocity, ft/sec 5.4-11.2
Recirculated absorbent, gal/min 700-1,400
L/G, gal/kft3 39-126
Absorber AP, in. H20 1.0-14.7
Absorbent basicity (absorber), % 11.0-35.2
The fixed-bed packing approaches flooding conditions when the absorber
is operated at 27,000 aft3/min (equivalent to 10 MW and an absorber super-
ficial gas velocity of 12 ft/sec). S02 removal efficiency did not remain
above 90% and steady operation was therefore not possible at a gas rate
equivalent to 10 MW.
After the superficial gas velocity was lowered to between 6 and 9
ft/sec, more stable operation and high S02 removal efficiency were
achieved. For example, 90% to 97% S02 removal was achieved while the
absorber operating conditions were 20,000 aft3/min flue gas rate, L/G
equal to 55, and pressure drop equal to 9.2 to 10.5 inches H20; 93% S02
removal was achieved while the absorber operating conditions were 20,000
aft /min flue gas rate, L/G equal to 82 gal/kft3, and pressure drop
equal to 9 inches of water; and 93% to 97% S02 removal was achieved while
the absorber operating conditions were 13,000 aft3/min flue gas rate, L/G
equal to 90 to 125 gal/kft3, and pressure drop equal to 1.5 to 2.5 inches
H2°r
320
-------
During the tests with 6 feet and 9 feet of packing in the absorber,
an excessive oxidation air stoichiometry was maintained to assure near
complete oxidation of the absorber liquor and to enhance S02 removal in
the absorber. The sulfite concentration in the recirculated absorber
liquor during the tests was ^0 to 60 mg/£. Data collected during the 6-
foot fixed-bed packing tests are plotted in Figures 3, 4, and 5. Data
collected with 9 feet of packing are being evaluated and will appear in
the Dowa project final report.
A typical gypsum byproduct composition during the absorption factorial
tests is as follows:
Component % by wt (dry)
Aluminum 0. 3a
Calcium 21.8
Carbonate Nil
Sulfite JfcO.O
Sulfate 53.8
Total solids (wet basis) 81.8
Acid insolubles Nil
a. Gypsum cake washing procedures
were not optimized. Lower aluminum
concentration is expected with im-
proved cake washing, such as 0.05%
Al achieved in commercial facilities
in Japan.
The final test was performed with flue gas taken from the duct
before the precipitator and thus the gas to the absorber contained full
fly ash loading, %A.O gr/aft^. No significant effect by the fly ash
upon the process was observed during this 1-week test.
SUMMARY OF RESULTS
The results of the Dowa test program are summarized as follows:
• Difficulties which were encountered during the Shawnee tests, such
as problems with analytical procedures for sulfite and preparation of
stable standard solutions for the basicity analyzer calibration,
undoubtedly had an adverse effect upon the removal efficiency of S02«
Also, the apparent poor gas distribution in the Shawnee TCA adversely
affected the test results. Although the Dowa process did not effec-
tively remove 90% of the S02 in the flue gas from a coal-fired boiler
while operating with a mobile-bed scrubber, the quantitative effect
of each of these problems upon the observed S02 removal efficiency
of the TCA is unknown.
321
-------
fiC
G
O
Z
UJ
o
SI
uu
Hi
CC
O
95
90
85
80
75
70
65
60
5.0 6.7 8.3 10.0 11,7 13.3
SATURATED GAS VELOCITY, FPS
Figure 3. Mass transfer characteristics: Polygrid packing - six foot data.
322
-------
o
z
o
Q.
H
LL
d
•
£
Q.
O
cc
c
CO
UJ
^
Q.
4.0
3.5
3.0
2.5
2.0
1.5
1.0 }-
0.5 *-
13CO
5.0 6.7 8.3 10.0 11.7 13.3
SATURATED GAS VELOCITY, FPS
Figure 4. Absorber pressure drop: Polygrid packing - six foot data.
323
-------
# 95
>
K
a
* 90
o
85
UL
IL
IU
80
75
Ut
ee
w 70
O
CO
65
60
I I
TESTING RANGE
GAS: 13,000-27,000 SCFM
LIQUID: 600-1300 GPM
I I
I I
I
I
50 100 150
LIQUID TO GAS RATIO, GAL/MSCFM
Figure 5. SC>2 removal efficiency versus liquid to gas ratio:
Polygrid packing - six foot data.
324
-------
• The SOp removal efficiency of the process is improved to greater
than 90% by providing better flue gas and absorbent contact with a
fixed-bed packed absorber. For example, with 6 feet of rigid
packing in the absorber, 93% S02 removal is achieved while operating
at 9 ft/sec superficial gas velocity, .82 gal/kft3 L/G, and 9 inches
H20 scrubber pressure drop.
• The neutralization and dewatering steps; of the process can effec-
tively produce a gypsum byproduct.
• High concentrations of fly ash in the process absorbent do not
affect the process performance (a preliminary result from a 1-week
test).
• There is no scale formation in the absorber.
CURRENT AND FUTURE TESTS
UOP is now performing laboratory studies and installing an integrated
pilot plant to further optimize the Dowa process. TVA and UOP have
independently proposed further Dowa process tests at Shawnee. These
further tests at Shawnee are contingent on completion of these laboratory
and pilot tests and an economic evaluation of the process.
325
-------
F.G.D. EXPERIENCES
SOUTHWEST UNIT 1
N. Dale Hicks, City Utilities, Springfield, Missouri
O. W. Hargrove, Radian Corporation., Austin, Texas
ABSTRACT
City Utilities of Springfield, Missouri, began commercial operation
of this F.G.D. system in September, 1977. Two turbulent contact absorber
modules are arranged in parallel and utilize a pulverized limestone slurry
for S02 removal. The scrubbers serve a 195 M.W. unit with a Riley Stoker
boiler burning 3.5% sulfur coal. Station design was by Burns & McDonnell,
with the Air Correction Division of Universal Oil Products, Inc. responsible
for the F.G.D. system on this new facility.
The absorber modules and various support systems have experienced a
variety of problems since initial start-up. The more severe problems
encountered have been: absorber and demister pluggage; failure of absorber
spheres; pipe breakage; control and instrumentation malfunctions; and
expansion joint, damper, and duct corrosion. Past and planned efforts
to rectify these difficulties, and to improve F.G.D. system reliability,
are discussed in detail.
A related problem area has been the continuous monitoring systems
for flue gas opacity and SO2 emissions. Original equipment has proven
unsuccessful and the investigation toward a solution, with the aid of a
consulting firm, is described.
The station is to be the host facility for an E.P.A. sponsored full
scale demonstration of adipic acid as an additive to wet limestone F.G.D.
systems. Anticipated results are enchanced efficiency and improved
operation of the pollution control facility. Also involved in the project
is the Radian Corporation and Universal Oil Products, Inc.
Preceding page blank
327
-------
FGD EXPERIENCES
SOUTHWEST UNIT 1
Southwest Unit 1 is a 194 MW fossil fueled unit owned and operated
by City Utilities of Springfield, Missouri. The unit was designed by
Burns & McDonnell. Commercial operation, of the generating unit began
in May, 1976 while the FGD Systems were not fully operational until
September 1977.
The steam generator was supplied by Riley Stoker Corporation and was
designed to burn high (3.5 to 4.0 percent) sulfur coal. Air Correction
Division of Universal Oil Products (UOP) furnished and installed the
electrostatic precipitator and the two-module flue-gas scrubber. Each
absorber module consists of a presaturator area; three TCA beds, each
containing spheres to enhance gas-liquid mixing; two chevron mist eliminator
banks; and a reaction hold tank. A common limestone preparation area and
sludge dewatering train serve both modules. The design was based on a
limestone composition of 98.7 percent CaC03, 0.7 percent MgCO3/ and 0.6
percent inerts.
The attached Figure 1 presents the process flow diagram. Basically,
the induced draft (.ID) fans pull the flue gas from the boiler through the
air heaters and the electrostatic precipitator (ESP) and discharge into
the scrubber inlets. Particulates (fly ash) are removed from the flue gas
by the ESP and conveyed to dry storage. SO_ removal is attained in the
scrubber with waste products of the reaction removed by continuous outflow
from the scrubbers to a thickener tank. The thickener separates the water
(supernatant) from the waste solids and recycles the water to the scrubber
system. The waste solids are drawn from the thickener and passed across
a travelling vacuum belt for further removal of water. This sixty to seventy
percent solids waste is then mixed with dry fly ash to produce a fixed
material which is landfilled on site.
Limestone is prepared for the scrubbing process by wet grinding in two
(2) ball-tube mills. There is no classification of the ballmill outputs so
fineness of grind cannot be readily controlled.
PROBLEM AREAS
Numerous problems have plagued the FGD systems at Southwest Unit 1.
Some problems have been solved, others are still being dealt with. The
following sections will detail the major classifications of problems
experienced and findings relative to their evaluations.
Pluggage
Pluggage in both the demister sections and the absorption areas of the
scrubbers originally hampered reliable operation of the FGD system. During
initial start-up and shakedown of the scrubbers it became readily apparent
that the demister wash system was inadequate. The system was designed to
operate as a closed loop as shown in Figure 2. Within two weeks of only
partial operation the demister chevrons were thickly scaled. Continuous
recirculation of the solids-ladened wash water caused a further scrubbing
action in the demister area resulting in scale formation and serious plugging.
328
-------
ELECTROSTATIC
PRECIPITATOR
co
N>
UD
AIR
HEATER
'B'l.D.
FAN
'A'l.D.
FAN
LIMESTONE
SLURRY
1 'B'ABSORBER
^-J-^ MODULE
Hi
\
T
SLUDGE
BYPASS DUCT
LIMESTONE
SLURRY
Y
r
SLUDGE
'A'ABSORBER
MODULE
STACK
70-1910-1
Figure 1 Southwest Unit 1 Process Flow Diagram.
-------
OJ
o
FROM
DRAW-OFF
SUMP
SUPERNATANT
TANK
MIST
ELIMINATOR
WASH TANK
LIMESTONE
RECYCLE
PUMPS
TO
THICKENER
FILTER CAKE
(MIXED WITH
FLY ASH FOR
LANDFILL
DISPOSAL
701911-1
Figure 2 Southwest Unit 1 Scrubber Flow Schematic.
-------
This situation was improved by a redesign of the demister spray and
presaturator spray systems as shown in Figure 3. Instead of a closed loop
system, the demister spray system was changed to utilize supernatant water
in a once-through flow sequence. This water was then collected and repumped
to provide the source for the presaturator spray system, and to wash the
underside of the trap-out-trays.
This modification has improved the operability with regard to the
demisters. Pluggage and scaling occurs much less frequently, but improve-
ments are still being sought by plant personnel.
Absorption area pluggage can be traced in part to the following factors:
sphere failure, inadequate limestone grinding, and on-off-on cycling.
Sphere failure has been a problem since the original system start-up. The
original sphere supplied by UOP was a seven gram white (TPR) sphere similar
in appearance to a ping-pong ball. The spheres were installed in only one
layer per module initially. Two levels per module were added prior to the
September, 1977 acceptance testing. It was evident after only a few weeks
of continuous operation that the spheres were failing. Many ruptured and
filled with slurry; others collapsed or dimpled losing their buoyancy. The
sphere layers were no longer completely fluidized and behaved as solids
filters resulting in severe absorber pluggage.
A random sampling of the spheres was conducted in September, 1978, to
determine the failure rate of the TPR spheres. It was determined that
virtually all of the spheres had either totally failed or were badly deform-
ed. Following discussions with UOP it was decided to replace the TPR
spheres with black foamed nitrile rubber spheres (eleven grams each). This
replacement was made in October, 1978, with a bed thickness of approximately
8 inches as prescribed. Upon scrubber start-up after this sphere replace-
ment, a significant increase in pressure drop through the scrubbers was
detected. Only about ninety percent of full load could be reached because
of the inability of the I.D. fans to make up the additional pressure drop.
Additionally the spheres began absorbing moisture, thus reducing their
buoyancy and creating the same type of pluggage problem that had existed
with the damaged TPR spheres.
In an effort to solve this situation, the sphere bed depths were
reduced from 8 inches per layer to approximately four inches per layer.
Some increase in load carrying capability was realized for short periods of
time, but the failure rate of the spheres was still rapid. Many split in
two; others shriveled and cracked and lost their buoyancy. Weekly cleaning
of the ball cages was sometimes not sufficient to prevent complete pluggage
of the sphere layers. It was not uncommon for plant employees to dig spheres
out of the pluggage with screwdrivers in attempts to clear the residue from
the cages.
In early 1980, it was decided to evaluate other spheres for possible
replacement. A sphere, manufactured by Puget Sound Trading Co., was select-
ed for testing. These spheres are green in color and approximately the same
size and weight as the original TPR spheres, but with cast ridges for addition-
al structrual strength. The spheres were installed during the summer of
331
-------
MAKEUP
WATER
FROM
DRAW-OFFlr-
TANK
MIST
ELIMINATOR
WASH TANK
PRESATURATOR
TANK
Co
co
ro
ABSORBER
STAGES
PRESATURATOR
TANK OVERFLOW
HOLDTANK
OVERFLOW
SLURRY
MAKEUP
HOLDTANK
SUPERNATANT
TANK
LIMESTONE
FILTER CAKE
(MIXED WITH
FLY ASH FOR
LANDFILL
DISPOSAL)
70 1912-1
RECYCLE
PUMPS
TO
THICKENER
Figure 3 Southwest Unit 1 Scrubber Flow Schematic.
Revised 5-77
-------
1980 and to date no substantial pluggage problems have been encountered.
Unit load-carrying ability has returned to original design levels.
City Utilities is currently evaluating the conversion of the scrubbers
from fluidized bed contacting to a tray type design. This conversion has
been completed at other locations and appears to have improved both the
economics and the availability of the scrubbers.
Funds to provide and install limestone classification equipment follow-
ing the ball mills have been budgeted. When this equipment is installed and
put into service, it is anticipated that limestone utilization will improve
and pluggage frequency decline.
Expansion Joints
The I.D. fan outlet expansion joints have been a source of considerable
scrubber downtime and expense. The joints originally installed by UOP
were manufactured of high-strength low-alloy steel. Within a few months of
operation, it was evident that the joints were failing. During the fall
1977 outage, the steel expansion joints were replaced by UOP with Viton
rubber joints. During 1978, over 3,000 hours of scrubber module downtime
resulted from numerous failures of these expansion joints.
In early 1979, plant maintenance personnel accepted the responsibility
of maintaining the joints from UOP. Joint life at that time could be
expected to range from two hours to two or three weeks. An analysis of
samples taken from failed joint specimens indicated an internal abrasion
failure mode. The presence of hardened fly ash and limestone in the insul-
ation boot on the flue gas side of the rubber Was determined to be the source
of the abrasion.
Further evaluation into the presence of the calcium material indicated
that the probable cause was presaturator spray nozzle pluggage. With a
nozzle plugged, flow of the spray could be directed into the duct counter
to the flow of the flue gas stream.
In May, 1979, the expansion joint was redesigned. The Viton material
was placed on the inside of the joint in contact with the flue gas. A
neoprene belting material was utilized for an external cover with insulation
fill between the two layers. Additionally, a small plate was installed
across the floor of the ductwork downstream of the expansion joint to halt
errant presaturator spray.
Since these modifications in mid-1979, the life expectancy of these
expansion joints has increased to at least six months. It is anticipated
that with slightly thicker belting material longer joint life can be achieved.
The evaluation of expansion joint performance is. ongoing.
333
-------
Piping
Most of the scrubber piping systems at' this installation are made of
fiber reinforced plastic (FKP) pipe. The pipe, manufactured by Fibercast
Co. (Div. of Youngstown Sheet & Tube Co.) and installed by OOP is resistant
to abrasive wear.
The Southwest Unit 1 Scrubber installation is not enclosed; i.e., all
piping systems are exposed to ambient weather conditions. The first winter
of scrubber operation demonstrated the vulnerability of FRP pipe to cold
weather failure. There were three types of failures: failure due to
freezing; failure due to pipe embrittlement; and failure of the joint
adhesive. Proper heat tracing and insulation of an FRP piping system is
most difficult because of its poor conductivity. There were instances of
FRP pipes with heat tracing and insulation that froze during the winter of
1978-79.
After evaluating various repair possibilities of the FRP piping systems
and researching other piping options, it was decided during the summer of
1979 to replace the mist eliminator trap-out-tray piping system with a lined
steel pipe system, manufactured by Peabody Dore. This replacement was
accomplished in October, 1979. The new piping system was heat-traced and
insulated in a proper manner. This piping system, historically the most
susceptible to freezing, did not sustain a single failure during the 1979-80
winter period.
Because of the improvements noted in the new piping system, it is
planned to replace all FRP scrubber piping with a lined-steel piping system
in the fall of 1980. With proper heat-tracing and insulation, pipe freezing
and breakage problems should be greatly reduced.
Corrosion
As with most scrubber installations, corrosion causes continuing and
extensive maintenance. Corrosion has caused deterioration of dampers, seal
strips, ducts, linings, and exposed metal surfaces in the outplant area.
Very shortly after initial scrubber start-up it became apparent that
material selections were not what they should have been. The chloride
concentration in the scrubber slurry has been measured as high as 2000 to
3000 ppm. Entrained mist that was not removed by the demisters collected in
the outlet duct, exposing the dampers, seal strips, and lining materials to
the high chloride liquid. In addition, continued contact of the liquid with
the flue gas resulted in a lowering of its pH to between 1.0 and 1.5 due to
further S02 absorption. The combination of this high chloride low pH
environment resulted in severe corrosion and rapid material deterioration.
The original outlet duct lining, Rigiflake 485, began peeling approximately
two weeks after initial scrubber start-up in April, 1977. Attempts were
made to spot-patch the failed liner areas. In October, 1977 the entire.
outlet duct surface was cleaned and relined again with Rigiflake. Within
a month of operation, the liner had again failed.
334
-------
Evaluations by the owner and OOP representatives led to the selection
of a different liner material for installation during a planned outage in
October, 1978. Two variations of the liner material, Placite 4005 and
Placite 4030, were used. After being in service for approximately one month,
the duct liners were inspected and found to be in very good condition.
After the second month of service, deterioration of the liner was quite
advanced. In April, 1979, the Placite application contractor returned to the
site to patch the failed areas and any other areas that were deteriorated.
After a few months of service, liner failure was again apparent. The
failure appearance was in most cases that of a blister. The Placite would
separate from the metal duct allowing corrosion of the metal beneath the
Placite. It was determined that improper metal surface preparation was
the probable cause of the failures.
During the month of October, 1979, the ducts were sandblasted to white
metal, and relined with Placite. During the process, all deteriorated
metal areas were patched and turning vanes replaced. Within three months
of operation, the Placite lining had deteriorated sufficiently in certain
areas to allow holes to be eaten through the one-quarter inch A-36 steel
duct material.
At this time other liner materials are being evaluated. Hastelloy
G (Cabot) appears to be a prospect for use but is so expensive that the
budget will not allow its use. Resista-Flake by Corrosioneering has some
applications which seem to have served marginally well. At the time of
this writing, no final decision had been made as to the material to be
selected for duct liner repairs in the fall of 1980.
Some gains have been made in the selection of materials for dampers
and seal strips. The original scrubber inlet and bypass damper seals were
of 304L stainless steel; the frames and blades were of A36 steel. Within
one month of operation, failure was evident. UOP then replaced the seals
with 316L stainless, the same material as the outlet dampers.
By the fall fo 1977 it was evident that the 316L material was not
suitable for the pH and chlorides present. UOP replaced the inlet and
bypass damper seals and the outlet dampers in December, 1977. The inlet
and bypass damper seal material used was Inconel 625 Huntington Alloy;
the outlet damper material used was Udeholm 904L, including seal strips.
To date, the Udeholm 904L outlet dampers have provided good service.
Some slight seal strip deterioration is evident and will be corrected.
The Inconel 625 inlet and bypass damper seals have not performed as well.
The seal strips have been completely corroded away and the carbon blades
and framework are badly deteriorated. An evaluation of materials to
replace these dampers is underway at the time of this writing.
The originally installed presaturator lining, Precrete Grout,
began failing soon after its application and before the scrubbers were
placed in service. Severe cracking appeared as if it were shrinkage induced.
The installation contractor made repairs on two different occasions in
an attempt to save the liner but to no avail. By late summer, 1977, holes
were eaten through the outer wall of the duct.
335
-------
During the outage that began in September of 1977, UOP removed the
Precrete material and relined the entire presaturator area of the scrubbers
with Udeholm 904L. This material served appreciably better than the Precrete.
The Udeholm lining required no maintenance until October, 1979, when several
sections had to be patched. Some degree of deterioration was evident over
most of the Udeholm lining indicating that a more resistant material was
needed. All repairs utilized Hastalloy G material in the presaturator area.
An inspection was conducted of the presaturator area during April, 1980.
Further deterioration of the Udeholm lining was evident. Several materials
were considered for repairs. Laboratory testing of Plastaloy (by Continent-
al Alloy Steel) indicated it possessed high abrasion resistance and high
corrosion resistance. It was decided to try some of this material in one
scrubber module. The sheets of the material were installed using special
nylon bolts. The material lasted less than one week in operation. Apparent-
ly the expansion of the material differed substantially from the metal caus-
ing the Plastaloy to twist and wrench its way apart from the duct.
Investigation is continuing into materials for future use in this area.
Instrumentation
Many of the problems initially encountered in the instrumentation
area were caused by long periods of inactivity while UOP performed needed
modifications on the FGD Systems. When UOP left the job site City Utilities
found itself without an adequately trained technical force to trouble-shoot
and maintain the systems. The maintenance staff has been expanded and train-
ing provided so that we now have good capabilities to deal with instrumenta-
tion and control problems.
Freezing problems have beset many of the instrument systems since the
first winter of operation. Instrument air drying capacity, as originally
installed, was sorely inadequate. The passage of this inadequately dried
air through small-diameter air control lines caused condensation in the
air lines and eventual freezing. Damage to transmitters and various other
instrumentation resulted. From initial scrubber operation until February,
1979, some 1500 hours of module downtime had occurred because of icing in
air lines and instrumentation.
A new instrument air dryer installation was funded by City Utilities
and the installation completed in February 1979. This dryer unit was of the
dessicant type and has served well. During the spring and summer of 1979,
the air lines and instruments were cleaned and purged to insure that no
moisture remained in the lines. As a result, there were no instrument air
line freezing problems during the winter of 1979-80. One plant air line
which supplies air to the limestone ball mill clutch control did freeze
last winter. A dessicant dryer assembly will soon be added to that air
system to correct the situation.
Another instrument freezing problem which has caused considerable
difficulty and module downtime is the pH monitoring system. Each scrubber
module has an on-line pH analyzing system with two glass electrode sensors.
The sensors are located in a small open tank through which the slurry flows
continuously. During extremely cold weather, ice forms around the sensors
336
-------
usually resulting in breakage of the sensors. Pluggage of the sensors by a
build-up of slurry requires frequent cleaning which eliminates the possible
use of a full enclosure. New in-line pH cells are now on order.
Continuous monitoring of flue gas opacity and SO has been totally
unsuccessful. The original opacity monitors were Lear Sielger RM-4 instru-
ments which were installed in the. I.D. fan outlet ducts. They never operated
reliably and failed soon after installation. The static pressure in the
duct where they were installed could easily reach 20 inches (W.C.). The
purge air fans on the instruments were inadequately sized to cope with this
pressure, thus making cleaning and other maintenance of the monitoring equip-
ment difficult because of flue gas infiltration. Corrosion of the equipment
eventually rendered it useless.
The original flue gas S02 monitoring equipment was a model 1268-2-21
Turnkey Gas Analysis System installed by Dynasciences Inc. This was an
extractive system with the sample obtained from the wet gas stream at an
upper stack platform (255 feet above ground level). The sample line, heat
traced and insulated, ran from this sample point down to the ground, then
to the scrubber control room where the monitoring equipment was located.
The total length of this sample line was over 400 feet. During operation,
sample line pluggage was nearly a daily occurence. If the sample probe
plugged, a technician was required to climb the stack ladder to the sample
probe level to clean and repair the probe. This was quite an unpopular duty,
and during wet or freezing weather was unsafe.
After discussions with various vendors, and continued lack of success
with further equipment modifications, it was determined that the existing
systems were not capable of operating in a reliable manner.
City Utilities evaluated consultants that specialized in the field of
flue gas monitoring and testing and selected Entropy Environmentalists, Inc.
to study the problems of the existing installations and to evaluate modifi-- .
cations or redesign of the monitoring equipment installation. Some of this
work has now been completed. A new, low pressure zone location has been
tentatively selected for the opacity monitors. The monitors would in fact
be located at the I.D. fan inlets in a negative pressure zone. The flue
gas S02 sampling point will be relocated to a lower stack platform just
above the scrubber outlet duct. A stairway is planned to connect the scrubber
to this platform and eliminate the need to climb the stack ladder to service
the equipment.
The specific equipment to be used is still being evaluated. Perfor-
mance specifications have been prepared. Entropy is not presently aware of
any monitoring equipment vendor who has successfully installed a system on
a wet stack which operated as reliably as required by the involved govern-
mental agencies. Few, if any vendors are willing to warranty their equip-
ment installation for any period of time after an acceptance test and their
people leave the site. Our experience has been that a vendor's serviceman
can get his equipment in. operation, leave the site, and his equipment would
again be inoperative before he would reach the airport.
337
-------
Thickener/Dewatering Systems
The waste slurry enters the thickener from the scrubbers with betweer
five and fifteen percent by weight solids content. Solids settling in the
thickener is aided by the addition of polymer. Frequent sampling of the
mixture to determine the settling rate is required to maintain proper inter-
face levels. The solids in the slurry must not settle out to more than
thirty to forty percent by weight or pluggage of lines and pumps occurs.
If the solids stay in suspension they overflow the thickener and are return-
ed as supernatant to the scrubber. These returned solids can contribute to
pluggage of demister nozzles.
One problem encountered in the thickener tank has been anaerobic
bacterial attack on sulfites and sulfates resulting in a change of color
of slurry and cake from cream to gray with a resultant smell of rotten eggs
(H S formation) . This problem occurs during warm weather and when the slurry
is retained in the thickener during breakdowns on the system lasting longer
than a day. The bacteria have been controlled by "shocking" the thickener
tank contents with swimming pool grade granular chlorine (usually four-hund-
red to five-hundred pounds broadcast into the thickener tank that has 750,000
gallons capacity).
The under flow from the thickener is pumped to an EIMCO Vacuum Filter
belt and discharged to conveyors with sixty to seventy percent solids. A
conveyor transports the dewatered sludge to a pug mill where it is mixed
50/50 (by weight) with dry fly ash. This produces a material which is
directly landfilled. During freezing weather, spillage of the material
causes pluggage and freezing of the conveyor tracks. Torn conveyor belts,
which results in temporary shutdown of the system, have been common.
Complete enclosure of this process has been comtemplated but funds have
not been available to accomplish the work.
A limiting factor in the dewatering operation is the pug mill. Its
capacity is such that at continuous high operating levels, sixteen to twenty
hours of operation per day are required to maintain the solids level within
the thickener tank at an acceptable level.
There is no redundancy of conveyors or the pug mill. If one breaks,
the system is down and the draw-off from the thickener must either be pumped
to the emergency pond or discharged to the ash pond. This type of break-
down would not immediately affect the operation of the scrubbers.
Other
Beneath each scrubber module in the hold tank is a large turbine agita-
tor manufactured by Lightnin Co. The drive shaft for the agitator is approx-
imately twenty-two feet (22') in length and is six inches (6") in diameter.
The shaft in the "B" scrubber module has now fractured twice. The first
break occurred in May, 1979. An analysis of the break indicated frequency
related flexural fatigue failure. While awaiting replacement material, the
338
-------
design of the hold tank and agitator shaft was evaluated to determine the
cause of flexure in the shaft. No conclusive determination was made so a
new shaft was installed and the scrubber returned to service in July, 1979.
This new shaft failed in May, 1980. This failure differed from the initial
one in that it occurred at a factory weld and there was no indication of
fatigue. A new shaft was fabricated locally and the scrubber returned to
service. The cause of this failure is still being diagnosed by the manufac-
turer.
Consideration has been given to two other additions which would undoubt-
edly improve overall scrubber availability. They are:
a) a spare (3rd) scrubber module
b) freeze-proof the FGD systems by totally enclosing them.
The cost of providing these improvements have prohibited serious considera-
tion of them.
ADIPIC ACID TEST PROGRAM
City Utilities is a progressive organization and is interested in
improving its operation. When it was learned that the Environmental Protection
Agency (EPA) was proposing to sponsor a full scale demonstration of adipic
acid addition to limestone scrubber operations, management was interested.
A contract now exists between Radian Corporation and EPA for this
demonstration program. City Utilities is providing the host site (Southwest
Power Station Unit No. 1) and other support services. The Air Correction
Division (ACD) of UOP is also participating in the program.
The addition of weak organic acids such as adipic acid to limestone FGD
systems has been shown to benefit both SC>2 removal and limestone utilization
and, also, to have a potential for improving the overall operability of a
limestone FGD system. Adipic acid has the effect of buffering scrubber
solutions, thereby enhancing liquid phase mass transfer. EPA has tested
adipic acid addition to limestone scrubbers at a 0.1 MW pilot plant in
Research Triangle Park, North Carolina and at the 10 MW prototype units at
Tennessee Valley Authority's Shawnee power plant near Paducah, Kentucky with
encouraging results. The program at Southwest Unit 1 is the final step in
demonstrating adipic acid as an additive for commercial FGD systems. At
the time of this writing, the demonstration program was in the second month
of a scheduled six-month program.
Test Program Objectives
EPA objectives in this program are to confirm the results of their
previous testing and successfully transfer this technology from the pilot
and prototype stages to a full-scale limestone FGD system operating in both
a forced oxidation and natural oxidation mode. City Utilities Southwest
Unit 1 represents a nearly ideal system for accomplishing these objectives
339
-------
for the following reasons:
the station fires high sulfur bituminous coal;
the FGD system already includes a thickener, a vacuum filter,
and a clay-lined landfill which are typical of current dewatering
and waste disposal techniques;
the potential exists to increase S02 removal, improve limestone
utilization, and increase the efficiency of the dewatering train;
the unit is a commercial scale power plant that will require
relatively minor modifications to perform the test program,
thereby resulting in the most efficient use of EPA funds.
City Utilities expects to gain valuable information on the operation
and performance of its scrubber. Its objectives in the program are three
fold:
aid in the successful completion of the demonstration
program by providing the host site;
evaluate the operating and cost advantages and
disadvantages of adipic acid addition and forced
oxidation at the SWPP scrubber;
investigate the ability of adipic acid addition
to keep the unit within compliance with the SO
New Source Performance Standards.
Anticipated advantages that City Utilities will see in their scrubbers
operation following adipic acid addition include:
increased SO removal, and
improved limestone utilization.
Increased SO2 removal has several potential benefits with respect to operation
of the scrubbers. First, the pH of the liquid in the outlet duct should
show a substantial rise above the normal range of 1.0 to 1.5. In fact, a
sample of this liquid has been tested during the recent preliminary adipic
acid testing and its pH has been found to have increased to 3.7. This is
due to the lower gas phase SO2 concentrations in the outlet duct. Liner and
duct corrosion rates should therefore be decreased for this reason. In
addition, potential exists for removing some of the ball charge in each
scrubber and still having high enough removal to keep the unit in compliance.
Removing balls would have two positive affects. First, the pressure drop
through the scrubbers would be decreased resulting in lower power costs for
the I.D. fans. Secondly, a smaller ball charge would mean less chance for
pluggage due to ruptured or deformed balls; thereby improving scrubber avail-
ability.
340
-------
The increased utilization of limestbne is also expected to result in
certain system operating improvements. First, there should be less dissolution
of limestone on the TCA beds and demisters, therefore decreasing the potential
for scaling. Also, fluctuations in SO removal efficiency caused by varia-
tions in routine process conditions may be reduced due to the buffering
capacity of the adipic acid. These advantages will be compared to the
additional operating cost incurred by adding the adipic acid.
Participants In Demonstration Program
The Environmental Protection Agency has funded the development work
on adipic acid as an additive to limestone FGD systems. This program
represents the agency's final step in upgrading the technology to commercial
status. John Williams is acting as the project officer for the EPA Indus-
trial Environmental Research Laboratory (IERL) at Research Triangle Park (RTF)
in this demonstration program. The EPA contracted Radian Corporation to
conduct the test program and provide the necessary FGD expertise to evaluate
the program results. Radian has subcontracted City Utilities of Springfield,
Missouri and UOP's Air Correction Division to provide the test site and
support for the necessary system modifications.
As prime contractor, Radian's responsibilities in this demonstration
effort include overall project management and coordination, conceptual
design of the forced oxidation and adipic acid feed systems, development
and.implementation of the test program, evaluation of results, and reporting.
Air Correction Division will prepare a detailed design and specifica-
tions for the forced oxidation and adipic acid feed systems; review quotes
and select vendors; and procure, install, and start-up the involved equip-
ment.
The primary responsibilities of City Utilities in this demonstration
program include power plant and scrubber operation, review of proposed site
modifications, coordination of site modifications and interfacing with both
Radian and UOP during onsite testing.
Proposed Test Plan
The adipic acid demonstration program will be divided into two test
phases. The first phase will be a series of tests in the natural oxidation^
mode (present equipment configuration),. Prior to the second test phase, V$\e
system will be modified so that air can be introduced into the reaction tank
for forced oxidation testing in Phase,II.
Within each of these phases, a one-month duration test without adipic
acid (baseline test) will be conducted followed by two months of testing
with adipic acid. Thus, the total program duration will be six months:
three months in the natural oxidation mode and three months in the forced
oxidation mode of operation.
341
-------
The current plan is to conduct this baseline natural oxidation test
during August, 1980. This will document the manner in which the system
presently operates. The scrubbing system will follow boiler load during
this period and several fairly short-term tests will be conducted at various
pH levels. However, most of this month will be used to monitor the system's
operation as City Utilities normally operates it.
Following this initial baseline test, adipic acid will be added to the
system, and the effects on system performance will be monitored. Two weeks
of adipic acid testing are planned between September 1st and 15th, when a
one month scheduled outage will begin. An additional six-weeks of testing
with adipic acid in the natural oxidation mode of operation will be conducted
after the outage. The forced oxidation baseline testing should be conducted
in December, 1980, with two months of adipic acid-forced oxidation tests
planned for January and February, 1981.
The adipic acid will be added to the limestone sump with the limestone
from the ball mills as shown in Figure 4. An on-off controller tied to the
limestone feed rate to the ball mills will insure that the desired amount of
adipic acid is maintained in the reaction tank. This desired adipic acid
concentration can be altered by changing the flow rate from a weigh feeder.
Adipic acid inventory in the weigh feeder will be maintained from a system
consisting of a vibratory hopper and screw feeder. The adipic acid concent-
ration in the reaction tank slurry will be analyzed periodically to insure
that the desired adipic acid concentration is being maintained.
The oxidation air will be introduced through a sparger network consist-
ing of PVC pipe in the reaction tank. The sparger pipe will be located fairly
close to the walls of the rectangular reaction vessel to minimize chances of
damaging the agitator shaft. The sparger ring will be installed during the
September outage to minimize downtime. Four 1800 SCFM compressors will be
utilized to supply air to both reaction tanks. An oxygen to SO sorbed
stoichiometry greater than 2.5 can be maintained at full load with this air
rate.
Since this program is a demonstration program rather than a research
program, only a minimum of parametric testing will be performed. However,
changes in scrubber feed pH, adipic acid concentration, and air/SO2
stoichiometry will be made to find the optimum operating conditions within
each test phase. Optimum performance will be evaluated by examining such
parameters as SO2 removal efficiency, limestone utilization, required adipic
aicd feed rate (unaccounted for losses of adipic acid), and sludge dewatering
properties.
Preliminary Test Results
The initial results of the adipic acid testing in early September
were very encouraging. Prior to the addition of adipic acid to S-l module,
its SO^ removal efficiency had averaged about 65 percent at the normal
operating pH of 5.5. After addition of adipic acid to a liquid phase
concentration of between 800 and 1000 ppm, the removal efficiency of the
module increased to above 90 percent with a high of 95 percent at high
342
-------
FLUE GAS
FROM BOILER
GO
£ COMPRESSORS
AIR FOR
FORCED OXIDATION
TO
DEWATERING
TO
STACK
i
V
HOLDTANK
THICKENER
OVERFLOW
RECYCLE
LIMESTONE
FEED
AIR
SPARGER
ADIPIC
ACID
r
BALLMILL
LIMESTONE
SLURRY SUMP
LIMESTONE
701913-1
Figure 4 Southwest Unit 1 Adipic Acid Test Program Flow Diagram.
-------
load conditions. Improvements in limestone utilization were also noticed.
Results of a later test at an operating pH of 5.0 and an adipic acid
concentration of 1500 ppm showed greater than 90 percent S02 removal
with limestone utilization of 99 percent. Testing is scheduled to begin
again after the maintenance outage.
CONCLUSIONS
City Utilities is continuing its efforts to improve the availability
and reliability of Southwest Unit 1 FGD Systems. Corrosion and absorber
area pluggage remain to be major problems to overcome.
The use of adipic acid as a limestone additive appears to provide
an interesting alternative to be considered while evaluating improvements
in system operation. Improved limestone utilization and SO2 removal will
provide an economic basis for comparison with other potential alternatives.
344
-------
ACKNOWLEDGMENTS
The authors would like to acknowledge several individuals who
contributed their efforts in this endeavor.
Bruce T. Stone, P.E.
Manager, Power Production and Electric Distribution
City Utilities; Springfield, Missouri
R. Dean Delleney
Program Manager
Research and Engineering
Radian Corporation; Austin, Texas
John E. Williams
Project Officer, Adipic Acid Test Program
Industrial Environmental Research Laboratories
U. S. Environmental Protection Agency
RTP, North Carolina
Walter C. Hauer, Operations Engineer
Douglas M. Kinney, Maintenance Engineer
0. C. Smith, Training Coordinator
and other staff and clerical personnel at
Southwest Power Station
City Utilities; Springfield, Missouri
345
-------
RESULTS OF THE CHIYODA THOROUGHBRED-121
PROTOTYPE EVALUATION
Thomas M. Morasky
Project Manager
Electric Power Research Institute
Palo Alto, CA
David P. Burford
Research Engineer
Southern Company Services Inc.
Birmingham, AL
0. W. Hargrove
Senior Engineer
Radian Corporation
Austin, Zexas
ABSTRACT
A ten-month evaluation of the Chiyoda Thoroughbred 121 Prototype ee
Process (CT-121) was conducted at the Sholz Electric Generating
Station of Gulf Power Company. The 23-megawatt CT-121 prototype
was modified from existing CT-101 process equipment at Scholz by
Chiyoda International Corporation, a subsidiary of Chiyoda Chemical
Engineering and Construction Company, Ltd. Chiyoda operated the
prototype, and the Electric Power Research Institute and Southern
Company sponsored technical evaluations of the prototype process
performance. This paper summarizes the findings of these evaluations
Detailed results of the gypsum stacking evaluation will be presented
with the Chiyoda Thoroughbred 121 presentation.
Preceding page blank
347
-------
RESULTS OF THE CHIYODA THOROUGHBRED-121
PROTOTYPE EVALUATION
INTRODUCTION
The Flue Gas Desulfurization Program area of the Electric Power Research Institute
(EPRI) is charged with responsibility of identifying, evaluating, and advancing
FGD technology to help the electric utility industry meet current sulfur dioxide
emission standards in the most efficient, reliable, and economic manner. The
Chiyoda Thoroughbred-121 (CT-121) system was reported to offer technical and eco-
nomical advantages over currently available flue gas (FGD) desulfurization tech-
nology. As a result, the EPRI and Southern Company Services sponsored a program
to have Radian Corporation of Austin, Texas evaluate the Chiyoda Thoroughbred-121
(CT-121) process. (The Southern Company is an electric utility holding company
operating in the Southeast. It includes Alabama Power Company, Georgia Power
Company, Gulf Power Company, Mississippi Power Company, and Southern Company
Services, Inc.) As part of this program, Chiyoda International Corporation, the
American subsidiary of Chiyoda Chemical Engineering and Construction Company of
Japan, built and operated a prototype (23 MW) CT-121 process at Gulf Power
Company's Scholz Power Station near Sneads, Florida with the cooperation of Gulf
Power Company. To a large extent, this system was constructed by modifying the
existing CT-101 demonstration equipment at Scholz. The CT-121 process at Scholz
is designated as a prototype because it was the first lar^e-scale application of
the CT-121 process.
SYSTEM DESCRIPTION
Figure 1 shows a schematic of the Scholz prototype CT-121 plant. This system was
designed to treat 53,000 standard cubic feet per minute (85,000 normal cubic
meters per hour) of flue gas (23 megawatts of electrical production at Scholz).
However, during the test program, gas flows ranging from 25,000-55,000 scfm were
studied.
As shown in the figure, the inlet gas was cooled and saturated liquid stream in a
venturi before entering the met bubbling reactor (JBR) where the bulk of the S02
removal occurred. Compressed air was injected into the JBR to completely oxidize
348
-------
LIMESTONE
SCRUBBED Q AS
TO MIST ELIMINATOR
Ftue a AS
V0
VENTURIPRESCRUBBER WITH
KNOCK-OUT TANK (S3 ai>)
OXIDATION AIR
Figure 1. Simplified process flow diagram of Scholz prototype CT-121
flue gas scrubbing system.
-------
the sorbed S02 and to assist the agitator in maintaining a good gypsum solids
suspension in the slurry. From the JBR, the gas passed through a mist eliminator
prior to exiting the system through a glass reinforced plastic stack. There was
no reheat provision in the prototype system. Powdered limestone was slurried and
added to the JBR to control pH. Limestone grinding facilities were not included
in the prototype. The gypsum produced during the evalutibn program was diposed of
in a gypsum stack, a disposal technique commonly used in the phosphate fertilizer
industry. Basically, a gypsum stack is a free standing body in which solid-liquid
separation is achieved by solids settling in a hollowed out section on top of the
stack. The supernatant liquid flows through the walls of the stack to form a
"moat" around the stack. This disposal was evaluated independently by Ardaman &
Associates of Orlando, Florida under EPRI research project 536-3 during the CT-121
demonstration. (A copy of the report was distributed at the EPA symposium as an
unpresented paper.)
The unique and central feature of the CT-121 process is the jet bubbling
reactor. Figure 2 shows a schematic of the prototype JBR configuration. SC^
removal, sulfite oxidation, limestone dissolution, and gypsum crystallization
reactions are all accomplished within this one vessel. This concept deviates from
the conventional limestone system which contains large recycle pumps, separate
absorption vessels and reaction tanks. Such a scheme can affect the capital cost
of a FGD system. In the JBR, the gas is dispersed several inches beneath the
slurry. This minimizes the liquid phase mass transfer resistance which can limit
S02 removal in conventional spray tower systems. The liquid pumping power
requirements are also low in the CT-121 system because large slurry recirculation
pumps are not used, however, the power required to overcome the high gas side
pressure drop tends to offset this savings somewhat. Figure 3 shows the physical
arrangement of the JBR, the inlet and outlet ducts and mist eliminator at Scholz.
TEST PLAN AND OBJECTIVES
The objective of the program was to evaluate the performance of the CT-121 system
under a wide range of operating conditions and to measure the reliability of this
prototype. By varying both site-specific and some non-site-specific parameters,
an "operating envelope" in which the CT-121 system can function successfully was
determined. This performance evaluation therefore provides a basis for cost
evaluation activities as well as for some of the design parameters required for
commercial units.
350
-------
c*>
FLUE GAS OUT TO
MIST ELIMINATOR
FLUE OAS IN FROM
VENTURI SCRUBBER
(SaS5,OOOSCFM)
PRESCRUBBER
SLOWDOWN
(0-30 QPM)
POND RECYCLE £
WATER 15 20 QPM)
pH METER
1
OVERFLOW TO
OYPSUM TANK
(a 10 QPM)
* LIMESTONE
TANK
LIMESTONE
SLURRY FEED
BOTTOMS TO GYPSUM
TANK (MO QPM)
Figure 2. Schematic of jet bubbling reactor (JBR).
-------
JBR-i
Mist eliminator —i
Outlet duct.
to stack
Figure 3. Jet bubbling reactor, gas ducts, and mist eliminator.
352
-------
The testing of the prototype CT-121 system was divided into four phases:
Q Phase 0 - Three-month duration startup, shakedown, and
initial parametric tests conducted by Chiyoda.
o Phase I - Two-month duration baseline tests conducted at
Chiyoda specified operating conditions to quantify
some of the control variable fluctuations that
might be encountered during routine operation.
o Phase II - Four-month duration test series conducted under a
variety of operating conditions (forced variable
perturbations) to evaluate system response under
operating conditions that may be representative of
a broad scope of utility applications.
o Phase III - Three-week duration tests conducted by Chiyoda
following modifications to the JBR internals to
simplify the JBR design and reduce capital cost.
In all, a total of ten months of tests were conducted over an eleven-month period
beginning in August 1978. Throughout the program locally hired personnel, oper-
ated the system. Chiyoda provided supervision only during the day shift. Chiyoda
conducted Phase 0 with no input from EPRI although operating data were transmitted
to EPRI. Phases I and II constituted the EPRI evaluation program. During these
phases, the test conditions were proposed by Radian and approved by EPRI, SCS, and
Chiyoda; an on-site Radian test crew conducted the tests and reviewed operating
conditions with Chiyoda personnel. During Phase III, Chiyoda performed the test-
ing independently, but Radian observed the testing as EPRI's and SCS's representa-
tive.
TEST RESULTS
Synopsis
When judged by five critical performance criteria: S02 removal efficiency, solid
waste'quality, limestone utilization, resistance to chemical scaling and reliabil-
ity, the performance of the CT-121 process throughout the EPRI evaluation program
was quite good. S02 removal efficiencies of 95 percent with an inlet flue gas
concentration of 3500 ppm S02 were achieved, and the gypsum produced throughout
the program settled rapidly and dewatered easily. The operation of the prototype
system was particularly outstanding from the standpoint of limestone utilization
and chemical scale control. Limestone utilization within the JBR averaged over
98% for the evaluation program. A detailed inspection at the conclusion of
Phase II revealed only minimal chemical scale deposition, none of which posed a
significance operating problem. This was after nine months of testing including
353
-------
three months of Chiyoda shakedown operation and six months of EPRI-sponsored
tests. These performance results are excellent in view of test conditions which
deviated significantly from Chiyoda's design operating set points. These results
thus indicate that the system is flexible and can withstand significant process
upsets. These results demonstrated that Chiyoda prototype CT-121 is an FGD capa-
ble of continuous, reliable, and efficient operation.
SQo Removal. The JBR overflow pH and JBR WP influenced the S02 removal efficiency
to the greatest extent in the test program. The S02 content of inlet flue gas
showed a marked effect on removal efficiency only at concentrations above
2200 ppm. The oxidation air stoichiometry and flue gas flow rate altered S02
removal characteristics of the JBR chloride levels up to 6000 mg/1 did not have a
measurable effect on S02 removal efficiency.
Three parameters, pH, WP, and inlet S02 concentration were fit to a theoretically
derived expression for S02 removal efficiency. The basic form of the mathematical
equation was initially derived by Chiyoda in 1978. Tests varying sulfur dioxide
(0/S02) stoichiometry and gas flow were fairly short term in nature and were not
varied in conjunction with variations in other process conditions. Because of
this these were not included in the mathematical model for predicting S02 removal.
The 229 data points used for this analysis were best fit by using two equations,
the first for inlet gas S02 concentrations less than 2200 ppm and the second for
higher S02 levels. The primary reason for using two equations is that nearly all
the testing was at an S02 concentration less than 2200 ppm (200 data points). A
single equation predicted accurate results for S02 levels less than 2200 ppm bu1:
did not adequately predict the removals observed at higher S02 concentrations.
Therefore, another set of equation coefficients were determined to better fit the
data at higher S02 concentrations.
354
-------
e
Equation 1 predicts the removal for inlet S02 levels less than 2200 ppm whi
Equation 2 describes the results achieved at the higher concentrations for the
prototype CT-121 system.
WP l'Q7
1 - exp (-3.49]\hir70 I
Fractional S02 removal = - ±2±2 - ' - - (1)
(for inlet gas S02 .,15.4, Min-pHmtr°2 „
concentrations less 1 + 56'9 ^wF° N1° ONTOOO°
than 2200 ppm)
1 - exp (-3.
Fractional S02 removal = - ±2i-i - - (2)
(for inlet gas S02 , S0
concentrations greater 1 + .84 N^~~
than or equal to 2200 ppm) w
where WP is the JBR pressure drop expressed as inches of water, S02 is the inlet
flue gas sulfur dioxide concentration in ppm and the pH is that measured at the
JBR overflow. Both of the equations are applicable only to the range of Phase II
test conditions at full load gas flow and with 0/S02 stoichiometric ratios greater
than 8. Figure 4 is a plot comparing the measured removal with the values gener-
ated by these two equations.
The equations show the importance of the pressure drop and S02 concentration pH,
on removal. As the pressure drop increases, the exponential term decreases, thus
predicting a higher S02 removal. Likewise, as the pH increases, increased S02
removal is predicted since both equations' denominators approach unity. Increases
in either pH or WP were expected to improve S02 removal efficiency. Since
increased pH results in decreased S02 back pressure in the froth zone, and
increased WP reflects longer gas-liquid contact time and/or more efficient flue
gas sparging. Figure 5 shows the effects of tradeoffs between WP and pH on S02
removal with the prototype unit. In most situations, it will be more desirable to
obtain'the required S02 removal by using higher pH due to the relatively low cost
of limestone. This should be evaluated on a case by case basis, and caution must
be used to ensure that increased limestone concentrations do not cause scaling
problems.
Further examination of Equations 1 and 2 show that S02 concentration had minimal
effect on S02 removal when the inlet S02 concentration remained below about
2200 ppm. Above this level, increases in S02 concentration caused a fairly rapid
decline in S02-term exponent in Equation 2 and is shown in Figure 6. This
555
-------
100
\ V
90"
-------
400
350 —
03
1
£ 300
o
-------
100
•=.- 95
CO
s
0
DC
CM
o
CO
E
o
s
o>
Q.
90
85
80
Predicted by Equations E-2, E-3
SI^'z3 95% confidence limits of equation
95% confidence region of single observations
0 400 800 1200 16QQ_ 2000 2400
Inlet SO2 Concentration (ppm)
2800 3200
Figure 6. Fractional SOo removal versus input SO2 with fixed pressure drop
(AP=11.5", pH = 3.5).
358
-------
drop-off in S02 removal occurs at an S02 concentration higher than in many spray
tower designs mainly because of the increased liquid surface renewal rate and
increased interfacial mass transfer area created by the JBR design.
Even though the flue gas flow rate and the oxidation air rate were not included in
the predictive equations, these variables had a measurable impact on the S02
removal rate. The boiler variable load tests in Phase I indicated that flue gas
flows lower than 30,000 scfm resulted in an average removal of 94% from a flue gas
containing 1000-1200 ppm S02 concentration. Flows of above 45,000 scfrn during the
variable load test period resulted in an average removal of 90 percent.
Unfortunately, there was not sufficient time to test the impact of low gas flows
at different pH's, WP's, and S02 concentrations.
The results of several short duration tests emphasized the importance of maintain-
ing rapid oxidation to achieve good S02 removal in the JBR. These short-term
tests quantified the effect of air-rate (stoichiometry) on S02 removal as shown in
Figure 7. While no difference between air rates of 1000 and 1300 scfm (1600 and
2090 Mm /hr) (0/S02 stoichiometric ratios ranging from 8 to 11) was seen in the
initial tests, Figure 7 shows a reduction in S0? removal efficiency to about
o *-
77 percent at an air rate of 480 scfm (770 Nm /hr)•(0/S02 stoichiometry of about
4). With the air shut off, the S02 removal dropped to below 40 percent. In addi-
tion to the 0/S02 stoichiometry, distribution of air in the JBR (which is
influenced by such factors as air sparger, agitator performance and specific craft
tube design) is also important in maintaining good sulfite oxidation
efficiencies. These design factors were not examined in detail in this evaluation
program.
Changes in limestone sources and increased chloride concentrations in the JBR
slurry had no measurable effect on S02 removal. The main difference between tne
Southern Materials Company (SMC) limestone and the Georgia Marble limestone was
the particle size since both were high calcium limestones. The specified SMC
limestone grind was 90 percent through 200 mesh (74)m) and the specified Georgia
Marble grind was 90 percent through 325 mesh (44)rn). ' The driving force for disso-
lution was sufficiently high at the low operating pH's in the CT-121 prototype for
limestone size to have nd effect on S02 removal. Likewise, spiking the system
with 6000 ppm chloride (added as CaCl2) had no effect on S02 removal.
359
-------
0)
cc
-------
Changes in the JBR configuration made by Chiyoda prior to Phase III appeared to
have only minor effects on S02 removal efficiency. Removal efficiencies of only
one to two percent lower than those calculated in Phase I and II were calculated
even though flue gas velocities through the sparger openings had increased
40 percent.
Limestone Utilization. Throughout the program the observed limestone utilization
in the CT-121 system was quite high. For both Phases I and II, the utilization
measured around the JBR remained above 98 percent. Changing the JBR overflow pH
from 2.5 to above 4.5, and the limestone grind from 90 percent less than 325 mesh
(44)m) to a grind of 90 percent less than 200 mesh (74)m) did not cause a
measurable change in utilization.
The utilization was also good even when one considers the limestone added to the
gypsum tank for final neutralization of the gypsum slurry to a pH of 6. Optimiza-
tion of this process step was not an objective of the program. The limestone flow
to the gypsum tank was only occasionally adjusted because there were no on-line ph
monitors or controllers on the gypsum tank. The samples taken during Phase II
indicated that the overall utilization including the neutralization tank, was
somewhat lower in Phase II (f93 percent) than in Phase I (f97 percent). However,
it appeared that the multiple changes in process conditions which occurred in
Phase II may have caused some pH upsets in the gypsum tank. This was probably the
primary cause of the lower utilizations.
During Phase III, Chiyoda tested JBR overflow pH set points approaching six. At
these conditions, the utilization in the JBr dropped to about 87 percent.
Solids Characteristics/Gypsum Scaling Tendency. The solids produced in the JBR
during the evaluation were generally greater than 97% gypsum. There were no sul-
fite solids measured since the pH was always low enough that calcium sulfite would
remain in solution until it was oxidized. Also, as discussed in the preceding
limestone utilization section, there were only small amounts of calcium carbonate
remaining in the JBR underflow slurry. The gypsum solids settled very rapidly
with no measurable differences in the free-fall characteristics between samples.
Figure 8 shows typical differences between solids formed when testing with lower
sulfur coal (nominal 1.8 percent sulfur) and those formed with higher sulfur coal
(nominal 3.2 percent sulfur). The crystals formed when cleaning the flue gas from
361
-------
JBR underflow solids 11/29/78
1.8 percent sulfur coal
JBR underflow solids 4/13/79
3.2 percent sulfur coal
Figure 8. Comparison of solids produced with two SO2 loading.
362
-------
the lower sulfur coal were long rod-shaped crystals. Many were over 400 microns
in length with length-to-diameter (L/D) ratios of from 10 to 20. The crystals
produced with higher sulfur loadings were less than 100 microns with L/D ratios of
from 2 to 5. This difference is consistent with what would be expected from oper-
ation with higher sulfate liquor loadings caused by the higher sulfur coal.
Operation with higher sulfur coal also increased the relative supersaturation of
gypsum in the JBR although scaling conditions were not noted during the program.
The maximum relative saturation measured, even during S02 spiking experiments
(3000-3500 ppm S02 concentration), was only 1.23. ;This is well beneath the crit-
ical level of 1.3 which has been identified as the threshold for incipient
scaling. '•',
Inspections at the conclusion of Phases I and II revealed little scale deposition
in the JBR. There were some random patches of gypsum scale on various surfaces,
but none of the depositions were threatening systemi performance and the scale
thickness was less than 1/16 inch (2 millimeters). Since the scale buildup was
minimal, infrequent routine cleaning might be necessary since the scale deposition
will be a continuing phenomenon. The duration between system cleanings was not
determined in the evaluation program, but it can be noted that nine months of
operation were logged and no operating difficulties-were experienced.
Gypsum Disposal. Throughout the program the gypsum was disposed of in a stack
which is a disposal technique commonly used for gypsum produced in the phosphate
fertilizer industry. Figures 9 and 10 show the stack during the initial fill
period and in the final configuration. Chiyoda also tested the product for use in
wallboard and Portland cement production. U.S. Gypsum and National Gypsum ooth
made successful production runs of over 100 tons each with gypsum produced by the
CT-121 prototype system. Laboratory tests also indicated that CT-121 gypsum could
be used successfully in Portland cement. Details of the gypsum disposal testing
will be the subject of a paper available with the handouts of this symposium.
.EPA Performance Parameters. The four performance parameters employed by EPA to
measure an FGD system's dependability are presented in Table 1. The overall
figures include both the Phase 0 shakedown and the Phase III test period. Both of
these periods involved some planned outages which penalized both the operability
and utilization factors. However, during the EPRI program, all four factors were
extremely good. There were only 22 hours of forced outages during the EPRI evalu-
ation program. Of this, 21 hours were due to limestone feeder problems.
363
-------
Figure 9. Stacking pond at start of Program 11/15/78.
364
-------
Figure 10. Filled stack, end of Phase II 5/22/79.
365
-------
Table 1
CT-121 VIABILITY PARAMETERS
Viability Parameters (percent)
Chiyoda Shakedown Phase
(Phase 0)
EPRI Evaluation Program
(Phases I and II)
Extended Chiyoda Testing
(Phase III)
Total Program Average
Availability3
99.2
99.3
99.5
99.3
Reliability5
99.1
99.3
99.1
99.2
0 per ability0
88.0
97.3
58.6
90.0
Utili:zationd
88.0
97.3
58.6
90. G
aAvailability - Hours the FGD system is available for operation (whether operated or
not), divided by the hours in the perid.
Reliability - Hours the FGD system was operated divided by the hours the FGD system
was called upon to operate.
C0perability - Hours the FGD was operated divided by the boiler operating hours in
the period.
Total Program = 6552/7276
"Utilization Factor - Hours that the FGD system operated divided by total hours in
the period.
Total Program = 6552/7276
366
-------
When inspection time was added to the total downtime, the fraction of the periocj
in which the FGD system operated was 97.3 percent as reflected in the utilization
for Phases I and II.
These performance parameters indicate the CT-121 prototype performed with
exceptional reliability during the evaluation program. These figures cannot be
used to accurately predict the performance of a commercial system, but the evalu-
ation program indicates that a properly designed CT-121 system could be expected
to operate with a minimum of process or mechanical problems.
Mist Eliminator Performance. The mist eliminator performance during the program
deserves special mention. The mist eliminator was composed of two banks vertical
Chevron blades mounted in a horizontal run of duct downstream of the JBR. This
mist eliminator was washed on an average of once a week for one minute with about
o
300 gpm 0.072 m/s of pond water. No signs of gypsum scaling or plugging were
noted during the program.
This excellent performance is attributed to two major factors. First, the super-
ficial gas velocity leaving the froth zone of the JBR was only about 2 ft/s
resulting in most of the entrained slurry being separated from the flue gas in the
interior of the JBR or in the JBR outlet gas chamber. Secondly, the slurry con-
tained very little solid phase alkalinity (CaCO, or CaSOo). Therefore, the dis-
0 O
solution of calcium solids and sorption of SC^ on the mist eliminator blades which
has caused scaling problems in many systems did not occur in the CT-121 prototype,
Overall System Controllability
The effective performance of the prototype system during the evaluation program
was .due to (1) the flexibility of the prototype to withstand process fluctuations,
and (2) the controllability of the prototype system. Two key process control
variables are monitored in the CT-121 system to ensure good performance: (1) JBR
overflow pH and (2) JBR overflow solids concentration. pH was used as the primary
control for S02 removal efficiency.
pH was continuously monitored with a dip-type sensor in the overflow weir, and
limestone feed rate was manually adjusted based on this reading. A neoprene wiper
was used, to keep a stagnant film from building up around the probe. This instru-
ment was checked daily and calibrated weekly.
367
-------
pH fluctuations remained within ±0.2 units even after flue gas flow rate
changes. Operation in the 3.0 to 4.5 pH range resulted in rapid limestone disso-
lution and good pH control.
The solids inventory in the JBR underflow was monitored by a nuclear density
meter, and the gypsum discharge rate from the JBR was adjusted based on the solids
concentration. Every four hours the operator checked the instrument by taking a
slurry sample and measuring the volume of the'settled solids. Although the solids
content did vary somewhat from the set point, deviation from this set point did
not cause scaling during the program. This was true even though the solids con-
tent was reduced significantly for a period of several hours during two different
short-term tests.
It is also noteworthy to mention that locally hired operators were employed to
actually run the process (2 operators per shift) and Chiyoda personnel were pres-
ent only during the day shift for most of the program. The process operated with
a minimum of problems or upsets using this approach to operator staffing.
SUMMARY
As a result of the independent evaluation program and related engineering activ-
ities, several CT-121 process design and operating features have been identified
which may result in improved operability and reduced operating costs relative to
existing lime/limestone systems:
o no large slurry recirculation pumps,
o no nozzles or screens,
o high limestone utilization,
o less dependence on limestone source and size on operation due
to the low operating pH,
o low slurry entrainment in the gas enhancing mist eliminator
performance,
o low scrubber profile which may lower capital costs and,
o the ability to operate successfully over a wide range of
operating conditions with a minimum of scale deposition.
363
-------
The concept of the JBR, therefore, represents a potentially attractive alternative.
to other currently available FGD technologies. The prototype at Scholz was suc-
cessfully tested over a ten month period and was shown to operate reliably and
efficiently under a variety of test conditions while treating flue gas from a
coal-fired utility boiler.
REFERENCES
1. Randall E. Rush and Reed A. Edwards. Evaluation of Three 20 MW Prototype Fl ue
Gas Desulfurization Processes. Final Report. Birmingham, AL: Southern
Company Services, Inc. March 1978. EPRI No. FP-713-SY, EPRI Project No.
RP536-1.
2. H. Idemura, T. Kanai, and H. Yanagioka. "Jet Bubbling Flue Gas
Desulfurization." Chemical Engineering Progress. February 1978. P. 46-5.0.
3. D. M. Ottmers, Jr., et al. A Theoretical and Experimental Study of the
Lime/Limestone Wet Scrubbing Process. Austin, TX: Radian Corp. 1974.
EPA 650/2-75-006, EPA Contract No. 68-02-0023.
36S
-------
Forced Oxidation of Limestone Scrubber Sludge
at TVA's Widows Creek Unit 8 Steam Plant
by
C. L. Massey, N. D. Moore, G. T. Munson,
R. A. Runyan, and W. L. Wells
Tennessee Valley Authority
Chattanooga, Tennessee
ABSTRACT
Tests on one module (140 MW) have been carried out to demonstrate the
feasibility of forced oxidation of limestone scrubber sludge to gypsum as a
viable technique for ultimate disposal of these waste materials. Both one-
tank and two-tank oxidation experiments were studied with data indicating the
two-tank runs more closely met test objectives. Equations to predict oxida-
tion were developed and expressed as a function of mass transfer and chemical
kinetics. Air stoichiometries of between 1.75 and 2.0 Ib atoms 0/lb mole SC>2
absorbed will consistently produce oxidation of ~95%.
As a result of the Forced Oxidation Test Program, this method is being
given consideration as one of the alternative methods of scrubber sludge dis-
posal for Widows Creek units 7 and 8. Additionally, Paradise Steam Plant
units 1 and 2 scrubber trains are being designed with a forced oxidation
option to produce a sulfate waste product.
Preceding page blank
371
-------
FORCED OXIDATION OF LIMESTONE SCRUBBER SLUDGE
AT TVA'S WIDOWS CREEK UNIT 8 STEAM PLANT
INTRODUCTION
TVA uses, or plans to use, limestone wet flue gas scrubbing as the method
of reducing S02 emissions at two of its twelve coal-fired steam-electric gen*
erating plants—Widows Creek units 7 and 8 and Paradise units 1 and 2. At the
remaining ten coal-fired plants, Widows Creek units 1-6, and Paradise unit 3,
TVA burns either low- or medium-sulfur coal, or washed coal to achieve the
required S02 emission limitations.
Forced oxidation will be utilized in the disposal of the sludge at the
Paradise plant. The scrubbers are scheduled to become operational by September
1982. Forced oxidation is presently being compared and evaluated with other
sludge disposal methods at Widows Creek. The total life-cycle costs of forced
oxidation sludge disposal will be compared with total life-cycle costs of the
alternatives of raw ponding, mixing the sludge with dry fly ash, and mixing
the sludge with dry fly ash plus additives.
The future role of forced oxidation depends primarily on the following:-
1. Technical feasibility of the process.
2. The total life-cycle costs of forced oxidation as compared with
other disposal methods.
3. The final requirements of the Resource Conservation and Recovery Act
(RCRA) on disposing of scrubber sludge.
BACKGROUND
For the last several years, TVA has been involved in an intensive research
and development project which was initiated to make a thorough and complete
assessment of its first full-scale scrubber system at Widows Creek unit 8,
located near Stevenson, Alabama. The research and development effort con-
sisted of six tasks which were designed to evaluate the scrubber system. This
paper will report results of the forced oxidation experiments at Widows Creek
unit 8.
The wet limestone scrubber system, designed and constructed by TVA, treats
flue gas from a 550-MW Combustion Engineering (CE) tangentially coal-fired
boiler. The flue gas desulfurization (FGD) system consists of four parallel
scrubber trains, each capable of scrubbing 25 percent of the flue gas. Only
one of the four scrubber trains, train D, was used for the forced oxidation
demonstration experiments. Assistance in identifying the design criteria for
the test program of the forced oxidation demonstration was obtained from the
studies performed at the Shawnee Test Facility.
TVA contracted with CE to install forced oxidation equipment on the FGD
system at the Widows Creek Steam Plant to demonstrate that forced oxidation of
FGD wastes is possible at this location as a processing scheme for waste
372
-------
disposal. The forced oxidation demonstration program began on April 2, 1979,
and continued until November 15, 1979. CE had the responsibility for the ini-
tial operational phase through June 30, after which Radian Corporation assumed
the operatipnal responsibility through November 15, 1979.
A flow schematic of Widows Creek Unit 8 Wet Limestone Scrubber System is
shown in Figure 1. The pressurized scrubber system consists of four
STEAM
TO ASH
DISPOSAL PONO
TO STACK PLENUM
2\
AIR HEATER COILS
AIR FROM
AIR HEATER FAN
POWER HOUSE
-o
-ENTRAPMENT
SEPARATOR
\ r
1
M
It
>>
*!
|.
, X
>
•1
1
•i
-V
\
^
m in FT
• f
/ — 3.
-fe/
RIVER
^ WATER PUMPS
FROM POND WATER
RECYCLE PUMPS
FROM
O
cc
FROM ESP
FROM B, CaD TRAIN
'VENTURI CIRC TANKS
TO SLUDGE
FAN
POND
VENTURI CIRC A0SORB£R CIRC
TANK 8 PUMPS TANK & PUMPS
EFFLUENT SLURRY
SURGE TANK & PUMPS
LIMESTONE
SLURRY
LS SLURRY STORAGE TANK
FEED PUMPS
SLUDGE POND
Figure I. Scrubber System Flow Diagram
identical trains located downstream of low efficiency (approximately 50 per-
cent) electrostatic precipitators. The principal components in each train A,
B, C, and D include a boiler I.D. fan, venturi scrubber, grid-type spray tower
absorber, Chevron-type entrainment separator, indirect steam reheat system,
venturi slurry recirculation system, and absorber slurry recirculation system.
Each module is capable of treating approximately 25 percent of the boiler flue
gas at full load.
The waste slurry produced by the FGD system currently is stored in a 110-
acre pond. Disposal of this slurry represents a major problem in continued,
long-term operation of the scrubber unit. It was decided to evaluate and
demonstrate the forced oxidation method for treating the sludge to decrease
the effective volume required for disposal or to improve the stability of
material in the disposal area. During the demonstration program, a forced
373
-------
oxidation system was installed on scrubber train D. Approximately 10 percent
(4,370 Ib/hr of solids) of the oxidized scrubber bleed stream was processed
through a 2-stage dewatering system consisting of a thickener and rotary drum
vacuum filter. A flow schematic of the demonstration unit is shown in
Figure 2. Initially, the filter cake from the vacuum filter was reslurried
COMPRESSED
ABSORBER
CIRCULATION
TANK
THICKENER
UNDERFLOW
PUMP
RESLURRY TRANSFER
TANK
EFFLUENT SLURRY
SURGE TANK
Figure 2. Forced Oxidation Dewatering Equipment
and disposed of in the pond. After the test objectives had been achieved and
conditions determined for producing a cake of consistent quality, arrangements
were made to initiate a landfill disposal project for long-term monitoring of
the final product. The gypsum produced at Widows Creek is unusual in that 30
percent of the solids is fly ash. The presence of this ash may have as yet
undetermined effects on the long-term stability of the final disposal material
Oxidizing air was introduced to both the venturi and absorber tanks by
means of a circular sparge ring, located just beneath the agitator impellers
as shown in Figures 3 and 4. Air was discharged through thirty-six 1-1/4-inch
holes on the outside of the sparge ring (Figure 5).
374
-------
on
Z <+
\
— I
— «*
•—
— — U-TYP.
•4'TYP.
^
•J
BAFFLE
C
1
T
> — 1
r
' -i fl"^l' »
L
I i
4*
0.0.
PIPE
8.
\
hi
I
;
i
I
33'-0"
-BAFFLE
16" O.D. PIPE
Figure 3. Venturi Tank Agitator Figure A. Absorber Tank Agitator
and Sparge Ring and Sparge Ring
(36VI 1/4" DIA. AIR DISCHARGE -
HOLES ON OUTSIDE
OFRINO
Figure 5. Absorber Tank Sparge Ring
375
-------
Objectives and Goals
The general objectives of the test program were as follows:
Demonstration of forced oxidation as a viable FGD waste disposal
option.
Acquisition of data applicable to the design of a forced oxidation
system for the Widows Creek units 7 and 8 FGD systems.
Three specifically quantitative goals of the demonstration program were
as follows:
Attainment of greater than 95 mole percent conversion of calcium
sulfite to sulfate.
Production of a waste product capable of being dewatered to greater
than 80 wt percent solids with a vacuum filter.
S(>2 removal efficiency of 88 percent or greater.
Originally, the test plan for the forced oxidation experiments was divided
into three separate test blocks. One test block (experiment B) was designed
to verify the hypothesis that the air stoichoimetry required for oxidation in
the venturi effluent hold tank only would be significantly less than in the
absorber tank. Results of this type have been observed at the Shawnee test
facility. However, such was not the case due presumably to excessive carry-
over of venturi-loop liquors into the absorber loop. This carryover appears
to be a function of boiler load (gas velocity) and results in either a leve-
lized pH in both tanks or inversions such that the absorber tank actually had
at times a lower pH than the venturi tank. Oxygen stoichiometry requirements
are closely related to pH.
A second test block (experiment C) involved oxidation in both the venturi
and absorber hold tanks simultaneously to determine if such dual tank oxidation
could be accomplished at a lower oxygen stoichiometry than single tank oxida-
tion. A third series of runs (experiment A) with oxidation in the absorber
tank only was cancelled because of the difficulties experienced with one tank
(venturi) oxidation as described in experiment B above.
Test Description and Results
Operating conditions were varied to meet the three goals of primary inter-
est. Oxidation, dewatering, and SOg removal were each studied with minimal
interference from the other parameters. To determine the minimum amount of
air necessary to achieve 95-percent oxidation, an initial air rate was chosen
which was known to give about 99-percent oxidation. The air rate was then
reduced stepwise until oxidation was consistently above 95 percent. After
tests at several air rates which gave between 91 and 99 percent oxidation, it
was possible graphically to determine the minimum air stoichiometry.
376
-------
The air stoichiometry was then held constant at this value while the
dewatering train was tested. This involved determination of the appropriate
combination of feed rate, filter area, drum speed, and filter pressure that
produced a product containing a minimum of 80 percent solids.
Since the FGD unit is a commercial system, it was difficult to vary a
large number of operating variables to increase S02 removal. In addition, TV*
was simultaneously involved in a comprehensive test program aimed at improving
the S02 removal of the "Widows Creek FGD system. " It was felt that the results
of this companion program could be applied to a forced oxidized system as well
as the existing system. Consequently, only two variables were examined in
regard to S02 removal: Ca/S ratio and the effect of additional packing in the
absorber. Testing of these two parameters was sufficient to produce the
desired 88-percent S02 removal. During the S02 removal tests, the forced oxi-
dation unit was operated to further refine the operating parameters. The
equipment was operated in combination to demonstrate that the FGD unit and oxi-
dation unit could produce oxidized sludge of the desired quality while meeting
the S02 removal requirements.
Operating parameters were set on the desired conditions and a period of 8
hours was used to allow steady state operation before sampling. Usually only
one sample per sample point per day was drawn and analyzed. An additional 12
to 14 hours of operation was required to verify steady state conditions and to
obtain a second set of chemical data. A summary of sample data and analytical
determinations is given in Table 1.
Table 1. Summary of Analyses
Saaple - Phase
Liquid
Solid
Slurry
Bleed Streaa
SO? Ca**
SO,
XCH
ci~
coT
soT
soC
COi
Mg
Anions Na+
K+
NHt
Anions Mg**
Cations
Cations
Acid Insoluble Quantity
Wt 2 Solids
PH
T«p
Settling Rate
Test
Thickener Underflow
so"
S0~
Acid
Ca++ )
( Cations
Anions Mg I
Insoluble Quantity
Particle Size Distribution
Wt Z
»H
Trap
Solids
Filter Uaf
Teat
Thickener
Overflow
X Solids
ull
Tenp
Vacuum
Filter Drum
Z Solids
Thickness on
l
i
Absorber bottoms
SOT Co"
SO^ Mg '
NOT Anions Na Cations
Cl Kf
COT Nut
soT j c*"
SOT Anlona MS" Citl°n6
Acid Insoluble Quantity
Wt 2 Solids
|)ll
Teap
The following methods were employed to determine quantitatively cations
and anions of both solid and liquid samples.
377
-------
Cation Analyses
Analyses for calcium, magnesium, sodium, and potassium were done by atomic
absorption spectrophotometry.
Anion Analyses
The analyses for the anions required a number of different analytical
methods. These methods included:
Ion chromatography for sulfate and chlorine determinations.
lodometric "back" titration with iodine for sulfite determinations.
• Nondispersive infrared (NDIR) analysis for C02 determinations.
A representative set of operational and chemical data collected during
the demonstration is shown in Table 2.
TABLE 2. BLEED STREAM ANALYSIS AND TEST RESULTS
Liquid Phase
Date
Time
Test
Condition
Temp
°C pH 80s
Milligrams/liter
SC-4 Cl" C03 Ca++ Mg++
12/20/79 0745
C5
41
5.7 19 1278 1021 144 770 163
Solid Phase
Milligrams/gram solid
Test Venturi
Date Time Condition pH
12/20/79 0745 C5 5.7
Test Results
Ca++ S03 S04 C03 Mg
177 3 303 91 3
Relative Relative % Residual
saturation saturation electroneu-
CaS03'^H20 CaS04-2H20 trality
0.2 0.7 -3.6
CaC03 S02 Oxygen
added absorbed added
Ib-mol Ib-mol Ib-atoms
min min min
Comments
Steady state
0/S02
ratio Percent
Ib-atoms oxidation
Ib-mol
++ Acid
insoluble
279
Ca/SO
ratio
Ib-mol
Ib-mol
1.4
Vacuum
filter cake
% solids
2.2
1.3
2.8
2.3
99
86.8
378
-------
CHEMISTRY OF FORCED OXIDATION
Proposed Reaction Mechanisms
6xidation may best be explained in terms of mass transfer and chemical
kinetics. The oxidation of sulfite to sulfate can be affected by several
chemical phenomena including gas absorption, reaction kinetics, dissolution
of solids and precipitation of solids. The overall reaction,
CaS03 + ^02 -»• CaS04 (1)
can be broken down into several steps, each a necessary link of the reaction
pathway and, as such, each a potential rate limiting step. These are shown in
Table 3.
TABLE 3. REACTION STEPS AND PHENOMENA INVOLVED IN SULFITE OXIDATION
Reaction Step Phenomena Involved
Dissolution of reactants Gas absorption, solids dissolution
Reaction Kinetics
Precipitation of products Solids precipitation
First, it is necessary for the reactants to be in the liquid phase. This
involves absorption of 02 by the scrubbing liquor and either dissolution of
solid calcium sulfite or absorption of S02 as indicated by reactions (2)
through (5).
02(g) -»• 02 (aq) (2)
S02(g) + H20 •* HSOs(aq) + H+(aq) (3)
CaS03 • &i20(s) -» Ca++ + 80s + %H20 (4)
80s + H20 •» HSOa + OH* (5)
The oxidation reaction then takes place between the dissolved reactants
according to reaction (6).
HSOs(aq) + V>2(aq) -> SO^Caq) + H+(aq) (6)
Finally, the reaction product, gypsum, is removed from the liquid phase
by precipitation.
SO^aq) + Ca*+(aq) + 2H20 •* CaS04-2H20(s) (7)
The entire sequence is illustrated schematically in Figure 6 and each of
these steps is discussed in detail in the following paragraphs .
379
-------
S02(g)
GAS \
V , -
LIQUID HS03 (aq) + &)2(aq) » S04~(aq) + H
/ V
SOLID / \
CaS03^H20(s) CaS04-2H20(S)
Figure 6. Mass Transfer and Reaction Sequence
Dissolution of Reactants
Bisulfite, HSOg, appears to be the reactive form of the dissolved
hence subsequent discussion will refer to bisulfite rather than sulfite which
is also a dissolved species.
Bisulfite can be dissolved in the liquid phase by two mechanisms, absorp-
tion of SOg from the flue gas and dissolution of solid calcium sulfite.
In forced oxidation processes where oxidation takes place outside the
scrubber circulation loop, dissolution of calcium sulfite solids is clearly
the predominant form of bisulfite generation. Since the bisulfite is being
removed by oxidation, the sulfite relative saturation is expected to be low
(<1) and calcium sulfite dissolves.
The Widows Creek demonstration involved oxidation within the scrubber
loop. In this case, bisulfite ions will be supplied by the absorber, and dis-
solution of solid sulfite will not be of any importance unless sulfite solids
are precipitating in some part of the system. During C-series testing, air was
sparged into both hold tanks. Figure 7 is a plot of the relative saturation
of calcium sulfite versus percent oxidation in the solid-phase for the B and C
series tests.
RS a
K&CaS03-%H20 = - £
where
a = activity of subscripted component
Ksp = s°lubility product of the precipitation reaction (temperature
dependent)
Note that for C-series tests (two tank oxidation) the relative saturation was
substantially greater than 1 at most oxidation levels, indicating that bisul-
fite ions absorbed in the scrubber were oxidized without first precipitating as
CaSOs'i^HgO- This means that at intermediate oxidation levels, oxidation com*
petes with precipitation for HSOs ions, and dissolution of solid sulfite is iiot
part of the reaction mechanism. The low relative saturations seen at extremely
high oxidation levels are most likely due to sulfite ion depletion, caused by
sulfate precipitation and not any dissolution mechanism.
380
-------
12
B SERIES TESTS - 0
ONE-TANK SPARGING
10 h c SERIES TESTS - A
TWO-TANK SPARGING
O
in
o
O
&
' 6
Ifl
>4
s
/
I
A ~ -
A
A
0 0 * * *
\A^
o ° o
A
*
0 A
J I I 1 1 1 1 1 $t
10 20 30 40 50 60 70 80 90 100
Mole % Oxidation = SO^/CSOa + 804) in solid phase
Figure 7- Relative Saturation of CaS03-%H20 vs Oxidation in
Venturi Hold Tank - One- and Two-Tank Air Sparging
Relative saturation data for B-series tests (one tank oxidation) indicate
that dissolution may have been a more important factor in these tests. With
air introduced only into the venturi circulation tank, CaS03*%H20 precipitat:.or
would have been expected in the absorber circulation tank, necessitating disso-
lution in the venturi tank. Also, since significant backmixing occurred
between the two tanks, the dissolution step has been made more pronounced as a
source of calcium bisulfite than for a true two-stage system with single-tank
oxidation.
In summary, this analysis indicates if the absorber tank is sparged, then
the source of bisulfite is S02 absorption and solids dissolution is not neces-
sary. Therefore, solids dissolution can be eliminated as a potential rate lim-
iting step in the C-series testing described in this report. However, this is
not true for the B-series testing performed earlier. When just the venturx
tank is sparged, solids dissolution is a necessary link in the reaction pathway
since the source of bisulfite resulting from S02 absorption is insignificant.
Another step of the reaction pathway involving reactant dissolution is
oxygen absorption. This is primarily a mass-transfer process, affected by both
chemical and mechanical factors. The system chemistry affects two parameters
381
-------
that are important to mass transfer: oxygen solubility and the diffusivity of
dissolved oxygen. Oxygen solubility decreases with increases in both tempera-?
ture and ionic strength. The diffusivity of dissolved oxygen will increase
with increasing temperature and decrease with increasing ionic strength.
Neither temperature nor ionic strength varied substantially during
demonstration testing.
pH also has an important effect on the oxidation reaction in that oxi-
dation of sulfite causes the pH to rise when in the presence of excess limer
stone. This increase in.pH will rapidly quench the reaction unless some
method of replenishing H ions is available. At Widows Creek, the pH level
for oxidation purposes was maintained by 862 absorption.
Under normal operating conditions at Widows Creek, with the pH at 5 to 6,
the sulfite species are 40 to 70 percent in the bisulfite form. This is suf-
ficient for forced oxidation. The pH's of the two hold tanks rarely differed
by more than 0.5, but if the backmixing problem were eliminated, the pH in the
absorber would be higher. Since bisulfite availability is not rate-limiting,
the effect this would have on a force-oxidized system may be small unless the
pH was over 6.5.
The data indicate that during the tests, pH values were in the proper
range to provide high concentrations of the bisulfite ion. Therefore, bisul-
fite ion availibility can be eliminated as a limiting step in the overall
reaction sequence. The potential rate-limiting steps are thus reduced to
reaction kinetics, product removal, or 02 transfer.
Reaction Kinetics
Thermodynamics indicate that the oxidation (equation 5) reaction is essen-
tially irreversible, with an equilibrium constant (25°C) on the order of 1040.1
Therefore, we need only consider the kinetics of the forward reaction.
Most forced-oxidation research has indicated that the reaction is rela-
tively fast. If this is true, then the site of the reaction is limited to the
gas/liquid interface. This indicates that mass transfer to the film between
the gas and the bulk liquid will be the limiting factor, and not reaction
kinetics.
Using film theory terminology, the phase interface and reaction zone of
the oxidation reaction can be shown by two models in Figure 8. Model I illus-
trates a fast second order rate reaction where the diffusion rate of 02 and
HS03 in the liquid film are not limiting the overall reaction and the reaction
is fast enough so that the reaction zone remains totally within the liquid
film.
Model II shows an instantaneous reaction first order rate expression where
the concentration of HS03 is relatively high and the reaction plane is moved
to the gas-liquid interface. The overall rate will be limited by the diffusion
of 02 through the gas film. The film models shown in the figure result in
mathematical models for the reaction rate that are essentially the same as
those for mass transfer without reaction, except for an "enhancement" term.
Reaction in the film tends to thin the film, and the enhancement term accouaU
for the corresponding increase in mass transfer.
382
-------
MOON. I:
SULK
QAS
PO,
QAS
FILM
LIQUID
RIM
BULK
LIQUID
Reaction only in film, general fast-irreversible reaction.
R
• reaction rate • K eaPOj
- Gas- and liquid-side oxygen mass-transfer coefficients
• Henry's Law constant for oxygen
• Complex ^function of C...-.-, C..
cloU 3 ^2
a • Interfacial area per unit volume
POa • Gas-side partial pressure 0:
For an infinitely fast reaction, the expressions become:
R • Kga E P02
\ I ^ S(
i -i
K K • v,
L o2g Oti J
- 1 + (constant) <
POi
.. ... 8ULX
ELII: GAS
PO,
GAS
flLM
.
\
\
\
>
LIQUID 1 8ULX
FILM | LIQUID
1
1
1
^- flEACTION ATINTEBFACE
.^ \
** \
\
\
I
CM483.(MiQW
Reaction only in film, high HSOj concentration, general fast-irreversible
reaction.
R - Kga ?02
K .1" _ 1_ + constant]
5 LK02g V^HSOj J
For an infinitely fast reaction, Che expressions become:
R - K a ?02
Figure 8. Models for Film Reactions2
383
-------
The film reaction models suggest that mass transfer of 02 is the limiting
step rather than kinetic limitations . This is evidenced by the presence of
the oxygen mass transfer coefficient in the expressions for reaction rate conr
stant. Therefore, in terms of the overall reaction rate, both sulfite and
reaction kinetics are not rate-limiting steps. Sulfate precipitation rate does
not affect the kinetics and, as will be shown in subsequent discussion, will
be a function of the reaction rate. The most probable rate-limiting step "in
the overall reaction sequence is then oxygen transfer. However, data are
needed to confirm this and to validate one. of the models shown in Figure. 11.
Removal of Products
An important aspect of forced oxidation is its effect on the relative
saturation of CaS04. The relative saturation for sulfate (RS) is defined
similarly to that for sulfite.
RS 3Ca++ '
KbCaS04-2H20 = -
SpCaS04-2H20
where
a = activity of subscripted component
K = solubility product of the precipitation reaction (temperature
dependent)
RS of calcium sulfate is important because it can affect scaling of the
scrubber. Generally, for 1 < RS < 1.3, crystal growth will occur on existing
gypsum crystals. At RS > 1.3 gypsum nucleation can occur which can result in
crystal growth or scaling on the scrubber internals. This can eventually
require a shutdown to remove the scale.
While forced oxidation increases the total amount of S04 or CaS04-2H20
present, the RS., 0/. 01I _ at Widows Creek was not increased and may have been
LabU4 •
decreased. This is attributed to the higher sulfate concentrations resulting
in an increase in gypsum solids surface area which will enhance the gypsum
precipitation rate.
After the bisulfite is oxidized, it is necessary to remove the sulfate
product from solution. This is accomplished via precipitation of the sulfate
as solid gypsum, CaS04'2H20. Gypsum precipitation has been studied extensive 0
with respect to flue gas desulfurization systems. The driving force for preci-
pitation is relative saturation, and three regimes can be considered as shown
in Table 4.
384
-------
TABLE 4. GENERAL EFFECT OF RELATIVE STATURATION OF GYPSUM PRECIPITATION
Relative Saturation Effect
<^ Dissolution of gypsum solids
!-° ~ I-3 Precipitation on existing solids
>l-3 Solids riucleation
A simple rate expression for gypsum precipitation can be written as
r = KafC (RS-1) (10)
where
r = precipitation rate (gram/liter-sec)
K = temperature dependent constant (gram/cm2-sec-unit driving force)
a = crystal interfacial area per gram of precipitation solid
(cm2/gram)
f = weight fraction of the precipitation species in the solid phase
C = total solids concentration in the slurry (grams/liter)
RS = relative saturation of calcium sulfate in the liquid phase
The precipitation rate of CaS04 is a function of both total crystal sur-
face area and relative saturation. At steady state the precipitation rate must
equal the oxidation rate. From an operating standpoint, it is desirable for
the relative saturation to be kept below 1.3 to avoid scaling of the scrubber
internals. In a design situation, the maximum relative saturation can be
limited by designing an appropriately sized reaction vessel. A large tank
volume will give an equivalent precipitation rate at a lower relative satura-
tion. However, in a retrofit situation, as exists at Widows Creek, the tar.k
volume is fixed, and either crystal surface area or relative saturation will
vary in order to provide the necessary precipitation rate.
The relative saturation of gypsum in the venturi hold tank was approxi-
mately 1 during most of the forced-oxidation testing at Widows Creek. This
indicates that a relatively small driving force (RS-1) was sufficient to keep
the crystallization rate equal to the oxidation rate. This result might have
been anticipated because the available gypsum crystal surface area (afC) is
high. The system may have benefited from the significant increase iu agita-
tion over nonforce-oxidized systems. The increased agitation may increase
available crystal surface area by shearing existing crystals into smaller
ones. Since a relatively small driving force Was sufficient to keep the
precipitation rate to equal the oxidation rate, this indicates that
precipitation was not a rate-limiting step.
385
-------
Oxygen Transfer
In a process consisting of several steps, the rate of the limiting step
is essentially.equal to that of the overall process. In this case, if the
oxygen transfer rate is assumed limiting and can be modeled, then the rate of
the overall process will be known. Furthermore, this analysis can determine
which are the significant variables affecting this critical path. This infor-
mation will allow accurate decisions to be made with respect to improving the
sulfite oxidation step and perhaps yield data useful in applying forced oxida-
tion to limestone scrubbers of different configurations. The analysis approach
and results are presented in the following section.
Data Analysis
The preceding discussion of chemistry and mass-transfer theories suggests
several variables that may have an effect on oxidation rates, since the reaction
is apparently mass-transfer limited, and the variables that most directly affect
mass-transfer should be of primary importance. These would include the partial
pressure of oxygen, the concentration of HSOs, and mass-transfer coefficients.
Also, pH, temperature, and ionic strength could be expected to affect the oxi-
dation rate because of effects on oxygen.solubility and sulfite concentration.
Table 5 presents some of the variables calculated in order to model the
forced oxidation process. The oxidation rate (R) calculation and the net air
stoichiometry (AS) calculation account for a 32 percent "baseline" or "natural"
oxidation that is not due to air sparging. The gross air stoichiometry (FAS)
calculation implicitly assumes 100 percent forced oxidation and no baseline
oxidation. Consequently, AS will be a somewhat higher number than FAS. FAS
is the air stoichiometry usually presented in the literature when forced
oxidation is discussed.
Two general types of models were examined, mass transfer and overall
oxidation rate. The first type involved correlating measured and calculated
variables with the corresponding calculated mass transfer coefficient (K a).
O \
The second type involved correlating measured and calculated variables with
the calculated oxidation rate (R).
Several data fits were made based on these models, in order to determine
the most accurate representation of the Widows Creek data. The results of the
correlations of the data obtained during the demonstration are inadequate to
model the rate of the oxidation process. This is not surprising when the
nature of the data is considered.
Attempts to model the forced-oxidation system at Widows Creek were compli-
cated by several problems. Modeling the reaction was not a program objective.
and, therefore, no attempt was made during demonstration testing to vary impor-
tant parameters systematically over a broad range. Also, there was no reliable
means of monitoring continuously the flue gas flow rate due to high grain load-
ings which plugged probes. Rather this was estimated by correlating fan amps
386
-------
Table 5. Variables and Units -.Oxidation Modeling
R
Variable
Oxidation Rate
("Forced")
Units
Ib-mole S
min
Source
Ox Net Solid-Phase
K a Mass-transfer
6 Coefficient
AS Net Air Stoichiometry
FAS Gross Air Stoichiometry
COi Partial Pressure 02
(log man)
POj In Partial Pressure Oi
(bottom of lank)
POi out Partial Pressure Oj
(top of tank)
mole Z
Uol
moles S0t
Ib-roole 0; F
•in-f>-(psi driving force) I (tank.
es, SO3 + moles SO.,
Air x (0.791)(scfm)
x 100"j ln 3Olid-phase analysis
Ib-atoa 0 in
Ib-mole S force-oxidized
Ib-atom O in
Ib-mole S removed
psi
psl
psi
volume f3) x Avg Pressure (psi)
[[Air x 0.209) scfa x 2 atom/mole Oa "1
379 f'/lb-mole x R Ib-mole S/min J
[(Air x 0.209) scfm x 2 atom/mole Os"|
Fliie Gas (scfm) x AS022 removal rate
calculations, oxidation rate calculations, and. air Stoichiometry calculations
and, as a result, errors or inaccuracies in the flow rate measurement impact
these other variables as well. Therefore, accurate flue gas flow rate measure-
ments are critical if an accurate model of oxidation rate is desired.
Oxygen partial pressure was not measured directly, and oxidation rate
could not be measured directly. Because of this, oxygen partial pressure, oxi-
dation rate, air Stoichiometry, and mass-transfer coefficients were calculated
from a mass balance involving flue gas flow rate, oxidizing air flow rate, SOa
removal, and oxidation. Because all these factors could only be calculated
from the same base data, it was not possible meaningfully to correlate them
against each other.
Figure 9 is a plot of solid phase oxidation versus air Stoichiometry.
While the rate does not appear to vary with Stoichiometry in any consistent
manner, it is possible to draw some conclusions concerning net oxidation. It
shows when gross air Stoichiometry was 2.0 or greater, net oxidation was con-
sistently 95 percent or greater. Thus, an air Stoichiometry of 2.0 to 2.1
could be considered a conservative guideline for this agitator/sparge ring
configuration.
387
-------
z
o
X
o
UJ
I
a.
o
CO
cr-
UJ
o
2
100.0
97.5
95.0
92.5
90.0
87.5
85.0
82.5
AAAAA A A A A
A A A A
A A
A A A A
A A
A A
A A
I
I
I
I
1.0 1.2 1.4 1.6 • 1.8 2.0 2.2 2.4 2.6 2.8
GROSS STOICHIOMETRY (FAS). Ib-atom O/Ib-mole S removed
Figure 9. Solid Phase Oxidation vs Air Stoichiometry
The analysis in the preceding section suggests that the oxidation reac-
tion is 02 mass-transfer limited. Although the data set was not suited to for-
mulating a generalized model, mass-transfer principles identify several factors
that will affect the process. Air sparging rate and agitation are both very
important mass-transfer parameters, but increases in either parameter will
result in increased capital and operating costs.
388
-------
RESULTS AND CONCLUSIONS
The following results and conclusions were reached at the conclusion of
this demonstration program. The process areas covered include system chemis-
try, the oxidation reaction, and oxidation and dewatering equipment. Most of
these conclusions are specific to dual-tank air sparging at Widows Creek.
The reaction appears to be limited by the rate of oxygen mass trans-
fer. Neither bisulfite dissolution, reaction kinetics, nor gypsum
precipitation is the rate-limiting step.
• During testing with air sparged in both hold tanks (C series), S02
pickup in the scrubber was the primary source of bisulfite ions for
the oxidation reaction. When only the venturi tank is sparged,
solids dissolution is a necessary link in the reaction pathway,
though not necessarily a rate limitation.
• During dual-tank oxidation tests, bisulfite ion availability was not
a rate-limiting factor, and the reaction was relatively insensitive
to pH. The pH should be maintained below 6.5 to keep sufficient
bisulfite available for the reaction,
S02 absorption in the scrubber is sufficient to maintain the pH in
the proper range when oxidation, is performed within the scrubber
loop.
• Forced oxidation did not cause gypsum scaling. Gypsum relative satu-
ration was only slightly greater than one during these tests, due to
high availability crystal surface area. Gypsum crystals existing on
packing will tend to grow in both oxidized and unoxidized systems,
resulting in scaling of the packing. Forced oxidation will not
eliminate maintenance or cleaning requirements for scrubber
internals.
• Forced oxidation nearly eliminated solid calcium sulfite in the FGB
waste at Widows Creek. For solids samples that met the 95 percent
oxidation criterion, mean calcium sulfite was less than I percent by
weight.
• The gypsum particles fell primarily in the 20 to 100 pm size range.
• The results of testing could not be modeled accurately, largely
because of inaccurate estimates of the flue gas flow rate.
• An air stoichiometry of 2.0 Ib-atoms 0/mole S02 absorbed provided
consistent sulfite oxidation of >95 percent.
Calculated thickener unit-area requirements for oxidized sludge are
1.7 square feet per ton per day of solids or less. A more conserva-
tive figure should be used in design. Settling-test results showed
the compression point was 40 percent solids or greater.
389
-------
• The small-scale thickener at Widows Creek was underloaded, and had
an unreliable rake mechanism. Its operation does not necessarily
predict the operation of a full-scale unit.
• The filter-sizing criterion for Widows Creek oxidized sludge is
approximately 200 pounds of solids per hour per square foot of clott
based on industry experience.
• The small-scale filter at Widows Creek was underloaded and its opera
tion does not necessarily predict the operation of a full-scale unit
• Filter cake product of 75 to 85 percent solids was attained with
forced oxidation.
• Forced oxidation did not significantly affect 863 removal.
pH is too insensitive to dissolved limestone concentrations to be an
effective control point for limestone feed rate.
• Forced oxidation can result in increased total dissolved solids (IDS)
in the scrubbing liquor as a result of water reuse.
• The effects of increased TDS on the Widows Creek system are not pres-
ently known and should be studied prior to applying forced oxidation
to these units.
REFERENCES
1. Dean, John A., Editor. Lange's Handbook of Chemistry, Eleventh Edition.
McGraw-Hill.
2. Levenspiel, Octave. Chemical Reaction Engineering, 2nd Ed., New York.
John Wiley and Sons. 1972.
ACKNOWLEDGEMENTS
The oxidation work was funded in part by the Environmental Protection
Agency, Industrial Environmental Research Laboratory, Research Triangle Park,
North Carolina.
Portions of this paper are taken from the Combustion Engineering Environ-
mental Systems Division's Final Report, "Demonstration of Forced Oxidation at
TVA's Widows Creek 8 FGD System," dated December 7, 1979, and Radian Corpora-
tion's Final Report, "Forced Oxidation Demonstration and Testing at Widows
Creek Steam Plant," dated July 18, 1980. Both corporations were under contract
for various phases of involvement pertaining to this TVA R&D project.
The contents of this paper do not necessarily reflect the view and poli-
cies of the Tennessee Valley Authority, nor does mention of any trade names or
commercial products constitute endorsement or recommendation for use.
The authors wish to express their appreciation to the following persons
without whom this report would not be possible: Jose DeGuzman, Jim M.
Cummings, and the Widows Creek Steam Plant Management and Staff.
390
-------
LA CYGNE STATION UNIT NO. 1
WET SCRUBBER OPERATING EXPERIENCE
by
Richard A. Spring
Superintendnent of Air Quality Control
La Cygne Station
Kansas City Power & Light Co.
In the late 1960's, Kansas City Power & Light Company and Kansas
Gas & Electric Company entered into a joint venture to construct
an 800 MW coal fired generating unit. An east central Kansas
location was selected for its ample coal reserves and adequate
water supply capabilities.
The coal reserves proved to be a low grade with an average of
5k percent sulfur and 24 percent ash. To make this coal an
acceptable boiler fuel a large scale air quality control sys-
tem was required. After considerable pilot testing on a
smaller generating unit burning similar coal, a venturi -
absorber scrubber using limestone as the scrubbing agent was
selected. Construction of the La Cygne Station Unit #1 started
in April 1969 and began commercial operation in June 1973.
This paper presents a review of the operating experiences,
0 & M cost trends, availabilities, modifications, manpower
and other supportive data relating to this limestone scrubber
system.
391
-------
STATION DESCRIPTION
The 820-megawatt La Cygne No. 1 Unit began commercial opera-
tion on June 1, 1973, as a joint project of Kansas Gas and
Electric Company and Kansas City Power and Light Company.
The companies share equally in ownership and output, and
the unit is operated by KCP&L. The 630-megawatt No. 2 Unit,
in service since being declared commercial May 15, 1977,
operates under an identical arrangement.
The plant site is located about 55 miles south of downtown
Kansas City, one-half mile west of the Missouri State line,
and was selected based on locally available coal, water,
and limestone. Construction of No. 1 Unit began in 1969
and erection of the Air Quality Control System was initiated
in mid 1971.
Water for cooling purposes is furnished from a 2,600 acre
reservoir constructed adjacent to the plant site. Fly ash
and spent slurry from the AQC system is piped to a 300 acre
settling pond located east of the reservoir.
Coal is delivered to the plant in off-the-road 120 ton
trucks from surface mines operated by the Pittsburg & Midway
Coal Mining Co. The nearby coal deposits are estimated
to contain 70 million tons. The fuel is low grade,
sub-bituminous with an as-fired heating value of 9,000 to
9,700 Btu/lb. and an ash content of 25 per cent and sulfur
content of 5 to 6 per cent (Table 1).
Limestone is obtained from nearby quarries and delivered
to the plant in off-the-road 50 ton trucks.
The boiler for No. 1 Unit is a cyclone-fired, supercritical,
once-through, balanced-draft Babcock & Wilcox unit, with
a rating of 6,200,000 pounds of steam per hour, 1,010
degrees F, 3,825 psig at the superheat outlet. The turbine-
generator was supplied by Westinghouse and is rated at 874
MW gross output with five per cent overpressure and 3,500
psi throttle pressure. Three auxiliary, oil-fired boilers
are used for plant start-up or for powering a 20 megawatt
house turbine-generator. The net plant output is 820 mega-
watts, adjusted to include 24 megawatts used by the FGD
system and 30 megawatts by plant auxiliaries.
FGD SYSTEM DESCRIPTION
The La Cygne wet limestone FDG system (AQC) consists of
eight venturi-absorber modules, connected together by a
common inlet and outlet plenum, capable of treating
2,760,000 ACFM of boiler flue gas at 285°F. The ductwork
design is such that flue gas cannot bypass the system, but
392
-------
Table 1
LA G.YGNE STATION
COAL AND ASH ANALYSIS
COAL
Proximate
Volatile
Fixed Carbon
Ash
Moisture
BTU/lb.
28.63
37.94
24.36
9.07
100.00
9421
Ultimate
Moisture
Carbon
Hydrogen
Nigrogen
Chlorine
Sulfur
Ash
Oxygen
3.60
51.93
3.43
0.94
0.027
5.39
24.36
5.33
100.007
Grindability
59.59
Analysis
Phosphorous Pentoxide
Silica
Ferric Oxide
Alumina
Lime
Magnesia
Sulfur Trioxide
Potassium Oxide
Sodium Oxide
Titania
Other
ASH
0.15
46.05
19.23
14.07
6.86
1.02
7.85
2.48
0.60
1.02
0.67
100.00
Fusion Temperature
Reducing I.D. 1957
Soft (H=W) 2045
Soft (H=W/2) 2169
Fluid 2321
Oxidizing I.D. 2156
Soft (H=W) 2338
Soft (H=W/2) 2415
Fluid
2520
393
-------
each individual module can be isolated from the system
for maintenance.
The on-site limestone processing facility is composed of
two 110 ton/hr. wet ball mills and two 260,000 gal. storage
tanks, capable of supplying up to 1,000 tons of slurry
per hour from 3/4 in. x 0 in. limestdne. This slurry
is processed to consist of 20 per cent solids by weight.
The unit is a balanced draft system with three 7000 hp
forced draft fans and six induced draft fans located
between the AQC and the 700 foot stack.
The spent slurry and fly ash is removed from the module
recirculation tanks thru rubber lined pipe to a 300 acre
settling pond at a rate of 3000 to 3500 tons per day.
Clear make-up water is pumped from this pond for slurry
make up, sump pump operation, and wash water thus allowing
a closed loop operation. (Figure 1)
The hot boiler flue gas first enters the venturi section
(Figure 2) of the module and is sprayed with limestone
slurry in a concurrent manner from 48 spray and 32 wall
wash nozzles. This results in agglomeration of up to 99
per cent of fly ash particles which is collected in the sump
below. The flue gas then makes a 180 degree turn up
through two layers of stainless steel sieve trays upon
which slurry is sprayed from 24 spinner vane nozzles.
At this point the SC>2 in the flue gas and the calcium
carbonate in the slurry react to form two relatively
insoluable salts, calcium sulfite and calcium sulfate,
which also fall to the sump. The scrubbed flue gas then
passes thru a series of demisters and is then reheated
before entering the induced draft fans.
OPERATING EXPERIENCE
As a result of the continuing modifications and improved
operating procedures, the module availabilities have
steadily improved. The annual averages (Table 2) have been
31% for 1973, 76.3% for 1974, 84.3% for 1975, 92% for 1976,
92.5% for 1977, 93.5% for 1978, and 95.1% for 1979. With
the addition of the eighth module in April 1977, continuous
daytime load capability has exceeded 800 megawatts without
appreciably affecting average module capability.
The results of a full load and stack emissions test on
August 26, 1977, (Table 3) indicated module gas flow was
still below crusing capability, the induced and forced
draft fans were loaded well beloW rating and most systems
were in good balance. Sulfur dioxide removal efficiency
394
-------
«D
.* C^y^its [-• r-'•*""
-------
LACYGNE FGD MODULE
4500
PPM
REHEAT
STEAM
550
1000 PPM '
x-*- S02
' ' 190° f
REHEAT COILS 1
(0 0 Q O Q O
J*
X
^
s
s
~\
-43"
H20
TO
FAN
HOT AIR
~ FROM
BOILER
INTERMITTENT
OVERSPRAY
2150 GPM
CONTINUOUS
UNDER SPRAY
140 GPM
VENTURI SPRAY
WALL
WASH
SPRAY
/\ /\ /\
V V V V V V V
VENTURI
THROAT
FLUSH
PREDEMISTER
ABSORBER
SPRAY
1_JL i_l __-_
ABSORBER
200-600 GPM
RECIRCULATION TANK
70 C/L
CASO3 35 C/L
CASO4 " 25 C/L
FLYASH 40 C/L
PH = 5 5- 6.0
8- 10% SOLIDS
SPENT SLURRY
TO POND
700 GPM
*3500 TONS/DAY
693000 TONS/ "EAR
453 ACRE FEET'YEAR
VENTURI
^ECIRC. PUMP
5000 GPM
ABSORBER
RECIRC. POMP
9000 GPM
FOR ALL MODULES
Figure 2
396
-------
Table 2
MODULE AVAILABILITY SUMMARY - 1973
MONTH
JANUARY
FEBRUARY
MARCH
APRIL
MAY
JUNE
JULY
AUGUST
SEPTEMBER
OCTOBER
NOVEMBER
DECEMBER
A
20
7
79
13
28
48
42
B
21
24
64
0
41
1
20
C
40
25
65
13
34
38
5
D
21
41
74
13
54
4
31
E
27
27
47
13
33
63
26
F
30
25
48
0
3
59
11
G
23
31
70
0
46
49
32
AVERAGE %
AVAILABILITY*
26
26
64
7
34
37
24
31%
NET MWH
87,529
90,669
250,319
20,073
117,106
104,255
61,013
BOILER
HOURS
294
303
699
95
452
463
339
GENERATION
LOAD FACTOR
15.2
15.2
42.1
3.5
19.7
18.1
10.3
17.7%
co
UD
*MODULE HOURS
HOURS IN MONTH
-------
Table 2 (Cont'd)
MODULE AVAILABILITY SUMMARY - 1974
MONTH
JANUARY
FEBRUARY
MARCH
APRIL
MAY
JUNE
JULY
AUGUST
SEPTEMBER
OCTOBER
NOVEMBER
DECEMBER
A
49
66
67
69
92
75
90
69
71
90
B
32
68
70
83
84
80
90
88
61
71
C
44
59
75
78
83
80
73
73
59
60
D
87
76
88
85
90
81
81
76
81
61
E
23
52
74
78
82
85
81
83
79
84
F
37
100
100
84
83
79
78
89
93
85
G
81
65
88
80
87
77
99
86
89
84
AVERAGE %
AVAILABILITY*
50
69
80
80
86
80
85
81
76
76
76.3%
NET MWH
35,862
85,256
83,880
157,949
185,473
110,122
231,382
209,127
230,302
130,128
BOILER
HOURS
364
364
332
500
480
313
571
606
662
386
GENERATION
LOAD FACTOR
6
16
15
27
32
19
39
36
39
23
25%
CO
l£>
00
*MODULE HOURS
BOILER HOURS
-------
Table 2 (Cont'd)
MODULE AVAILABILITY SUMMARY - 1975
YIONTH
JANUARY
A
FEBRUARY
S1ARCH
kPRIL
YlAY
JUNE-
JULY
AUGUST
SEPTEMBER
OCTOBER
SIOVEMBER
DECEMBER
82.4
94.6
87.8
7 8 . 4
74.5
78.4
66.2
92.9
90.7
B
Turbii
Turbii
96.0
Generc
85.1
85.4
89.7
88.1
83.6
77.3
Gener<
90.8
Generc
87.4
C
le Gene
le Gene
89.5
itor Re
94.2
83.9
89.6
87.3
.84.4
46.3
ator Re
80.2
ator Re
80.9
D
:rator
>rator
76.6
E
Repair
Repair
93.0
:pair 25 Days
89.5
84.9
83.7
78.0
84.7
73.6
?pair '.
93.2
jpair !
85.2
89.8
84.1
85.4
92.4
78.8
71.9
.5 Days
96.1
L7 Days
86.9
F
91.5
89.3
86.1
87.4
85.0
77.8
73.1
89.4
88.6
G
96.0
83.4
88.6
85.2
83.1
74.2
64.7
93.9
83.7
AVERAGE
AVAILABILITY*
89.33
89.4
85.8
85.6
84.07
80.25
67.57
90.83
86.19
84.3
NET MWH
7,886
244,873
23,014
332,526
324,952
297,870
294,402
239,954
74,660
165,058
278,597
BOILER
HOURS
694
683
667
590
630
610
231
346
597
GENERATION
LOAD FACTOR
41.1
3.4
55.9
56.4
50.0
49.5
41.7
12.5
28.7
46.8
38.6
ID
*WORK-ING HOURS + RESERVE
HOURS IN MONTH
-------
Table 2 (Cont'd)
MODULE AVAILABILITY SUMMARY - 1976
MONTH
JANUARY
FEBRUARY
MARCH
APRIL
MAY
JUNE
JULY
AUGUST
SEPTEMBER
OCTOBER
NOVEMBER
DECEMBER
A
85.8
93.9
92.3
92.3
96.5
93.3
95.6
94.1
97.4
94.7
86.8
B
84.6
90.3
89.7
90.5
Schedi
92.1
C
90.7
85.8
88.4
88.7
D
71.8
91.2
93.0
97.1
E
83.9
91.7
94.2
95.8
F
82.3
93.1
91.3
98.0
aled Outage 24 Days
93.5
95.7
89.4
Scheduled Outage 9 Days
94.1
95.0
93.1
94.0
91.9
91.8
95.0
92.9
93.4
92.3
93.0
91.8
95.3
93.5
93.7
90.4
Turbine Repair, Stack Relininc
Turbine Repa
96.7
97.5
iir, St
89.0
:ack Re
96.1
Turbine Repair, Stack Re
93.3
88.5
93.7
81.0
95.3
93.5
94.2
93.6
ilininc
"
96.1
:lininc
91.3
94.7
G
84.3
94.6
91.4
94.8
96.2
90.6
94.0
87.6
AVERAGE
AVAILABILITY*
83.3
91.5
91.5
93.9
94.1
93.3
93.7
91.7
8 Days
30 Days
96.1
18 E
93.6
91.4
95.6
ays
94.0
89.9
92.0
NET MWH
301,641
308,361
337,468
76,810
223,048
320,701
359,028
275,014
88,925
342,236
358,338
BOILER
HOURS
620.5
594.5
643.0
143.0
436.3
656.0
688.3
521.0
255.8
626.8
706.3
GENERATION
LOAD FACTOR
50.6
55.4
56.7
13.3
37.5
55.7
60.3
46.2
14.9
59.4
60.2
46.4
O
O
^WORKING HOURS -I- RESERVE
HOURS IN MOiri'H
-------
Table 2 (Cont'd)
MODULE AVAILABILITY SUMMARY - 1977
MONTH
JANUARY
FEBRUARY
MARCH
APRIL
MAY
JUNE
JULY
AUGUST
SEPTEMBER
OCTOBER
NOVEMBER
DECEMBER
A
94.2
93.4
94.0
96.1
95.0
88.9
93.2
90.7
93.1
B
90.0
93.0
92.2
93.7
C
95.0
92.6
85.9
97.0
D
95.1
93.8
94.3
94.2
E
94.5
93.3
91.4
95.2
Generator Repair And
Stack Relining - 63 Days
92.8
55.2
93.7
95.6
96.3
Turbii
94.4
93.2
89.1
89.3
93.4
ie Rep a
94.8 J94.6
93.1 J89.7
90.0
94.2
94.2
ir Nov
92.8
93.4
92.2
. 15 -
F
91.6
93.9
94.0
96.1
94.9
92.8
95.0
93.5
92.5
Dec.
G
89.8
88.0
90.1
94.5
95.4
92.9
91.7
88.5
95.5
25
H
95.4
93.3
93.0
93.0
95.1
AVAILABILITY*
92.9
92.5
91.7
95.2
94.6
87.4
92.3
92.3
94.0
92.5%
NET MWH
255,822
310,748
295,420
178,226
213,334
253,605
287,701
173,979
118,439
BOILER
HOURS
539
590
558
384
485
501
524
457
234
LOAD FACTO]
43.0
57.8
49.6
30.9
35.8
42.6
49.9
29.2
20.6
39.9
*WORKING HOURS & RESERVE HOURS
HOURS IN MONTH
-------
Table 2 (Cont'd)
MODULE AVAILABILITY SUMMARY - 1978
MONTH
JANUARY
FEBRUARY
MARCH
APRIL
MAY
JUNE
JULY
AUGUST
SEPTEMBER
OCTOBER
NOVEMBER
DECEMBER
A
90.2
92.4
95.3
91.4
88.9
87.9
92.1
96.1
95.9
91.7
93.9
B
94.8
93.4
95.2
92.1
91.5
OUTAG
97.2
92.5
96.0
95.5
94.9
92.9
C
94.6
95.1
90.4
92.8
91.6
D
95.1
94.3
95.4
90.8
93.1
E
93.4
90.6
94.4
90.2
91.5
F
93.5
96.9
94.7
91.8
90.6
E 6-8-78 thru 7-17-78
91.9
95.0
96.3
98.3
94.3
94.0
93.9
95.7
95.8
97.0
93.3
95.0
88.4
92.7
95.9
97.0
93.6
94.7
92.8
94.3
95.7
97.6
93.0
90.5
G
94.4
95.5
88.6
90.6
93.1
93.1
94.7
95.3
96.7
94.3
94.4
H
94.0
93.4
93.3
90.5
85.6
95.3
95.3
96.6
96.3
96.1
94.7
AVAILABILITY*
93.8
94.0
93.4
91.3
90.7
92.6
94.0
96.0
96.8
93.9
93.8
93.5
NET MWH
332,033
334,897
264,961
330,571
291,651
160,847
307,378
390,826
138,126
386,402
91,744
BOILER
HOURS
582
594
593
620
582
14
340
579
720
255
720
239
GENERATION
LOAD FACTOR
54.2
60.5
43.2
55.7
47.6
0
26.2
50.1
65.9
22.5
65.1
15
42.2
*WORKING HOURS & RESERVE HOURS
HOURS IN MONTH
-------
Tabi-e 2 (Cont'd.)
MODULE AVAILABILITY SUMMARY - 1979
MONTH
JANUARY
FEBRUARY
MARCH
APRIL
MAY
JUNE
JULY
AUGUST
SEPTEMBER
OCTOBER
NOVEMBER
DECEMBER
A
95.6
95
96.1
95.5
96.5
86.8
96
95.3
B
96.5
94.6
96.0
95.7
96.3
Outag
95.9
96.1
95.8
C
97.2
92.6
93.2
94.4
96.7
e May
96.3
95.6
94.7
Outage Oct<
I
D
96.3
93.5
95.6
91.4
95.3
26 - A
96.3
94.3
92.7
3ber 19
E
90.7
95.1
96.5
95.5
95.4
ugust
95.9
96.7
94.4
- Dec
F
97.2
94.3
94.8
96.2
95.7
16
96.2
96.1
94.9
. 31
G
97.2
94.1
95.7
95.9
96.3
88.5
96.0
94.7
H
95.4
93.8
93.4
95.7
95.5
96.9
96.9
94.5
SYSTEM
AVAILABILITY*
95.8
94.1
95.2
95.0
96.0
94.1
96
94.6
95.1
NET MWH
46,538
141,322
147,645
342,240
222,924
83,169
321,108
207,639
BOILER
HOURS
205
342
314
638
452
230
618
455
GENERATION
LOAD FACTO
7.82
26.29
24.81
59.43
37.45
13.97
55.75
34.89
32.55
o
CO
*WORKING HOURS & RESERVE HOURS
PERIOD HOURS
-------
Table 2 (Cont'd)
MODULE AVAILABILITY SUMMARY - 1980
MONTH
B
D
SYSTEM BOILER GENERATION
H AVAILABILITY* NET MWH HOURS LOAD FACTOR
JANUARY Outage January 1 - February 20
FEBRUARY 98.2 98.2 97.4 99.1 98.2 99.1 99.1 99.6 98.6
MARCH 94.6 96.2 96.1 96.1 95.8 94.7 93.3 95.7 95.3
APRIL 96.3 95.1 95 96.7 95.3 92.5 97.0 97.0 95.6
MAY 96.4 94.6 95.7 95.9 96.1 96.5 96.1 96.8 96.0
JUNE 98.2 98.0 97.4 98.1 98.1 98.3 98.7 99.3 98.3
JULY
AUGUST
SEPTEMBER
OCTOBER
NOVEMBER
DECEMBER
52,768 157 9.48
1,187 32 2.00
206,936 472 35.93
324,478 689 55.19
195,974 370 34.02
*WORKING HOURS & RESERVE HOURS
PERIOD HOURS
-------
o
01
Table 3
LA CYGNE STATION UNIT NO. 1
FOUR HOUR FULL LOAD & STACK EMISSION TEST
DATE: August 26, 1977
TIME: 11:00 A.M. - 12:00
LOAD RANGE: 800 + MW
AMBIENT TEMP: 94° F
MODULES A B
GAS FLOW INDICATED
THROAT POSITION
REHEAT TEMPERATURE
VENTURI P
REHEATER P
ABSORBER DEM. P
REHEAT OUTLET
DAMPER POS.
ID FAN AMPS
ID FAN INLET
DAMPER POS.
FD FAN AMPS
LAB pH
SULFITE g/1
CARBONATE g/1
S02 EFFICIENCY %
INLET (PPM)
OUTLET (PPM)
400
OPEN
170
5
2.5
6.5
50
380
42
490
5.45
60.4
50.3
80.0
4600
920
350
OPEN
190
5.5
5.5
5.5
100
420
42
470
5.7
72.4
75.6
82.1
4600
825
NOX EMISSION: 0.81 # mm BTU
Midnight AVERAGE SO2 REMOVAL: 77%
PARTICULATE EMISSION: .213 # mm BTU
C D E F G H
380
OPEN
150
5
4.5
10
96
380
32
430
5.55
101.0
53.1
74.9
4600
1150
400
OPEN
190
5
4-5
7.5
38
400
36
5.7
74.1
54.4
64.3
4600
2285
352
OPEN
185
5
5
7.0
100
470
36
5.58
70.0
59.4
76.4
4600
1085
380
OPEN
180
5
2.55
6.5
52
470
40
5.77
43.9
83.8
72.1
4600
1285
370
OPEN
160
5
4.5
8.0
100
(540
366
OPEN
170
5
5.5
7.0
100
MAX)
( % OPEN)
(540
5.72
43.9
68.1
73.1
4600
1235
MAX)
5.29
63.6
42.5
74.8
4600
1160
-------
Table 3 (Cont'd)
CONDENSER VAC (IN. HG) 2.5
WINDBOX FURNACE DIFF. PRESS (IN.H20) 32_
SCRUBBER OUTLET PRESS (IN.H20) -39"
FURNACE PRESS (IN.H20) -2
F.D. FAN DISCHARGE (IN.H2O) 41
PEND. REHEAT GAS PRESSURE (IN.H2O) -5
AIR FLOW (%) 85
BOILER EXCESS O2 (%) 2.2
BAROMETRIC PRESSURE (IN.Hg) 29.01
STACK GAS TEMP (°F) 209
FLUE GAS MOISTURE (%) 13.66
STACK GAS VELOCITY Ft/Sec 103.15
PRIMARY SUPER GAS PRESS. (IN.H2O) -8
HORZ REHEAT GAS PRESS. (IN.H2O) -9.5
ECON OUTLET GAS PRESS. (IN.H2O) -11.5
FEEDWATER PRESSURE (PSI) 4200
THROTTLE PRESSURE (PSI) 3400
THROTTLE TEMP. (°F) 1000°
HOT REHEAT TEMP. (°F) 1300
FUEL FLOW % 68
FUEL HEATING VALUE (MTB) 9800
FLUE GAS VOLUME (MCFM) 2998
STACK CO2 % 13.4
STACK O2 % 5.4
O
en
-------
averaged 77% with individual modules averaging from 65 to
80%. Although particulate emissions from the plant have
met EPA and Kansas State requirements, research and
development work continues in an endeavor to reduce further
the particulate emissions from Unit #1.
Limestone utilization has greatly improved with improved
pH control. In the past, it has been almost insurmountable:
to maintain inline glass cells without caking the limestone
during shutdown or abrading the cells during operation
with the high concentration of fly ash. By centralizing
the pH monitoring equipment and backflushing the pH cells
with water for 5 minutes every eight hours "straight
line" pH is resulting in approximately 30% less limestone,
better control of scaling and has eliminated one more
variable which hinders analysis in other areas.
Demister pluggage or scaling is no longer a problem at
La Cygne. By eliminating the intermittent wash and moving
the continous wash (140 GPM) from below to above the first
demister with increased number of nozzles (230 GPM),
the chevrons operate "squeaking clean". Further experi-
mentation may allow a reduction in these nozzles and per-
haps sequential washing to reduce excess water.
Hard scale on the reheater tubes has been eliminated by
the addition of a second layer of demisters in each of the
modules. Scaling of the reheaters continues to be a pro-
blem, however, it is soft and can be removed using fire
hoses. The previous hard scale required high pressure
water to remove the deposits.
MAINTENANCE
Cleaning schedules continue to call for taking one module
out of service each night on a rotational schedule and
keeping all modules available for the daytime peak loads.
This allows a complete checkout of module internals to
clean steam reheater pluggage, check nozzles for debris
or loose rubber pluggage, to clean sump accumulation and
to inspect for any other maintenace that could reduce
reliability during the week. Module inspection and
cleaning is now reduced to six hours or less with re-
heater pluggage the greatest problem. Scaling is not one
of our chief problems and we ordinarily ignore soft scale
that forms on walls, on beams, or on the outside of
nozzles.
Carryover to the induced draft fan blades continues to
require regular washings. Each fan now requires cleaning
once very four to seven days. A "spinning" process
407
-------
with low pressure hoses has been very effective for cleaning
the spare fan while out of service. The washings are
usually done on a preventative basis, but must be taken
out of service if bearing vibrations exceed 12 mils.
Rubber pipe linings and rubber-lined pumps have been an
increasing maintenance problem. After several years
operation, some materials that haven't been modified are
wearing out. Rubber linings that tear out cause damage
in other piping or pumps, plug nozzles and allow the steel
pipes to wear through. This problem would not have been
classified as serious, but this very abrasive slurry in
practically continuous operation can be detrimental in
trying to attain higher module availability, so a preventa-
tive maintenance program to change the piping in critical
areas has been initiated.
Corrosion of carbon steel in the ductwork, dampers, in-
duced draft fan rotors and housings, breeching and stack
liner is and will continue to be our greatest concern.
A replacement program has been underway since the fall of
1979. This program began at the outlet of the demister
section where the walls from this point to the reheater
section were replaced with 1/4" 316L stainless steel.
In the reheater section replacement continued with the
duct from the reheat bundles to the module outlet dampers
being replaced with a coated carbon steel. New module
outlet dampers have also been installed. Future plans
include replacement of all the ducting from the module
outlet dampers to the induced draft fans, the induced
draft fans inlet and outlet dampers, induced draft fan
housings, and the ductwork from the induced draft fans
to the stack.
MANPOWER REQUIREMENTS
The scrubber operating and maintenance force has been
increased to 54 people by adding one electrician for a
total of two and a maintenance foreman to supervise both
electricians and technicians. The remaining personnel
will remain the same (Table 4).
Also worth noting are the increased demands on present
maintenance personnel to accumulate, record and evaluate
operating data on water saturation trends,limestone
utilization, draft fan wear rates, reheater bundle failures,
lined pump failures, rubber lined pipe replacements,
nozzle replacements, spare parts, etc. for preventative
maintenance programs. The operators are also busy up-
dating and extending operating instructions, special
instructions and reviewing safety and training procedures.
408
-------
Table 4
LA CYGNE AIR QUALITY CONTROL
MANPOWER REQUIREMENTS
OPERATORS PER SHIFT
3 Attendants 13
3 Clean-up 14
1 Shift Foreman 5
1 Process Attendant (Chemist) 1
33
MAINTENANCE
Mechanics 8
Apprentice Mechanics 2
Welder 1
Electrician 2
Technician 2
Plant Helpers 2
Foreman 2
19
ADMINISTRATIVE
Superintendent 1
Engineer 1
TOTAL 54
409
-------
COSTS
The total cost of the FGD system to date has increased to
$55.1 million or 25% of the $216.3 million total Unit #1
cost.
The production costs for the La Cygne FGD system (Table 5)
in 1977 was 1.7 mills/KWH and for 1979 it was 4.9 mills/
KWH. This drastic rise is due to the increase in main-
tenance materials to repair the "cold end" corrosion
mentioned earlier. Discounting escalation, future
production costs associated with the operating labor,
operating materials, and maintenance labor should re-
main the same or trend downward while the maintenance
materials will increase slightly.
CURRENT PROGRAMS AND PROJECTS
The major project concerning the FGD system, at present,
is the systematic replacement of the cold end duct
work mentioned previously. To help combat this corrosion
problem, studies are continuing for increasing the reheat
steam supply.
Analytical programs to investigate the mechanics involved
with the scale formation at the various levels of the
scrubber modules and collection of sub-micron flyash
have been implemented. Results of these programs are
currently under scrutinization.
Other areas of hopeful improvement are in the demister
section with the addition of a third layer, and on line
incline reheat tube cleaning.
410
-------
Table 5 - LA CYGNE UNIT #1
FGD SYSTEM OPERATING EXPENSE
OPERATING
LABOR
OPERATING
MATERIALS
MAINTENANCE
LABOR
MAINTENANCE
MATERIALS
LIMESTONE
1973
DOLLARS-MILLS/KWH
$ 162,934 - 0.223
3,480 - 0.005
189,400 - 0.259
441,737 - 0.604
264,514 - 0.362
1974
DOLLARS-MILLS/KWH
1975
DOLLARS-MILLS/KWH
1976
DOLLARS-MILLS/KWH
$ 284,541 - 0.223 $ 601,029 - 0.265 $ 683,939 - 0.229
67,032 - 0.053
401,414 - 0.315
335,486 - 0.263
195,926 - 0.086
416,206 - 0.184
386,397 - 0.171
415,226 - 0.139
358,941 - 0.129
93,292 - 0.031
780,297 - 0.613 1,256,048 - 0.554 1,717,949 - 0.574
TOTAL
1,062,065 - 1.453 1,868,770 - 1.467 2,855,606 - 1.260 3,269,347 - 1.102
OPERATING
LABOR
OPERATING
MATERIALS
MAINENANCE
LABOR
MAINTENANCE
MATERIALS
LIMESTONE
1977
DOLLARS-MILLS/KWH
$ 679,628 - 0.313
253,662 - 0.117
476,724 - 0.219
1,083,167 - 0.493*
1,202,005 - 0.553
1978
DOLLARS-MILLS/KWH
1979
DOLLARS-MILLS/KWH
1980 (Jan. - June)
DOLLARS-MILLS/KWH
$ 755,500 - 0.250 $ 733,016 - 0.485 $ 331,654 - 0.423
453,140 - 0.150
414,355 - 0.137
537,172 - 0.355
561,624 - 0.371
757,951 - 0.251 4,398,066 - 2.90
1,452,792 - 0.482 1,183,169 - 0.782
69,632 - 0.089
295,094 - 0.376
2,399,063 - 3.056
289,265 - 0.368
TOTAI
3,695,186 - 1.695 3,833,738-1.270 7,413.,047 - 4.89
3,384,708 - 4.312
: 6 00,0 00 Pond Dredg J n9
-------
CATIONS
Calcium (Ca)
Magnesium (mg)
Sodium (Na)
Potassium (K)
Table 6
LA CYGNE SCRUBBER WATER ANALYSIS
COOLING SETTLING
LAKE POND
126.4 808.0
16.3 106.0
31.0 52.5
5.1 41.6
ANIONS
Bicarbonate Alk (AS HC03)
Chloride (CI)
Sulfate (SO4)
Sulfite (803)
Silica (Si02)
112.2
44.9
295.2
* ND
1.12
79.3
314.0
1995.1
* ND
52.0
OTHERS
pH (pH UNITS) 7.7
Conductivity in Michromhos 820.0
Solids, Suspended 5.0
Dissolved 610.0
7.5
3500.0
5.0
3450.0
*ND - Not Detected
412
-------
ACKNOWLEDGEMENT
This paper is based upon presentations by:
Mr. C. F. McDaniel
EPA Symposium on Flue Gas Desulfurization (1974, 1976, 1977)
and
Mr. Terry Eaton
EPA Symposium on Flue Gas Desulfurization (1979)
413
-------
ONE BUTTON OPERATION
START-UP OF THE ALABAMA ELECTRIC COOPERATIVE FGD SYSTEM
Royce Hutcheson
Chief Environmental Results Engineer
Alabama Electric Cooperative
Leroy, Alabama
Carl ton Johnson
Product Sales Manager, FGD Systems
Peabody Process Systems, Inc.
Stamford, CT
ABSTRACT
In September of 1978, Alabama Electric Cooperative started up a
limestone FGD system for its 255 MW Tombigbee Station Unit #2,
Leroy, Alabama. Since the start-up, the operating experience of
the system has been extremely successful.
A sophisticated control system has been provided for the FGD
system which permits operation from the control room by means of
a single button. Start-up of the FGD system consisted of pushing
this button. The unit has been on stream since that time.
The FGD system is designed to remove 85% of the S02 in the flue
gas generator from the combustion of 1.8% coal. Under performance
test conditions the absorber gas load and inlet S02 content were
20% and 35% respectively greater than design. Despite the greater
than design conditions a S0£ removal efficiency of 94% was achieved.
A limestone stoichiometry of 1.01 was obtained, probably the lowest
ever achieved in the FGD industry.
After a year of operation, the system has exhibited a high degree of
reliability. Based upon actual measured hours, the system avail-
ability has been 91.6%.
The FGD system for Unit No. 3, a duplicate of Unit No. 2, has recently
been started up. Preliminary results indicate similar performance
to that obtained with Unit No. 2.
It is the purpose of this paper to discuss in detail the process
chemistry, system description and controls which have permitted the
successful operation of this unit.
Preceding page blank
415
-------
INTRODUCTION
Alabama Electric Cooperative's Tombigbee Power Station
is located on the Tombigbee River approximately 70 miles
due north of Mobile, Alabama. The most recent expansion
at this site was the addition of Units No. 2 and No. 3.
Each unit has a rated capacity of 255MW and is designed
to burn Alabama and Kentucky coals with a maximum sulfur
content of 1.8%. To meet the emission standard of 1.2 Ibs.
S02 per million BTU, flue gas desulfurization was required.
In September 1975 Peabody Process Systems was awarded a
contract to furnish a limestone FGD system for both units.
The FGD system for Unit No. 2 was started up in September
1978. Unit No. 3 was put in service July 1979. In com-
missioning both units, start-up was achieved by the
pushing of a single button located in the control room.
The pushbutton start-up was simple. However, the ease
and simplicity of the start-up was not an accident. It
was the result of careful attention to process design,
mechanical design and pre-commissioning check out of the
system. Since the initial pushbutton start-up, superior
operating results have likewise confirmed the importance
of giving proper attention to these design details. In
the sections that follow, the details which contributed
to the success of the Alabama Electric Cooperative System
will be discussed as well as the performance history for
both units.
416
-------
SYSTEM DESCRIPTION
The system design criteria are shown in Table No, 1.
TABLE NO. 1
DESIGN BASIS PER UNIT
Unit Generating Capacity 255MW
Coal Sulfur Content 1.8% S.
S02 Emission Standard 1.2 Ibs, S02/mm BTU
Flue Gas Volume 953,000 ACFM
Percent of Flue Gas Scrubbed 70%
No, of Absorbers 2
No. of Recycle Pumps/Absorbers- 3
Absorber S02 Removal Efficiency 85
Alkali Limestone
Waste Solids Disposal Method Ponding
Reheat Method By-Pass Gas
The flue gas entering the FGD System has been cleaned of
particulate by means of a hot side precipitator, Two I,D,
fans, providing the draft for both the boiler, precipitator
and FGD system, are located ahead of the absorbers, It was
the Owner's preference that two absorbers be used. Each
absorber has a 22* diameter and is designed for 85% S02 removal,
Seventy percent of the gas is scrubhed and 30% is by-passed for
use as reheat. Each absorber consists of six spray banks throug;
which a slurry containing calcium sulfite, calcium sulfate and
unreacted limestone is sprayed countercurrent to the gas flow,
The gas, as a result of being contacted with the slurry, is
cleaned of S02= After leaving the absorption zone, entrained
slurry in the flue gas is removed by means of a two stage mist
elimination section, The first stage is a weeping sieve tray
deluged with a chemically non-reactive slurry produced by means
°f hydroclones The hydroclones are used to classify the
417
-------
absorber recycle slurry by particle size, Unreacted limestone
particles are ten CIO) times larger than the reacted product,
Hydroclones permit removing the unreacted limestone from the
reacted products because of the particle size difference.
Decarbonated slurry is used to deluge the weeping sieve tray
and thus prevent a plugging chemical reaction, Final de-
entrainment, particularly of gas entrained water, is accomplished
in a second stage which is a Chevron type mist eliminator. The
clean gas then leaves the absorber where it is mixed with by-pass
gas to provide reheat, At the ductwork juncture where the by-pass
gas and the scrubbed gas meet, a mixing baffle is used to ensure
a uniform gas temperature prior to entering the stack.
A single limestone preparation system is common to both units.
Limestone rock is crushed on site to approximately a 3/4" size
and stored in a silo. A weigh feeder conveys the limestone to
a ball mill where it is ground to proper size and stored in a
tank, The limestone is fed to the individual units as a 35%
slurry via a recirculation loop. Limestone slurry is fed to
the individual recycle tanks as required.
Per boiler, both absorbers are supported at grade and share a
common recycle tank. The recycle slurry is recirculated from
the recycle tank to each absorber. Each absorber has three
recycle slurry pumps - one pump is dedicated to two absorber ,
spray headers,
Waste slurry overflows from the recycle tank to a waste slurry
sump which also collects all drainage and water used for system
flushing. The waste slurry is then transported from the sump
to a pond in which the solids are allowed to settle. The water
reclaimed from the slurry is recycled back to the FGD system
for reuse. The system operates on a totally closed loop water
balance basis. Fresh water is added to the system to make up
for losses resulting from evaporation and water bound with the
waste solids,
418
-------
FACTORS EFFECTING OPERATING AND MAINTENANCE' COSTS
There are many factors which contribute to the success of an
operating FGD system. The whole is truly the sum of its parts,
and design detail, no matter, hew small, if ignored can adversely
effect system performance, The following areas, which will be
discussed, highlight some of the many features applied to
Alabama Electric Cooperative which have contributed to the
FGD System's successful operating experience.
The following factors can and do effect the operability of
an FGD System:
1. Absorber design
2, Mist eliminator chemistry control
3. Simplicity of design
4, Materials of construction
5. Slurry piping design
6. Adaptability to actual operating conditions
Limestone FGD systems are frequently approached with the idea
that reacting limestone with S0£ is simple high school chemistry.
This is very far from the truth; the chemistry is very complex
One of the unique features of this chemistry is that the reaction
products produced can result in significant scaling and plugging
in an absorber. Consequently, the type of absorber used is
very significant. Industry experience has shown that in many
systems an absorber with complicated internals, for example,
tray type absorbers, packed towers, etc,, offer great potential
for solids to deposit thus hindering the operability of the
absorber. For the Alabama FGD system, a spray tower absorber
was used as a basis for the absorber design, A significant
feature of the spray tower is that its internals are minimal
and thus minimizes the opportunity for solids to deposit,
Selection of the spray tower design significantly increases
the operability of a FGD system, However, the mist elimination
section still provides the complicated surfaces which can create
plugging problems. In any absprber design, de-entrainment is
419
-------
a factor which, must be considered to avoid particulate emission'
through the stack, A standard design mist eliminator can
provide the opportunity for solids to deposit on the surfaces;.
This is attributed to the fact that entrained slurry from the
absorber zone contains unreacted limestone. This unreacted
limestone can then react with the remaining S02 in the flue
gas and cause solids to deposit on the complex surfaces of
the mist eliminator. Attempts are frequently made to avoid this
plugging problem by washing with water. However, a system
designed for a closed loop water balance usually does not have
the necessary quantity of water available under all load conditions
and all sulfur coals to adequately preclude a plugging situation
in the mist eliminator.
An alternate is to control the chemistry by preventing entrained
limestone from reaching the mist eliminator. This is achieved
in the Alabama Electric Cooperative design by using hydroclones
to remove unreacted limestone from the process slurry and then
using the limestone free slurry to provide a liquid barrier
below the mist elimination zone, This liquid barrier prevents
entrained slurry, with, the unreacted limestone, from reaching
the critical mist elimination area, This technique insures
that the mist elimination area operates with, a non-plugging
chemistry regardless of the load or sulfur content of the coal
being burned. It is thus another step in improving the relia-
bility of the system.,
Hydroclones are also used to screen all of the waste solids
prior to discharge from the system. This removes all of the
unreacted limestone from the waste slurry such that the FGD
system efficiency of limestone utilization is almost 100%
Simplicity of design is another important factor which adds
to minimal maintenance. Generally, the fewer the number of
components of a system the less the probability of having
problems. This philosophy has been utilized in the. control
concept for the Alabama Electric Cooperative FGD System.
Control valves in slurry service, which create both abrasion
420
-------
and plugging problems, have been eliminated completely. The
only exception is a small limestone slurry feed control valve.
The elimination of valves is -made possible by employing gravity
overflow where possible. An example of this is the main recycle
tank and wash tank. The quantity of slurry to the spray headers
within the absorber is controlled by turning off pumps rather
than modulating the slurry flow. This eliminates both a control
valve and the plugging of the slurry pipe line which would occur
under low flow conditions.
In slurry services which require operating over a wide range of
flows, various approaches are taken to prevent plugging of the
slurry pipe lines.
Simplicity of slurry piping design is achieved by having the
absorber recycle slurry pumps feed a dedicated spray header
system. Thus, two levels of spray banks are dedicated to a
single recycle pump. Pipe manifolds and isolation valves in
the discharge pipe of the recycle pump are eliminated. Regu-
lation of slurry flow to the absorber is effected on a step-
wise basis by turning individual pumps on or off as required to
meet emission standards based upon the actual sulfur content of
the coal being burned, This concept also eliminates plugging
problems due to the creation of dead pockets in slurry pipe
systems and abrasion problems of valves in the discharge piping.
In the limestone feed system, limestone is circulated via a
distribution loop such, that regardless of how much slurry is
required by the FGD system, (Q to 100% of design) the lime-
stone transfer system will always have velocities sufficient
to prevent settling of solids and the resulting plugging which
would ensue. Likewise, in the waste solids transport system,
a long distance between the absorber and the pond is very common.
This line must also be capable of transporting varying quantities
of waste solids resulting from fluctuating gas load and coal
sulfur content. A plugging condition resulting from insufficient;
slurry velocities in the waste solids- transport system will
exist when less than design quantities of waste solids are
421
-------
produced during normal operation. An alternate is to design
the system to operate on a constant velocity basis at all
times and thus eliminate the. plugging problem.
As operating conditions vary, the quantity of waste slurry over-
flowing from the recycle tank to the waste slurry sump will like-
wise vary. Reclaimed water from the pond is added to the waste
slurry sump. The quantity of water added reflects the difference
between the quantity of waste slurry produced and design capacity
of the system. This insures that the transfer system has a
slurry velocity sufficient to prevent settling of solids and
prevent plugging under all operating conditions.
The FGD industry has evaluated many materials of construction
with varying degrees of success. For the Alabama Electric
Cooperative System, linings have been used extensively.
Consider, for example, the materials of construction selected
for the absorber, All wetted parts of the system are subject
to corrosion, In addition, the spray absorbtion zone must
contend with abrasion, To effectively remove the SC>2 from the
gas, all of the gas must be contacted with the slurry. To
insure proper gas/slurry contacting and prevent short circuiting
of the gas, the spray pattern must be designed such that the
slurry impinges on the absorber wall which creates a sand
blasting situation. To withstand both the corrosion and the
abrasion, the spray absorber zone has a rubber lining. The
internal spray headers are carbon steel, rubber lined, rubber
covered, to withstand abrasion internally and externally. All
connections are flanged and are rubber covered with high alloy
bolting and backup rings to insure that the bolting does not
destroy the integrity of the lining. Also critical is the
selection of the spray nozzle material which in this case is
a cast silicon carbide, The nozzle has no internals and has a
minimum opening of 1" which makes it insensitive to plugging
because of trash material in the system,
422
-------
In the absorber area above and below the absorption zone,
abrasion is not a problem and only corrosion must be considered.
In these sections, a vinyl ester flake glass lining is used,
Regardless of how good the lining material selected, the liner
is no better than the manner in which it is installed, Quality
Control during installation becomes critical and directly effects
the maintenance requirements of the system, Apparently minor
details such as how the rubber lining sections are lapped can
effect the success of the lining installation, For the Alabama
design, where two dissimilar lining materials are joined, a
full body flange on the absorber module is used to mechanically
join the dissimilar materials. Though more expensive, the
mechanical joint eliminates the problems associated with chemically
bonding dissimilar materials, Chemical bonding has generally
proved unsuccessful and will create a maintenance problem,
In any FGD system, the design condition specifications rarely
reflect the actual operating conditions of the plant. Sulfur
contents in the coal vary and load conditions vary, The absorber
design provided has no limitations with regard to minimum gas
flows and yet has the capability of achieving minimum operating
costs by turning off recycle pumps when less than'design sulfur
coals are burned, This permits achieving the desired S02 emission
level at the lowest possible operating cost.
423
-------
OPERATING RESULTS
Various concepts have been discussed with, regard to. ensuring
the reliable performance of the Alabama Electric Cooperative .
FGD Systems, Operating results are the proof as to how success-
ful these concepts have been, Therefore, it is important to
review actual performance,
In September of 1978, Unit #2 was placed on stream, In July of
1979 Unit #3 was placed in service. The operating results which
are discussed here reflect the experience which Alabama Electric
Cooperative has had over more than twenty (20) months of FGD
system operation.
Start-up
The design concept Alabama Electric Cooperative chose for its
control system is a fairly sophisticated one. By means of
programmable controllers the total start-up and shutdown sequence
of the FGD system is accomplished by the pushing of a single
button, As part of the start-up, the total system had been
checked out mechanically and electrically such that all sub-
systems were proven, Having done this, the units were started
up on flue gas by means of pushing that single button, The
single button start-up was achieved for Unit #2 and duplicated
for Unit #3 ten (10) months later,
The coals which the Tombigbee Station burn are from four or
five different mines located in Alabama and Kentucky. Though
the maximum design was a 1,8% sulfur coal, the actual sulfur
content of the coals being fired range from 0,88% to 3,6%
Csee Table No, 2),
424
-------
PROXIMATE ANALYSIS
MOISTURE %
ASH - 7»
VOLATILE MATTER -
FIXED CARBON - %
SULFUR - %
HEATING VALUE -
BTU/lb,
TABLE NO. 2
TYPICAL COALS BURNED
5.82
13.98
% 31,61
48,60
0,88
11,805
7.40
15,10
31.40
46,11
1.06
11,169
4.64
12.62
35.32
47.42 i
3.62
12,199
The sulfur content variation experienced during the month of
December 1979 can be considered typical. This is shown in
Figure No, 1,
Alabama Electric Cooperative
Tombigbee Station
Sulfur Content in Coal Burned In December 1979
I
S
1
• i ^I^^^^^^^^^^^^^^^^^^^^^^^^^^^^T^^^^^^^^^^^^^^^^^^^^^T^^I
10 11 ia U 14 11 16 17 1t 18 20 21 22 23 24 25 2t 27 28 28 30 31
Day of Month
425
-------
SC>2 Removal and Limestone Utilization
As indicated earlier, the absorbers had been designed for 85%
S02 removal while burning a 1,8% sulfur coal, A limestone
stoichiometry of 1.1Q moIs of calcium carbonate/moIs of S02
absorbed had been guaranteed, Performance tests were performed
by the Owner*s engineers, Burns & McDonnell, For Unit #2f a
93,5% 502 removal was obtained while burning a 2,7% sulfur coal
at an absorber gas flow 24% greater than design. When the test
was repeated for Unit #3 - a 97% removal efficiency was obtained
while burning a 2% sulfur coal. In both tests, summarized in
Table No. 3, a limestone utilization was very close to the theoreti
quantity, which, reflects 1QQ7» utilization of the limestone, This
is attributable to the use of the hydroclones for removal of
limestone from the waste slurry.
TABIE NO. 3
PERFORMANCE TEST RESULTS
% S in Coal
Gas VolunE/Absorber-ACFM
Inlet S02 Conc,-ppm
Outlet S02 Gonc.-ppm
% S02 Renoval
Limestone Stoichiometry
(moles CaCoo/
moles S02 Absorbed)
Design
Values
1,8
270,000
1106
166
85
1.10
Unit No. 2
Test Value
. 2.7
335,000
1614
105
93,5
1,01
% of Des,
+50%
+24%
+46%
Unit No. 3
Test Value
2.0
270,000
1250
36
97,1
1.02
% of Des.
' +11%
+ 1%
+13%
Power
The power requirements for the system are low and are summarized
in Table No, 4. The power consumption shown for the FGD system
under design conditions is less than 1% of rated generating
capacity, However, the capability of the system to save power
when operating at less than design sulfur coals (1,1% S normal
vs 1,8% S design) is evidenced by the fact only 0,6% of rated
generating capacity is required,
426
-------
TABLE NO. 4
FULL LOAD POWER REQUIREMENTS
Desiga % S Coal Noimal % S Coal
No.of Absorbers/Unit 2 2
No. of Operating Recycle Pumps/Absorbers 3 1
Power - FGD and I,D, Fan
KW Required/Unit 4275 3496
% of Rated Capacity 1,68 1.37
Power - FGD Only*
KW Required/Unit 2342 1564
% of Rated Capacity 0.92 0.61
*Includes Flue Gas Pressure Drop for FGD System
Manpower
The manpower requirements for the two operating FGD units
(total 510 MW) are low - very low. Alabama'Electric Cooperative
employs two operators per shift, on a four shift basis. With
regard to maintenance, all work is performed on a work release
basis. In terms of maintenance manhours actually expended, 40
hours per week are required for instrumentation and 20-30 hours
per week for mechanical work is required, These numbers are
contrary to the 50 or 60 operators which are frequently cited
for FGD systems.
427
-------
Availability
Since start-up, both, units 2 and 3 have been characterized by-
high, availability. As illustrated by the graphs shown below
as the operators learn to run the system properly, the availa-
bility, improves significantly and availabilities of 9070 or
greater have been consistently achieved. With the learning
experience having already been gained on Unit #2, start-up of
Unit #3 was virtually trouble free and this is reflected in the
high availabilities achieved right from the start. Except for
the first month of start-up, Unit $3 has consistently achieved
monthly availabilities in excess of 97%,
Alabama Electric Cooperative
Tombigbee Station
Unit No. 2 Availability
SOMOJFUAUj JASONDJ
Alabama Electric Cooperative
Tomblgbee Station
Unit No. 3 Availability
J*SO*'OJI=MAMJ.IASONO
428
-------
PROBLEM AREAS
Like any system, problems have been experienced in the operation
of the FGD system. Fortunately, the problems encountered were
correctable and non-recurrent,
The ball mill system had two problem areas, A seal at the feed
end of the ball mill had not been installed - this resulted in
leakage of the limestone s-lurry, Likewise a ball retention helix
at the discharge of the ball mill was not sufficiently deep to
retain large particles. Peabody corrected both problems and the
system now functions adequately.
No system would be complete without damper problems. In the
Tombigbee Station these were also encountered, Double guillotine;
dampers were used in which seal air was injected between damper
ii
blades. A spare blower was provided for each seal air system.
When the unit was started up two problems were encountered. The
operators were undersized and would not move the damper, When
the dampers were in the open position, flue gas containing 862
would leak into the seal air blower system, condense and create
a corrosion problem. The operator problem was corrected by
installing larger motors, The seal air system was corrected by
installing an isolation valve between the blowers and the damper
such, that flue gas would not flow back into the blower system
and condense. With these problems corrected, the dampers are
operating satisfactorily.
Two problems were encountered in the instrumentation area. The pH
sensing probe is emersed in the slurry in the absorber recycle tank.
Problems were encountered with slurry leaking into the preamplifier
which caused failure on several occasions. The preamplifier wa.s
changed to a different type which was enclosed in a seal housing
which prevented leakage. This eliminated the problem. Gas flow
measurement by means of an anubar was a total failure. Under
429
-------
Low flow conditions, it was not possible to get a meaningful
signal, Measurement of gas flow to an absorber was not critical
and therefore attempts at this measurement were abandoned,
Trash material has caused the spray wash nozzles under the
interface tray to plug. Placing an in-line strainer in the
suction of the wash pump which feeds the spray nozzles has
eliminated this problem.
The slurry transfer line from the waste s-ump to the pond is made
of FRP pipe. Rupture of this line has occured several times
because of inadequate pipe supports and also water hammer resulting
from switching waste slurry pumps on and off. Pipe supports have
been redesigned. The method of operating the waste slurry pumps'
has been modified by inclusion of a timer to provide a delay
time when switching from an operating waste slurry pump to a
spare pump, The object of this is to minimize the effect of
water hammer.
430
-------
SUMMARY OF OPERATING AND MAINTENANCE EXPERIENCE
The operating experience of Alabama Electric Cooperative has been
unique and is characterized by:
1) Push button start-up
2) High availability
3) High SC>2 removal efficiency
4) High limestone utilization
5) Low manpower requirements
6) Low maintenance costs
431
-------
OPERATION AND MAINTENANCE EXPERIENCE
OF THE WORLD'S LARGEST SPRAY TOWER SOV SCRUBBERS
n '••"" " ' •£••••' -•
: ROBERT A. HEWITT - TEXAS UTILITIES SERVICES, INC.
and A. SALEEM - CHEMICO AIR POLLUTION CONTROL CORP.
The 750 MW Monticello boiler #3 of Texas Utilities Services, fir-
ing lignite coal, is equipped with three large spray towers, de-
signed by Chemico Air Pollution Control Corporation. Each spray
tower is sized to handle over one million cubic feet per minute of
flue gas. This flue gas desulfurization system uses pulverized
limestone slurry for scrubbing and includes a flue gas bypass as
well as external steam flue gas reheat system. The FGD system
went into operation in mid 1978 and has since logged consistently
very high availability as well as high S02 removal efficiency.
The extreme simplicity of the spray tower system has resulted in
only modest increase in the power plant's operating and maintenance
staff. A recent inspection of the system revealed no major pro-
blems with the tower and duct liners or the tower internals. A
few isolated spots on the internal slurry pipes showed wear due to-
close proximity to the sprays. Failure of the rubber lining on
the side mounted agitators and slurry recycle pumps has been the.
primary source of problems with the system. The experience with
this system in general has been very satisfactory and Texas Utilitie
has purchased two essentially duplicate systems for the Twin Oak
Power Station.
Preceding page blank
433
-------
INTRODUCTION
Texas Utilities Company is investor owned and includes three electi
utilities, two fuel companies, a generating company, a service com-
pany and two non-utility companies engaged in energy related activi
ties. The total generating capability at the end of 1979 was
17,430 megawatts. Units utilizing Texas lignite as a fuel account
for 5300 megawatts of this capability and in 1979 about 50% of the
total generation of the system was from the lignite fired units.
Monticello #3 is a lignite fired unit rated at 750 megawatts locatec
at a site near Mt. Pleasant, Texas. A typical fuel analysis is
shown in Table I. Units 1 and 2 were placed in service in 1974 and
1975. No flue gas desulfurization (FGD) systems were required for
these units. Emission regulations applicable to the #3 unit are a
maximum 2 hour average particulates emission of 0.1 Ibs per 10 BTU,
maximum opacity of 20% and a maximum 2 hour average S02 emission of
1.2 Ibs per 106 BTU.
After evaluation of bids Chemico Air Pollution Control Corporation
was awarded a contract to supply the electrostatic precipitators,
I.D. fans and ductwork to the chimney along with major engineering
and design for the SO2 removal system with an option of provision jf
the total FGD system. Included in this contract was the construc-
tion and operation of a 4000 ACFM pilot plant utilizing flue gas
from one of the existing units. The objectives of the pilot plant
study were to determine:
a. Reactivity of available limestone.
b. Optimum stoichiometry, L/G and recycle solids.
c. S02 removal efficiency at full and partial load.
d. Limestone consumption.
e. Susceptability of system to plugging and scaling.
434
-------
TABLE 1
TEXAS UTILITIES-MONTICELLO NO. 3
TYPICAL LIGNITE FUEL CHARACTERISTICS
Proximate Analysis-% Typical Range
MOISTURE 32.68 27.5-36.3
VOLATILE MATTER 30.17 25.0-34.7
FIXED CARBON 23.88 20.8-27.5
ASH 13.27 6.4-18.9
SULFUR .72 .36-1.79
BTU AS RECEIVED 6689 6068-7302
-------
In February, 1976 the option of having Chemico supply the necessan
additional design, engineering and material for the complete FGD
system was taken. The system was to be completed and ready for
trial operation by January of 1978.
This FGD system is unique in that it involves the world's largest
spray towers. The system has been in service since August 1978 and
performance and availability have been satisfactory. The main
focus of this paper is to review the operating and maintenance ex-
periences to date.
FGD SYSTEM DESCRIPTION
The lignite fired boiler generates about 3.4 million ACFM of flue
gas at full load. The FGD system is designed to keep the SC>2 emis-
sions to less than 1.2 Ibs/MMBTU. The S02 removal is accomplished
by scrubbing with an aqueous slurry of pulverized limestone in
three large spray towers. The spent slurry containing calcium sul-
fur salts is disposed of in an onsite pond from which the reclaimed
water is recycled to the FGD system. A simplified flow diagram of
the FGD system is shown in Figure 1. The general arrangement O'l
the major equipment is shown in Figures 2 and 3.
After passage through the electrostatic precipitators for particu-
late removal, three centrifugal boiler I.D. fans drive the flue gas
into a common inlet manifold from which it is equally distributed
into three spray towers for SO- removal. The scrubbed gas leaving
the spray towers is again collected into a common outlet manifold
for discharge into the stack.
Partial or full bypass of flue gas around the scrubbers is possible
with two bypass ducts which are equipped with split louvre dampers
for gas flow control. The partial bypass, up to a maximum of 50%,
436
-------
FIGURE 1
CSWiPLSPlEB FLOW DIAGRAM OF THE FGD SYSTEM
AT MONTICELLO #3 BOILER OF TEXAS UTILITIES SERVICE
SERVICE
WATER
BLEED TO
RETENTION POND
RETENTION
POND WATER
MAKE-UP
6-
<>
TOP WASH SPRAYS
BOTTOM WASH SPRAYS
~t"\ BLEED
JJ FROM
PUMPS
J-101 B. C. D
"-©-
BLEED
SPRAY
4"
RETURN BY PASS
T—TT—I I I
i V I I I i
ABSORBER
SPRAYS
GAS OUT
MIST ELIMINATOR
TOWER MAKE-UP—'
GAS IN
12"
^EMERGENCY
OVERFLOW
\O
"
PUMP
J-101A SUCTION
RECYCLE PUMP r~
TO
HEADERS ^_
B. C. D
o
CO
CC
CE
UJ
1
LIMESTONE SLURRY
FROM GRINDING SYSTEM
cr
01
y
0-00
8"
^
LIMESTONE SLURRY RECYCE
LOOP
10"
LIMESTONE
SLURRY FEED
TOR-101
POND WATER
FOR
DILUTION
SEAL POT
R-101
SPRAY TOWER
SUCTION
J-002A G-003
LIMESTONE SLURRY LIMESTONE SLURRY
PUMP STORAGE TANK
-------
FIGURE 2
GENERAL ARRANGEMENT OF FGD SYSTEM
AT MONTSCELLO #3 BOILER SHOWING ELEVATION
CHIMNEY
SPRAY TOWERS
R-101.201 &301
PRECIPITATOR
OUTLET
INLET & OUTLET
MANIFOLDS
BY-PAS
DUCT
PRECtFiTATORS
FAN ROTOR
REMOVAL &
ACCESS AISLE
I.D FANS
K-001 A. BAG
CO
FROM
AIR PREHEATER
REHEATER E-001 A & B
REHEATER F.D. FANS K-002 A &
FAN
RECYCLE PUMPS
J-101 A, B. C& D
J-201 A. B. C&D
J-301 A. B, C& D
-------
FIGURE 3
GENERAL ARRANGEMENT OF FGD SYSTEM
AT MONTICELLO #3 BOSLER SHOWING PLOT PLAN
'
T
FAN SUCTION MANIFOLD
INLET MANIFOLD
PRECIPITATOR
R-001B
PRECIPITATOR
R-001A
SPRAY TOWERS
86'-ir
OUTLET MANIFOLD
i
.41'-6" I 47'-6'
o
DC
;. CHIMNEY
i
-------
is automatically controlled to maintain a predetermined S02 level
in the stack. The control signal is provided by the S02 analyzer
monitoring the stack gas. A supplemental reheat system is also
provided for use when bypass gas is not sufficient to give the mini
mum superheat of 25°F. The supplemental reheat system consists of-
two parallel steam heat exchangers for heating the ambient air to
300°P which is injected into the outlet duct leading to the stack.
The ambient air is driven by two centrifugal fans. Model tests for
gas mixing were conducted to determine the location of bypass con-
necting ducts as well as point of injection of the supplemental
steam heated air. The bypass gas is injected into breachings of
the outlet manifold while the hot air is injected through four
opposing ports in the outlet duct leading to the stack. (See
Figures 2 and 3).
The SO2 scrubbing is accomplished by three self-supporting spray
towers with integral slurry recycle tanks. Each tower is equipped
with a single blade guillotine damper on the inlet and a single
louvre damper on the outlet. The single louvre damper can be used
for flow balancing if required. A profile of the spray tower is
shown in Figure 4. Each tower is 44 feet in diameter in the area
of gas flow and expands to 55 feet to accomodate the recycle tank.
The limestone slurry in the recycle tank is kept in suspension by
four side mounted agitators. Each spray tower is equipped with
four spray headers which are fed separately by dedicated centrifuga.
pumps of about 16,000 GPM nominal capacityc The slurry in each
tower is sprayed through 200, 3 inch size hollow cone nozzles made
of silicon carbide.
440
-------
FIGURE 4
VERTICAL PROFILE OF THE SPRAY TOWER
SHOWING MATERIALS OF CONSTRUCTION
44'-0" l.D.
GAS
OUTLET
MIST
ELIMINATOR
WASH
SPRAYS
ABSORPTION
SPRAYS
/\ /\ /\
/\
x\
b
i
x\
\ /\ / \ /\
ACID PROOF
CEMENT
GAS
INLET
55'-0" l.D.
GLASS-FLAKE
FILLED POLYESTER
COATING
GRIT-FILLED
FIBERGLASS-REiNFORCEL
POLYESTER COATING
GLASS-FLAKE
FILLED POLYESTER
COATING
-------
Each spray tower has an integral four pass, open louvre vane mist
eliminator which can be washed from both sides by spraying makeup
water. The bottom side is continuously washed in sequence by
actuating sprays in 12 pie shaped segments. The top is infrequent]
washed in a similar fashion as required. Prior to construction
of the spray towers, gas distribution model tests had revealed that
resistance imposed by the sprays was sufficient for uniform gas dis
tribution, hence no additional gas distribution aids are installed.
Limestone quarry tailings are received by rail car and stored under
covered shed. From the storage pile the limestone is conveyed into
feeder hoppers for the wet ball mill grinders. One operating and
one spare mill is provided, each with a capacity of about 30 tons
per hour with product consistency of 90% minus 200 mesh. The
ground limestone slurry is stored in a day tank and pumped to the
spray towers as required for pH control in each operating spray
tower. The pH is automatically controlled by signals from pH meters
sensing the pH of slurry leaving the recycle tank.
The spent slurry bleed from each tower, taken under level control,
is disposed of in an on-site pond. The reclaimed water from the
dispsoal pond is recycled to the FGD system.
The density of the recirculating slurry is held at 8-10% solids by
the use of nuclear type density analyzers, which control the amount
of reclaimed water added to the towers.
MATERIALS OF CONSTRUCTION
All equipment and duct work upstream of the spray tower dealing
with hot gas is carbon steel. Spray towers and the downstream duct
work up to the stack are fabricated from carbon steel which is pro-
tected against corrosion and erosion by various linings. Each spray
442
-------
tower inlet duct starting from the guillotine damper and leading
some distance into the tower is lined with acid proof cement
(Prekrete G8). The base of the tower up to the liquid level is
lined with glass filled polyester (Ceilcote 103). Above the liquid
level and up to the top spray level is lined with 1/8 inch thick
glass filled polyester lining (Ceilcrete 2500 AR) which incorporates
special grit to provide abrasion resistance. The remaining portion
of the tower above the top spray level up to the outlet isolation
damper is glass filled polyester lined. The ductwork downstream
of the outlet isolation damper, the outlet manifold and the duct
leading up to the stack are lined with acid proof cement (Prekrete
G8).. The slurry recycle pumps and associated piping are rubber lined.
The outside of the spray piping inside the tower is lined with
abrasion resistant polyester lining (Ceilcrete 2500 AR). The spr^y
nozzles are made from silicon carbide to provide abrasion proof
service. The mist eliminator is constructed from fire retardent
polypropylene and supported on FRP beams. The support beams for
the slurry spray piping are rubber lined. The agitators for the .'re-
cycle tank are also rubber lined.
The limestone slurry preparation section is carbon steel, except
for the ball mills, recycle pumps and piping, cyclone classifiers
and agitators which are rubber lined to guard against abrasion.
The dampers upstream of the spray towers including the bypass louvre
are carbon steel construction. The expansion joints in this area
are carbon steel and layered asbestos fabric construction. The
dampers at the outlet of the tower are 316L stainless steel while
the expansion joints are asbestos filled viton.
443
-------
SYSTEM PERFORMANCE
The FGD system was first placed in service on August 18, 1978.
Since that time performance of the system has been good. The syste
easily met EPA compliance requirements during testing conducted
during June, 1979. The SO, removal efficiency of the towers has
generally been 95% or better. As a result of this efficiency, it
has been possible to operate through the unit load range with two
towers under most conditions; utilizing the third tower as a stands
by unit. However, due to the recycle pump situation, to be discus-
sed later, and the desire to maintain the highest possible compliant
with regulatory agency standards, it has been necessary to operate
with all three towers in service for the last year. This has re-
sulted in satisfactory operation even though two and often three
of the towers have been operated with only two recycle pumps in
service for extended periods.
During the operation of the scrubber to date the sulfur content of
the fuel has varied from 0.4% to 1.7%, with a typical range of
0.6% to 0.8%. The E.E.I, availability of the scrubber has been
99.5% or greater. Texas Utilities Generating Company utilizes a
"compliance factor" as a better indication of the true performance
of an FGD system. The "compliance factor" is determined by divid-
ing the number of hours of operation within SO2 compliance limits
by the total hours of generation. This information is shown in
Figure 6. The compliance factor reflects non-compliance excursions
resulting from all factors. The low compliance readings during the
first few months of operation as indicated in Figure 6 resulted
from problems outside of the FGD system primarily related to the
precipitator performance. Figure 6 also shows the limitation facto.
which is a measure of generation loss due to FGD system. During
444
-------
FOR THE MONTICELLO #3 FGD STSTEM.
Stofchlometry defined as pound moles of limestone used per pound mole of SO, removed as a
function of recirculating slurry pH.
en
-------
FIGURES
100
90
80
70
60
50
40
30
20
10
0
cr
o
o
LLJ
O
_J
Q_
5
O
o
o
CO
LOW COMPLIANCE DUE TO
FACTORS OUTSIDE OF FGD
ID
u.
u.
O
COMPLIANCE = HOURS IN COMPLIANCE x 100
FACTOR HOURS GENERATOR ON LINE
SCRUBBER AVAILABILITY FOR 26
MONTH PERIOD WAS 99.5%
TIII!TTIII||III|IITTI~~IIIIIIIII
ASONDIJFMAMJJASONDIJ FMAMJ JASO
1078
1979
1980
o
o
T~
X
Of
Rf
0
-?
0
^
1
«„¥
I
T^"
90 -
80 -
70 -
60 -
50 -
40 -
30-
20 -
10 -
n -
UJ
-z.
H
U-
LL
O
fc
"Z.
111
u.
O
LIMITATION ==
FACTOR
FOH + EFOH
HOURS GENERATOR ON LINE + FOH + EFOH
FOH = FULL UNIT OUTAGE ATTRIBUTED TO SCRUBBER
EFOH - EQUIVALENT UNIT OUTAGE ATTRIBUTED TO SCRUBBER
(INCLUDES LOAD CURTAILMENT TO MAINTAIN COMPLIANCE)
A s
A S r<
' T
n
-------
the 26 month period shown here, the power generation loss attri-
butable to the FGD system has only been a fraction of a percent-
age.
POWER REQUIREMENTS
Full load auxiliary power consumption of the S02 removal system
including I.D. fans is approximately 10 MW. This is based on
the assumption that 35% of the I.D. fan power consumption is due
to the scrubber operation. When the limestone grinding system
is in service, the auxiliary load is increased by 0.6 MW. The
limestone grinding system has had a duty cycle of 5 - 6 hours per
day when the unit is operating at near full load and the sulfur
content of fuel is in the range of 0.7 - 0.9%.
REAGENT REQUIREMENTS
The limestone utilization is pH dependent as shown in Figure 5.
When the system is operated within a pH range of 5.5 - 6 the lime-
stone stoichiometry based on absorbed SC>2 is between 1 to 1.10.
MANPOWER REQUIREMENTS
The spray tower system has been relatively easy to operate and
maintain, consequently the manpower requirement for operation has
been modest. Since the flue gas controls are intergrated into
the boiler train, the BTG operator can also control the flue gas
flow to the spray towers.
The following is the list of personnel dedicated to scrubber
operation:
447
-------
System Area Personnel Per Shift
Scrubber Control 1
Limestone Handling and Milling 1/2
Chemical Technician 1/3
Environment & Instrument Technician 1/4
Mechanical Maintenance 1 1/2
Electrical Maintenance 1/4
Total 3 5/6
In order to have 24 hour a day seven days a week coverage, a total
manpower of 15-1/3 men is dedicated to the scrubber operation.
OPERATION EXPERIENCE
During the first four months of operation several breaks in the
fiberglass line that supplies reclaim water from the sludge disposal
pond to the towers were experienced due to poor make-up of joints
during original installation and vibration due to inadequate support
and restraint of piping in the area. This has been the only problem
that resulted in the removal of a scrubber when the generator was
on the line. The problem was corrected with the replacement of the
fiberglass line with carbon steel pipe in the areas of failure.
Difficulty has been experienced in moving limestone from the storage
pile to the grinding system due to pluggage of the hoppers and
mechanical failure of the feeders and associated equipment. Plug-
gage is a problem due to the nature of the limestone used, which
is a by-product of crushed limestone production and contains a high
percentage of fine material and moisture. This results in pluggage
in the reclaim hoppers especially when wet or when stacked high.
It has been necessary to feed the reclaim hopper with a front end
loader.
448
-------
Several problems have been experienced with the tower inlet guil-
lotine dampers. The bottom seals have been damaged due to ash and
sludge accumulation in the seal trough. Several of the jack screws
and pushrods have been damaged due to binding of the dampers.
Minor linkage problems were experienced on the by-pass dampers;
otherwise the tower outlet and by-pass dampers have performed well.
Since no internal maintenance has been required while the scrubbe.r
has been in service, it is not known whether damper leakage would
permit safe entry while on the line.
Several minor instrumentation problems have been experienced. Ex-
cessive drift has been a problem with the density control instru-
mentation. The 0-14 range of the pH instrumentation originally
supplied was too wide to allow good control in the narrow range of
5.6 - 5.8. The scale was expanded and the pH system has performed
satisfactorily. The system is easily operated manually with the re-
sult that instrumentation problems have not had any appreciable im-
pact on the operation of the scrubber.
The rubber lining of the side mounted agitator blades has failed
at the tips allowing erosion damage to all agitator blades.
The most significant problem experienced with the FGD system has
been repeated failures of the rubber lining of the slurry recycle
pumps. Although this problem has not resulted in the loss of avail-
ability of the FGD system or noncompliance with emission limits,
it has resulted in a very significant maintenance expense. Efforts
are continuing to resolve this problem by reducing the speed of the
pump. A different manufacturers' pump has also been installed for
testing.
449
-------
Another problem was experienced when an attempt was made to use
ash water on one tower for mist eliminator wash rather than fresh
water. The high levels of calcium sulfate in this water resulted
in extreme fouling of the mist eliminator packing material. The
high velocity of gas through the unplugged areas combined with the
increased load on the other towers resulted in slurry carryover into
the outlet duct and chimney.
With the exception of the recycle pump problem, the operating ex-
perience with the system has been relatively good. There have been
no lining problems in the towers. Spray nozzle plugging has not
been a problem and no significant problems have been experienced
with the internal piping other than external erosion due to imping-
ment in isolated areas. It has not yet been necessary to remove a
tower from service for internal maintenance.
OPERATING AND MAINTENANCE COSTS
Operating and maintenance costs are summarized in Table 2. Over
the first 22 months of operation the FGD system operating labor
cost has averaged $7,222 per month. The maintenance material and
labor cost for the scrubber has averaged $65,396 per month with
material only averaging at $35,341. The maintenance material and
labor cost for the limestone preparation system has averaged
$15,388 per month.
A significant portion (estimated at 40-50%) of the scrubber main-
tenance material and labor cost has been due to the recycle pump
problems. Resolution of this one problem will significantly .reduce
maintenance cost.
450
-------
TABLE 2
OPERATING AND MAINTENANCE COSTS
FOR MONT1CELLO #3 FGD SYSTEM
Maintenance
Labor and Material
Calendar
Month
A/78
S
O
N
D
J/79
F
M
A
M
J
J
A
S
O
N
D
J/80
F
M
A
M/80
Total
Monthly Average
Scrubber
Operation
$ 7,000*
7,000*
7,000*
7,000*
7,000*
7,855
7,855
7,855
7,855
4,309
8,028
6,863
8,117
6,890
7,088
7,763
9,695
7,521
7,891
6,695
7,110
4.488
158,873
7,222
Scrubber
Area
$ 5,651
28,085
10,066
12,670
27,307
15,149
24,007
30,730
32,506
78,383
24,597
44,128
60,128
51,222
44,225
127,713
158,096
49,484
119,228
70,494
294,694
130,144
1,438,707
65,396
Limestone
Area
$ 3,598
7,439
5,312
12,351
12,376
10,881
6,703
15,587
9,988
3,821
11,866
19,933
19,147
6,625
17,610
26,351
7,836
12,199
45,223
36,621
15,873
30,185
338,525
15,388
'Estimated Costs
ADDITIONAL ESTIMATED MONTHLY COSTS
AVERAGED OVEF! 22 MONTH PERIOD
Chemical Tecnician
Instrument Technician
Supervisory
Total Additional Costs
1,000
500
500
2,000
-------
DUAL ALKALI DEMONSTRATION PROJECT INTERIM REPORT
by
R. P. Van Ness
Manager of Environmental Affairs
Louisville Gas & Electric Co.
Louisville, Kentucky
Norman Kaplan
Industrial Environmental Research Laboratory
Office of Research and Development
Environmental Protection Agency
Research Triangle Park, North Carolina
D. A. Watson
Project Manager
Bechtel National, Inc.
San Francisco, California
ABSTRACT
This paper will discuss the results of the recently performed acceptance
test on the dual alkali system serving Louisville Gas and Electric Com-
pany's Cane Run Unit 6 boiler. The acceptance test was conducted to
measure the system performance with respect to the guarantees offered
Louisville Gas and Electric by Combustion Equipment Associates. The results
of the testing were as follows:
• SO? removal averaged 94% and 143 ppm outlet concentration
• Soda ash consumption averaged 0.042 mole soda ash per
mole sulfur dioxide removed
• Lime consumption averaged 1.04 mole CaO per mole sulfur
dioxide removed
• Power consumption averaged 1.05% of generation
• Filter cake solids averaged 52.2 wt % insoluble solids
• There was no net particulate matter addition
Various problems attributable to the boiler, the FGD system, and the quality
and quantity of the carbide lime supplied to the system delayed the accept-
ance testing until July 1980. The year-long demonstration period was offi-
cially started in May 1980. The nature of the problems experienced and
their solutions are discussed.
Preceding page blank
453
-------
NOTES
1. Company Mames and Products
The mention of company names or products is not to be considered an
endorsement or recommendation for use by the U.S. Environmental Pro-
tection Agency.
2. Units of Measure
EPA policy is to express all measurements in Agency documents in metric
units. When implementing this practice will result in undue cost or
difficulty in clarity, IERL-RTP provides conversion factors for the
non-metric units. Generally, this paper uses British units of measure.
The following equivalents can be used for conversion to the Metric System:
British Metric
5/9 (°F-32) °C
1 ft0 0.3048 m
1 ft; 0.0929 nt
1 ft* 0.0283 m3
1 grain 0.0648 gram
1 in. 2.54 cm
1 in.; 6.452 cmi:
1 in.-5 16.39 cm3
1 Ib (avoir.) 0.4536 kg
1 ton (long) 1.0160 m tons
1 ton (short) 0.9072 m tons
1 gal. 3.7854 liters
454
-------
DUAL ALKALI DEMONSTRATION PROJECT INTERIM REPORT
INTRODUCTION
The Dual Alkali Demonstration Project is a joint effort by a number of
organizations under the sponsorship of the Environmental Protection
Agency. The process being demonstrated is a sodium based concentrated
mode using carbide lime as a regenerant. Louisville Gas and Electric
Company (LG&E) is the owner-operator of the dual alkali system serving
their Cane Run Unit 6 boiler, which is a nominal 280 MW high-sulfur
coal-fired boiler (3.5-4.0% S). The design was developed by Combustion
Equipment Associates (CEA) and Arthur D. Little, Inc. (ADD. The system
was erected by the construction department of LG&E under the Guidance of
CEA/ADL at total cost of about $22 million (1976-1980 dollars) or about
$79 per kW installed generating capacity (including waste disposal).
A process flow schematic of the dual alkali process at Cane Run 6 is
depicted in Figure 1. Flue gas from the boiler passes through the
electrostatic precipitators and is fed to two absorbers. A recycling
sodium sulfite solution, flowing countercurrent to the flue gas across
two stainless steel perforated plate trays, absorbs SO according to
the following reaction:
1 S03=+ S02 + H20 -**2HS03"
In addition, due to the absorption of sulfur trioxide from the gas and
due to the oxidation of sulfite ion in solution, sulfate (S04~) is
formed in the absorbent liquor:
2 H20 + S03->H2S04-»*2H + S04
3 S03= + 1/2 02-*-S04=
The scrubbed flue gas is reheated by combustion gases from a direct oil
fired reheater and is ducted to the stack.
Sodium carbonate is added to either the thickener or the absorber to
make up for losses of sodium in the system. Bleed streams of the
spent aborbent solution from the absorbers are sent to the regenerator
Reactor trains where carbide lime is added to convert the bisulfite
(HS03~) in the spent absorbent, to sulfite (S03=) in the regenerated
absorbent, precipitating a mixture of calcium sulfite and sulfate
solids:
4 2HS03" + Ca(OH)2-»*CaS03{ + S03= + 2H20
5 S04= + 2HS03' + Ca(OH)2 ->• CaS04| + 2H20 + 2 S03=
The mixed solids actually can be designated as: x CaS03 . y CaS04 .
z H20 where the ratio x:y is usually greater than 4 and z represents
455
-------
TO ATMOS
FIGURE 1
DUAL ALKALI SYSTEM AT CAME RUN 6
-------
some amount of water of hydration. No pure gypsum phase is formed.
The sol Ids are separated from the liquor in a thickener and are
removed from the system on washed vacuum filters. The filter cake is
mixed with fly ash and quicklime in a system designed by I.U. Conversion
Systems. After fixation the solids are trucked to a landfill site for
disposal. The clear liquor overflowing from the thickener is returned to
the absorber recycle loop.
A comparison between the design basis and observed operation is given
in Table 1. The design basis is taken from the design manual produced
under this project, one of the sources of information to which the
reader is referred for additional detail (References 1, 2, and 3).
The system is designed to operate with a liquid to gas ratio of less
than 10 gal./l(r acf including liquor feed to the tray and spray
recycle (typical lime or limestone slurry process are designed for
about 50 gal./10J acf). The design flue gas pressure drop from the
booster fan to the stack entrance is 8.5 in. of water.
Bechtel National, Inc. is under a separate contract with EPA to provide
an independent test program to assess the operation of the system with
regard to its performance guarantees, and to provide a demonstration
program designed to characterize the system and monitor its performance
over a year-long demonstration period.
Construction was completed in March 1979 and the system was initially
charged and started up in April 1979. Various problems attributable to
the boiler, the FGD system, and the quality and quantity of lime supplied
to the system delayed the acceptance testing until July 1980. The year-
long demonstration period officially started in May 1980. The problems
and solutions are discussed later.
The acceptance test was conducted from July 17 to July 28, 1980. With
one minor exception (filter cake quality), the system proved to be
capable of successfully meeting its performance guarantees.
ACCEPTANCE TEST RESULTS
The 12-day acceptance test was conducted to measure the performance of
the dual alkali system with respect to the guarantees provided to
Louisville Gas and Electric Company by Combustion Equipment Associates.
Seven guarantees concern the operation in the following areas:
Sulfur dioxide removal
Carbide lime consumption
Soda ash consumption
Particulate matter emissions
Power consumption
Filter cake quality
Year-long system availability
457
-------
TABLE 1
Performance Conditions
Design Observed
Coal (Dry Basis)
Sulfur 5.0% S 3.7% S (ave.)
Chloride 0.04% Cl 0.02% Cl (ave.)
Heat Content 11,000 Btu/lb 10,650 Btu/lb (ave.)
Inlet Gas:
Flow Rate (Volumetric) 1,065,000 acfm 1,045,000 acfm (max.;
Temperature 300°F 280°F (max.)
S02 3471 ppm 2323 ppm (ave.)
Oo 5.7% 6.7% (ave.)
Particulate 0.10 lb/106 Btu 0.84 lb/106 Btu (ave.)
Outlet Gas:
S0? <200 ppm 143 ppm (ayej
Particulate TJ.10 lb/106 Btu 0.10 lb/106 Btu (ave.)
Boiler Operation:
Generation 280 MW 240 MW (max.)
458
-------
Table 2 summarizes the guarantees offered and the corresponding results
of the acceptance test. A brief discussion of each of the guarantee
tests performed during the acceptance test follows.
Sulfur Dioxide Removal
The primary method of determining S02 removal relied on the continuous
Lear Siegler monitor installed in the stack. This analyzer was certified
in December 1979 by an outside contractor according to the procedure
specified in the Federal Register. During the acceptance test, as a
backup to the continous monitor and as an ongoing confirmation of the
analyzer accuracy, wet chemical tests of the'stack effluent according to
EPA Method 6 were also performed daily, in conjunction with the particulate
tests.
Preliminary results from the wet chemical analysis showed a discrepancy
between these measurements and the continuous monitor readout. After
an extensive check of the system, a burned ground wire was discovered
in the signal line of the Lear Siegler continuous S02 monitor. From
the data on the calibration sequences of the analyzer prior to, during,
and after elimination of the grounding problem, it was concluded that
the signal from the analyzer was offset on the low side by 30 ppm by
the malfunction. Therefore the continuous stack S02 monitor readings
for the first 7 days of the tests were corrected by 30 ppm. With
this correction applied-to the early readings, and subsequent to the
repairs to the ground in the analyzer, the two techniques were in good
agreement.
Both measurements showed that the system could meet the 200 ppm S02
outlet concentration guarantee. Table 3 summarizes the 24-hour average
S02 results for the 12-day acceptance test. Table 4 summarizes the
simultaneous wet chemical and continuous monitor measurements (the
Method 6 tests were conducted only for the first 10 days).
Lime Consumption Guarantee
The lime consumption guarantee was specified as "not [to] exceed 1.05
moles of available CaO in the lime per mole of S02 removed from the
flue gas". Lime consumption was determined by analyzing representative
samples of filter cake collected as the cake was discharged from the
filters prior to fixation. The cake was analyzed for total calcium and
total sulfur. The total calcium represented the Time used, and the
total sulfur represented S02 removed from the flue gas. A portion of
the calcium entering the system with the carbide lime is present as
carbonate and therefore does not represent alkalinity available for
regeneration. Each time the lime day tank was filled, a sample of lime
was analyzed for available alkalinity and total calcium. From these
results, a correction factor was developed to account for unreactive
calcium in the carbide lime feed. During the 12-day acceptance test
the calcium consumption, corrected for available alkalinity as described
above, averaged 1.04 moles of available CaO per mole of S02 removed,
thus meeting the guarantee which required less than 1.05 moles/mole of
sulfur removed. Table 5 summarizes the analyses performed on the filter
cake samples.
459
-------
TABLE 2
Performance Guarantees and Acceptance Test Results
GUARANTEE
TEST RESULTS
S02 Removal
200 ppm dry basis (D.B.)
without additional air dilution
Calcium Consumption
1.05 moles available CaO
per mole S02 removed
Soda Ash Consumption
0.045 moles Na2C03 per
mole S02 removed
Net Particulate Addition
Mo net particulate addition
by FGD system
Power Consumption
System will consume (excluding
reheat) not more than 1.2% of
power generated at peak capacity
Filter Cake Properties
Filter cake will contain a mini-
mum of 55 wt % insoluble solids
143 ppm (D.B.) without additional
air dilution
1.04 moles available CaO per
mole S02 removed
0.042 moles Na2C03 per
mole S02 removed
Met particulate removal averaging
88% efficiency
System consumed 1.05% of power
generated
Filter cake averaged 52.2 wt
insoluble solids
460
-------
TABLE 3
Acceptance Test Continuous S02 Analysis
Acceptance Test
Day
1
2
3
4
5
6
7
8
9
10
11
12
Average
-
24 Hour Continuous S02 Analyzer Results
(ppm, dry basis)
A Inlet
2444
2674
*
*
2265
2567
2113
2116
2395
2372
2292
2167
2340
B Inlet
2418
2570
2390
2290
2315
2515
2021
2088
2339
2315
2233
2166
2305
Stack
130
129
130
152
157
140
132
124
146
171
156
130
141
% Removal
94.7
95.0
.94.6
93.4
93.1
94.5
93.6
94.1
93.8
92.7
93.1
94.0
93.9
* Analyzer printout malfunction
461
-------
TABLE 4
Acceptance Test Continuous Monitor and EPA Method 6 Analysis
Acceptance
Day
1
2
3
4
5
6
7
8
9
10
Hours
1400-
1700
1100-
1300
1000-
1300
1200-
1500
1000-
1300
1000-
1300
1600-
1900
1100-
1300
1100-
1400
0900-
1200
S02 Concentration, ppm, dry basis
A Inlet
DuPont
Analyzer
2434
2434
2592
2836
2656
2716
2337
2395
2864
2690
Method 6
2330
2150
2210
2390
2330
2350
2040
2120
2530
2410
B Inlet
DuPont
Analyzer
2516
2423
2670
2674
2606
2418
2250
2330
2721
2624
Method 6
2330
2180
2290
2480
2410
2480
2030
2100
2450
2360
Stack
LSI *
Analyzer
119
122
117
155
136
184
**
113
122
160
Method 6
124
163
154
159
13?
212
137
130
133
137
* Analyzer readings for days 1-6 corrected for the effect of the burned out
ground wire
** Analyzer out of service for repairs to ground wire
462
-------
TABLE 5
Acceptance Test Daily Average Filter Cake Analysis
Test Day
1
2
3
4
5
6
7
8
9
10
11
12
Average
As F
Na
wt %
0.55
0.58
0.70
0.48
0.45
1.11
0.58
0.50
1.07
0.45
0.77
0.62
.eceivecTT:
Ca
wt %
14.88
14.60
15.64
15.72
15.15
15.20
14.35
14.44
14.05
14.43
14.85
13.74
lasis
Total
Sulfur
wt %
31.35
31.58
32.60
32.68
31.92
33.20
31.80
32.32
32.58
33.07
33.48
31.64
Insoluble
Solids
wt %
52.65
52.20
52.60
53.72
53.92
51.90
50.70
51.40
51.00
52.43
52.82
50.58
52.16
Mole NaoCO^
Mole S02
0.037
0.038
0.045
0.031
0.029
0.070
0.038
0.032
0.068
0.028
0.047
0.041
0.042
Mole CaO j
Mole S02
1.139
1.109
1.151
1.154
1.139
1.099
1.083
1.072
1.035
1.047
1.064 ;
1.042
1.095
Calcium consumption corrected for available alkalinity (1.095 x 0.95 = 1.040)
* Correction factor developed from analysis of incoming carbide lime for
mole of available alkalinity per mole of total calcium
463
-------
Soda Ash Consumption
The soda ash consumption was determined by analysis of total sodium
and total sulfur in the filter cake. According to this analysis the
consumption of soda ash averaged 0.042 moles of Ma2C03 per mole of
sulfur dioxide removed and therefore met the guarantee requirement of
0.045.
Particulate Matter Emission
The system was guaranteed not to make any net addition of particulate
matter to the gas stream prior to discharge. Particulate tests, following
EPA Method 5, were conducted on the inlets to the absorber modules and
in the stack (downstream of reheaters) during the acceptance test.
The results of 10 simultaneous tests showed convincingly that there
was no net addition of particulate matter across the system. Actually,
the absorber performed as a particulate removal device averaging 88%
net removal of incoming particulate. Table 6 displays the results
of particulate matter tests performed during the test program.
Although the FGD system met the guarantee requirements, the test was
not very stringent due to the low level of performance by the electro-
static precipitator during the acceptance test period. The FGD system
was originally designed to process an incoming flue gas stream containing
the equivalent of 0.1 Ib of particulate matter/10" Btu or less. :
During the acceptance test, however, the level of incoming particulate
matter was almost an order of magnitude higher. Thus it is not surprising
that the absorbers functioned to remove particulate matter even at the
relatively low pressure drop at which they operated. The particulate
matter emissions from the stack, however, were on the order of 0.1
lb/10 Btu as required for the Cane Run Unit 6 FGD system under the
appropriate requirements to control particulate matter emissions.
Power Consumption
The system, excluding reheat, was guaranteed not to use more than 1.2%
of the total power generated by the boiler/turbine unit at gross peak
load. During the acceptance test the peak generation was 240 megawatts
(MW). Correspondingly, the power consumed during peak generation was
2.5 MW, or 1.05%. The guarantee was met based on peak generation and
also based on average generation over the the test period. During the
12-day test, the average load was 178 MW and the average power consump-
tion by the FGD system was 2.05 MW, or 1.15%.
Waste Filter Cake Properties
The system was guaranteed to produce a waste filter cake containing a
minimum of 55 wt % insoluble solids. The filter cake averaged 52.2 wt %
insoluble solids during the acceptance test. While this fell slightly
464
-------
TABLE 6
Acceptance Test Particulate Test Results
Acceptance Test Day
1
2
3
4
5
6
7
8
9
10
Average
Partlculate (lb/106 Btu)
A Inlet
0.5320
0.6590
0.9470
0.9440
1.1100
0.9900
0.5890
0.6250
0.7890
0.9620
0.8147
B Inlet
0.7120
0.3620
1.0700
0.8060
0.9200
1.4900
0.8470
0.6490
1.2000
0.5930
0.8649
Stack
0.0895
0.0932
0.1110
0.1030
0.1020
0.1020
0.1020
. 0,0893
0.1100
0.1020
0.1004
% Removal
85.6
81.7
89.0
88.2
90.0
91.8
85.6
86.0
88.9
86.9
88.0
465
-------
short of guarantee, the product discharged to the IUCS process was
uniform in moisture content and was suitable for working into a stable
and manageable product through the fixative process. Optimization of
filter cloth selection and filter cycle will continue with the goal of
showing that compliance with this guarantee can be met during the
demonstration year.
System Availability
System availability, as defined by the Edison Electric Institute (available
hours divided by the total hours in the period under consideration), was
guaranteed to be greater than 90% for the demonstration year. While it
is too early to report such a figure, through the first 4 months of
the demonstration year (May-August), the availability of the system
has averaged 99.8%.
OPERATING AND MAINTENANCE PROBLEMS
Up to the time of the acceptance testing there were a number of mechanical
problems and a few chemical problems which affected system performance
and led to cumulative delays in executing the program. None of the
problems have been insurmountable, but their solutions have been
time consuming. It is important to report the nature of these obstacles
so that future installations of this or similar technology can benefit
from the experience.
Recycle and Thickener Return Pumps
There have been two ma.ior problems with the high-capacity low-speed
pumps for recirculation of absorbent liquor to the trays, and return of
thickener overflow liquor to the absorbers. The first problem was the
mechanical shearing of the impellers at the hub. The original pump
impellers were manufactured in two parts: a body and a separate hub for
attachment to the shaft. The hub was welded to the body. All of the
impeller failures were on this welded seam. This problem was elimin-
ated when the pump vendor supplied a one-piece molded impeller body.
The second major problem involved the rapid failure of the suction
side of the pump liner. As a result of close tolerances between the
casing liner and the impeller, the two surfaces were rubbing; the
resulting abrasion destroyed the liner. After completely dismantling
the pumps, it was discovered that a finishing step appeared to have
been omitted at the factory, leaving about 1/4-in. excess length on
each shaft. Milling each shaft to its design size eliminated this
problem.
Mist Eliminator Collapse
Within a few months the startup, both absorber modules experienced
high pressure drop problems. Inspection of the internal structure
revealed that the mist eliminator sections had sagged or collapsed
structurally. The problem was solved by replacing the mist eliminator
sections with those of a different manufacturer. Since the replacement,
466
-------
In August 1979, there has been no further problem with the mist
eliminators.
Tray Pluggage
One of the most perplexing problems was the pluggage of the absorber
trays due to deposition. At first the observed deposit was thought to
be carbonate scale resulting from pH upsets in the modules. Careful
analysis showed the precipitate to be an aluminum-hydroxy-silicate
complex. The mechanism of dissolution and subsequent deposition was
traced to the operating pH of the reaction train. Aluminum was found
to be entering the system with the carbide lime. At the operating pH
in the reactor, above 11.5, the aluminum compound is soluble in the
liquor. When the thickener overflow recycle combined with the recircu-
lating absorbent, the resultant drop in pH caused the aluminum to
precipitate on the absorber trays.
Reducing the operating pH of the reactor to between 10.0 and 10.8
reduced the solubility of the aluminum within the reactor and thickener.
This change ahead of the absorber minimized the pluggage problem. At
the reduced pH set point, however, there is less buffering and control
of reactor pH is more difficult.
Water Balance
The system initially experienced a severe water imbalance. This was
partly due to a lack of familiarity with the system, and partly because
of low-solids concentration in the carbide lime feed. The other lime
slurry systems at Cane Run can tolerate an occasional open-loop excursion.
However, the dual alkali process must operate in a closed-loop at all
times, since the high concentration of solubles in the scrubbing liquor
makes disposal unacceptable for both environmental and economic reasons.
The system was designed to accommodate 70% water (30% solids) in the
incoming carbide lime slurry. Initially the water content was
consistently in the 82-85% range. At this concentration the system
was receiving twice the design input water flow. After only a few
hours of operation the volume of water in the system had accumulated to
the point where the lime feed had to be cut off. The absorbers continued
to function as evaporators until the water level dropped low enough
to resume normal operation.
Strict control of the incoming lime concentration from the supplier and
the addition of a ball mill-hydroclone system to remove oversize particles
alleviated the problem.
Soda Ash Silo Pluggage
Soda ash is added to the system by a dry weigh feeder which feeds dry
solids from a storage silo to a mix tank where it is mixed with absorbent
liquor. Vented moisture vapor from the hot mix tank backs up into the
weigh feeder screw conveyor and causes the soda ash to form lumps which
prevent the smooth flow of feed to the system. The system had a small
fan to blow the moisture-laden air back into the mix tank; however,
467
-------
It proved to be under-designed. Although a larger fan was installed
to improve the situation, the soda ash feed system still remains a
relatively high maintenance item.
Thickener Blockage
In mid-January of 1980, the thickener rake seized during a boiler outage
for repair and ultimately required a shutdown and major overhaul of
the thickener. This did not occur during normal operation, but rather
during the transient period in which the boiler and FGD were being
shut down for maintenance. The stoppage was postulated to have resulted
from an overloading of the thickener with washings of accumulated
solids (including fly ash) from the bottom of the absorber. Lacking a
bottom drawoff, the absorber allowed fly ash to be trapped and accumu-
lated in its lower portion. The problem could apparently have been
avoided if the solids from the bottom of the absorber had been slowly
pumped to the thickener while the thickener and filters continued in
operation until the absorber bottom was purged of solids.
Correction of the problem took about 3 weeks, during which about 2
million gallons of liquid and solids had to be removed from the thickener
(liquid was temporarily stored, and solids were impounded off site).
To accomplish this, large access entrances were cut in the thickener
sides to allow entry by personnel and equipment to dig out the compacted
solids.
Overloading of the thickener has not recurred. The solids in the
bottom of the absorbers are still not subjected to mechanical agitation,
but they are no longer washed into the thickener in large slugs.
Sulfur Dioxide Monitoring
Sulfur dioxide measurement in the inlet to and the outlet from the
absorbers is performed by continuous DuPont UV Model 460 SO? analyzers.
In the stack, sulfur dioxide concentration of the scrubbed gas is
measured by a Lear Siegler S02 analyzer.
Three problems have occurred in the measurement of S02 using the DuPont
analyzers supplied with the dual alkali system:
• Plugging of the sample probe
» Maintaining a steady calibration of the instruments
e Stratification of scrubbed gas across the absorber exit duct
The first two problems have been minimized by daily inspections to
determine if calibration or cleaning of the probes is required. An
attempt to alleviate the last problem will be made by moving the SO?
probes downstream of the reheaters, which should also help reduce the
first two problems.
Failure of FRP Piping
The FRP (fiberglass reinforced plastic) piping in slurry service (i.e.,
468
-------
thickener underflow and filter feed) has been a major maintenance Item.
Some failures have been spectacular, some minor. Late In the fall of
1979 a flush connection on the underflow line snapped off. Slurry
from the thickener flooded the access tunnel below the thickener before
the break could be Isolated. Routinely, elbows In the line from the
thickener to the filter have required repairs because of erosion
damage and failure of the connection bond. Gradually all the underflow
FRP piping is being replaced with mild steel. While mild steel has a
limited life span in this service, failures will be less catastrophic.
pH Control
Reliable and accurate pH measurements for pH control in the reactors
and in the scrubber bleed stream have been particularly bothersome.
The pH related problems are attributed to:
• Inability to keep the probes clean
• Poor responsiveness of the probes
• Pluggage of the sample lines
• Poor calibration techniques
Experimentation with different instrument designs and sampling methods
is gradually alleviating the first three problems. Detailed calibration
instructions and cross checking of the results by two operating departments
have minimized the last. On-line pH readings are compared daily with
pH measurements taken with a portable pH meter by the LG&E scrubber
laboratory personnel. If these readings are in disharmony by more
than 0.3 pH units, the on-line probes are recalibrated.
All the original L&N pH probes have been replaced with Great Lakes
models. To measure the pH of the primary and secondary reactors, a
Great Lakes Model 60 submersible probe is placed in the overflow chute
from primary to secondary reactor, and in the secondary reactor below
the liquid level near the overflow. The pH of the bleed and thickener
return streams is measured by Great Lakes Model 60 flow-through pH
probes with ultrasonic cleaners.
Filter Operation
There have been two major concerns with the rotary vacuum drum filters.
First, the cake quality has varied between 45% and 55% solids. Second,
it has not always been possible to properly wash the cake to meet sodium
consumption guarantee.
Prior to the acceptance test, experimentation with different filter
cloths led to installation of a new filter cloth. The original cloth
was a polypropylene cloth supplied by National Filter Media of Hamden,
CN. During the acceptance test this cloth was replaced with a multi-
filament nylon cloth supplied by Thoerner Products Corp., of Pittsburgh,
PA. The new cloth produced a more consistent quality cake but had a
tendency to blind. During the acceptance test, cake washing was
sufficient to meet the soda ash consumption guarantee, but the blinding
detrimentally affected the percent solids of the filter cake. There
469
-------
is also some concern that poorer quality solids may be produced in the
reactors at the lower pH levels required to control dissolution of
aluminum and silicon compounds.
Proper cake washing on the filters is subject to a number of considerations.
The wash water rate (as limited by the water balance), the quality of
solids produced, the thickness of the cake (controlled by drum angular
velocity), the wash spray configuration, and the quality of the filter
cloth (blinding characteristics) are all important parameters. Therefore,
experimentation with different filter cloths and varying operating
parameters is continuing.
DEMONSTRATION PROGRAM PLANS
Although the results reported here have focused on startup and acceptance
testing, it seems appropriate to outline the major work underway and
planned as part of the demonstration program.
Commercial Grade Lime Testing
A month-long test, using commercial-grade lime in place of carbide
lime, will be conducted as part of the demonstration program to confirm
the interchangeability of the two materials for use in a lime dual
alkali system. The carbide lime contains silicon and aluminum compounds
that are potentially detrimental to the operation of the system, as
previously noted. Bench-scale tests have already shown that commercial-
grade lime is more reactive than carbide lime, and further improvement
in lime consumption is expected during this test. Conversely, carbide
lime is thought to contain an oxidation inhibitor not present in commer-
cial lime. Much of the success of this system relies on the process
liquor remaining subsaturated in calcium sulfate. During the month-long
test oxidation levels in the system will be closely monitored for any
observed difference in oxidation levels.
Materials Evaluation
Sample coupon racks containing several polymer- or rubber-coated specimens
and various stainless steel coupons have been installed in numerous
locations throughout the system. Additionally, pipe spool samples
have been installed in the bleed stream and thickener underflow line.
These spools are constructed of various steels with polymer or rubber
linings.
Some of these corrosion samples will be removed after 6 months and
the remainder at the end of 1 year. Recommendations for materials
for future installations will be based on the analyses of these samples.
Sludge Disposal
A study of the effects of the long term disposal of the sludge generated
by the dual alkali process has been developed. For the test program,
unfixed sludge and two different combinations of sludge, fly ash, and
quicklime will be placed with and without mechanical compaction in six
separate impoundments (each 50 ft x 10 ft x 5 ft deep) for close study.
470
-------
Test work will include leachate sampling, as well as a number of engineer-
Ing and durability tests to characterize the sludge and sludge-mixture
properties.
Centrifuge Evaluation
Because of the problems associated with operation of the filters in
conjunction with a thickener, a pilot size centrifuge will be installed
in early fall for experimentation. The centrifuge will be tested to
determine its ability to produce an acceptable waste while separating
sodium compounds from the cake. It will also be evaluated and compared
with the rotary vacuum filters in terms of reliability and maintenance
requirements.
CONCLUSION TO DATE
As indicated by operation since March 1980, and the successful completion
of the acceptance test in July, the dual alkali process is capable of
achieving greater than 90% SO? removal with an availability of more
than 99% while processing a flue gas generated in a high-sulfur (>3.5%)
coal-fired, full-size (280 MW) utility boiler. Consumption of raw
materials and power was less than expected (guaranteed) while the S02
removal was over 94% on the average for the 12-day acceptance test.
Most of the problems initially encountered were mechanical and have
been solved or greatly reduced in the operation at Louisville Gas &
Electric's 280 MW Cane Run Unit 6.
Further investigation of filter operation, reactor operation, filter
cloths, materials of construction, and major process component
characterization is underway -
471
-------
REFERENCES
1. Van Ness, R. P. et al., "Full-Scale Dual Alkali Demonstration System
at Louisville Gas and Electric Co. - Final Design and System Cost,"
EPA-600/7-79-221b, NTIS No. PB80 146715, September 1979.
2. Van Ness, R. P- et al., "Project Manual for Full-Scale Dual Alkali
Demonstration at Louisville Gas and Electric Co. - Preliminary
Design and Cost Estimate," EPA-600/7-78-010, NTIS No. PB278722,
January 1978.
3. Kaplan, N., "Summary of Dual Alkali Systems," in Proceedings:
Symposium on Flue Gas Desulfurization Las Vegas, Nevada, March 1979,
Volume II, EPA-600/7-79-167b, NTIS No. PB80-133176, July 1979.
472
-------
OPERATING EXPERIENCE WITH THE FMC DOUBLE ALKALI PROCESS
By
Thomas H. Durkin, P.E., Plant Manager, A. B. Brown Station
and
James A. Van Meter, Director of Power Production and Procurement
Southern Indiana Gas and Electric Company
Evansville, Indiana
and
L. Karl Legatski, Manager Process Technology
FMC Corporation
Itasca, Illinois
This paper reviews the design and initial operating experience with the
flue gas desulfurization system at Southern Indiana Gas and Electric
Company's (SIGECO's) A. B. Brown Station Unit #1, a 265 MW steam electric
station burning up to 4.5% sulfur coal in a pressurized, pulverized coal
boiler.
After initial checkout in the spring and summer of 1979, the FGD system
began routine continuous operation. Overall operating results for sulfur
dioxide collection, chemical consumption, availability, maintenance
requirements, and operating costs are presented. The problem areas
that contributed significantly to maintenance requirements or non-avail-
ability of the system are discussed in detail. Not counting the scheduled
outage, the system has enjoyed a 96% availability overall in its first
year of operation on a high sulfur coal application. Sulfur dioxide
removal of over 90% has been routinely demonstrated. Overall operating
costs on an annual revenue requirements basis are close to the original
projections.
473
-------
OPERATING EXPERIENCE WITH THE FMC DOUBLE ALKALI PROCESS
SYSTEM DESCRIPTION
The FGD system at Southern Indiana Gas and Electric Company's (SIGECO)
A. B. Brown Station Unit #1 utilizes FMC's patented concentrated double
alkali process for sulfur dioxide control. Figure 1 is a schematic
representation of the process. The central purpose of this paper is a
reliability analysis, for which we have chosen to divide the system
into five major systems, some of which are overlapping:
A. Sulfur dioxide absorption
B. Lime chemical addition
C. Regeneration
D. Soda ash chemical addition
E. Sludge removal and disposal
Design Criteria
The A. B. Brown Unit #1 is a 265 MW steam electric station burning up
to 4.5 percent sulfur coal in a pressurized, pulverized coal boiler.
Make-up water to the FGD system comes from a collector well located
adjacent to the Ohio River. Coal is transported to the site by rail car.
For equipment sizing and redundancy purposes the design basis is keyed
to 23,788 m /minute (840,000 acfm) of flue gas at 138°C (280°F) for gas
handling purposes and 85% collection of 9227 kg/hr (20,300 Ib/hr) of sulfur
dioxide for chemcial capacity.
Sulfur Dioxide Absorption
In the double alkali process, sulfur dioxide is absorbed according to
the following reaction:
Na2S03 + S02 + H20 —> 2NaHS03
sodium sulfite + sulfur dioxide + water —> sodium bisulfite
An important additional reaction is the oxidation of sodium sulfite:
Na2S03 + 1/2 02 —> Na2S04
Sodium sulfite + oxygen —> sodium sulfate
The sulfate ion, which is not active in absorbing sulfur dioxide, can
be partially precipitated by reaction with calcium hydroxide. The
remaining sodium sulfate is purged from the process through the entrainment
of solution in the dewatered calcium sulfite/sulfate solids sent to the
landfill.
The sulfur dioxide absorption is accomplished in the vendor's
proprietary absorber. This absorber is designed to allow high sulfur
dioxide collection efficiencies at a relatively low pressure drop without
474
-------
SODIUM MG6ENERATION
ZN»HSOj»CalOH)2-«-C«SO3«
Figure I. Double Alkoii process schematic including duct arrangement.
-------
the use of spray nozzles. Collection efficiencies above 92 percent
have been demonstrated while operating at less than 12.7 cm (5 inches)
of water pressure drop.
The pH of the scrubbing solution is controlled at 6.5. At this pH,
the scrubbing solution contains equimolar concentrations of sodium
sulfite and sodium bisulfite. This equimolar solution is highly buffered
and can accept rapidly changing flue gas inlet conditions caused by swings
in boiler load and/or changes in coal composition without upsetting the
process control. As sulfur dioxide is absorbed, the ratio of bisulfite
to sulfite increases causing a decrease in pH. A bleed stream from the
.absorber recirculation loop is directed to the lime reactor, and the
absorber reservoir is replenished with regenerated sodium sulfite which
maintains the scrubbing solution pH at 6.5. Maintaining pH levels in a
range of 6.2 to 6.8 is important for several reasons. At a pH above 7.0,
carbon dioxide absorption becomes significant and can lead to carbonate
scaling. At a pH below 6.0, the vapor pressure of sulfur dioxide increases
dramatically and can lead to equilibrium-inhibited sulfur dioxide collection.
Each absorber is about 9.14 m (30 feet) in diameter and 21.9 m (72 feet)
tall to the outlet duct. Superficial gas velocity is approximately 2.7
m/sec (9 feet/sec) at design conditions. There are three stages of discs
and doughnuts in each absorber. A schematic of the absorber internals can
be seen in Figure 1. The bottom 2 m (7 feet) of the absorber comprise
an integral reservoir for the recirculation liquor- To minimize wet/dry
interface corrosion problems, the bottom disc and inlet plenum are made
of Hastelloy G. The rest of the absorber internals are carbon steel lined
with a glass flake polyester resin mastic. The absorber reservoir is
additionally lined with acid resistant brick up to the bottom doughnut for
thermal protection of the lining. A single stage thermoplastic chevron mist
eliminator is provided downstream of the last absorption stage.
The recirculation liquor liquid-to-gas ratio at design condition is
approximately 1.34 L/m (10 gallon/1000 ACF). Recirculation liquor flow
from the integral reservoir to the top of the last disc is provided by
a rubber-lined centrifugal pump, which also provides bleed flow to
the lime reactor.
A small slipstream from the recirculation liquor line is passed through
a pH electrode to monitor and control recirculation liquor pH by
controlling the regenerated liquor return flow. All of the recirculation
liquor piping is fiberglass reinforced polyester for corrosion and abrasion
resistance.
One of the interesting features of the system is the open bypass
arrangement in which the ducting is designed to direct the gas to the
system booster fans or through an undampered bypass duct directly to the
stack. The advantage of this "open" bypass is threefold. First, it
allows upsets in gas flow through the system to occur without affecting
the boiler draft controls. Second, due to the high collection efficiency
of the absorbers, it allows partial bypass of flue gas while maintaining
compliance emissions; this minimizes chemical consumption while providing
up to 11°C (20°F) reheat. Third, gas flow changes can be more readily
accommodated because of the minimization of the number of dampers. Each
476
-------
absorber module is provided with guillotine isolation dampers in the
inlet and outlet ducts. A louver damper is provided at the inlet to
each booster fan far gas flow control. Each booster fan is capable of
providing 14,273 m /minute (504,000 ACFM) at 17.8 cm (7.0 inches) of
water pressure.
Lime Chemical Addition
Two pebble lime storage silos, each 9.14 m (30 feet) in diameter and
26.5 m (87 feet) tall and fabricated from carbon steel, provide 14
days supply at design conditions. Pebble lime is transported to the
site by rail or truck. It is transferred from either storage silo via
a pressure pneumatic system to any of three use bins, one above each
slaker. Each use bin is fabricated from carbon steel and holds
approximately one-half hour's supply at design use rates.
Regeneration
Calcium sulfite is precipitated by lime addition to regenerate sodium
sulfite for use in the absorber according to the following reaction:
2NaHS03 + Ca(OH)2 —> CaS03 ' 1/2 H20 + Na2$03 + 1 1/2 H20
Sodium bisulfite + calcium hydroxide --->
Calcium sulfite + sodium sulfite + water
The regeneration is accomplished in a low-residence-time continuously
stirred tank reactor, which is controlled at a pH of 8.5, the titrametric
endpoint of sodium bisulfite. The sensitivity of the pH control system
is excellent at this set point resulting in effectively stoichiometric
consumption of lime but a relatively wide control band.
The reactor is agitated with a vertically mounted top entry turbine
agitator. The lime is fed to the reactor from two paste-type slakers each
capable of feeding nearly 4990 kg (11,000 pounds) of Ca(OH)2 per hour as
approximately a 20 weight percent slurry. A third installea slaker provides
a 100 percent spare for one of the other two slakers. Each slaker has an
•integral grit removal chamber. The reactor is provided with two immersion-
type pH electrodes (one serves as an installed spare) which monitor and
control the lime reactor overflow pH by controlling the feedrate to the
slakers.
The lime reactor overflows to a 30.48 m (100 foot) diameter thickener tank
where gravity settling of the calcium sulfite slurry takes place. The
thickener concentrates the 1 to 2 weight percent solids in the feed slurry to
20 to 30 weight percent in the thickener underflow.
The regenerated liquor overflow from the thickener flows to the surge
tank by gravity. Water is added to the tank by level control to
maintain system water balance. Regenerated liquor is returned to the
absorbers by a centrifugal pump. There is a 100 percent installed
spare regenerated liquor return pump.
477
-------
Soda Ash Chemical Addition
Soda ash is stored in a wet system. It is unloaded from a truck at
a maximum rate of 9072 kg/hr (20,000 Ibs/hr) into a proprietary apparatus
for converting dry soda ash into the monohydrate crystal form in preparation
for use as a saturated solution. Specifically, saturated soda ash solution
from the tank is sprayed at about 1893 L/min (500 GPM) to wet the incoming
dry soda ash. The wetted soda ash drops into the tank forming a bed of
sodium carbonate monohydrate crystal in a saturated sodium carbonate solution.
As soda ash solution is used in the process, fresh make-up water is
added to the tank, dissolving the crystal bed to maintain a saturated
solution. The advantage of using a saturated solution for chemical
make-up is that it allows sodium addition to the absorbers to be controlled
by the volume of saturated liquor delivered to the absorbers because the
concentration is constant. Soda ash solution from the soda ash storage
tank to the absorbers flows continuously in a loop to minimize concentration
and temperature gradients within the solution layer above the crystal
bed in the storage tank. This also helps prevent crystallization in the
soda ash transfer lines. In addition, all soda ash piping is heat
traced and the tank is also provided with steam plate coils in order to
maintain solution temperature. Flow through the transfer loop and to
the absorbers is provided by one of two centrifugal pumps.
Sludge Dewatering and Disposal
The sludge dewatering equipment consists of three rotary vacuum filters,
each sized for 33 1/3 percent of total capacity required when burning
the maximum sulfur coal (4.5 percent). When burning the nominal coal
(3.7 percent sulfur) each filter is essentially a 50 percent filter.
Thickener underflow is pumped to the filter vats by an air operated diaphragm
pump. There are two full flow underflow pumps installed per filter. The
rotary vacuum filters are primarily of carbon steel construction. The
filters are knife-discharge type, and the 50 to 60 percent solids cake is
discharged directly into dump trucks for transportation to the on-site
landfill area. Each filter is equipped with a wash belt compression assembly
for applying wash water to the cake to enhance sodium recovery.
SYSTEM AVAILABILITY
Overall system availability as defined by PEDCO for the first 13 months
of routine operation beginning in Augusts 1979, is summarized in Table 1.
While we feel that the definitions of some of the PEDCO parameters leave
something to be desired from a utility point of view, they at least
provide a consistent basis for comparison.
Table 2 shows the incidents that contributed to system unavailability for
the same period. The total scrubber forced outage rate was 3.3%. At SIGECO,
a forced outage rate for the boiler and turbine of 1% is considered good, and
the goal at A. B. Brown. There exists here a good comparison of boiler-turbine
to scrubber state-of-the-art design. Forced outages on the scrubber occur
at a frequency of three plus times what we strive for on the remainder
of the unit. In addition to the normally scheduled annual outage, we feel
that an additional outage is required each year for inspection of ducts,
linings and breechings, due to the possibility of corrosion.
478
-------
TABLE 1 PEDCO INDICES
Availability Operability Reliability Utilization
August, 1979
September
October
November
December
January, 1980
February
March
April
May
June
July
August, 1980
96
99
88
97
81
98
99
45
100
100
100
92
95
93
76
99
96
81
98
70
65
100
100
100
93
95
93
88
99
97
81
98
98
68
100
100
100
93
95
93
76
86
88
81
98
69
27
96
83
98
88
95
TABLE 2 SOURCES OF UNAVAILBILITY
August 1, 1979 to August 31, 1980
Unavailability Cause
Recirculating Pump Failure (2)*
Miscellaneous Electrical Trips (4)*
Slaker Feeder Controls
Thickener Rake Stall or Overload (3)*
Isolation Dampers
Lime Reactor Overflow Elbow
Thickener Underflow Pump Suction Plugged (2)*
Lime Transfer System Plug-up
Rotary Filters Unavailable
Low Water Pressure
Subtotal
Scheduled Outages
Total
Hours
Period Hours 9528
Availability 90.5%
Discounting Scheduled Outages 96.7%
*Denotes number of incidents, if greater than one.
479
-------
The following section provides a more detailed chronological discussion
of the incidents that contributed to unavailability.
Month-By-Month Performances
The period of March through July, 1979, constituted the initial check-out
and debugging of the system. There were some areas which required
modifications for mechanical improvements to cope with situations not
envisioned during the original design phase. Details on these modifications
were presented in an earlier paper by Durkin et al (1).
The system availability began to improve dramatically in August as
operator attention and awareness increased and as mechanical problems
became fewer. The system was down the first day of August for a recirc-
ulation pump impeller lining failure. The only other system outage was for
the removal of tramp metal from the lime transfer system. These two
items account for 32 hours of down-time.
September's record shows 48 hours down for an FGD system water filter
tie-in, necessitated by high suspended solids in the service water.
A failure of a 460 volt power cable to one of the recirculation liquor
pumps caused a 600 amp main feeder to overload and trip. It took only
two hours to reset the feeder and restart all of the systems. The spare
recirculation liquor pump was put in service and full FGD system operations
resumed. For two separate periods totalling 51 hours, the FGD system was
bypassed so that precipitator particulate emissions could be tested.
There was one short booster fan trip caused by a mis-wired relay which
was not to be diagnosed until several weeks later. At the end of the
month, the filter cake quality got so poor that the dump trucks could
not handle it. We elected to shut down the FGD system so that the filter
building could be cleared of spilled cake.
The filter building clean-up continued into the first day of October.
Again, a short duration booster fan trip occurred. On October 28, the
FGD system and boiler were taken off-line for a scheduled precipitator
and FGD system inspection.
The inspection outage continued through the first three days of
November. Once put on-line, the only system outage was due to electrical
problems of the slaker control printed circuit boards. It took about
12 hours for the cause to be diagnosed and repairs made.
The most significant outage occurred in December. After a short duration
boiler trip, the thickener rake was discovered to be stopped. Numerous
attempts were made to restart the rake to no avail. Finally, the thickener
had to be drained and the solids removed to free the rake. The system
outage lasted 146 hours.
January's record shows only 12 hours of downtime. The thickener rake
torque instruments indicated a steady increase almost to the point of
motor overload. Since we were fearful of another stall, the system was
shut down. After several hours of filtering, the torque dropped and the
system put back on line.
480
-------
We experienced a period of poor quality filter cake in February, similar
to September's situation. We elected to take the FGD system off-line
for building and roadway clean-up.
March downtime was split between two outages on the balance of plant
and the completion of roadway work begun in February. Availability was
45% based on 444 boiler hours.
May downtime was due to an outage on the make-up water line to the cooling
tower during which some internal scrubber inspection work was done.
July downtime was caused by two scrubber recirculating pump impeller liner
failures in three days.
August downtime was caused by pluggage of the thickener underflow pump
suction header, and a couple of other minor problems.
Overall, since the FGD system reached stable operation early in August
of 1979, it has operated at 90% availability and a 3.3% forced outage
rate. The FGD system's longest continuous run has been 54 days, and it
had run 71 days between partial forced outages.
RELIABILITY ANALYSIS
While the total availability record at A. B. Brown has been good compared
to most other FGD systems now in operation, our goal is to achieve a
forced outage rate of 1%. To better understand how we can improve our
availability in this and future installations, we have performed a
reliability analysis on each of the five major systems described in
the first section of this paper. The systems are not mutually exclusive,
meaning that an element may be included in more than one system. The
thickener tank is an example, since it is included in both the regeneration
system and the sludge removal system. Also, the systems do not necessarily
follow the flow of one fluid, but generally follow the series of events or
reactions that must occur to insure availability and compliance.
Figure 2 shows the block diagram of the five systems. All of these
systems are integral to scrubber availability and the failure of any
system will result in scrubber downtime. Reduced capacity of a system
may-result in unavailability, but most likely will be considered available
at low load. With a total FGD system availability of 96%, shown at the far
right side of Figure 2 the contribution of each system is shown in the
.lower right hand corner of each box. The decimal1 in the lower left hand
corner is the sizing of each part, compared to the boiler full load gas
flow or maximum sulfur dioxide collection.
Sulfur Dioxide Absorption (A)
Figure 3 illustrates the elements of the sulfur dioxide absorption system.
The gas contact systems have performed at a reasonable level of reliability,
99% plus, but not without some problems. The absorbers have operated
perfectly, with no pluggage and only minimal lining deterioration. The
single stage chevron type mist eliminator has given us some problems. In
481
-------
00
ro
3000 PPM
302
FROM BOILER
SLUDGE REMOVAL
AND DISPOSAL
1.0 "E" 97%
CHEMICAL ADDITION
SODA ASH"D"
1.0 100%,
3000 PPM
S02
FROM BOILER
GAS CONTACT NORTH
0.6*
99%
400 PPM
REGENERATION
"c"
1.0
98%
CHEMICAL ADDITION
LIME "B"
1.0
99%
GAS CONTACT SOUTH
"A"
0.6
99%
400PPM SO?,
GAS TO STACK
96%
* DENOTES SUBSYSTEM FRACTION OF
TOTAL SYSTEM DESIGN CAPACITY.
Figure 2. Reliability Block Diagram of FGD major systems.
-------
oo
OJ
TO SYSTEM V
REGENERATION
RECIRC.
PUMP
0.6 98%
SCRUBBED GAS
TO STACK
SYSTEM "A"
99 % AVAILABLE
BOOSTER
FAN
ABSORBER
0.6 IOO%
CONTROL
VALVES
GAS FROM
PRECIPiTATO
RECIRC.
PUMP
0.6 98%
FROM SYSTEM D"
iSOLATION
DAMPER
SODA ASH ADDITION
QUENCH
SPRAY
FROM SYSTEM V
REGENERATION
0.6 100%
System MA\ gas e©rstaet (ont sids only)
-------
June of this year, recirculating liquor pumps in the south module began
to fail at an alarming rate. Cavitation was the suspected problem,
caused by pluggage of the pump suction. Inspection revealed that two
sections of the mist eliminator had collapsed into the scrubber sump,
blocking the suction pipe. The mist eliminator sections appeared to slip
off the 7.5 cm (3 in) wide shelf supporting them, and the Inconel tie-wires
were not adequate to hold them.
The mist eliminators seem to perform adequately when in place. However,
in recent months there has been evidence of increasing losses of sodium
due to entrapment. The problem appears to coincide with the collapse of
several sections of the mist eliminator. The losses are primarily in the
form of high sodium levels in the drainage from the stack. Sodium levels in
the exit gas, as determined by analysis of EPA Method 5 particulate catches,
have been consistently low, in the range of 0.005 to 0.006 Ib/mm BTU, or
about 10% to 15% of the total particulate. This suggests that the entrained
drops are large enough that they fall down the stack. Stack drain losses of
sodium may amount to 1 to 2% of the SOp collected, or up to more than
one-half of the total excess sodium consumption, which will be discussed in
more detail in a later section. Steps are now being taken to repair the
existing mist eliminator and add a second level of eliminators to reduce
losses.
The rubber-lined recirculating liquor pumps have performed about as
expected. During engineering, pump lining and impeller life were
estimated at one year and the pumps were spared accordingly. Disregarding
cavitation problems and some sub-par work on original installation, our
experience has been reasonable.
The system design calls for three levels of protection for the scrubber
lining: 1) cooling with recirculated liquor with the recirculation
pumps; 2) isolation dampers; 3) water quenching of inlet gas. Plant
experience with the guillotine isolation dampers has been poor. Original
materials of construction caused some problems, and of late, seemingly
minor problems such as a stuck relay in the entry door logic have led
to major problems. In addition, the lead time for fabricated pieces
of a material suitable for a wet flue gas environment is truly ridiculous.
As a result, the plant relies on water quenching of the incoming gas to
protect scrubber internals more than we rely on dampers.
The outlet duct from the absorber to the outlet isolation damper is
coated with a flake glass lining. In the outlet duct from the damper to
the stack, which is lined with an epoxy vinyl ester coating, the lining
has failed dramatically on one module, subjecting the carbon steel structure
to severe corrosion. This duct segment is exposed to cool 54°C (130°F)
saturated gas when the scrubber is in service and to hot 149°C (300°F) gas
those few times it is off. An appropriate low carbon, high molybdenum
stainless steel will be installed at the next scheduled outage as a
replacement of the corroded duct. The corresponding duct segment on the
other module will be lined with a new experimental lining we want to try.
This duct section was patched from the outside with the unit on line, and
has caused no downtime.
484
-------
Lime Chemical Addition (B)
The lime chemical addition system shown in Figure 4 has been responsible
for less than 1% of tne FGD system 3.3% forced outage rate, but this 99%
availability is not an accurate measure of the strength of the design and
operating capabilities of this system. There have been numerous instances
of reduced capacity because of inadequate lime supply to regeneration.
However, the surge capacity of the thickener tank has helped us minimize
any lost time or non-compliance. The system can run for several hours
before a reduction in scrubbing capacity, due to the storage of 3,028
kiloliters (800,000 gallons) of regenerated liquor in the thickener.
Circulating low pH (6.5) liquor through unlined carbon steel piping
and tanks does cause some accelerated corrosion and at every outage great
care is taken to ascertain material integrity with a view toward possible
future replacement. To date, no problems have been seen, but we are
installing additional monitors to record the pH in these unprotected
areas. As our experience grows, we will be establishing some lower range
cutoff points below which we will not operate, based on engineering
judgments.
Interruptions in lime supply are caused by failure of the lime transfer
system or by foreign material in the lime. The transfer system itself,
while it has performed reasonably well, is recognized by plant personnel
as a weak link. The system has only one blower and one feed line going
from the two storage silos to the slaker use bins. A malfunction of
any one component will shut down the entire system, and result in low
pH incidents described previously. We are engineering some redundancy
for lime transfer to the slakers. As a system, lime transfer has been
only 98.5% available.
Regeneration (C)
The second largest cause of scrubber forced outages is the regeneration
system. Figure 5 shows the elements of the system. Slakers are Included
in both system B, lime addition, and C, regeneration. The regeneration
area has not caused much downtime on the scrubber, but as with the lime
addition system that is not a fair measure of how well it performed.
Our incidence of reduced capacity due to regeneration problems has been
significant.
As designed, pH at the lime reactor discharge is fed back to the slaker
feeder controls to control lime addition. The pH normally swings plus
or minus one unit from the setpoint. The control problems in this
situation are obvious. As a result, slaker feeders are constantly
varying in response to both load changes and normal swings. This causes
greater than normal wear on the rotary air lock feeders, and they overcharge
the slakers occasionally, causing plug-ups. Our early experience with
lime feeder controls was very poor. Slaker control problems are often
attributable to the dusty, wet environment of the slaker building.
All local control panels are mounted adjacent to the slakers, and subjected
to the same ravages of steam, caustic, and water as the slaker itself.
NEMA dustproof rated enclosures have brought these problems to a manageable
level.
485
-------
00
o>
UNLOADING a
STORAGE
1.0 100%
«fr
pt
LIME
TRAfdCFfrp
1.0 94%
tot
p
WATER
ClBppJ V
1.0 99%
|->
... fih
IF
- feft
SLAKER 1
0.5 94%
SLAKER 2
0.5 94%
SLAKER 3
0.5 94%
Figure 4. System B , lime chemical addition
-------
SLAKER I
0.5 94%
SLAKER 2
0.5 94%
SLAKER 3
0.5 94%
SYSTEM "A"- NORTH
BLEED VALVE
0.6 99%
LIME
REACTOR
1.0 99%
SYSTEM "A"- NORTH
BLEED VALVE
0,6 99%
pH CONTROL
1.0 99%
REGENERATION
SUR6E
1.0 100%
THICKENER
1.0 98%
Figure 5. System "C",
-------
The lime reactor overflow elbow has been patched several times and is
scheduled for replacement in the near future. Failure analysis questioned
the suitability of carbon steel in a service where it is exposed to
abrasion from lime grit and corrosion from occasional pH excurisons.
The remainder of the regeneration system has operated well with minimal
problems, and has offered the level of reliability that is expected
from power plant machinery.
Soda Ash Chemical Addition (D)
The soda ash chemical addition system shown in Figure 6 has not contributed
to any FGD system downtime, although low sodium concentration in the system
has resulted in some non-compliance. This is not seen as a problem though,
and with fourteen days' inventory available 1n the tank, no changes are
foreseen*
Sludge Removal and Disposal (E)
As shown in Figure 2, the sludge removal and disposal system has been the
weakest of the five systems, due to;both mechanical and process difficulties.
This weakness would be even more costly in terms of availability if the
system did not have some surge capacity in the thickener tank, which allows
the plant to run as long as eight hours at full load without filtering.
However, this is not a preferred operating mode, and it is not without
some detrimental system effects, so in deciding to utilize the thickener
surge capacity the value versus the consequences of continued operation
must be carefully considered.
The detrimental side effects of utilizing the thickener surge capacity
are as follows. First, as part of our filter cake quality testing, it
was shown that increasing the inventory of solids in the thickener tanks
tends to increase the pH of the underflow, apparently due to the continuing
reaction of small amounts of alkali. This increase in underflow slurry
pH coincides with a deterioration in cake quality. Second, the thickener
is designed with a pivoted rake and no powered rake lifting mechanism.
The rake and drive are protected from overload by a spring loaded clutch,
designed to trip the rake drive when it reaches its torque rating. If
the rake were to trip with a high inventory of solids in the tank, it
would sink into the mud and be impossible to restart. This would require
a lengthy outage to pump out the thickener tank with the loss of $50,000
to $100,000 worth of chemicals. Third, operation of the gas contact system
without filtering will increase the total volume of material -- solids
and liquids -- in the thickener tank and result in liquid losses because
of overflowing the surge tank capacity. This costs money in chemical
losses, amounting to 10 to 20% of the excess sodium consumption.
Figure 7 shows in greater detail the reliability analysis of the sludge
removal and disposal system, and shows elemental availabilities much lower
than the 97% for the system. Redundancy is the key, along with high
maintenance requirements and quick response to problems. Availabilities
of the filters have been in the mid eighties, but the problems have been
generally unrelated to design. Rotary vacuum filters are high maintenance
488
-------
&
pp
SLUR-0-LYZER
TANK
1.0 100%
fe,.
i
k r
TRANSFER
PUMP i
1.0 !00%
tot
t*
PIPING LOOP
8 HEAT TRACE
1.0 100%
£
1.0
!00%
TRANSFER
PUMP 2
1.0 100%
Figure 6. System UD",
Soda ash chemical addition.
-------
•JD
O
PUMP H
0.5 80%
TRUCK I
0.3 50%
0.5 88%
THICKENER
FROM LIM
REACTOR
0.3 50%
SURGE
CAPACITY
1.0 8 HRS.
TRUCK 5
0.3 50%
HAUL ROAO
TO LANDFILL
SYSTEM V
% AVAILABLE/
Figure 7. System E , sludge removal and disposal
-------
items. Cloth life is only one to three months, and between cloth changes,
maintenance is required regularly to keep caulking ropes in place and to
repair holes in the cloth. A large portion of our filter problems concern
the vacuum pumps and filtrate return pumps becoming overloaded with solids
earned over from the filter, generally through holes in the cloth. The
solids appear to be mainly grit discharged from the slakers. The lime
supply at 88% - 90% available CaO, has a reasonably high amount of grit which,
no doubt, contributes to this condition.
Another problem area is the underflow pumps, which have had only 80%
availability. Here the problems are basically design related. The pumps
are air operated double diaphragm type. Air supply to the pumps is
controlled by the filter vat liquid level. As vat level drops, the
pumps are energized to refill vat level. This constant on-off operation
allows the pump, the suction lines, and the discharge lines to lie full
of thickener underflow slurry. They will eventually clog. Various
operating procedures to flush the system on startup and shutdown have
lessened the problem but not eliminated it. It appears that any design with
low or intermittent flow in this critical area is a weak one.
The final element in the sludge removal and disposal system is the trucks.
A. B. Brown is using tandem axle dump trucks to transport filter cake
to the landfill. Two to four trucks are required for full load operation
depending on the combination of filters in service. Our experience has
been very poor in this area. With five trucks assigned to filter cake hauling,
we have had three or more available for load only 75% of the time. This
is another area where the surge capacity of the system comes into effect.
It gives us enough time to repair either the right combination of filters or
trucks to maintain availability and compliance.
Truck problems have been in two general areas: drive train (transmission
and axles) and tailgate. Both of these are contributed to very heavily by
the condition of the sludge. Wet, soupy material is not only hard to
handle in the landfill but exceptionally hard to drive through, and it
exerts a great deal of hydraulic pressure on the tailgates.
Trucks frequently must be taken out of service to clean beds. This
situation occurs not just in winter when cold weather causes some freezing
to beds, but also during hot weather. We have experimented with plastic
liners and feel that they are an improvement but not a cure-all. Many
of our truck problems could probably be solved with more suitable trucks.
However, the existing filter building layout limits our choices.
In addition to the mechanical aspects of the sludge removal system
performance, there are process considerations. There have been extreme
variations in cake quality that are not entirely due to mechanical
conditions. While difficult to quantify, it is clear that poor cake
quality increases the cost of operation and maintenance and negatively
impacts availability.
A large effort in the last six months has been devoted to understanding the
variations in quality, and while we're not completely satisfied that we
understand what's happening, we have found some general truths. First,
491
-------
it Is clear that cake quality is directly related to crystal size and
shape; moisture content is not so important, i Second, as noted above,
elevated pH generally deteriorates cake quality, at least in this application.
Third, transient process conditions are bad; our worst cake has always
been associated with situations in which we changed solution chemical
composition rapidly.
Finally; we believe there is a relationship between sodium sulfate
concentration and cake quality. Sodium sulfate concentration seems to
affect the size and shape of the calcium sulfite crystals formed in the
regeneration section. All other things being equal, the greater the
sulfate concentration the larger the size and the more irregular the
surface of the crystal, as shown in Figure 8.. Crystals with an irregular
porcupine-like surface are called radial crystals. The larger, more
radial crystals result in a filter cake with better handling properties.
As noted earlier, sodium sulfate formed in the system is purged from the
process through the entrainment of solution in the dewatered filter
cake. Since the fuel sulfur level is high, there is a large amount of
cake formed in this system. This means that the concentration of sodium
sulfate in the scrubbing solution does not need to be very high to
maintain equilibrium in the system. (The amount of sulfate formed is
relatively fixed for a given load. Thus, the more cake formed the lower
the concentration of sulfate in the cake.) The situation is aggravated
by transient losses of sodium sulfate through carryover into the gas
stream or spills.
There are three general categories of sodium loss to consider: entrainment,
filter cake, and miscellaneous spills and leaks. Entrainment, or stack
losses, was discussed earlier, and we hope to correct that problem during
our next outage. Filter cake washing, while steadily improving, is still
not up to design. We plan to increase our hot water capacity in the FGD
area in an attempt to improve our washing. Spills, a problem in early
operation, have been brought under control in recent months, but we still
are considering increasing our surge tank capacity to allow for greater
fluctuations in the water balance. We are hopeful, but still not certain,
that these improvements will allow us to simultaneously reduce sodium
consumption and improve cake quality.
While we endeavor increasingly to implement the above improvements in
our routine operation, there are aspects of cake quality that we still
do not understand to our satisfactaion. Continuing research projects
at FMC's Central Engineering Laboratories and Purdue University, together
with experiments at A. B. Brown, will hopefully eventually lead to a
scheme for consistent production of easily handleable cake and a better
fundamental understanding of the numerous parameters that effect the cake
quality.
Conclusion
Typically, in the operating life of mechanical equipment, there is a
break-in time during which breakdowns are frequent; a useful operating
life during which breakdowns are at a low, manageable level; and finally
a wear-out period when failures increase dramatically. Without changes
492
-------
ZOOOx, low sulfate
10,000x, "low sulfate
2000x, medium sulfate
10,000x, medium sulfate
2000x, high sulfate"
105000x, high sulfate
Figure 8. Effect of increasing sulfate concentration on crystal structure.
493
-------
and improvements, we feel that the availability of the equipment has
reached its high point. With that in mind and remembering that our
scrubber forced outage rate is three plus times what we would like, we
are engineering system improvements in these areas:
1. Filter cake quality
2. Underflow pump recirculation
3. Duplicate lime transfer system
4. pH controls
We hope that these improvements will affect the wear-out period and
allow us to improve our availability. Some other problem areas addressed
in this paper are locked into the system because of original design,
layout, or available space. Frankly, we feel that it is unrealistic to
expect the first utility installation of a new FGD technology, even
with the process advantages of double alkali, could ever achieve the level
of reliability and availability that the balance of the system achieves.
However, we feel that the problems we have had are primarily mechanical,
and correctable, in future installations.
OPERATING COSTS
The variables in the FGD system operating costs are: operating manpower,
maintenance expenses (labor and materials), lime usage, soda ash usage,
power and landfill costs. In analyzing these costs for equation back
to the manufacturer's original estimate, upon which the selection of the
double alkali scrubber over a lime or limestone system was based, several
adjustment factors must be applied to the costs incurred in order to
put them on the same basis as the specification and proposal.
The system is designed to handle 265 MW gross of flue gas at 143°C (290°F)
and 5.0% Op. Coal burned was to be a maximum of 4.5% sulfur, 26,749,000
joules per kilogram (11500 Btu per pound) and 0.05% chlorides. Cost analysis
was based on 70% load factor. Actual experience has been off-evaluation
on some key items: sulfur has averaged much closer to 3.5%; the unit load
factor has been 58% rather than 70%; and the flue gas volume and excess
air have been higher than anticipated. All these items significantly impact
the cost comparison of actual to guarantee. Rather than go through a
laborious explanation of each variable with all applicable adjustments, it
is most likely more informative to qualify each.
Operating manpower was based on one additional man per shift. Our
experience has been very favorable in this area. In all but extremely
unusual circumstances, one local FGD system operator has been sufficient
to operate all equipment. The scrubber control board is located in the
main control room, adjacent to the boiler-turbine-generator control panel,
and this design feature has been a big manpower saver. Existing control
room personnel operate the panel.
Maintenance costs (including electrical and instrument) were predicted
to be 1.5% of capital costs per year. Both labor and material were
included in the 1.5% figure. Our experience has shown this to be low
by about 50% after allowing for inflation. Future maintenance costs
494
-------
are expected to increase as we approach the "wear out" period on some of
our major pieces of equipment. The 1.5% maintenance was a concession
made during bid analysis to the claimed features of the double alkali
process versus 2.5% of capital for lime systems and 4.0% of capital for
limestone systems. While our experience has not been quite that good,
experience by others would indicate that our predictions for lime and
limestone were also low.
Excess lime usage was predicted to be 1% or less (stoichiometric ratio -l.Oi;,
Our experience has shown excess lime usage as low as 0.02% (stoichiometric
ratio = 1.002) to be attainable under normal operating conditions.
Soda ash usage was predicted to be 2.5% of the moles of S0? collected
plus 1% allowance for the chlorides in the coal. As alreaay discussed,
our usage has been higher than expected and the increase over design has
been attributed by FMC to inability to wash the filter cake with adequate
amounts of hot water and to system losses through spills and stack drainage.
It is anticipated that improvements in soda ash usage will be made. Soda
ash prices have jumped sharply in the past year, and the price now is
almost double what was budgeted in 1976. However, soda ash is not a
large percentage of total operating cost and we are hopeful that recent
price increases, which were caused by some unexpected closings of obsolete
synthetic soda ash plants, will not recur.
Power requirements were predicted to be about 0.8% of net generation
at full load. Our experience has been favorable in this area, with the
FGD system using slightly less than the predicted amount.
Landfill costs were predicted at $2.00 per ton of material. Our experience
to date with the landfill operation has not met our expectations. The
double alkali filter cake has generally been of a poorer quality than we
anticipated, contributing to the problems and expense at the landfill.
The two dollars per ton (1977 dollars) would nave been adequate were it
not for the difficulties resulting from the cake quality.
The double alkali system was purchased based on a lower evaluated cost.
In cents per kilowatt hour, FMC's process was estimated to cost 0.269
vs. 0.306 for the next closest system, as reported by Wagner (2). Applying
all appropriate correction factors to our experience, and inflating the
other systems costs by actual reagent costs and the consumer price index
for other items, the FMC process still exhibits the lowest cost for this
installation, although not by as wide a margin.
CONCLUSION
In conclusion, SIGECO feels that the Double Alkali System installed at
the A. B. Brown station can be successfully operated by utility personnel
and can meet the requirements of Federal New Source Performance Standards
while burning high sulfur midwestern coal.
495
-------
REFERENCES
1. Durkin, T. H., Jaslovsky, G. S., and Boward, W. L., Operating
Experiences with a Concentrated Double Alkali Process, American
Power Conferences, April 23, 1980.
2. Wagner, N. P., Adams, L. J., and Ramirez, A. A., Technical and
Economic Factors for Evaluating Flue Gas Desulfurization Technologies,
American Power Conferences, April, 1978.
496
-------
STATUS REPORT ON THE WELLMAN-LORD/ALLIED
CHEMICAL FLUE GAS DESULFURIZATION PLANT AT
NORTHERN INDIANA PUBLIC SERVICE COMPANY'S
DEAN H. MITCHELL STATION
E. L. Mann
Northern Indiana Public Service Company
Hammond, Indiana
and
R. C. Adams
TRW Inc.
Research Triangle Park, North Carolina
ABSTRACT
The Northern Indiana Public Service Company and the U. S. Environmental
Protection Agency entered into a cost-shared contract in June of 1972
for the design, construction, and operation of a regenerable flue gas
desulfurization (FGD) demonstration plant. The system selected for the
project was a combination of the Wellman-Lord S0? Recovery Process and
the Allied Chemical S0? Reduction Process. The FGD plant was to be
retrofitted to NIPSCO's 115 MW pulverized coal-fired Unit No. 11 at the
Dean H. Mitchell Station in Gary, Indiana. NIPSCO entered into contracts
with Davy Powergas, Inc., for the design and construction of the FGD
plant and with Allied Chemical Corporation for operation of the plant.
The FGD plant acceptance test was successfully completed on September
14, 1977. The plant completed a two-year demonstration test period
during which information was collected and reported regarding pollution
control performance, secondary effects, economics, and reliability of
the system. TRW, Inc. was the independent evaluator for the EPA through
October, 1979. A follow-on EPA/NIPSCO contract of seven and one-half
months has recently been completed. Operation of the plant continues.
497
-------
PROCESS DESCRIPTION
Wellman-Lord S02 Recovery
The Wellman-Lord process consists of three major operating sections -
S02 absorption, purge treatment and S02 regeneration.
In the S02 absorption section, residual fly ash in the flue gas is
removed by water scrubbing. S02 is then, removed from the flue gas by
scrubbing with a solution of sodium sulfite. The chemicals contained in
this solution remain completely dissolved throughout the absorber. Flue
gas scrubbing with a clear solution, free from suspended solids, plugging
and scaling, is a fundamental reason underlying the exceptional on-
stream reliability experienced in the commercial operations of the
Wellman-Lord process.
The purge treatment section selectively removes inactive oxidized sodium
compounds from a sidestream of the absorbing solution and converts this
material into a dry granular product which is marketed.
The third section of the Wellman-Lord process involves thermal regenera-
tion of the absorbing solution to release the absorbed S02 as a concen-
trated gas stream and return of the reconstituted solution to the absorber.
The concentrated S02 gas may be converted to liquid S02, sulfuric acid
or elemental sulfur. NIPSCO elected to use the Allied Chemical S02
Reduction Process to convert to elemental sulfur.
Allied Chemical S02 Reduction to Sulfur
Sulfur is recovered by Allied Chemical's S0? reduction process which
consists of two principal operating sections.
In the primary reduction section, more than one-half of the entering S02
is converted to elemental sulfur. A key feature of this section is the
effective control of chemical reactions between S02 and natural gas over
a catalyst developed by Allied Chemical for this purpose. Heat generated
by these chemical reactions is recovered and utilized to preheat the S0?
gas stream entering this section.
Packed bed regenerative heaters provide a rugged and efficient means for
achieving this heat exchanger function. The process gas flow through
498
-------
the regenerators is periodically reversed to alternately store and
remove heat from the packing; hence, the overall section is thermally
self-sustaining.
Automatic control of the flow reversing cycles and other process con-
ditions achieves optimum performance in the system, with high sulfur
recovery efficiency and reductant utilization at all operating rates.
The gas leaving the primary reactor system is cooled in a sulfur condenser,
for condensation and recovery of sulfur product. The remaining gas,
containing proper proportions of S02 and H2$ is processed through a
Claus conversion system for recovery of additional sulfur product. The
Claus system off-gas is incinerated and recycled to the Wellman-Lord S02
absorber (see Figure 1). Since startup in 1977, 5843 long tons of
sulfur have been produced.
PROCESS CHEMISTRY
The Wellman-Lord 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 SCL, converting sodium sulfite to sodium
bisulfite:
(a) S02 + Na?S03 + H20 -* 2 NaHS03
The sodium bisulfite solution is regenerated by thermal decomposition.
Application of heat simply reverses equation (a):
(b) 2 NaHS03
The SCL is recovered in a concentrated stream.
The concentrated stream of S02 gas is then reduced to high purity elemen-
tal sulfur in the Allied Chemical Process. This conversion is carried
out in two steps. In the first step, a portion of the SCL in the feed
gas. reacts with natural gas, yielding a mixture of elemental sulfur,
hydrogen sulfide, carbon dioxide and water vapor:
(c) 2CH4 + 3S02 "» S + 2H2S + 2C02 + 2H20
In the second step, H2S formed in the first step reacts with the remaining
S02 yielding additional elemental sulfur and water vapor:
499
-------
CM
Secondjry
Condenser
in
O
o
Figure 1.
Wellman-Lord Recovery/Allied Chemical S02Reduction
General Schematic Flow Diagram
NIPSCO Mitchell Unit No. 11
-------
(d) 2H2S + S02 ±-— 3S + 2H20
The tail gas from the sulfur plant is incinerated and recycled to the
Wellman-Lord absorber.
TEST PROGRAM RESULTS
This section includes an analysis of the test results from the EPA
evaluation that was conducted by TRW. The analysis focuses on the last
thirteen months of the two-year demonstration period.
Description of Test Program
The test program as originally designed consisted of three major test
phases:
(1) a baseline test
(2) acceptance testing
(3) a one year demonstration test and evaluation
1 2
The initial baseline and acceptance tests have been described in detail.
The acceptance test was successfully completed in September of 1977 and
the scheduled one year of operation for demonstration testing followed
immediately. During the demonstration year, operating experience was
limited due to both boiler and FGD related operating problems. Operating
experience and operating problems were described at the FGD Symposium
held in March of 1979. The test results have been reported. These
test results were inadequate for fully evaluating the FGD process
because of those upsets caused by the boiler and thus external to the
FGD plant. Modifications were begun during the latter half of the
demonstration year that prompted the decision to continue with a demon-
stration test program for another full year. In this report, we will
focus on the operating and S02 removal performance of the Well man-
Lord/Allied Chemical FGD unit during the second year of demonstration.
The period covered is from October 1978 through October 1979. It was
preceded by a second baseline test that provided up to date performance
and operating data on the boiler while the FGD plant was down and com-
pletely isolated from the boiler.
501
-------
As we stated earlier, modifications were begun during the latter half of
the first year of demonstration. Except for insulation of the inlet
ductwork, these modifications were completed during a scheduled boiler
shutdown in September 1978. Our data show that the modifications ulti-
mately enhanced the performance of the FGD unit and of the boiler. The
boiler was utilized during 93% of the second year of the demonstration.
With a dependable supply of flue gas to feed the FGD plant, conditions
were quite favorable for gathering test data. The modifications that
provided substantial improvement were as follows:
(1) Use of Captain coal. Coal feeding problems were minimized
when this coal was used. Other corrective action for improving
coal feeding were to enlarge the coal mill feed chutes and to
increase capacity of the coal mills.
(2) Elimination of a boiler feedwater problem.
(3) NIPSCO agreed to remove a part of the heat transfer surfaces
from the Ljungstrom air preheaters at some penalty in boiler
efficiency. With this modification, flue gas temperatures
were maintained above the dew point and booster blower problems
caused by wet operation were eliminated.
(4) Electrification of the FGD evaporator circulating pump.
Conversion from steam turbine to electrical drive reduced the
startup time.
The test program demonstrates performance of the Wellman-Lord/Allied
Chemical FGD process in these four major areas:
(1) Dependability of the FGD unit
(2) SOp removal performance
(3) Energy and raw material consumption
(4) Cost
The TRW test installation provided 10 or more measurements per hour of
flue gas composition, steam and electrical energy consumption, and the
boiler operating parameters of interest. One hour averages computed
502
-------
from these data served as the primary data base for most of the data
interpretations. The amount of raw materials, natural gas and soda ash,
and sulfur production were measured less frequently.
Summary of Results
The test program was designed to demonstrate guaranteed performance of
the Wellman-Lord/Allied process and its ability to meet these performance
criteria in a long term dependable manner and relative to the specific
flue gas conditions at the host site. Since the FGD plant was designed
and sized for a specific load factor and specific flue gas characteris-
tics, the test also evaluated its operability over the normal range of
load variation and flue gas composition experienced during the second
demonstration year. The results are summarized as follows:
1. Reliability of the FGD unit, hours operated/hours called upon
to operate, was 61%. The reliability record was established
with virtually no redundancy built into the FGD unit. Also,
the evaporator was designed for only 80% of full boiler load.
With limited surge capacity within the regeneration loop, the
FGD plant was not able to operate to effect complete S02
recovery during evaporator or reduction unit shutdowns.
2. The major sources of interruptions were
the reduction unit
the evaporator circulating pump
the booster blower
the evaporator
startup time
3. Twenty four-hour average S02 removal efficiences of 85% to 92%
were typical. The pounds of S02 emitted per million Btu of
heat input varied from 0.25 to 0.94.
4. S02 removal was attained at boiler loads in the range of 53
MWe to 85 MWe of the 115 MW boiler. Some operation was achieved
up to 93 MWe. The lower limit was set by the limiting turndown
capability of the reduction unit. The upper limit was set by
503
-------
the capacity limitation of the evaporator as designed. This
would not have been a limitation had the evaporator been
designed to match the full SOr, removal capability of the
absorber. Since a substantial amount of energy largely as
boiler main steam was consumed by the FGD plant, the generating
potential of the boiler was actually about 95 MW at the FGD
maximum capacity limit of 85 MWe.
S02 removal was attained from flue gas with the following
characteristics relative to design:
• SCL feed in excess of the expected plant capacity of about
5400 Ib/hr was successfully treated for sustained periods
of 24 hours and greater without loss of SOp removal effi-
ciency. Overall for the second demonstration year, S02
feed averaged 4700 Ib/hr.
e Flue gas flow rates were usually higher than the expected
flow rate of 320,000 acfm by a substantial amount. All
flue gas flow rates in this report are at the design basis
of 300°F and one atmosphere.
• Inlet temperature, following modifications to the air
preheater of the boiler to obtain higher temperatures,
averaged 305°F. Design basis temperature was 300°F.
The steam consumed by the FGD plant amounted to about 11% of
the boiler input energy. Boiler derating averaged 9%.
Raw material consumption was as follows:
« Soda ash average daily consumption was 9.9 tons. Moles of
sodium consumed averaged 10.6% of the moles of SOo removed
from the flue gas.
• Natural gas was consumed at a rate of 7.1 cubic feet per
pound of sulfur produced.
The production of sulfur as a byproduct averaged 17-1 tons per
day of full operation. The product was sold locally.
504
-------
FGD Dependability
There were various modes of operation depending on the availability of
equipment and on flue gas availability. The principal operating modes
were as follows:
Mode 1 - boiler operating, FGD not operating.
Mode 2 - integrated operation of the absorber/evaporator and reduc-
tion units without bypass of any of the flue gas.
Mode 3 - operation of absorber/evaporator with the flue gas bypass
damper open. Bypassing of the flue gas may or may not
occur depending on the booster blower speed setting.
Mode 4 - operation of absorber/evaporator with the reduction unit
not operating. Recovered SCL (from evaporator overhead)
is vented after dilution with flue gas from other units.
The FGD unit was considered to be fully operable only during Mode 2,
although that also may have been the case during some of the Mode 3
operation. The exact operating status was difficult to determine while
the bypass damper was open. However, failure to include any part of Mode
3 operation as fully operable time did not penalize the FGD process
unfairly because the amount of accumulated full operation time with the
bypass damper open was very low. Mode 2 or full operation status does
not take into account the operation or performance of the purge treatment
unit. The purge unit may or may not have been operating during Mode 2
operation and problems with the purge unit will be discussed later.
Figure 2 shows the reliability, hours of Mode 2 operation/hours called
upon to operate, for the thirteen months of the second demonstration
year. Called upon hours are those boiler operating hours when the
boiler is delivering flue gas and steam within the design range. Figure
2 shows the FGD unit reliability factors plotted for each of the thirteen
months. The overall average reliability was 61%. The ups and downs of
operating performance shown here may be summarized as follows:
• Best reliability was achieved during October and November,
1978. For a 57 day period, October 16 to December 11, inter-
ruptions were minor and shortlived. FGD reliabiltiy was 99%
during November.
505
-------
CD
o
TOO
90
80
70
60
50
40
30
20
10
0
Oct
Nov
Dec
1978
Jan
1979
Feb Mar Apr May June July Aug Sep
PERIOD
Oct
Figure 2. FGD Reliability Index
-------
• Operation was limited to 66% reliability during December to
clean the evaporator heater and for reduction area repairs.
• The FGD plant went down on January 10 for 43 days to repack the
evaporator circulating pump and to retube a sulfur condenser.
• Full operation was limited during late February and early March
due to numerous mechanical problems and leaks. The major
problem during March was an outage of 16 days to overhaul and
realign the evaporator circulating pump.
• From April until October, FGD plant reliability averaged 73%.
Recurring problems with the booster blower turbine speed
control and with the reduction unit were the primary limitations
to better reliability.
It should be noted that during the seven month period, April to October,
month by month reliabilities were fairly consistent, were primarily in a
range of 70-75%, and were 13% higher than the thirteen month average
reliability.
Table 1 identifies the equipment items that gave the most problems. The
highest percentage of downtime was due to numerous interruptions of the
SOp reduction unit. Since surge capacity for the scrubbing solution was
minimal, any interruption of reduction unit operation required that
either the evaporator be shutdown or recovered S02 be vented. Usual
practice was to vent the SO^.
The major interruption of the reduction unit was a 35 day shutdown to
retube a sulfur condenser. Without this interruption, the reduction
unit limited full operation during 7% rather than 17% of the called upon
hours. The most frequent outages were those due to sulfur deposition,
leaks, and valve repairs.
With electric instead of steam turbine drive for the evaporator circu-
lating pump, emergency shutdown of the FGD plant was accomplished several
times without difficulty. There were eight boiler shutdowns while the
FGD plant was at full operation. Four of the shutdowns were without
warning. Two of the shutdowns were after a short warning period of less
than one hour. The other two occurred with adequate advance notice and
the FGD shutdowns preceded the boiler shutdowns by 6 hours and 9 hours.
507
-------
Table 1. REASONS FOR INTERRUPTION TO OPERATIONS
en
0
CD
Equipment or
reasons
Reduction unit
Evaporator circulating pump
Booster blower
Startup and shutdown
Other equipments including absorber
Evaporator
Days of
interruption
59
28
18
17
10
8
% of called
upon time
17
8
5
5
3
2
-------
Startups were a different story. Table 2 shows the startup time required
for the same eight boiler shutdowns. Usual sequence after boiler startup
was to start the absorber/evaporator loop and flue gas flow first followed.
by the reduction unit after which the bypass damper was closed. The
startup record of Table 2 indicates, perhaps, that more time is required
for startup after the more lengthy shutdowns. Otherwise, some of the
startups seem to be unnecessarily long.
The other reasons for FGD plant interruptions may be summarized as
fol1ows:
• The evaporator circulating pump was down three times for
repacking, for overhaul and for realignment of the motor
shaft. It was also down once to replace a seal attributable
to interruption of the steam supply from the boiler. Without
steam, the condensate used for seal water was lost.
• The booster blower was down for relatively short periods but
frequently. Most of the problems centered around the turbine
governor and the gear reducer. There were no problems associated
with the internal surfaces of the fan itself.
• There was one interruption for cleaning the evaporator heater
after it had plugged.
• The absorber operated essentially trouble free. There was
only one six hour interruption caused by an obstruction in the
process water valve.
t Other problems accounted for less than 3% of the called upon
time. They include frequent replacement of the S02 superheater
with an overhauled spare and repairs to the S02 compressor.
These items of equipment are in the line feeding SO^ to the
reduction unit. Other interruptions were for instrument and
duct leak repairs.
S0£ Removal
Removal efficiencies averaged 90% overall for the 13 months of operation.
These removal efficiencies were obtained during an accumulated 211 days
509
-------
Table 2. FGD PLANT STARTUP TIME
Boiler
down hrs .
3
3
34
100
3
138
11
3
FGD startup
time hrs
1
0.2
31
84
38
70
11
4
Startup
Absorber/evaporator
0
0
9
8
3
24
2
0
sequence, hrs.
Reduction
0.3
0.2
22
47
31
21
4
3
Bypass damper
0.7
0
0.2
29
4
25
5
1
295 239 46 129 65
-------
of Mode 2 operation of the FGD unit. Thirty day average removal efficien-
cies varied only from 88% to 93%. The range of values increases as
averaging times decrease. Figures 3 and 4 are frequency distributions
of 24-hour averages and one-hour averages, respectively. These data
show that S02 removal performance compared to design (90%) and to the
operating control point of 89% was as follows:
% of time
S02 Removal 24-hr Avg. 1-hr Avg.
90% and greater 60 52
89% and greater 84 78
85% and greater 97 97
It is seen that removal efficiencies of less than 85% occurred infre-
quently.
One-hour averages were accumulated to determine removal performance at
the longer averaging times. Figure 4 shows that S02 removal was less
than the limit of 89% for 22% of the time. The plant operated at removal
rates of 85% or better for 97% of the time. The absorber was operated
to achieve about 89% or higher removal rather than 90% removal. This
was because operating control for S02 removal was set to reduce the S02
concentration on a diluted basis by 90%. Since dilution amounted to
about ten percent, primarily as moisture, the actual removal was about
one percent less than the reduction in concentration. The S02 removal
data presented in this report have been corrected for dilution.
The flue gas being treated had the following characteristics. The
characteristics of primary interest are: S02 feed, flue gas flow rate.
and temperature.
S00 Feed. Sulfur level in the as received coal averaged 3.09 wt%.
~t
Confidence limits for this mean value at the 95% level were 2.99 wt% to
3.19 wt%. A coal of 3.16 wt% sulfur was used for design of the FGD
plant. The estimated range, based on a distribution of two standard
deviations from the mean, was 2.6 wt% to 3.6 wt%.
511
-------
100
90
80
70
60
<£.
Q
VI
UJ
to
ce
UJ
a.
50
40
30
20
10
70
. • I il
80
88 90
100
S02REMOVAL, %
Figure 3 . Cumulative Frequency Distribution of 24-Hour Average S02 Removal
512
-------
100
90
80
70
u 60
Q
Ld
£ 50
>
VI
u
10
2 40
UJ
U
o:
LJ
0.
30
20
TO
70
Figure 4.
.mil
80 88 90
S02REMOVAL, %
100
Cumulative Frequency Distribution of One Hour Average SCLRemoval
513
-------
Concentration of S02 was expected to be 2185 ppmv, wet basis, at a coal
sulfur level of 3.16 wt%. The equivalent S02 feed rate for this concen-
tration is 4842 Ib/hr at the design flue gas flow rate of 320,000 acfm
and the design equivalent boiler load factor of 80%. Actual flue gas
flows were somewhat higher due to higher than expected excess air caused
by excessive inleakage believed to be at the air preheaters. The absorber
was designed to receive a volume of flue gas equivalent to 100% load
factor and to remove the corresponding amount of S02 at inlet concen-
trations of at least 2185 ppmv. While the capacity to handle flue gas
flow rates and S02 feed rates equivalent to 100% load factor was demon-
strated for short periods, sustained operation was only possible at load
factors slightly better than 80% because of the limited capacity of the
evaporator and limited surge capacity in the regeneration loop. This
limited the capacity of the boiler during high demand that was in addition
to the derating effect from FGD plant steam consumption. While perform-
ance at high load was not fully demonstrable, minimum sustainable operat-
ing rates were demonstrated during turndown tests. Full operation at 53
MW was sustained for four days. Table 3 summarizes the inlet conditions
of this test.
Table 3. FGD PLANT MINIMUM'SUSTAINABLE
LOAD TEST RESULTS
Length of test 4 days
Average load 53 MWe
Flue gas flow 237,000 acfm
S02 feed 3,381 Ib/hr
S02 feed 2,069 ppmv wet
Oxygen in flue gas 7.5 vol. % dry
Below this operating level, the FGD plant is limited by the turndown
capability of the reduction unit. The absorber/evaporator was operated
down to 44 MWe. It must be remembered, however, that these minimum
loads are generator output after derating due to the steam requirements
514
-------
of the FGD plant. Equivalent loads relative to flue gas flow and boiler
heat input are 50 MW generating potential at the 44 MWe level and 60 MW
generating potential at the 53 MWe level.
Boiler Load. The FGD plant operated'with S02 removal rates of 89%
or better at 24-hour average loads from 53.MWe to 93 MWe. However, 85
MWe was usually the upper limit for operation. Figure 5 presents the
load factors for each 30-day period of the second demonstration year
during Mode 2 operation of the FGD plant. The FGD plant was expected to
operate at a boiler load potential of 92 MW and did indeed meet or
exceed this capacity during three of the 30-day periods. For the rest
of the time, the boiler was demand limited and the load factors remained
below the 92 MW level.
Flue Gas Flow Rate. While boiler load was generally below the
design level, the reverse was the case for flue gas flows. This was due
to higher than expected excess air in the flue gas. Figure 6 shows flue
gas flow rates as a function of both the actual generator output and the
boiler load potential. Flue gas flows of 320,000 acfm, the design
level, were attained at 71 MW of generator output or 80 MW of load
potential. At a FGD load limit of 85 MW, the flue gas flow rate was
over 360,000 acfm. At the load potential of 92 MW, the design point,
the flue gas flow rate was nearly 360,000 acfm. Since the absorber was
designed for full load, the greater volume of flue gas presented no
apparent problems for the booster fan or the absorber.
Excess Air. The high flue gas flow rates are explained by the
higher than expected amount of excess air that can be attributed primarily
to inleakage air believed to be entering at the air preheaters. Oxygen
levels in the flue gas averaged 8.0% by volume compared to an expected
qxygen level of 5.6%. On average, the additional amount of air would
increase the total quantity of flue gas by about 17%. At higher than
average loads, the excess air averaged a little less than the overall
average and would add a little less than 17% to the quantity of flue
gas.
Temperature^ Inlet temperatures averaged 305°F during the second
demonstration year. However, 38% of the hourly average temperatures
515
-------
o
«r
o
to
120
100
80
60
40
20
Generated Load
Load Potential
FGD Expected Capacity
I 1
XJ
flJ
4J
Oct Nov Dec Jan Feb
1978 1979
Mar Apr May June July Aug Sep. Oct
PERIOD
Figure n Boiler Load and Load Loss During FGD Operation
-------
*/>
g
LU
r>
U-
100
Actual Load
400
Load Potential
<0
o
O
O
o
re
o
o
o
300
200
50
60
70
80
90
GROSS LOAD, MW
6. Flue Gas Flow Rates at Actual Load and Load Potential
-------
were below 300°F but virtually all of these temperatures were above
280°F. It should be noted that these are single point temperatures well
within the flue gas stream and do not reflect temperatures at and near
the duct surfaces. There were, nevertheless, no problems attributable
to wet flue gas.
Energy Consumption
A significant amount of the steam produced by the boiler was consumed by
the FGD plant, primarily for operation of the evaporator for recovering
the S02 and regenerating the scrubber solution. Boiler main steam from
the superheater at 1800 psi and 1000°F was let down and desuperheated to
obtain steam for the FGD plant at 550 psig and 750°F. This steam was
used in steam turbines to drive the booster blower, S0? compressor and
the evaporator circulating pump. However, before the start of the
second demonstration year, the evaporator circulating pump drive was
electrified to eliminate the startup and shutdown problems that occurred
when high pressure steam was interrupted by unscheduled boiler shutdowns.
The turbine exhaust steam along with additional 550 psig steam that had
been let down through a pressure reducing valve was desuperheated further
and used for process heat, primarily at the evaporator.
Actual steam consumption (at 550 psig, 750°F) varied from 52,000 Ib/hr
to 59,000 Ib/hr during the second demonstration year. In Btu's, this
was equivalent to 11% of the boiler input energy derived from fuel and
derated the 115 MW boiler by 8% at the average boiler load of 77 MWe.
In addition to steam consumption, about 700 kW of electricity was con-
sumed, exclusive of the evaporator circulating pump motor. This increases
the total energy requirement to about 12% of the boiler heat input
derived from fuel. This derated the boiler another 0.6%. The total
derating is equivalent to an electric production loss of 10 MW of gener-
ator output. Power to the evaporator circulating pump was not metered
but is estimated at 330 kW,
Raw Material Consumption
Soda ash is used as makeup sodium carbonate for the scrubbing process.
Usage is determined by buildup of inactive constituents in the absorber/
518
-------
evaporator loop, such as sulfate and thiosulfate, that have to be purged.
Any loss from the system due to leaks would also require soda ash makeup.
High soda ash consumption during the first demonstration year were due
to leaks at the bottom collector tray of the absorber that were repaired
before start of the second demonstration year. These leaks effectively
aborted the estimation of purge rates during the first year.
For the thirteen months of the second demonstration year, 2273 tons of
soda ash were consumed, for an average daily consumption of 9.4 tons per
day, using the total operating days of the absorber/evaporator as the
time base. The performance guarantee for acceptance was 6.6 tons per
day at the design levels of flue gas flow and inlet S02.
Natural gas is used as the reductant for converting the SCL to elemental
sulfur. It is also the fuel used to incinerate the tail gas emitted
from the reduction process. The tail gas is returned to the inlet of
the absorber after incineration. It was necessary that the incinerator
continue to be operated during shutdowns for destruction of the reduced
sulfur forms that desorb from the reduction unit refractory materials.
Thus, there is a corresponding improvement in unit consumption of natural
gas with improvement in reliability. Table 4 shows that slightly over 7
cubic feet of natural gas was consumed per pound of sulfur produced.
Table 4. NATURAL GAS CONSUMPTION
Annual consumption, million cf 54.1
For process use, % 87.9
For incineration use, % 12.1
Average consumption during
mode 2 operation, cf/hr 9745
For process use, % 92.5
For incinerator use, % 7.5
Process gas/sulfur produced,
cf/lb 6.2
(continued)
519
-------
Table 4 (continued)
Total gas/sulfur produced,
cf/lb 7.1
Consumption during shutdown,
% of total 5.6
This meets the design expectations. Average consumption was 9745 cf/hr.
of which 7.5% was consumed by the incinerator. In contrast, the inciner-
ator consumed over 12% of the total gas overall for the second demonstra-
tion year and is a consequence of the 61% FGD plant reliability factor.
Purge Treatment Limitations
The purge unit as initially designed was to have treated a small purge
stream removed from the regeneration loop, to effect a separation of
sodium sulfate from most of the sulfite/bisulfite components, and to dry
the sodium sulfate to produce a salable product. The "wet" end of this
purge treatment system performed satisfactorily but the dryer had a
capacity of only about 50% of that needed. The requirements on the
purge unit and the drying problem will be discussed in turn.
The amount of purge to be treated is a function of the formation of
sulfate and possibly thiosulfate during absorption. Attempts were made
by TRW to determine the amount of sulfate formation during absorption
but these efforts were frustrated by inability to obtain correct flow
measurements and uncertainties about the specific water balance across
the absorber. However, the data seem to indicate that sulfate formation
is a function of oxygen concentration in the flue gas. Since excess air
levels were higher than design expectations, higher than design purge
rates might be necessary. Purge rates were not measured directly.
However, an average purge rate for the period April through October has
been estimated at between 10.6 and 12.4%. Purge rate is the ratio of
moles sodium in purge to moles SOp removed from flue gas, expressed as a
percentage. The estimate was determined from soda ash consumption and
the calculated amount of SO,, removed. A purge rate of about 10% was the
520
-------
value indicated during the design phase of the project. In the aggregate,
the above information seems to point to actual purge rates higher than
design, the magnitude of which is unclear. As stated before, the process
up to drying seemed to perform satisfactorily.
Dryer tests performed by Davy McKee determined that the dryer would not
work, even at design rates. There had always been a question of whether
the sulfate dryer actually had heat duty design capacity. In May of
1979, tests verified that the dryer did not have design capacity. The
purpose of the test was to demonstrate the heat capacity of the dryer
with a water feed onto a sodium sulfate bed in the dryer. If the maximum
rate could be reached with water, the dryer was then to be tested with
the sodium sulfate solution recovered from the purge solution.
The maximum dryer capacity achieved during the test was approximately
.66% of the design heat duty. Capacities of 59-66% of design were main-
tained for 2-1/2 days. However, after operating for 2-1/2 days at 59-
66%, the motor tripped out several times because of an amperage overload.
When the motor tripped out, a buildup of solids was observed at the
discharge end of the dryer. To prevent overloading the motor, the water
feed rate was reduced, so during the last 2-1/2 days of the test, the
dryer capacity dropped to 45-50% of design. At this point, further
testing was abandoned.
Possible solutions include:
1. A more concentrated feed fed to same dryer.
2. A dryer of different design.
3. Addition of an antioxidant to the absorbing solution.
Replacement of the equipment capable of attaining design capacity would
cost approximately $500,000. Alternates to the equipment replacement
were sought. Antioxidants to reduce sulfate formation were considered.
Tests of EDTA (Ethylene Diamine Tetra Acetic Acid) were run April - May of
1980. The tests were terminated after five days of an intended two-week
test because of unrelated equipment failure. The too brief test period
indicated a possible 50% reduction in sulfate formation. Additional
testing of EDTA is planned.
521
-------
OPERATING AND MAINTENANCE COSTS
NIPSCO UNIT NO. 11 FGD PLANT
Operating, Maintenance and Improvement Costs are listed from the
period of January 1, 1979, through April 30, 1980.
Operation and Maintenance - Offsites Facilities $ 154,160
(including booster blower, flue gas ductwork
and dampers, utilities system)
Operation and Maintenance - FGD Process 6,061,205
(includes by-products storage and loading,
raw materials unloading and storage, and
Allied Management Fee)
Total FGD Costs before By-Product Credit 6,215,365
Credit for By-Products 104,963
Total FGD Operation/Maintenance Costs after $6,110,402
By-Product Credit
No amortization costs are included above.
522
-------
COAL DATA AND UTILITIES - January 1, 1979, through April, 1980
High Sulfur Coal Burned, Tons 349,121
Average BTU/# 10,586
Average Sulfur % 2.85
Steam Used, Pounds 455,000,000
Boiler Feedwater Used, Pounds 29,128,000
Condensate Returned to Gen. Sta., Pounds 333,120,000
Condensate Dumped, Pounds 151,008,000
Electric Power, Kwh (Including 500 HP 6,472,000
Circ. Pump Motor)
Natural Gas, Ft3 57,642,000
Service Water, 1000 Gallons 2,784,000
Elemental Sulfur Sold, Long Tons 2,668
Sulfate Sold, net tons (no dry sulfate was 139.6
produced in 1980)
523
-------
NORTHERN INDIANA PUBLIC SERVICE COMPANY
DEAN H. MITCHELL STATION
UNIT NO. 11 S02 DEMONSTRATION PLANT
OPERATING HOURS GRAPH
1. Solid line indicates Unit No. 11 or S02 Plant is in opera-
tion.
2. Definition of Unit No. 11 being in operation is: Unit
synchronized on line regardless of megawatt load.
3. Definition of S02 Plant being in operation is:
a. Receiving all of flue gas from Unit No. 11.
b. No S02 bypassed to the Unit 6-11 stack.
4. Unit 11 operating conditions required for S02 Plant Operation
are:
a. Unit 11 operating on high sulfur coal at 46 GMWE (min.).
b. Sufficient main steam available (530 PSIG minimum).
c. Sufficient demineralized make-up water available.
d. Unit 11 supplied utilities available (electricity,
boiler feed water).
524
-------
OPERATING HOURS
mi — _
,^— .*•. • Bijj ..-...- , ,,,.
^iH
1
1
1
1
I6;30 AM **"* Repair evapor
I pump. Repair
| tank agitator
I 1st sulfur co
sulfur con den
ator circulating
dump dissolve
Tube leak in
ndenser. Retube
ser.
w
pL
f':'$
~~~~||
1
k
P;
" "m
1
— .(.,.
r-: - '
$
F
F&fJ
1
r:
fe
.'f.1
lift
fe
if
f',"'
v •
!*".
!f.'
! ?•
>.(
I
1"- — p
li...-
f
1
&
I
fi
1
)1
•OV,''!
Unit No. 11
71
12
13
14
15
17
119 i
22
2.3
28
29
30
525
-------
OPERATING HOURS
Plant
Retube sulfur condenser.
Absorber & evaporator
start-up. Expansion
joint leaks in S02
reduction area.
Unit No. 11
9
no
11
12
13
14
15
16
17
18
19
20
iAIJ° ™- BH V
___2; 10 PM £laus expansipn, joint leak.
12:20 PM
-aT/tf-W.."**" S0o superheater repair.
Replace evaporator circ.
pump packing due to loss of1
seal water caused by boiler
Trin & lac|c o_j? emers. steam.
Nozzle steam leak in S02
• — . 5-e.duct:laa, ja^^-
_
m
ii
6:31 PM Vacuum pump trip. 1
7.42 PM ™iSiffl
ffil^S^
fe »f»i
p^1 1
p^'-fc'.^( 1 .
us
22 ;
23
2i
*•• t
*~ -' ;
26
27
28
!
i
526
-------
OPERATING HOURS
S02 Plant
Unit No. 11
I
ll
1
1
1
1
1
1
.
.
.
.
_
S(>2 superheater
&
1:55 AM
£-.30 PM1™ Glaus convertei
joint repair.
L2:28 PM|g||
s
H
m
P^epair evapora
pump shaft .
i
\
Booster Fan Tr
: repair.
r expansion
!
tor circulating
k
IP-
t |
>'i
' 1
i!
i *l
')
i
-I
j
ni
t 1
f»
%
i t
l!
L ^
/• «
i
t j,
, 'i*
T''
i ^
r
1
r
fl ,
]
^
i "- 1-
I ;
r
f
i
i
i
i
i
4
f.
,
\
2
3
4
1 „. i
"" \
i
' i
!
t
8
9
10
11
12
13
14
15
16
17
'id
19
20
i ^
i--
w
r>
tc
i— '
^
• uj
i
Clean tail gas piping.
25
!26
30
:
527
-------
OPERATING HOURS
?C? Plant
HI
ii
7:55 AM""" S02 superheater leak.
JH:30 AKJpSl
||
]||
[itJl Plugged tail gas line.
tiff
||
ii
H
„ iiral
H
fU
83
il
3:15 ?Mp|!i| Incinerator Trip - Condensa-
^J^CLJ'SLs!'*''35' tion pj3 burner Controls .
H*'1'''?'
US
|j|| Bypass louver dampers tripped
^•*m open on loss of control air.
H
~|i
m
Louver dampers tripped open -
K',f, High duct pressure.
9:00 AM ^^ Evaporator heat exchanger
tube pluggage.
High vibration on evap . circ.
piimn mqt-pr.
Unit No. 11
US
if
M
m
II
H
•
m
M
it
if1*
^, ,,44i
|||
fe
"^S
IS
|f;
Si
|i
ii
II
11
ii
ii
te
fc
B
fc
IK
1
2 !
3
'->
J
5
6
7
8
9
10
11
12
13
14
15
16
17
(
•*- -^ 1
„ . _..!
20 !
21
*". <*»
' 2.L
"•>'•
^
25
26
27
28
£ y
30
528
-------
OPERATING HOURS
S02 Plant
Unit N'o. 11
P Continue Rvap. CIrc. Pump
|_ . EleC. MoJinr^Pjopp I r
_^_ ^
V
ISO. Plant Start-Up
12:10 PM |H
IS
II
H
ir niRB t\fciuctxon r> mituoWn J.TI
11:45 AM ^^ Preparation for boiler outage
:J:15 PM Repair SOo Superheater
: K
li
1 s
1 ^:r>-'
1 .,..., f
fl
^
!
1 ll
i t
I
\
S
>
\
i
?.
t
k.^
9:03 PM Ji Boiler Tube Leak
«->:43 AM r...
' •',•'•'*."•
. ' ,''i
i •• • ' ""j
w?«.;
|:iif:
ps
V ' - J
J. w
i ;
2
3
4
r-
6
7
8
9
10
!•
1
11
12
13
14
13 i
:io
i (
17
18
t n
vO
-J
"suTator TaTTTirr ^~~~\
not '
11.55 P"M close.
-------
OPERATING HOURS
S02 Plant Unit No. 11
j
i
SO Plant in Start-Fp Mode
11:15 AM F^
JJ
H
H
Hi
M
m
Brief Main Steam Supplv
Interruption
8 '
il
fat"!
m?, ,
w
S§
&
H:l5 $>f fete Plugged Makeup Water Valve
12:30 PM S-v"
£•/??,
• 5:00 AM P^'S Reduction Unit dovm duo to
j i-&Sfl low inlet S02 to FGD Plant
| Pepair leaks in inlet flue
P.TK duct.
SO Plant in start-up mode.
6:10 PM
8 : A 5 AM Repair Expansion Joint.
bi
12:AO PM |j^
P-eheat Stop Valve Repair
4:14 AMH
1
9
I
r^vyt^s!!
H
11
fes
P"^"^
Ifl
I
fstv&yt
|
Bl Failure of 11-F. F.T). Fan
H Motor
Si
IP
1
*•
i Trial ^-'un on mixture1 01
I Hir,h and Medium Sulfur Goal.
1 Unsatisfactory
\ let S07 to FGT1
due to low In-
Plant.
B
II
if
s
IPS
2
3
4
5
6
1
8
9
10
11
1 0
13
1A
15
i
16 i
17
1
To
19
20
21 ;
'2: :
i
13
2A !
25
26
M I27
Si
p
iff
1
28
.29
30
530
-------
OPEEAT1NG HOURS
SO, Plant
Inspect booster fan speed
reducer
Inspect booEffer fan speed
reducer
Repack main steam pressure
reducing valve . Testing and
inspection of reduction unit
mixed gas systen.
11:25 AM
Reduction uni~t dovm~fo
conduct turndovm tests.
Pluggago in mixed pas
cooler and sulfur condenser.
Unit No. 11
-fp-
—!i
10
11
13
14
15
16
20
23
25
26
i">,
28
1 Los
t Northeast fUts.
29
in'
531
-------
OPERATING HOURS
SOj
1:20 PM r Plant in
Start-up' mode.
1:40 PM
1
8:00 AH m.
S
« Change R02 Superheater.
L___fe
•T
'
•
1
i
033
I
i
w
'-&
i
-i
i
•i
9:51 PM iii East Unit Aux.
4 KV
Breaker failure. Cold re-
heat line water hammer on
start-up. Repair and inspec
tion of cold reheat line.
1:5P Pit T
i
1:
[>
i
1
L
^
i \
^1
\n
not
/
1
2
3
4
^
6
7
8
" S
10 S.
!
11
12
13
, J
14
15
16
17
18
19
21
•C i-
'.|23 |
i
5:24 AM*8"" Precipitator Repair,
1 ] : 50 PM
P
K
P
•2*1
2> i
26
27
26
i
29
4:57 PM iiLb Hi^h ^en^rator voltape. j-5 ;
'i:26'AM trip - rhebftCat fallnrCT"
31
532
-------
OPERATING HOURS
SO2 Plant
TAIL GAS INCINERATOR
MALFUNCTION
BROKEN GOVERNOR LINKAGE
ON BOOSTER FAN
BOOSTER FAN TRIPS -
SPililD INDICATOR ERR01
CAUSED OVERSPEED TRIP
LEAK IN BOOSTER FAN
TURBINE GOVERNOR STEAM
5:00 am
7:51 am
Unit No. 11
130 KV YARD FAULT
11
12
13
14
--J
16
i7
18
21 :
22
23 1
124 ;
:1
t1
••£
i
1
4
27
28
29
!_
O "^
oO
533
-------
OPERATING HOURS
SO.-, Plant
Unit No. 11
- • — \
4 ,
5
6
5:20 pmHll BROKEN BOOSTER FAN
F mm TURBINE GOVERNOR VALVE
10
5:30 pm
STEM
11
12
9:15 am
BOOSTER FAN SPEED
REDUCER MALFUNCTION
13
14
15
16
17
•'2G
J22 ;
23
11:27
UNIT 11 TURBINE
BLADE FAILURES
-ML
!25
26
27
29
30
534
-------
OPERATING HOURS
SO, Plant
Unit No. 11
CONTINUED SCHEDULED
CONTINUED OUTAGE FOR
1 i
OUTAGE WlTH UNlT 11"
TURBINE BLADE REPAIR AND
SCHEDULED TURBINE AND
BOILER MAINTENANCE
5
ac
12 !
13
14
15
17
18 I
19 i
1
20
22
24
25
26
27
28
29
30
535
-------
OPERATING HOURS
S02 Plant
'
CONTINUED SCHEDULED
OUTAGE WITH UNIT 11.
L__
Unit No. 11
CONTINUED SCHEDULED
OUTAGE FOR BOILER AND
TURBINE MAINTENANCE.
12:10 PM — - UNIT START-UP
l P • 1 7 PM
2:53 PM 3:21 PM
9 : 11 PM ~ 9:53 PM
8:15 PM
11:30 PMw
8: 35 PMp,™, /:
12:13 AM
HIGH EXHAUST H001) .
6:^5 PM,^ TEMP. & VIBRATION.
12:32 AM BUM 5.50 m
12:50 AM TRIP CHECKS.
12 : 40 AM *•* k : 30 AM
1
2
*
4
5 }
6
7
8
9
10
11
12
13
14
15
16
17
18
.!•">
J20
21
22
23
J2/,
I
2^
2b
27
23
• 29
3')
:, 31
8
8
?
T''
JL
UD
1
i
1
i
536
-------
OPERATING HOURS
S02 Plant
Unit No. 11
.bUUbThK J?'AW STAftl'-'UP MD
BALANCING.
FGD PLANT START-UP IN
PROGRESS .
BOOSTER FAN BALANCING
COMPLETE .
CONTAMINATED CONDENSATE
RETURN. SHORTAGE OF
BOILER FEEDWATER.
CONTAMlWA'm; UUWUKWbATK
RETURN .
t
BOOSTER FAN SHUT DOWN
TO ALLOW REPAIR OF
ORIFICE CONTACTOR.
1 BOOSTER FAN START-UP.
If
-- if
ss
w
\%
4
!
II
fe
.1:30 AM ^
1 : 30 AM D
i
— f
i
hn
1*
fe
li
f,
jft
i»
1
8:19 PM «
1:28 AM I
I
t
:
8: 11 "AM"
11:27 AM
1
i
&
&
•* a
1
a
1^s PRECIPITATOR REPAIR.
sis
mat -•" -
1
ii
*m
m
if
~ hti
S»
^B
if
2H
•lufej
^:-:l
ial CONTROL VALVE MALFUNCTION.
?"'Poi
?;*f'
i*^'";
MS,^
*„,-•#'>".}
^''•""-'7l'Lf .,..„.,._ ...„,„„ ...„,„,*, „ ,ui
?v4.y
^i
Pi
v^
'y:-^";fe ,
gft-l'S
COLLECTOR RING BRUSH HOLDER
B SHIFTED - LOST EXCITATION.
1
P
?'"*•*«
1
ill
i :^$
*-..fts
^-.V':^
S:V. 4
»toa
^
3
4i
'
• 1
7 1
8 '
9
LO
11
12
13 !
14 1
5
15 H
^
16 c
17
,, i
19 ;
20
i
21
22
•
23
24
25
26 i
27
28
29
I
30
31
537
-------
NORTHERN INDIANA PUBLIC SERVICE COMPANY
FGD PLANT OPERATING HOURS
DATE
l
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
FEBRUARY
FGD PLANT
Orifice
Contactor
Repairs
On Tl:28am
Off l:45pm *•"
Piping Leak In
Reduction Area
On l:45pm g^gg
Off 10: 25am Leak In Shell Of "A"
..^JK.egenerator
8 : 00am
S02 Plant Available
But Condensate Quality
Problem in Generating
Station
On. 2 : 03pm
Off 4:I5pm a""™1
Leak In Regenerator
On 9; 48am
Off 8:IZam «»""Leak~~tn Regenerator
On 10:29am plffl
B
s
OFF 10" tiOain"™ Replace Leaking SO? S.
On 2: 00pm *
B
S am #11 To Come Off Line
, 1980
UNIT NO, 11
in
H
||
H
|p
Off 3: 42pm H|Turbine Control Valve
On 4: 45pm ^^Malfunction
B
n
I
il
H
,
n
H
S
1
1
B
s
fe
H
if
m
IP
Htr* H
H
Off 10: 40pm ^Kenerator H2 Leak
DATE
X
2
3
4
5
6
7
6
9
10
11
12
13
14
15
16
17
18
T) J
20
21
22
23 |
, , , — .
2, ]
25 |
26
27
2e
29 J
' j
- •: 1
538
-------
NORTHERN INDIANA PUBLIC SERVICE COMPANY
FGD PLANT OPERATING HOURS
DAIh
i
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
"31
MARCH
FGD PLANT
SO/, Compressor
Problems
]
.
1
s
i
On 0:10pm ,,,,„„ 1
Off 2: 20pm S02 Leak In Reduction
~ , Area Piping
On 0:45pm mi&
iff j
11
H
Off 4: 22pm KB Incinerator Control
Failure
On 10: 20am ___.
Off 12:01pm^™Unit 11 Scheduled To
Be Off
On 1 1 : 45am mmm
Off 3: 50pm *"HH" Sulfur Blockage In
Reduction Area
, 1980
UNIT NO, 11
H
11
B
n
S
H
m
11
ill
— — II
H
iS
gi
H
«li
-|;«
I'M
~~A| »^;.;;'' " " ~"
Off 7?09pmifiiM Remove Turbine Stop
Valve Screens And
Check H2 Seals
On 1 • 02pm j^p'fgi
Off 10:56pmHlFalse Trip - Ground
s-fj,™ Protect ion Relay
On 6: 22am fe^f
|||
|||
irt
DATE
1
2
3 \
4 i
5 !
6 \
7
8
9 1
10 :
11 j
12
13
14
15
r~nn
•*. 1* *
i? i
18 |
H
19 |
_£fH
~2l!
22 ",
23- -;
24
25
26
27 .
I __»
i w :
29 !
30 !
31
a
539
-------
NORTHERN INDIANA PUBLIC SERVICE COMPANY
FGD PLANT OPERATING HOURS
APRIL, 1980
DATE,
1
FGD PLANT
Sulfur Blockage
In Reduction Area
On ll:AOam
Off l:40pm
UNIT NO, 11
DATE
.'2
Sulfur Blockage
In "B" Claus Reactor
10
11
10
11
12
13
14
15
16
17
18
19
On 10:15am
12
13
14
15
16
17
IS
19 i
20
21
22
23
24
Off IrAOpm
Unit 11 Off
Off 3:48pm
Condenser Leaks
20
21
2-3
24
On 9:56am
25
26
27
28
S02 Plant
Start Up
On 11:20am
26
27
28
29
30
•29
•30
540
-------
NORTHERN INDIANA PUBLIC SERVICE COMPANY
FGD PLANT OPERATING HOURS
MAY, 1980
DATE
FGD PLANT
Repair
Bearing
UNIT NO. II
DATE
Failure
In The
Evaporator
'alse Trip - Genera-
or Protection Relay
10
Pump Motor
l:15pm
Plant Available
1:25pm
.But No Steam From Unit
11
10
11
12
S02 Plant
Start Up
11
12
13
14
15
On 10:30am
Off l;45pm
Reduction Area S09
Leak & SC>2 Compressor
Turbine Governor
13
14
15
16
17
Steam Leak
SC>2 Plant
pff 2:32am
Dispatcher - Low
16
17'
18
Available
System Load
19
S02 Plant Start Up
On 9:28am
Dn l:05pm
20
21
Dff 7:00pm
22
23
24
SC>2 Plant Available
25
26
Low System Load
)ff 2:27am
Dispatcher - Low
'26
27
System Load
Zh 9:27am
Off A;29ptn
27
28
recipitator Problem
28
29
1:40pm
"Available
30
Dispatcher - Low
31
System Load
541
-------
REFERENCES
1. R. C. Adams, T. E. Eggleston, J. L. Haslbeck, R. C. Jordon, and
Ellen Pulaski. Demonstration of WeiIman-Lord/Allied Chemical
FGD Technology: Boiler Operating Characteristics, EPA Contract
No. 68-02-0235 (1977).
2. R. C. Adams, S. J. Lutz, and S. W. Mulligan. Demonstration of
Wellman-Lord/Allied Chemical FGD Technology: Acceptance Test
Results, EPA Contract No. 68-02-1877 (1979).
3. F. A. Ayer. Proceedings: Symposium on Flue Gas Desulfurization
Las Vegas, Nevada, EPA Contract No. 68-02-2612 (1979).
4. R. C. Adams, J. Cotter, and S. W. Mulligan. Demonstration of
Wellman-Lord/Allied Chemical FGD Technology: Demonstration Test
First Year Results, EPA Contract No. 68-02-1877 (1979).
542
-------
MAGNESIUM FGD AT TVA: PILOT AND FULL-SCALE DESIGNS
by
E. G. Marcus
Chemical Engineer - Gaseous Emission Control
Tennessee Valley Authority
Chattanooga, Tennessee 37401
T. L. Wright
Mechanical Engineer - Mechanical Engineering Branch
Tennessee Valley Authority
Knoxville, Tennessee 37902
W. L. Wells
Program Manager - Gaseous Emission Control
Tennessee Valley Authority
Chattanooga, Tennessee 37401
ABSTRACT
This paper discusses pilot and full-scale magnesium flue gas desulfurization
(FGD) designs by TVA.
The full-scale (600-MW equivalent) magnesium FGD design is for operation at
high and low load factors for a high sulfur coal. After a process and system
chemistry (magnesium sulfite/bisulfite) description, the paper describes the FGD
equipment and system operation which includes an onsite acid plant. The second
part of the paper discusses information on a test program and schedule of a pilot
plant being considered by TVA to verify the magnesium FGD design.
543
-------
MAGNESIUM FGD AT TVA: FULL-SCALE DESIGN
INTRODUCTION
TVA's involvement with magnesium flue gas desulfurization (FGD) systems
has included some testing of MgO on the 1-MW cocurrent scrubber pilot plant at
Colbert Steam Plant in 1976, a 10-MW cocurrent prototype system at Shawnee Steam
Plant in 1978, and various economic studies.1'2'3'4 The use of lime or lime-
stone FGD systems requires large land areas for disposal of the calcium sludge
produced. A regenerable FGD system eliminates the calcium sludge, provides a
salable byproduct, and regenerates the absorbent for S02 removal. United Engi-
neers and Constructors (UE&C) assisted TVA in the process design of the FGD
system and provided detailed engineering support, especially in the magnesium
regeneration area.
PROCESS DESCRIPTION/SYSTEM CHEMISTRY
The magnesium FGD design uses magnesium sulfite-bisulfite chemistry for
S02 absorption, operating at a pH of approximately 6.0. After absorption of
S02 by magnesium sulfite to form magnesium bisulfite, magnesium oxide is added
to the absorber recycle tank to react with the bisulfite and precipitate mag-
nesium sulfite. The sulfite can exist as tri- or hexa-hydrate depending on
startup and operating conditions. The equipment is being sized to operate with
either compound. In the trihydrate mode, the magnesium salts (10 percent solids)
will be dewatered to 70 percent solids and then dried to remove all free water
and most of the bound moisture (see Figure 1). The use of a sulfite storage
silo provides the ability to operate the absorption and dewatering areas com-
pletely independent of the downstream regeneration area and acid plant. The
calciner is designed to decompose the magnesium sulfite (with an average of
1/2 mole of bound water per mole of magnesium sulfite) and sulfate (with 7
mole of bound water per mole of magnesium sulfate) into magnesium oxide and
S02. At a calciner operating temperatures of 1800°F, only 60 percent of the
sulfate is decomposed. In actual operation the thermal decomposition of sul-
fate would be optimized to give the highest necessary percentage decomposition
at the lowest temperature. The offgas from the calciner is designed for an
S02 concentration of 17 volume percent for feed to a single contact acid plant
Since the acid plant tail gas goes back to the main plenum, the S02 emissions
from the acid plant are part of the overall plant S02 emissions.
Both the Philadelphia Electric Company's (PECo) experience at Eddystone
and the brief TVA experience at Shawnee indicates that there is little, if any,
solid magnesium sulfate formed as a solid in the absorber recirculating material.
The dissolved magnesium sulfate level in the absorbing slurry would be about
30 percent (by weight) using only the liquor in the trihydrate centrifuge cake
to the dryer as a purge stream. However, this 30 percent is based on 10 percent
oxidation of the absorbed S02 to sulfate, 60 percent decomposition of sulfate
544
-------
en
-P*
01
UOUirV'SLURRY
SOLID
GAS
(S) SPARE
Figure i: FGD System Process Flow Diogram
-------
in the calciner, and a trihydrate mode of operation (70 percent solids in the
centrifuge cake). The PECo experience shows less than 10 percent oxidation
but at a higher operating pH in the absorber. Operation at the lower pH may
produce more oxidation and a hexahydrate crystal reducing the amount of liquor'
evaporated in the rotary dryer. The hexahydrate cake is easier to dewater thafl
the trihydrate. This operation would raise the dissolved sulfate level in the
slurry to well-in-excess of 30 percent by weight.
In order to provide flexibility of operation and for thermal protection
of the baghouse used for final magnesium particulate collection before the
acid plant, a spray dryer was placed between the calciner and the equipment
used for final magnesium particulate removal. The spray dryer will use cen-
tra te from the centrifuge containing dissolved magnesium sulfite, bisulfite,
and sulfate. In this manner, magnesium value in the form of MgSOs and MgSC>4
can be recovered. This sulfite/sulfate mixture will be recycled to the sul-
fite silo for feed to the calciner. Since there is a 60 percent sulfate
decomposition in the calciner, this use of the spray dryer allows for control
of the dissolved sulfate in the absorbing slurry, depending on actual operating
conditions. For the design case, the sulfate level will be 24 percent MgS04
(by weight).
FGD SYSTEM EQUIPMENT
In the absorption area, particulate, chlorides, and SOz are removed by
four (4) venturi-type prewash and absorber modules. The venturi with its own
two-stage mist eliminator removes chlorides, 863, and fly ash in a separate
liquor loop. Any chlorides which are not removed in the venturi loop will
form magnesium chloride in the absorber loop and could cause corrosion problems
in the regeneration area and acid plant feed gas clean-up system. The blowdown
from the venturi recycle tank is neutralized in a separate facility with hydrated
lime and is then pumped to the plant disposal area. The venturi liquor loop is
not neutralized and, therefore, operates at a pH of less than one. In addition
to the S03 and chlorides removed in the venturi, additional dilute sulfuric
waste acid from the acid plant feed gas humidification/cooling towers and wet
ESP's are added to the venturi recycle tank. This acid blowdown from the acid
plant feed gas clean-up system results from the water quenching of the calciner
offgas which contains about 0.6 volume percent 80s.
To raise the venturi liquor pH to about 1.0, consideration is being given
in the design to pumping the acid plant blowdown directly to the neutralization
facility or to the absorber recycle liquor tank. In theory, the waste acid
stream would form more magnesium bisulfite in the recycled liquor and require
more magnesium oxide in the absorber recycle tanks. This would increase the
load on the regeneration area and acid plant but increase the amount of sul-
furic acid recovered and decrease the lime required for neutralization. Of
the anticipated lime requirement for neutralization, 50 percent is required
for the acid plant feed gas clean-up system blowdown. In either case, the low
pH and high chloride concentration will require corrosion protection of the
546
-------
equipment. For the venturi-type prescrubber, besides Inconel 625 for high
abrasion areas and FRP internals, the mist eliminators could be factory-
installed on special Inconel strips to avoid field installation of the blades.
An organic lining could be used to protect the carbon steel venturi shell.
Efforts would have to be made to ensure that there is minimal interior work
after lining installation (such as factory-installed mist eliminator blades).
The S02 absorbers and their two-stage mist eliminators are designed for
operation in a cocurrent gas and slurry mode at 4.6 meters per second (15 fee:
per second) and an L/G of 4.5 liters per actual cubic meter (34 gallons/1000
acf)j and for mist elimination with horizontal gas flows of 6.1 meters per
second (20 feet per second). The cocurrent absorbers are based on TVA's ear-
lier test .work with cocurrent absorbers at Colbert (1 MW) and Shawnee (10 MW).
Gas/slurry flows in the absorber/mist eliminator, especially the wet elbow
(180° turn) where the bulk of the entrained slurry is removed (see Figure 2),
can cause large variations in the gas velocity profile and produce solids
deposition on the mist eliminator. The absorbers could be 316L stainless
steel; a need for higher grades of alloys is not anticipated due to the low
concentration of chlorides projected for the absorber slurry.
There is no reheater in this design since saturated flue gas can be mixed
with linscrubbed flue gas after the scrubber for the flue gas reheat. The scrub-
ber .fans have Inconel 625 rotors and 316L housings and the bypass fans, if needed,
would be Corten construction. The lining for this very corrosive flue gas would
require extensive corrosion resistance due to the mixture of water vapor (scrub-
bed gas) and SOs/chlorides (bypassed gas), below the average boiler flue gas
acid dew point.
The absorber recycle pumps and tanks are designed for normal slurry serv-
ice, similar to a limestone FGD system except for the use of a waterless seal
to assist in water balance maintenance. Since the liquor in the centrifuge
cake (going to the dryer for evaporation) is only 35 GPM and any seal water
for these large 10,000 GPM recycle pumps would be at least 10 to 15 GPM per
pump (for a total of 60-90 GPM for the six (6) operating pumps), the use of
seal water for the slurry would require a purge of magnesium liquor to main-
tain the absorber loop water balance. With all evaporation of water for
quenching the flue gas taking place in the venturi portion of the process, the
major water loss from the absorber-regeneration section is through the centri-
fuge cake into the drying system. Since each GPM lost to blowdown to allow
for water in-leakage from the pump seals was worth $200,000 in makeup magensia
value over the life of the plant, TVA designed for pumps that did not require
seal water. These types of pumps had been used successfully at Philadelphia
Electric Company (PECo) to help maintain a water balance.
The regeneration area required several difficult decisions. Since the
calciner offgas is at 1800°F, economic considerations dictated some heat be
recovered. Based on PECo's experience, the original design of a shell and
tube heat exchanger to preheat fludizing air would present major operating
difficultues caused by the fouling of tube side heat transfer surfaces. The
547
-------
Figure 2. Cocurrent downflow absorber "wet elbow" design
548
-------
th^h Off8as contains traces of S03 and the latter apparently recombines
TVA H re8enerated magnesium oxide to form a solid sulfate. Although the
IVA aesign has MgO product recovery cyclones which are more efficient than
those at i^Co there is some indication that the recombination is S03 control-
lea, or a aitrusion limited reaction. Therefore, no matter how efficient the
product recovery cyclones, there will always be recombination and potential
subsequent pluggage. The method chosen to avoid this problem in the design is
to preheat the magnesium sulfite solids being fed to the calciner in a manner
which can be called "suspension heating." The sulfite solids at ambient tem-
perature from the silo are injected into the 1800°F calciner off gas after the
recovery cyclones. The sulfite solids are preheated and then recovered in the
sulfite cylones. Although the inlets of these high efficiency cyclones are
small, very little pluggage would be expected due to the high gas velocities
ia the cyclones. This type of solids preheating is now used in the cement
industry and its successful operation in magnesium FGD requires that sulfite
solids not be heated to the decomposition temperature during suspension heating.
The other difficult problem in the regeneration area concerned the use of
either an electrostatic precipitator (ESP) or a fabric filter for final mag-
nesium oxide particulate removal before the off gas passed to the acid plant.
Any of this particulate which reaches the acid plant is lost from the regen-
eration loop and may poison the acid plant catalyst if not removed by the
humidification or gas cooling towers of the acid plant feed gas clean-up sys-
tem. The TVA design required 99 percent removal of the particulate remaining
after the three sets of upstream cyclones. Originally for the design concept
it was desirable for the final MgO particulate collection device to operate at
temperatures in excess of 600°F, collect greater than 98 percent of the MgO
and provide reliable operation. At first an ESP was selected over a fabric
filter. The fabric filter's maximum operation is about 500°F; therefore, it
is not suitable for this application. However, this selection was before the
addition of a spray dryer for sulfate control to the regeneration area and
before investigations were carried out on ESP collection of magnesium oxide.
The change from an ESP to a fabric filter was finally resolved based on the
operating characteristics of ESP collection of magnesium oxide and the reali-
zation that once the spray dryer was added, there was no reason to worry about
high temperature excursions (greater than 500°F) affecting fabric material.
PECo operated a pilot ESP on the regeneration offgas and verified the informa-
tion that had been accumulating from ESP manufacturers and magnesium oxide sup-
pliers. Effective ESP operation on magnesium oxide requires temperatures in
excess of 600°F and would be preferable at 700°F. In other systems this may
have been acceptable but the proper operation of the spray dryers requires the
largest temperature differential possible for maximum flexibility in control-
ling sulfate concentrations. A fabric filter would allow operation at the
original design temperature of 450°F for final particulate removal.
549
-------
FGD SYSTEM OPERATION
The magnesium FGD system design has an onsite acid plant containing two
trains and with a total capacity of 350 tons per day. There are four 750,000-
gallon storage tanks which can provide up to 60 days of storage. The acid plant
can produce either 93 or 98 percent acid, depending on market conditions. The
key to the system operation is the fact that storage silos for magnesium oxide
and magnesium sulfite allow decoupled operation of the regeneration area/acid
plant from that of the power plant. The absorber and dewatering/dryer areas
are sized to treat the required flue gas at full load. Although the future
yearly capacity factor for this FGD system would be significantly less than 60
percent, these systems must be sized to accommodate full load operation during
peak periods.
The regeneration area and acid plant are not designed to follow load as
are the absorbers (three operating and one spare), the centrifuges (four), and
the rotary dryers (two). The regeneration area is gas flow dependent and the
acid plant is 862 dependent. Therefore, a sulfite storage silo will be located
after the rotary dryer. Correspondingly, there is a magnesium oxide storage
silo for the regenerated magnesium after the recovery cyclones. Since the sea-
sonal electric loads will probably not correspond with the seasonal industrial
acid demand, typical operation will probably be as follows: high load on the
absorbers (filling the sulfite silo) with low load on the regeneration/ acid
plant (drawing down the oxide silo); or low absorber load (drawing down the
sulfite silo) with high load on the regeneration/acid plant (filling the oxide
silo).
The second section of this paper provides information on a test program
and schedule of a pilot plant being considered by TVA. The research and devel-
opment areas of concern are identified for each of the major process steps. A
number of research and development tasks which would provide the base technology
for the successful operation of the full-scale magnesia FGD system are discussed.
550
-------
MAGNESIUM FGD AT TVA: PILOT DESIGN
INTRODUCTION
As mentioned earlier, land availability around existing and future fossil
fuel steam plants could limit the disposal of wastes for any FGD system that
has a "throwaway" product. However, it is generally agreed that magnesium FGD
technology is not as well developed, particularly insofar as the regeneration
portion of the process is concerned, as that of lime/limestone scrubbing.
PILOT OBJECTIVES
The operation of the pilot plant will simulate the full-scale TVA magne-
sium FGD design in every practical way for the purposes of this study. Pri-
mary objectives of the pilot plant program are as follows:
1. Develop valuable in-house experience with the MgO process.
2. Anticipate potential problems with the TVA MgO process chemistry or
equipment.
3. Evaluate the long-term effects of process contaminants such as fly
ash, chlorides, and trace elements which evolve from the burning of
coal by reliability runs of several months' duration.
4. Study the formation and properties of magnesium sulfite hexahydrate
versus trihydrate crystals with respect to potential solids handling
problems.
5. Study the regenerated MgO absorption capability after several cycles.
Secondary objectives are to develop process design improvements and innova-
tions such as sulfur production, and the use of coal for drying and calcining
through additional studies not yet completely defined. All these studies will
be done in conjunction with EPA. The secondary objectives will be met in such
a way as not to impact the schedule for fulfillment of primary objectives.
Most of these objectives are process related and have general application to
magnesium FGD technology.
TVA has retained the magnesium FGD design architect/engineer, United Engi-
neers and Constructors, Inc., and had the magnesium FGD design scaled down to
the 1-MW level. The pilot plant would not include the acid production facilities.
Sections of the pilot plant (primarily the regeneration and drying sec-
tions) will be "skid" mounted to provide maximum flexibility in operation.
Skid mounting will allow the operation of the pilot plant to be broken down
into a Phase I and Phase II operational scheme. Phase I operation will be
551
-------
chiefly concerned with the primary objectives of the pilot plant as previously
listed. Phase II operation will be concerned with additional studies.
These additional studies will develop process design improvements and inno-
vative concepts. Examples of such improvements and innovative concepts may
include but not be limited to the following:
1. Using coal instead of oil in the drying and calcining of the
magnesium sulfite.
2. Spray drying for SC-2 absorption instead of wet scrubbing.
3. Testing of a sulfur-producing technology using coal as a reductant
instead of natural gas or oil.
The current schedule calls for the pilot plant to start up in the fall of
1981, with the intention of providing operational experience and some solutions
to the problems mentioned earlier.
RESEARCH AND DEVELOPMENT AREAS OF CONCERN
Identified in the following subsections for each of the major processing
steps are more details of a number of research and development tasks (outlined
above) which would provide base technology for the successful operation of a
full-scale magnesium FGD system.
PRESCRUBBER
The pH of the prescrubber solution has been calculated to be lower than
1.0 due to dissolved HC1 and H2S04- In the particulate scrubber, the hot flue
gas is contacted with a slurry of fly ash and river water. Most of the particu-
late (fly ash) and hydrogen chloride, and a variable fraction of the sulfur
trioxide, are removed in the scrubber liquor. The blowdown slurry from the
particulate scrubber, therefore, is acidic and the water may contain high con-
centrations of dissolved solids, trace metals, toxic organics, and radionuclifle;
either leached out from the fly ash or absorbed from the flue gas.
In this acidic environment only expensive alloys such as Inconel 626 or
Hastelloy G, or an organic lined metal alloy, may be suitable. Should the
prescrubber be made of materials such as 316L stainless steel, addition of a
neutralizing agent such as caustic soda or lime/limestone would be necessary
to raise the pH to at least 3.0. This addition would be undesirable since it
introduces additional dissolved solids into the system and may complicate the
disposal of the prescrubber blowdown as described later. Various materials of
construction (through the use of coupons) will be tested in the pilot plant
venturi-type prescrubber to identify those alloys and organic lined metals able
to withstand the low pH environment without neutralization to a higher pH.
552
-------
S02 ABSORPTION
For this portion of the FGD process there remain two concerns: (1) the
chemical effects of chlorides, fly ash, and dissolved magnesium salts on sul-
fite oxidation and (2) magnesium sulfite trihydrate (MgS03»3H20) vs. hexahy-
drate (MgS03'6H20) formation. Experience has shown that the trihydrate
crystals can be obtained during steady-state operation, but during startup,
shutdown, or during nonsteady-state operation hexahydrate crystals have been
observed. The crystals of the two hydrates have widely different handling
properties and this fact can introduce difficulties in the solids separation
and drying steps of the process. The operating parameters affecting the above
items will be investigated.
As an alternative to wet scrubbing TVA is considering spray drying and
subsequent fabric filter collection for S02 absorption as part of the Phase II
operational scheme. Although definite plans have not been formulated, a spray
dry/fabric filter would be investigated to determine optimum operating condi-
tions. The regeneration section from the Phase I tests would be used to decom-
pose the MgS03 to MgO and S02. This FGD system would have the advantages over
the wet scrubbing system of no flue gas reheat and elimination of the drying
step.
DRYING
Two control schemes described below will be investigated to determine
optimum operating parameters with regards to (1) sulfite to sulfate oxidation
during drying, (2) MgS03 decomposition, and (3) economical operating conditions.
With the first scheme, two variables—the airflow (primary combustion air
and dilution air) and fuel flow—will be controlled in series to maintain a
constant dryer gas discharge temperature. Fuel flow control follows combustion
chamber temperature, thereby increasing lag time for response and decreasing
control accuracy. The dryer discharge gas temperature control point senses
the gas temperature and demands an inverse gas flow change before a fuel adjust-
ment has fully responded to its initial demand. Also, the refractory lined
combustion chamber is an excellent heat sump which further increases lag time
for fuel adjustment.
For the second control scheme, common practice in other industries is to
maintain constant gas flow at a maximum rate that has an acceptable dew point
and carry-over dust load; the discharge gas temperature then controls the fuel
flow directly. Such a control method will eliminate one of the two control
variables and may reduce control lag time to a minimum.
553
-------
REGENERATION
There are two concerns about the fate and possible accumulation of minor
elements such as calcium, sodium, and chlorine in the regeneration portion of
the FGD process: (1) Does the chlorine leave the process perhaps as HC1, or
does it remain in the process tying up magnesium salts, and (2) Does a molten
liquid phase of MgCl2, NaCl, or CaCl2 exist in the calciner offgas?
A predicted composition of the calciner feed stream for the magnesium FGD
design is summarized in Table I.5 At the desired operating temperature (1800°F)
the output from the calciner was thermodynamically calculated and is shown in
Table 2.6 Thermodynamically, at 1800°F the decomposition of MgS04 to MgO, S03,
S02> and 02 is favorable; calcium is present in the solid phase as CaS04, and
a liquid sodium sulfate/chloride melt is predicted. Most of the chlorine in
the calciner feed leaves the system in the gas phase as HC1.
TABLE 1. CAICINER SOLIDS FEED
Compounds Weight Percent
MgS03
MgS04
MgCl2
Na2S04
CaCl2
Ash
70.0
8.9
9.8
0.2
0.4
10.6
Total 99.9
Experience with PECo's prototype calciner (actually an entrained-bed reactor
rather than a traditional fluid-bed reactor) at its Essex Chemical test facility
has revealed that approximately 30 percent of the MgSOs feed forms hard, che&ji--
caily unreactive MgO pellets. These pellets do not discharge with the MgO fines
overhead in the offgas but rather eventually fill the calciner bed and have to
be ground before reuse in S02 sorption. Examination of the high density MgO
pellets indicates a possible double salt, Mgs Ca(S04)4, as the root cause of
pellet formation.7 The MgO fines generated overhead are very fine (3-30
microns) and difficult to handle, thereby causing transportation and storage
problems.
As discussed earlier, in the full-scale design fouling of the she11-and-
tube air preheater exchanger by MgS04 was also experienced at PECo's regenera-
tion test facility at Essex Chemical. ,The plugging was rapid and rendered the
554
-------
heat exchanger unusable. The magnesium sulfate is thought to form in the cal-
ciner offgas by recombination of MgO and the relatively small amounts of highly
reactive 803. The recombination should be avoided not only because of heat
exchanger plugging, but also because it introduces a recirculating load of MgS04
in the system that is inactive for S02 sorption.
The regeneration study will determine whether the predicted sodium sulfate/
chloride melt and CaS04 in the gas phase are the cause of the problems at Essex
Chemical and, if so, will determine the most economical solution to the shell
and tube preheater fouling. Specifically, we propose to examine, as a function
of the operating parameters (1) MgS03 and MgS04 decomposition, (2) MgO pellet
formation, (3) MgS04 recombination, (4) optimum product yield, and (5) chlorine
composition and purge rate.
TABLE 2. OUTLET CALCINER GAS COMPOSITION AT 1800°F
Mole Percent
Gases
C02 9.4
H20 5.5
N2 53.7 ,
NO 8.75 X 10"-3
HC1 3.4 „
Cl" 1.40 X 10
C12 2.11 X 10
S02 11.6
S03 0.3 o
MgCl2 1.60 X 10"J
02 1.7
Liquids
Na2S04 1.16 X 10"^
MgCl2 1-26 X 10 \
NaCl 3.75 X lO""3
Solids
MgO 13.6
MgS04
CaS04 0.7
100.0
555
-------
DESIGN IMPROVEMENTS AND INNOVATIVE CONCEPTS
In addition to spray drying and subsequent fabric filter collection for
S02 collection, TVA is considering the following: (1) the use of coal or a
coal-oil mixture in place of oil in the drying and calcining of the magnesium
sulfite and (2) the direct production of elemental sulfur from the decomposi-
tion of MgS03. Although no definite plans have been formulated, minor modifi-
cations to the Phase I drying and regeneration equipment is all that is
necessary to test the coal-fired option (excepting, of course, the addition of
coal handling facilities). For the production of elemental sulfur, TVA has
obtained the services of P. S. Lowell to expand upon an earlier EPA study on
this topic.8
SUMMARY
Thus, we have identified proposed research and development studies on a
number of potential problems. These problems, it is felt, are soluble with
current state-of-the-art engineering knowledge. The magnesium FGD design
offers promise as a technically viable, economically feasible process for
recovering SOa as a useful product—sulfuric acid.
556
-------
REFERENCES
1. Robards, R. F.; Cole, R. M.; Morasky, T. M.; "TVA's Cocurrent Scrubber
Evaluation," Winter Annual Meeting of the Air Pollution Control Division
of ASME, Atlanta, Georgia, 1977, 77-WA/APC-8.
2. Robards, R. F.; Moore, N. D.; Kelso, T. M.; and Cole, R. M.; "Cocurrent
Scrubber Evaluation: TVA's Colbert Lime-Limestone Wet Scrubbing Pilot
Plant," EPRI FP-941, Research Project 537-1; January 1979.
3. Marcus, E. G., and Wells, W. L., "Magnesium Oxide Testing on the iO-MW
Cocurrent Scrubber at the Shawnee Steam Plant," in press.
4. McGlamery, G. G.; Torstrick, R. L.; Simpson, J. P.; and Phillips, Jr.,
J. F.; Conceptual Design and Cost Study; Sulfur Oxide Removal from Power
Plant Stack Gas, EPA-R2-73-244, May 1973.
5. Lowell, P. S., "Technical Memorandum 009-02-02, Equilibrium Program and
Thermodynamic Data Base for TVA Reducing Calciner," May 1980.
6. Lowell, P. S., "Technical Memorandum 009-02-08, Thermodynamic Analysis of
the TVA MgS03 Colbert Pilot Calciner," May 1980.
7- Kelmer, A. D., "Magnesia Flue Gas Desulfurization—Status of Development
and Needed Technological Support, Letter from Charles D. Scott (Assoicate
Director, Chemical Technology Division, ORNL) to J. Frederick Weinhold
(Director, Division of Energy Demonstrations and Technology, TVA),
February 21, 1980.
8. Lowell, P. S.; Corbett, W. E.; Brown, G. D.; and Wilde, K. A.; Feasibility
of Producing Elemental Sulfur from Magnesium Sulfite, EPA-600/7-76-030,
October 1976.
ACKNOWLEDGEMENTS
The authors wish to thank and acknowledge the assistance of the following
persons without whose help this paper could not have been written and assembled:
B. A. Anz and C. C. Thompson, Jr., of United Engineers and Constructors. We
also wish to acknowledge the financial assistance of the Environmental Protection
Agency in the pilot plant program.
The contents of this paper do not necessarily reflect the views and policies
of the Tennessee Valley Authority or the Environmental Protection Agency, nor
does mention of any trade names, commercial products, or companies constitute
endorsement or recommendation for use. This is a government publication and
not subject to copyright.
557
------- |